0% found this document useful (0 votes)

100 views

Based

Uploaded by

Intern

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

100 views

Based

Uploaded by

Intern

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 446

Task-Based Language Teaching

Task-based language teaching (TBLT) is an approach that differs from traditional

approaches by emphasizing the importance of engaging learners’ natural abilities
for acquiring language incidentally through the performance of tasks that draw
learners’ attention to form. Drawing on the multiple perspectives and expertise of
five leading authorities in the field, this book provides a comprehensive and bal-
anced account of TBLT. Split into five parts, the book provides an historical
account of the development of TBLT and introduces the key issues facing the area.
A number of different theoretical perspectives that have informed TBLT are pre-
sented, followed by a discussion on key pedagogic aspects – syllabus design, the
methodology of a task-based lesson and task-based assessment. The final parts
consider the research that has investigated the effectiveness of TBLT, address
critiques and suggest directions for future research. TBLT is now mandated by
many educational authorities throughout the world and this book serves as a core
source of information for researchers, teachers and students.

Rod Ellis is a research professor in the School of Education, Curtin University in

Perth, Australia, as well as a visiting professor at Shanghai International Studies
University and an emeritus distinguished professor of the University of Auckland. His
most recent publication is Reﬂections on Task-Based Language Teaching (2018).

Peter Skehan is an honorary research fellow at Birkbeck College, University of

London. His most recent publications include Processing Perspectives on Task
Performance (2014) and Second Language Task-Based Performance (2018).

Shaofeng Li is an associate professor of second and foreign language education at

Florida State University and an honorary professor at Zhengzhou University and
Guangdong University of Foreign Studies. His main research interests include
TBLT, language aptitude, working memory and form-focused instruction.

Natsuko Shintani is a professor in the Faculty of Foreign Language Studies, Kansai

University. Her work has been published in leading journals and she is author of
Input-Based Tasks in Foreign Language Instruction for Young Learners (2016).

Craig Lambert is an associate professor of applied linguistics and TESOL at Curtin

University in Perth, Australia. His research on TBLT has appeared in leading journals
and he is author of Referent Similarity and Nominal Syntax in TBLT (2019).

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:31, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689
THE CAMBRIDGE APPLIED LINGUISTICS SERIES
The authority on cutting-edge Applied Linguistics research
Series Editors 2007–present: Carol A. Chapelle and Susan Hunston
1988–2007: Michael H. Long and Jack C. Richards
For a complete list of titles please visit: www.cambridge.org
Recent Titles in This Series:
Feedback in Second Language Writing Practice in a Second Language
Contexts and Issues Perspectives from Applied Linguistics and
Edited by Ken Hyland and Fiona Hyland Cognitive Psychology
Language and Television Series Edited by Robert M. DeKeyser
A Linguistic Approach to TV Dialogue Task-Based Language Education
Monika Bednarek From Theory to Practice
Intelligibility, Oral Communication, and the Edited by Kris van den Branden
Teaching of Pronunciation Second Language Needs Analysis
John M. Levis Edited by Michael H. Long
Multilingual Education Insights into Second Language Reading
BetweenLanguageLearningandTranslanguaging A Cross-Linguistic Approach
Edited by Jasone Cenoz and Durk Gorter Keiko Koda
Learning Vocabulary in Another Language Research Genres
2nd Edition Exploration and Applications
I. S. P. Nation John M. Swales
Narrative Research in Applied Linguistics Critical Pedagogies and Language Learning
Edited by Gary Barkhuizen Edited by Bonny Norton and Kelleen
Teacher Research in Language Teaching Toohey
A Critical Analysis Exploring the Dynamics of Second Language
Simon Borg Writing
Figurative Language, Genre and Register Edited by Barbara Kroll
Alice Deignan, Jeannette Littlemore and Understanding Expertise in Teaching
Elena Semino Case Studies of Second Language Teachers
Exploring ELF Amy B. M. Tsui
Academic English Shaped by Non-native Criterion-Referenced Language Testing
Speakers James Dean Brown and Thom Hudson
Anna Mauranen Corpora in Applied Linguistics
Genres across the Disciplines Susan Hunston
Student Writing in Higher Education Pragmatics in Language Teaching
Hilary Nesi and Sheena Gardner Edited by Kenneth R. Rose and Gabriele
Disciplinary Identities Kasper
Individuality and Community in Academic Cognition and Second Language Instruction
Discourse Edited by Peter Robinson
Ken Hyland
Research Perspectives on English for
Replication Research in Applied Linguistics Academic Purposes
Edited by Graeme Porte Edited by John Flowerdew and Matthew
The Language of Business Meetings Peacock
Michael Handford Computer Applications in Second Language
Reading in a Second Language Acquisition
Moving from Theory to Practice Foundations for Teaching, Testing and
William Grabe Research
Modelling and Assessing Vocabulary Carol A. Chapelle
Knowledge
Edited by Helmut Daller, James Milton and
Jeanine Treffers-Daller

Rod Ellis
Curtin University, Perth

Peter Skehan
Birckbeck, University of London

Shaofeng Li
Florida State University, Tallalhassee

Natsuko Shintani
Kansai University, Osaka

Craig Lambert
Curtin University, Perth

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108494083
DOI: 10.1017/9781108643689
© Rod Ellis, Peter Skehan, Shaofeng Li, Natsuko Shintani and Craig Lambert 2020
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2020
Printed in the United Kingdom by TJ International Ltd, Padstow Cornwall
A catalogue record for this publication is available from the British Library.
ISBN 978-1-108-49408-3 Hardback
ISBN 978-1-108-71389-4 Paperback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.

Rod Ellis is a research professor in the School of Education, Curtin University in

Peter Skehan is an honorary research fellow at Birkbeck College, University of

London. His most recent publications include Processing Perspectives on Task
Performance (2014) and Second Language Task-Based Performance (2018).

Shaofeng Li is an associate professor of second and foreign language education at

Natsuko Shintani is a professor in the Faculty of Foreign Language Studies, Kansai

University. Her work has been published in leading journals and she is author of
Input-Based Tasks in Foreign Language Instruction for Young Learners (2016).

Craig Lambert is an associate professor of applied linguistics and TESOL at Curtin

University in Perth, Australia. His research on TBLT has appeared in leading journals
and he is author of Referent Similarity and Nominal Syntax in TBLT (2019).

Rod Ellis
Curtin University, Perth

Peter Skehan
Birckbeck, University of London

Shaofeng Li
Florida State University, Tallalhassee

Natsuko Shintani
Kansai University, Osaka

Craig Lambert
Curtin University, Perth

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

Rod Ellis is a research professor in the School of Education, Curtin University in

Peter Skehan is an honorary research fellow at Birkbeck College, University of

London. His most recent publications include Processing Perspectives on Task
Performance (2014) and Second Language Task-Based Performance (2018).

Shaofeng Li is an associate professor of second and foreign language education at

Natsuko Shintani is a professor in the Faculty of Foreign Language Studies, Kansai

University. Her work has been published in leading journals and she is author of
Input-Based Tasks in Foreign Language Instruction for Young Learners (2016).

Craig Lambert is an associate professor of applied linguistics and TESOL at Curtin

University in Perth, Australia. His research on TBLT has appeared in leading journals
and he is author of Referent Similarity and Nominal Syntax in TBLT (2019).

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:30, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689
THE CAMBRIDGE APPLIED LINGUISTICS SERIES
The authority on cutting-edge Applied Linguistics research
Series Editors 2007–present: Carol A. Chapelle and Susan Hunston
1988–2007: Michael H. Long and Jack C. Richards
For a complete list of titles please visit: www.cambridge.org
Recent Titles in This Series:
Feedback in Second Language Writing Practice in a Second Language
Contexts and Issues Perspectives from Applied Linguistics and
Edited by Ken Hyland and Fiona Hyland Cognitive Psychology
Language and Television Series Edited by Robert M. DeKeyser
A Linguistic Approach to TV Dialogue Task-Based Language Education
Monika Bednarek From Theory to Practice
Intelligibility, Oral Communication, and the Edited by Kris van den Branden
Teaching of Pronunciation Second Language Needs Analysis
John M. Levis Edited by Michael H. Long
Multilingual Education Insights into Second Language Reading
BetweenLanguageLearningandTranslanguaging A Cross-Linguistic Approach
Edited by Jasone Cenoz and Durk Gorter Keiko Koda
Learning Vocabulary in Another Language Research Genres
2nd Edition Exploration and Applications
I. S. P. Nation John M. Swales
Narrative Research in Applied Linguistics Critical Pedagogies and Language Learning
Edited by Gary Barkhuizen Edited by Bonny Norton and Kelleen
Teacher Research in Language Teaching Toohey
A Critical Analysis Exploring the Dynamics of Second Language
Simon Borg Writing
Figurative Language, Genre and Register Edited by Barbara Kroll
Alice Deignan, Jeannette Littlemore and Understanding Expertise in Teaching
Elena Semino Case Studies of Second Language Teachers
Exploring ELF Amy B. M. Tsui
Academic English Shaped by Non-native Criterion-Referenced Language Testing
Speakers James Dean Brown and Thom Hudson
Anna Mauranen Corpora in Applied Linguistics
Genres across the Disciplines Susan Hunston
Student Writing in Higher Education Pragmatics in Language Teaching
Hilary Nesi and Sheena Gardner Edited by Kenneth R. Rose and Gabriele
Disciplinary Identities Kasper
Individuality and Community in Academic Cognition and Second Language Instruction
Discourse Edited by Peter Robinson
Ken Hyland
Research Perspectives on English for
Replication Research in Applied Linguistics Academic Purposes
Edited by Graeme Porte Edited by John Flowerdew and Matthew
The Language of Business Meetings Peacock
Michael Handford Computer Applications in Second Language
Reading in a Second Language Acquisition
Moving from Theory to Practice Foundations for Teaching, Testing and
William Grabe Research
Modelling and Assessing Vocabulary Carol A. Chapelle
Knowledge
Edited by Helmut Daller, James Milton and
Jeanine Treffers-Daller

Rod Ellis
Curtin University, Perth

Peter Skehan
Birckbeck, University of London

Shaofeng Li
Florida State University, Tallalhassee

Natsuko Shintani
Kansai University, Osaka

Craig Lambert
Curtin University, Perth

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

Rod Ellis is a research professor in the School of Education, Curtin University in

Peter Skehan is an honorary research fellow at Birkbeck College, University of

London. His most recent publications include Processing Perspectives on Task
Performance (2014) and Second Language Task-Based Performance (2018).

Shaofeng Li is an associate professor of second and foreign language education at

Natsuko Shintani is a professor in the Faculty of Foreign Language Studies, Kansai

University. Her work has been published in leading journals and she is author of
Input-Based Tasks in Foreign Language Instruction for Young Learners (2016).

Craig Lambert is an associate professor of applied linguistics and TESOL at Curtin

University in Perth, Australia. His research on TBLT has appeared in leading journals
and he is author of Referent Similarity and Nominal Syntax in TBLT (2019).

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689
THE CAMBRIDGE APPLIED LINGUISTICS SERIES
The authority on cutting-edge Applied Linguistics research
Series Editors 2007–present: Carol A. Chapelle and Susan Hunston
1988–2007: Michael H. Long and Jack C. Richards
For a complete list of titles please visit: www.cambridge.org
Recent Titles in This Series:
Feedback in Second Language Writing Practice in a Second Language
Contexts and Issues Perspectives from Applied Linguistics and
Edited by Ken Hyland and Fiona Hyland Cognitive Psychology
Language and Television Series Edited by Robert M. DeKeyser
A Linguistic Approach to TV Dialogue Task-Based Language Education
Monika Bednarek From Theory to Practice
Intelligibility, Oral Communication, and the Edited by Kris van den Branden
Teaching of Pronunciation Second Language Needs Analysis
John M. Levis Edited by Michael H. Long
Multilingual Education Insights into Second Language Reading
BetweenLanguageLearningandTranslanguaging A Cross-Linguistic Approach
Edited by Jasone Cenoz and Durk Gorter Keiko Koda
Learning Vocabulary in Another Language Research Genres
2nd Edition Exploration and Applications
I. S. P. Nation John M. Swales
Narrative Research in Applied Linguistics Critical Pedagogies and Language Learning
Edited by Gary Barkhuizen Edited by Bonny Norton and Kelleen
Teacher Research in Language Teaching Toohey
A Critical Analysis Exploring the Dynamics of Second Language
Simon Borg Writing
Figurative Language, Genre and Register Edited by Barbara Kroll
Alice Deignan, Jeannette Littlemore and Understanding Expertise in Teaching
Elena Semino Case Studies of Second Language Teachers
Exploring ELF Amy B. M. Tsui
Academic English Shaped by Non-native Criterion-Referenced Language Testing
Speakers James Dean Brown and Thom Hudson
Anna Mauranen Corpora in Applied Linguistics
Genres across the Disciplines Susan Hunston
Student Writing in Higher Education Pragmatics in Language Teaching
Hilary Nesi and Sheena Gardner Edited by Kenneth R. Rose and Gabriele
Disciplinary Identities Kasper
Individuality and Community in Academic Cognition and Second Language Instruction
Discourse Edited by Peter Robinson
Ken Hyland
Research Perspectives on English for
Replication Research in Applied Linguistics Academic Purposes
Edited by Graeme Porte Edited by John Flowerdew and Matthew
The Language of Business Meetings Peacock
Michael Handford Computer Applications in Second Language
Reading in a Second Language Acquisition
Moving from Theory to Practice Foundations for Teaching, Testing and
William Grabe Research
Modelling and Assessing Vocabulary Carol A. Chapelle
Knowledge
Edited by Helmut Daller, James Milton and
Jeanine Treffers-Daller

Rod Ellis
Curtin University, Perth

Peter Skehan
Birckbeck, University of London

Shaofeng Li
Florida State University, Tallalhassee

Natsuko Shintani
Kansai University, Osaka

Craig Lambert
Curtin University, Perth

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

List of Figures page vii

List of Tables viii
Series Editors’ Preface xi
Authors’ Preface xiii

part i introduction 1
1 The Pedagogic Background to Task-Based
Language Teaching 3

part ii theoretical perspectives 27

2 Cognitive-Interactionist Perspectives 29
3 Psycholinguistic Perspectives 64
4 Sociocultural Perspectives 103
5 Psychological Perspectives 129
6 Educational Perspectives 155

part iii pedagogical perspectives 175

7 Task-Based Syllabus Design 179
8 Methodology of Task-Based Language
Teaching 208
9 Task-Based Testing and Assessment 241

part iv investigating task-based programmes 281

10 Comparative Method Studies 283
11 Evaluating Task-Based Language Teaching 303

part v moving forward 331

12 Responding to the Critics of Task-Based Language
Teaching 333
13 Questions, Challenges and the Future 353

Endnotes 371
References 374
Index 412

1.1 Outline of the task-based learning framework

(based on Willis 1996, p. 52) page 15
2.1 Cognitive-interactionist model informing TBLT 31
2.2 Model of non-understanding routines
(Varonis and Gass 1985) 35
2.3 The explicit/implicit continuum
(slightly modiﬁed from Lyster and Saito 2010, p. 278) 53
4.1 Activity theory model (Engeström 1999) 110
6.1 Maehr’s theory of personal investment
(from Lambert 1998) 161
9.1 An outline model for task-based testing 245
11.1 Types of evaluation 308

vii

1.1 A typology of task types (based on Prabhu 1987,

pp. 46–7) page 8
1.2 Criteria for defining a task-as-workplan
(based on Ellis and Shintani 2014) 10
1.3 Features of different tasks 11
1.4 TBLT and CLT/CLIL compared (based on
Ortega 2015, p. 104) 18
2.1 Indicator types in corrective feedback
(based on Lyster and Ranta 1997) 36
2.2 Task design and implementation variables
investigated in interaction studies 40
3.1 Potential sequences in the SSARC model 83
3.2 A comparison between the LAC approach and the
CH/SSARC model 87
3.3 Malicka and Sasayama’s analysis of CH-linked
variables 92
4.1 Basic constructs in sociocultural SLA 106
6.1 Instruction task discourse based on LGC 165
6.2 Narrative task discourse based on LGC 166
6.3 Opinion task discourse based on LGC 166
6.4 Strengths of LGC and TGC tasks within the L2
curriculum 168
7.1 Prabhu’s criteria for task sequencing 183
7.2 Ellis’ criteria for task sequencing (Ellis 2003) 198
8.1 Summary of pre-task methodological options 211
8.2 Summary of within-task methodological options 223
8.3 Summary of post-task methodological options 229
8.4 Participatory structure of task-based interaction 236
9.1 Overview of meta-analytic results 248
9.2 Level of difficulty linked to Conceptualizer and
Formulator influences 263
10.1 ‘Need’, ‘search’ and ‘evaluation’ in PPP and TBLT
(Shintani 2015, p. 58) 294

viii

10.2 General programme comparisons between TBLT

and PPP 296
10.3 Focused comparison studies 298
11.1 Factors determining the success of innovations
(from Ellis 1997, p. 29) 305
11.2 A selection of evaluations of TBLT programmes 310
12.1 The characteristics of TBLT according to Swan
(2005a) 334
12.2 Critiques of TBLT 352

Task-based language teaching (TBLT) has been enormously inﬂuential

since the 1980s, when it inspired a generation of language teachers
seeking to engage productively with Communicative Language Teach-
ing. Since then it has developed as an approach to methodology,
assessment and syllabus design. As TBLT has grown in popularity it
has also diversiﬁed, incorporating a number of theoretical stances
towards how languages are learnt.
This book provides a substantial overview of the current position of
TBLT in the language-teaching world. It covers both pedagogic and
research perspectives, arguing that the two activities are complemen-
tary and mutually supportive. In terms of research, the book provides
a detailed account of the theoretical approaches that underpin TBLT.
Those theories relate to a number of perspectives: cognitive, psycho-
linguistic, sociocultural, psychological and educational. Under those
headings, the book includes comprehensive and authoritative assess-
ments of research into such issues as: the roles of interaction and
feedback; measures of complexity, accuracy and ﬂuency; the import-
ance of classroom phenomena such as scaffolding and individual
variables such as motivation; the relation between psychological vari-
ables and language learning; and the intersection between educational
practice in general and language teaching in particular.
The pedagogic chapters are more practically oriented, but also draw
extensively on research into the effectiveness of TBLT. They provide a
wealth of information on how to design a task-based course, what
methods are used in such courses and why, and how task-based learn-
ing can and should be assessed. What comes across strongly is the
degree of variation within TBLT: there is no one syllabus design and
no one methodology that takes precedence over others. The authors
argue convincingly that this is a positive feature of TBLT, in that it can
be adapted to suit a variety of contexts and learning styles. In short, the
authors do not present TBLT as an approach wherein a centre imposes
action on a periphery. Rather, the principles that lie behind TBLT are an
inspiration for many kinds of classroom and assessment contexts.

The ﬁnal part of the book presents an honest appraisal of task-based

language teaching in relation to language teaching more generally.
Research that addresses this issue is summarized and a balanced
conclusion presented. TBLT is not a ‘magic bullet’, and research still
needs to be undertaken to establish the extent of its efﬁcacy. The
chapters in this part indicate how this research can be done, and what
challenges are involved in carrying it out. To date, the effectiveness of
TBLT is apparent in situations in which it is the dominant paradigm
and also in those where it exerts an inﬂuence on teaching and assess-
ment approaches that prioritize attention to meaning and interaction.
The authors present TBLT as a major development in language
teaching, and a crucial part of current pedagogic practice. The message
of this book is that in TBLT research and practice form a continuous
whole. It is a welcome addition to the series.

Interest in task-based language teaching (TBLT) has burgeoned over

the last thirty years. It can now be considered one of the mainstream
approaches to teaching second/foreign languages as reflected in the
growing number of publications intended for teachers (e.g. Willis
1996; Willis and Willis 2007; Ellis 2018a) and an expansive body of
research that has investigated the effect of task design and implemen-
tation variables on the performance of tasks and on L2 acquisition
(e.g. Ellis 2003; Van den Branden, Bygate and Norris 2009; Robinson
2011; Long 2015; Skehan 2018).
This book aims to provide a comprehensive survey of the pedagogic
and the research literature. It has three aims:
1. The general aim is to provide a broad-based and accessible state-of-
the art account of TBLT by considering the pedagogical aspects of
this approach and by reviewing relevant theories and research that
have informed the design and implementation of task-based
courses. While these two perspectives are inter-related they have
led to somewhat different justifications for designing and imple-
menting task-based courses.
2. The second aim is to examine the effectiveness of TBLT in relation
to other mainstream approaches to language teaching. One of the
criticisms levelled at TBLT is that there is insufficient evidence to
demonstrate that TBLT is more effective in developing L2 learners’
communicative abilities than other more traditional approaches.
A number of comparative method and evaluation studies enable us
to examine the validity of this criticism and to demonstrate that
TBLT is effective.
3. The third aim is to examine the criticisms of TBLT that have been
advanced by advocates of traditional language teaching and then to
identify a number of ‘real’ issues that need to be addressed. To this
end, we will consider the problems that teachers face in introducing
TBLT into their classrooms and how these problems can be
addressed.

xiii

There are two general principles that inform the positions we have
taken in the book:
1. Task-based pedagogy and task-based research are complementary.
There is perhaps no area of language teaching where pedagogy and
research have been so closely intertwined. The practice of TBLT in
real classrooms has raised questions that are not just important for
teachers but also of interest to researchers. For example, teachers
have expressed concern about learners’ use of their ﬁrst language
(L1) when they are performing speaking tasks while researchers
have investigated speciﬁc ways in which the use of L1 can facilitate
both the performance of a task and second language (L2) learning.
Research-directed activity has also fed into the practice of teaching.
For example, the usefulness of having learners plan before they
perform a task has been clearly established through the research
that has investigated pre-task planning. As Pica (1997) noted
teachers, methodologists and researchers have a shared interest in
the use communication tasks. This shared interest is what informs
the book.
2. We view TBLT as an approach, not a method. That is, TBLT is
based on a set of general principles that inform how a language is
best taught and learned but it is not prescriptive of either how to
design a task-based course or how to implement tasks in the
classroom. Nor is the approach monolithic. There are different
versions of the approach. We acknowledge these differences and
consider how TBLT can be adapted to take account of the needs of
teachers and learners in different instructional contexts. This
acknowledgement of the diversity in TBLT is a key feature of the
book that distinguishes it from the narrower, more circumscribed
view of TBLT found in some other publications.
Each part of the book approaches TBLT from a different angle while
always maintaining the interface between pedagogical concerns and
research and acknowledging the diversity within TBLT. Part I provides
the general background to TBLT and serves as a foundation for
subsequent parts. Part II focuses on the theories and research that
have informed task-based research. It examines a number of different
perspectives by addressing the theoretical constructs that underlie each
perspective and the research methodologies that have been utilized in
investigating them. In Part III the focus switches to pedagogy, drawing
on relevant research and emphasizing the diversity in TBLT. It
addresses the principles that inform the selection and sequencing of
tasks in a task-based course, the methodological principles that

underlie proposals for implementing a task in the classroom, and the

kinds of assessment that are compatible with TBLT. Part IV looks at
the research that has investigated complete TBLT courses. It considers
whether the claim that TBLT is more effective than traditional, struc-
tural approaches to language teaching is justified and reports on
evaluation studies that have examined the viability of introducing
TBLT in different instructional contexts. Part V concludes the book
by first examining the criticisms of TBLT that have been made and
suggesting the lines of research needed to further understanding of the
relationship between tasks and learning. Finally, we return to con-
sidering how task-based research and task-based teaching can most
profitably interface.
The primary readers of this book will be researchers, postgraduate
students and teachers who are interested in using TBLT in their
classrooms. It seeks to be accessible to readers who are not familiar
with the research and theory that inform TBLT but it is not a ‘how-to-
do-it’ book. Our aim is to survey the field in order to provide a wealth
of information that can inform the design of task-based courses, the
planning of task-based lessons, the assessment of learning and the
evaluation of courses.

Task-based language teaching (TBLT) constitutes an approach to

language teaching that prioritizes meaning but does not neglect form.
It emphasizes the importance of engaging learners’ natural abilities for
acquiring language incidentally as they engage with language as a
meaning-making tool; it thus contrasts with structural approaches that
emphasize language as an object to be systematically taught and
intentionally learned.
The purpose of the chapter in Part I of the book is to provide a
general introduction by outlining a number of key issues that will be
addressed more fully in subsequent parts. We begin by providing a
historical sketch of TBLT, showing its pedagogic origins in communi-
cative language teaching (CLT) and its theoretical foundations in
second language acquisition (SLA) research and principles of sound
education. We then trace the development of TBLT from its early
days, pointing to the multiple influences that have helped to shape its
evolution. We address key issues such as how to define ‘task’, how
tasks have been classified, how they can be sequenced into a syllabus,
how a complete lesson can be built around a task, the use of tasks
in computer-mediated (CM) language teaching, and task-based assess-
ment. We introduce the key construct of ‘focus on form’ and
explain its importance in TBLT and consider the difference between
‘task-based’ and ‘task-supported’ language teaching.
TBLT constitutes a major innovation in those instructional contexts
where language has been taught through a structural syllabus. For this
reason, the evaluation of task-based courses plays an important role in
understanding how TBLT can be made to work efficiently and effect-
ively in different contexts. TBLT has not always been welcomed by
members of the language teaching profession. We are aware of the
critiques that have been mounted against TBLT and briefly address
them. We point out that these are often based on misunderstandings of

TBLT, but we also acknowledge the need to demonstrate that TBLT is

in fact more effective than traditional approaches.
As noted in the Preface, the position we have taken in this book is
that TBLT is not a monolithic, tightly deﬁned approach but quite
diverse. There are many issues relating to the design and implementa-
tion of task-based courses that continue to be debated. It is appropri-
ate, therefore, that the chapter ends with a set of questions rather than
a summative statement about TBLT. These questions are addressed in
subsequent chapters of the book.

The overall purpose of the chapter is to introduce key issues in task-

based language teaching (TBLT), which will be taken up in subsequent
chapters. We ﬁrst consider initial proposals for a task-based approach
in the 1980s. We then examine how TBLT subsequently developed,
focusing on the design of a task-based syllabus and the methodology
for implementing tasks. We brieﬂy consider how TBLT has been
adapted to computer-mediated (CM) environments and also look at
task-based assessment. We discuss what evaluation studies have
shown about the effectiveness of TBLT and the problems that teachers
face in implementing it. The chapter concludes with a brief discussion
of the criticisms that have been levelled at TBLT.

Starting Points
The importance of including tasks in a language curriculum was
established in the communicative language teaching (CLT)
movement of the 1970s and 1980s. TBLT grew out of this movement,
with further input from early research in second language acquisition
(SLA), which led to a questioning of the structural approach to teach-
ing languages where a language is broken down into bits to be taught
sequentially one at a time.

CLT
CLT drew on theories of language that emphasized communicative
competence (Hymes, 1971) and that viewed language as functional in
nature (Halliday, 1973). These theories led to the recognition that
‘there is more to the business of communicating than the ability to
produce grammatically correct utterances’ (Johnson, 1982) and to the
idea of replacing a traditional structural syllabus with a notional
syllabus (Wilkins, 1976). In other words, there was a move away from

a ‘synthetic’ way of teaching founded on an inventory of grammatical

structures to an ‘analytic’ approach based on language functions such
as ‘expressing agreement or disagreement’ and semantic notions
such as ‘time’ and ‘space’. However, the language teaching materials
based on a notional syllabus (e.g. Abbs and Freebairn, 1982) did not
differ greatly from those based on a structural syllabus. That is, the
linguistic forms for expressing each notion were mainly presented
in situations and then practised in controlled exercises. Thus, while
the organizational framework of a language course had changed, the
methodology had not.
There was, however, a growing recognition of the need for a com-
municative methodology. Johnson (1982), for example, advocated
what he called the deep-end strategy, where ‘the student is placed in
a situation where he may need to use language not yet taught’ so as to
activate ‘the ability to search for circumlocutions when the appropri-
ate language item is not known’ (p. 193). This called for communi-
cative tasks where the learner’ use of language was judged not in terms
of whether it was grammatically correct but in terms of whether the
communicative outcome of the task was achieved.
CLT never developed into well-defined ‘method’. Howatt (1984)
distinguished a weak version, where teaching content was defined in
terms of the linguistic realizations of notions and functions, but the
methodology remained essentially the same as in the traditional struc-
tural approach, and a strong version, where the content of a language
programme was specified in terms of communicative tasks and the
methodological focus was on fluency. TBLT grew out of the strong
CLT approach.

SLA Research
The SLA research that started in the 1960s and 1970s fed into the
emergence of TBLT. Cross-sectional studies of learners acquiring a
second language (L2) naturalistically (e.g. Dulay and Burt, 1973)
provided evidence that there was an acquisition order that was
common to all learners irrespective of their ﬁrst languages (L1) or
their age. Furthermore, a very similar order was found in classroom
learners, suggesting that instruction did not have a major impact
on the developmental route learners followed. Longitudinal studies
(e.g. Cancino, Rosansky and Schumann, 1978) showed that learners
passed through a series of stages involving ‘transitional constructions’
en route to the target form. Progress was gradual and often very slow,
and at any one stage of development considerable variability was
evident in those constructions that had been acquired up to that point.

Furthermore, it was clear that learners did not set about achieving
target-like use of grammatical structures in linear fashion. Rather, they
worked on several structures concurrently. This research led to the
claim that there was a ‘natural route’ for mastering the grammar of a
language and that learners had their own ‘built-in syllabus’ for learn-
ing it (Corder, 1967).
Drawing on this research, Krashen (1985) argued that true profi-
ciency in an L2 depends on ‘acquisition’, defined as ‘the subconscious
process identical in all important ways to the process children utilize in
acquiring their first language’ and not on ‘learning’, defined as ‘the
conscious process that results in “knowing about” language’ (p. 1).
The Natural Approach (Krashen and Terrell, 1983) constituted an
attempt to apply Krashen’s ideas about how languages were ‘acquired’
to pedagogic practice. It emphasizes activities that focus learners’
primary attention on meaning and caters to incidental acquisition.
TBLT is based on the same principle.

Early TBLT Proposals

‘Tasks’ figured in both early CLT and the Natural Approach but in
neither were they conceived of as the units around which a complete
language course could be built. It was not until the mid- to late 1980s
that the first proposals for a task-based approach appeared. These
early proposals (Long, 1985; Candlin, 1987; Breen, 1989) were
largely programmatic in nature. They focused on the rationale for a
task-based syllabus and outlined how to design and evaluate a task-
based curriculum. Prabhu (1987) provided the first complete account
of a task-based course while Nunan (1989) gave practical advice
about how to design tasks.1

Rationale for TBLT

These early proposals were based on:
• research in SLA (Long, 1985);
• general educational principles (Candlin and Breen);
• dissatisfaction with structural-based teaching and the intuition that
the development of grammatical competence was best achieved
through the effort to cope with communication (Prabhu);
• the utility of ‘task’ as a unit that integrates what learners will learn
(i.e. the syllabus) with how they learn (i.e. methodology) (Nunan).
From the start, therefore, there were multiple inputs into the rationale
for TBLT.

• Drawing on research in SLA, Long (1985) argued that ‘there is no

reason to assume that presenting the target language as a series of
discrete linguistic or sociolinguistic teaching points is the best, or
even a way to get learners to synthesize the parts into a coherent
whole’ (p. 79). He saw an approach based on tasks as providing an
‘integrated solution to both syllabus and methodological issues’
(p. 89).
• Candlin (1987) critiqued traditional approaches from an educa-
tional standpoint. He argued that they failed to ‘emphasize educa-
tional goals . . . in their pursuit of cost-effective training’ (p. 16).
Along with Breen (1989), he emphasized the importance of teachers
and students jointly negotiating the content of a course and argued
that tasks provided the best means for achieving this. Candlin
claimed that an approach based on tasks would enable learners ‘to
become more aware of their own personalities and social roles’
(p. 17), foster self-realization and self-fulﬁlment and enhance their
self-conﬁdence.
• Prabhu’s (1987) starting point was dissatisfaction with the
Structural-Oral Situational Method which was dominant in his
particular teaching context (India) at that time. He argued that
‘the development of competence in a second language requires not
systematization of language input or maximation of planned prac-
tice, but rather the creation of conditions in which learners engage
in an effort to cope with communication’ (p. 1) and that this could
be best achieved by having students perform tasks.
• Nunan (1989) sought to provide teachers with a practical introduc-
tion to the design and use of tasks. He claimed that basing teaching
on tasks avoided the traditional distinction between syllabus and
methodology. Traditional syllabuses did have a role, but as check-
lists rather than as directives about what to teach. Thus the starting
point was the selection of the task(s) for a particular lesson.

Defining ‘Task’
The early proposals for task-based teaching all provided definitions of
a ‘task’ but these varied in a number of ways. Breen’s (1989) definition
was the most encompassing. A task is ‘a structured plan for the
provision of opportunities for the refinement of knowledge and cap-
abilities entailed in a new language and its use during communication’.
According to this definition, a task could be both a brief practice
exercise and ‘a more complex workplan that requires spontaneous
communication’. Other definitions emphasized four important aspects
of a task:

• A task is a meaning-focused activity. It requires learners to focus on

meaning rather than form (Nunan, 1989).
• A task does not specify the exact meaning-content to be addressed as
this will be subject to modification when it is performed. The language
needed to perform a task is negotiable as the task is performed.
• A task should bear some resemblance to a task that people perform
in real life. Long (1985) defined tasks as ‘the hundred and one things
people do in everyday life, at work, at play and in between’ (p. 89).
• A task should have ‘a sense of completeness’ and ‘stand alone as a
communicative act in its own right’ (Nunan, 1989, p. 10).
One of the problems with these early definitions is that they conflated
two senses of ‘task’ – task-as-workplan and task-as-process (Breen,
1989). It was the failure to make this crucial distinction that led to
the claim that the traditional distinction between ‘syllabus’ and ‘meth-
odology’ loses relevance. We will argue later, however, that this dis-
tinction is very relevant to TBLT and that it is best to define task as a
workplan.

Classifying Tasks
We ﬁnd a mixed bag of suggestions for distinguishing different types
of task in these early proposals. Candlin commented that it is not
possible to ‘offer anything other than implicit suggestions that tasks
might be catalogued under several distinct types’ (1987, p. 14) and
that as a result ‘a typology is bound to be fuzzy-edged and at most a
managerial convenience’ (p. 15). Long distinguished ‘target tasks’ (i.e.
real-life tasks such as ‘selling an airline ticket’), ‘task types’ (i.e. general
tasks such as ‘selling an item’), and ‘pedagogic tasks’ (i.e. the actual
tasks that teachers and students work with). Nunan presented a
number of task typologies drawn from different sources, the most
useful of which is Prabhu’s (see Table 1.1). This is based on how the
information in a task is handled by the participants.

Grading and Sequencing Tasks

The early proposals for TBLT identiﬁed a number of criteria for
determining the difﬁculty of pedagogical tasks:
• The linguistic complexity of the input provided by a task.
• The amount of input provided in the task.
• The number of steps involved in the execution of a task.
• The degree of structure in the information presented or required by
the task.

Table 1.1 A typology of task types

Type of task Deﬁnition

Information gap This type involves ‘a transfer of given information from one
person to another – or from one form to another, or from
one place to another’.
Reasoning gap This type involves ‘deriving some new information from given
information through the processes of inference, deduction,
practical reasoning, or a perception of relationships or
patterns’.
Opinion gap This type involves ‘identifying and articulating a personal
preference, feeling, or attitude in response to a given
situation’.

Source: Based on Prabhu (1987, pp. 46–7).

• The number of objects, events or people that need to be distin-

guished when performing the task.
• The extent to which the task requires reference to present or past/
future events.
• The extent to which reasons for actions or decisions need to
be given.
• The intellectual challenge posed.
• The learners’ familiarity with the topic of the task.
It should be immediately apparent that while such factors can clearly
inﬂuence the difﬁculty of individual tasks, they cannot be easily used
to grade tasks. It is not evident, for example, how one factor should be
balanced against others. Prabhu found that the grading and
sequencing tasks in the Communicational Teaching Project was more
a matter of intuition than precise measurement and therefore largely a
matter of trial and error.2

Evaluating Tasks
The importance of evaluating tasks was also recognized in these early
proposals for TBLT. Long made the point that the success of a task
needs to be judged in terms of task accomplishment rather than target-
like linguistic production. He suggested that specialists should assess
whether learners had mastered the ability to perform a ‘target task’.
Candlin proposed three general areas to be considered in evaluating
the utility of a task – its diagnostic value, its implementability in the
classroom and the extent to which it ﬁts in with and leads to other
tasks. Nunan offered the most detailed proposal in the form of a

checklist of questions to be asked about a task (see pp. 135–7). This

list includes questions relating to the design of the task (e.g. ‘Is there an
information-gap?’), its implementation (e.g. ‘What type of language is
stimulated by the task?’), and the learners’ affective response to the
task (e.g. ‘Does the task engage the learners’ interests?’). As with the
other aspects of TBLT, these suggestions were insightful but clearly
programmatic.

Subsequent Developments
Over time, the issues raised in the early proposals were built on and
new issues emerged. The rationale for TBLT was further expanded to
incorporate general educational principles. The thorny issue of the
deﬁnition of a task was revisited. The assumption that the traditional
distinction between syllabus and methodology was no longer applic-
able in TBLT was challenged as it became clear that the issues relating
to the design and implementation of tasks remain distinct and thus
warrant separate consideration.

Broadening the Rationale for TBLT

We have seen that the underpinnings of TBLT lay in CLT (the ‘strong
version’) and in SLA research and theory. With the exception of
Candlin (1987), little attention was initially paid to broader educa-
tional principles. One of the major developments that followed was an
attempt to align TBLT with general theories of education. Samuda and
Bygate (2008) drew on Dewey’s (1938) critique of the traditional
classroom with its view of learning as the mastery of ready-made
products and his emphasis on the importance of learning that connects
with experience of the real world. They pointed to Bruner’s (1960)
emphasis on ‘learning for use’ where the learner is positioned not just
as a ‘student’ but as a ‘practitioner’. TBLT is highly compatible with
the holistic, experience-driven pedagogies advocated by these promin-
ent educationalists.

Defining ‘Task’
Definitions of tasks have proliferated over the years. Van den Branden
(2006) reviewed a total of seventeen different definitions which
he divided into two groups, depending on whether they were
viewed as tasks in terms of language learning goals or educational
activity. We do not find this proliferation of definitions helpful
and argue that there is a need for a definition that is applicable across
contexts and purposes.

The problem in arriving at such a deﬁnition originates in the failure

to distinguish task-as-workplan and task-as-process. This is evident
in the meaning attached to the word ‘activity’, which figures in many
of the definitions. This term is ambiguous as it can refer to both
the actual materials that constitute a task (i.e. the workplan) or
to the language use resulting from the performance of the task (i.e. the
process). We argue that a task cannot be defined in terms of process
as this is, to some extent, unpredictable. Moreover, from the perspec-
tive of course design as well as language testing and research, the
starting point needs to be the task-as-workplan, namely the design
materials that will create a context for the communicative use of the
L2. Whether this is in fact achieved (i.e. whether the task-as-workplan
results in the activity intended) is an important question which can
only be answered by investigating the task-as-process.
We propose, therefore, a definition based on criteria that can be
used to distinguish whether a given workplan is a task or not a task
(i.e. an ‘exercise’). We nevertheless acknowledge that some workplans
may satisfy some but not all the criteria and therefore can be more or
less ‘task-like’. The criteria are listed in Table 1.2.

Table 1.2 Criteria for deﬁning a task-as-workplan

Criteria Description
The primary focus is The workplan is intended to ensure that learners are
on meaning primarily concerned with comprehending or/and
producing messages for a communicative purpose
(i.e. there is primary focus on meaning-making).
There is some kind of gap The workplan is designed in such a way as to
incorporate a gap which creates a need to convey
information, to reason or to express an opinion.
Learners rely mainly Learners need to draw on their existing linguistic
on their own linguistic resources (potentially both L1 and L2) and their
and non-linguistic non-linguistic resources (e.g. gesture; facial
resources expressions) for comprehension and production.
There is therefore no explicit presentation of
language.
There is a clearly deﬁned The workplan speciﬁes the communicative outcome
communicative of the task. Thus task accomplishment is to be
outcome assessed not in terms of whether learners use
language correctly but in terms of whether the
communicative outcome is achieved.

Source: Based on Ellis and Shintani (2014).

Issues Relating to Task Design

TASK TYPES

There is still no generally accepted way of classifying tasks. By and

large, pedagogical accounts have continued to distinguish tasks in
terms of the operations learners are required to carry out when they
perform them. Willis (1996), for example, distinguished six types –
listing, ordering and sequencing, comparing, problem solving, sharing
personal experiences and creative. Other ways of classifying tasks
have emerged from research that has investigated the communicative
and cognitive processes involved in performing different tasks leading
to a set of features (see Table 1.3) that may impact on the language a
task elicits. Any particular task can be described in terms of the speciﬁc
features it incorporates. For example, an information-gap task that
requires one learner to provide detailed descriptions of a set of pictures

Table 1.3 Features of different tasks

Task type Description

One way versus two In a one-way information-gap task, one participant holds
way all the information that needs to be communicated and
thus functions as the information-provider while the
other functions primarily as the receiver of the
information but may interact if communication
becomes problematic. In a two-way task, the
information is split between the participants so both
need to function as the providers and receivers of the
information.
Monologic versus A monologic task places the burden of performing the task
dialogic entirely on a single speaker and therefore involves a
long, uninterrupted turn. A dialogic task is interactive
and thus necessitates interaction between the
participants and typically results in shorter turns.
Closed versus open In a closed task there is single (or very limited set of )
possible outcomes (i.e. solutions). In an open task there
are a number of possible outcomes. A closed task is
typically an information-gap task whereas an open task
is typically an opinion-gap task.
Convergent versus Opinion-gap tasks can require learners to converge on an
divergent agreed solution to the task or can allow learners to
arrive at their own individual solutions.
Rhetorical mode The task can involve describing, narrating, instructing,
reporting or arguing.

in order for another learner to identify the objects referred to is one-

way, monologic, closed, convergent and descriptive. An opinion-gap
task where learners are given information about four people who need
a heart transplant and have to decide which person will be given the
one heart available is two-way, dialogic, open, potentially divergent
and argumentative.
Another important distinction is between real-world and pedagogic
tasks. The former are based on target tasks and so have situational
authenticity. An example might be a task where two students take on
the roles of hotel receptionist and prospective guest where the latter
has to make a booking for a room based on the information provided
by the former. A pedagogic task lacks situational authenticity but must
still display interactional authenticity (i.e. result in the kind of natural
language use found in the world outside the classroom). An example is
the picture-description task described in the previous paragraph. An
issue of some debate (considered below) is whether a task-based
course should consist only of real-world tasks or whether pedagogic
tasks also have a place.
A task can be input-based, requiring learners to simply process
the oral or written information provided and demonstrate their
understanding of it (for example by drawing a picture or making
a model), or it can be output-based, requiring the learner to speak
or write to achieve the task outcome. This distinction is important
because, as Prabhu (1987) noted, beginner learners cannot be
expected to use the L2 productively so task-based learning must
initially be input-driven.
Tasks can also be unfocused or focused (Ellis, 2003). An unfocused
task is intended to elicit general samples of language. In the early
proposals for TBLT it was generally assumed that tasks would be
unfocused. A focused task must satisfy the general criteria for a task
but is designed to orientate learners to the use of a particular linguistic
feature – typically but not necessarily a grammatical structure. This
possibility was explored in an important article by Loschky and Bley-
Vroman (1993). They suggested that a task could be designed in such
a way that it made the processing of a particular grammatical
structure:
1. ‘natural’ (i.e. the task lends itself, in some natural way, to the
frequent use of the structure (p. 132),
2. ‘useful’ (i.e. the use of the structure is very helpful for performing
the task) or
3. ‘essential’ (i.e. successful performance of the task is only possible if
the structure is used).3

The incorporation of focused tasks into a task-based curriculum need

not result in a return to a structural approach if there is no attempt to
teach the target structure directly, only to create a communicative
context for its use. Some proponents of TBLT (e.g. Skehan, 1998;
Long, 2015), however, favour a curriculum consisting only of
unfocused tasks. Focused tasks, though, have a role in directing
attention at those speciﬁc linguistic features that learners have shown
they have difﬁculty in using accurately. Also focused tasks have been
used frequently in researching tasks.

TASK SELECTION

Long (1985) proposed that the tasks to be included in a course should

be needs-based, that is, the starting point is the target tasks that a
specific group of learners need to ‘function adequately in a particular
target domain’ (p. 91). Once identified these target tasks can be
grouped into task types. The obvious advantage of such an approach
that it ensures the relevance of a task-based course. However, it may
prove very difficult to identify the target task needs of some groups of
learners (e.g. learners in foreign language settings). Cameron (2001),
for example, argued that for young foreign language learners a needs-
based syllabus is not feasible.4 Van Avermaet and Gysen (2006) also
questioned whether any transfer of learning from the performance of
one task to another task of the same type can be expected. It does not
follow, for example, that because learners can ‘buy a railway ticket’
then can also ‘buy an airline ticket’ even though both belong to the
same task type (i.e. ‘buying a ticket’).
Arguably, what is needed for general purpose learners are peda-
gogic tasks that draw on interesting and familiar content. Estaire and
Zanon (1994), in one of the earliest attempts to provide practical
guidance in how to plan a task-based course, suggested that task
selection should be based on ‘themes’, which they classified in terms
of how close or remote these are to the lives of the learners – the
students themselves, their homes, their school, the world around
them and fantasy and imagination.5 They suggested that those
themes closer to their everyday lives would be more appropriate for
beginner-level learners and more remote themes for more advanced
learners. However, there are dangers in materials writers or teachers
deciding what their students will find familiar, relevant or interesting.
Park (2015), for example, reported a marked gap between the topics
that Korean middle school teachers considered ideal and the
topics preferred by their students.

The early TBLT proposals identiﬁed a number of factors that inﬂu-

ence the complexity of a task but gave no guidance as to how these
factors could be applied in the practical business of grading tasks. In
Chapter 3 we will examine what light theories of task complexity and
the research they have generated shed on the problem of grading
tasks. There is, however, little evidence that these theories have had
much influence on the design of task-based courses. Willis and Willis
(2007), for example, offered a list of variables for assessing task
difficulty but then, like Prabhu, concluded that teachers have to rely
on their own intuition. They suggested that teachers will in general
have an idea about whether a particular task is suitable for their
students but that referring to a list of variables can help to sharpen
their intuitions.
However, there have been attempts to develop explicit guidelines for
determining task complexity. Duran and Ramant’s (2006) ‘complexity
scale’ for input-based tasks distinguishes three categories of task
complexity: (1) the world represented in the task, (2) the processing
demands required for task performance and (3) the linguistic input
features. Parameters relating to each of these categories are identified
and arranged on a three-point scale (from simple to complex). For
example, for (1), the parameters are ‘level of abstraction’ (i.e. whether
the topic is concrete or abstract), ‘degree of visual support’ (i.e.
whether visual support is provided and supports task performance)
and ‘linguistic context’ (i.e. whether the linguistic context is available
and supports task performance). There have also been attempts to
investigate the effects of specific variables predicted to influence the
complexity of a task on both learners’ actual performance of a task
and on their subjective appraisal of its difficulty. We will consider this
research in Chapters 3 and 7.
Research may lead to a theory of task complexity that can inform
the grading and sequencing of tasks. However, tasks are conglomer-
ates of variables and complexity is therefore influenced by the inter-
section of countless variables in ways that may make codification
difficult if not impossible. Also, complexity depends on how the task
is implemented (e.g. whether there is opportunity for learners to plan
before they perform the task) as much if not more than on the design
of the workplan. The grading and sequencing tasks remain a major
challenge in TBLT. Perhaps the best that can be done, as Prabhu and
Willis and Willis have suggested, is for teachers and course designers
to rely on their experience and intuition while loosely guided by what
research and theory has shown can affect task complexity.

Methodological Issues
The early proposals had little to say about how a task should be
implemented and, with the exception of Prahbu, even less about how
to plan a task-based lesson. Subsequently, however, greater attention
has been paid to methodological issues in TBLT.

THE TASK-BASED LESSON

In the Communicational Language Project, a task-based lesson con-

sisted of a pre-task, which served as a preparation for a main task of
the same kind. The pre-task was performed in a whole-class context
while the main task was completed by the students working individu-
ally. In other words, there was no small group work. In the pre-task
the teachers guided learners’ performance of the task by simplifying,
repeating and paraphrasing their input to make it comprehensible and,
where necessary, by reformulating the learners’ own attempts to use
the L2 in a target-like way. Prabhu rejected group work on the
grounds that it would expose learners to poor models of English.
Willis (1996) proposed a very different framework for a task-based
lesson, one that prioritized learner–learner interaction. This framework
is shown in outline in Figure 1.1 and an example of a lesson plan based
on it can be found in the Appendix to this chapter. It established the
standard format for a task-based lesson, namely a pre-task stage, a
main-task stage and a post-task stage. Willis prioritized small group
work in the main task phase (called the ‘task cycle’) but allowed for
teacher-centred activity in the pre-task and language focus stages.

Pre-task

Introduction to topic and task

Task cycle

Task → Planning → Report

Language focus

Analysis; Practice

Figure 1.1 Outline of the task-based learning framework

Source: Based on Willis (1996, p. 52).

Willis (1996) advised teachers to ‘stand back and let the learners get
on with the task on their own’ (p. 54) and argued they should resist the
temptation to provide language support or correct learners’ produc-
tion while they are performing a task. She suggested that a concern
for accuracy would arise naturally in the reporting stage of task cycle
and could be addressed directly in the language focus stage. Long
(1991), however, argued that there was a need to draw learners’
attention to form during the performance of a task. He coined the
term ‘focus on form’ to refer to a teaching strategy that ‘overtly draws
students attention to linguistic elements as they arise incidentally in
lessons whose overriding focus is on meaning or communication’
(pp. 45–6).
Long (2015) saw focus on form as essentially reactive but in fact it
can take place both pre-emptively (e.g. when a teacher or student
anticipates the need for a specific linguistic item as they perform the
task) and reactively in response to students’ comprehension or pro-
duction problems. It can also be very implicit, as when the teacher
quickly recasts a learner utterance, or very explicit, as when the
teacher points out an error and corrects it. In other words there are
a variety of strategies available to teachers to attract learners’ attention
to form while they are performing the task (see Ellis, Basturkmen and
Loewen, 2002).
The recognition that task-based teaching does not necessitate an
exclusive focus on meaning but also allows for (indeed requires in the
opinion of many commentators) attention to form during the perform-
ance of a task constitutes one of the major developments in TBLT.
Nevertheless, the belief that teachers should not intervene either pre-
emptively or reactively in a ‘fluency’ activity still holds sway in popular
teacher guides. Hedge (2000), for example, observed that the teacher
notes accompanying course books frequently instruct teachers to
avoid correcting learners until the end of a fluency activity. There is,
however, growing evidence that focus on form facilitates acquisition
(see Ellis, 2015a).
According to Willis (1996), the point of the pre-task stage of a
lesson ‘is not to teach large amounts of new language and certainly
not to teach one particular grammatical structure’ (p. 43). Tomlinson
(2015) took an even stronger stance, arguing against the pre-teaching
of any language on the grounds that it ‘risks changing the task into a
language activity’ (p. 329). These commentators adhere to the general
principle of task-based teaching, namely that there should be no direct
teaching of the language needed to perform a task. However,

opportunities for introducing a focus on form in the pre-task phase are

available. One possibility is to give learners the opportunity to plan
before they perform a task. This will involve them in both conceptual-
izing what they wish to communicate and formulating the language
they will need. Pre-task planning places the burden of working out
how to perform the task squarely on the learner and thus is compatible
with a key principle of TBLT, namely that the learners should be free
to choose from their own linguistic repertoires. See Chapters 3 and 8
for research on planning in TBLT.
The post-task stage offers the clearest opportunities for form-
focused activities including traditional ones. Willis and Willis (2007)
suggested that when the task cycle is complete the teacher is free to
isolate specific linguistic forms for study and work on these forms
outside the context of the communicative activity. Selection of the
linguistic forms for instruction can be based either on the task work-
plan – for example, by identifying specific items from the texts
included in a workplan and preparing activities to practise or develop
awareness of the use of them – or on linguistic features the learners
experienced actual difficulty with when they performed the task.
The methodology of TBLT is now well articulated but there is no
consensus about which methodological procedures are appropriate.
There is a growing consensus that attention to linguistic form is
needed as long as the primary focus remains on meaning. There are
differences in opinion, however, regarding whether a focus on form is
desirable during the performance of the task and also what strategies
should be used to draw attention to form.

Content-Based Language Teaching and TBLT

Content-based instruction (CBI) and content-integrated language
learning (CLIL) share with TBLT the assumption that a language is
best learned when learners are primarily focused on using language. In
CBI and CLIL learners learn language through the process of
mastering the content of (typically) academic subjects (e.g. history,
science, mathematics) and this can include completing subject-relevant
tasks. It might seem, then, that CBI/CLIL and TBLT are just versions
of the same overall approach. Ortega (2015), however, points out that
‘the two fields are pre-occupied with quite distinct issues’ (p. 103).
Table 1.4 summarizes the differences Ortega identified. These differ-
ences are by and large contextual in nature, reflecting the importance
of context and pedagogic purpose in shaping meaning-oriented
approaches to language teaching. However, the differences are histor-
ical, reflecting how the two fields have evolved, rather than

Table 1.4 TBLT and CLT/CLIL compared

Task-based language teaching Content-integrated language learning

Emphasis on college-level learners Mainly implemented with school-level
learners
Easier to implement in second Common in foreign language contexts
language contexts
Experimental research carried out Descriptive research of intact classrooms
in laboratories
Emphasis on transfer of learning Emphasis on demonstrating balanced
from pedagogic tasks to real-life gains in language learning and content
(target) tasks. learning

Source: Based on Ortega (2015, p. 104).

fundamental. For example, there is growing recognition that TBLT is

highly relevant for foreign language contexts and for young children.
Lyster (2007) provides an example of the kind of task that figures in
a CBI. Students were asked ‘to create a continent, identifying its name
and illustrating its geographical features on a map, which they then
presented to their teacher and classmates with a detailed explanation
of how the various geographical features influence the continent’s
overall climatic conditions’ (p. 74). This task illustrates one advantage
that CBI has over TBLT: the choice of topics is determined by the need
to follow the syllabus for a particular academic subject. However,
CBI/CLIL do not rely exclusively on tasks to provide language-rich
content. Teachers may engage in types of classroom interaction
(e.g. initiate-response exchanges) that TBLT is designed to replace.
This reflects the final point in Table 1.4, namely that in CBI/CLIL
content learning is of equal importance to language learning and that
tasks are not the only (or in some cases perhaps not even the best) way
of teaching content.

Technology-Mediated TBLT
One of the major developments in the last thirty or so years has been
the use of technology in language teaching – micro-computers in
particular, but also mobile phones, telecommunication systems and
social media sites. Computer-mediated language learning (CALL)
appeared on the scene in the 1980s at much the same time as the early
proposals for TBLT. While the initial proposals for TBLT had the
face-to-face classroom very much in mind, it was not long before
suggestions appeared for CM task-based teaching. Developments in

CALL mirrored those in language pedagogy in general. There was a

structural/behaviourist phase that gave way to a communicative phase
and ﬁnally to a more integrative stage with the ‘centrality of task-
based authentic learning moving increasingly into the foreground’
(Thomas and Reinders, 2010, p. 6).
Technology-mediated TBLT has a number of advantages. Lai and Li
(2011) emphasized the natural synergy of technology and TBLT:
On the one hand, technology facilitates and enhances TBLT both in terms of
its effectiveness and its contribution to our understanding of TBLT; on the
other hand, TBLT serves as a useful pedagogical framework and set of
principles that can enrich and maximize the use of technology for language
learning. (p. 499)
Technology affords multi-modal opportunities for presenting complex
workplans (aural, written and visual) and for performing them syn-
chronously and/or asynchronously. Appel and Gilabert (2002)
describe a task that involved planning a route and budget for a one-
night trip that required email exchanges, the use of web pages and
synchronous communication. Technology allows the input materials
for a task to be fed into the performance of the task in steps. This is
also possible in the face-to-face classroom but is much easier in a
technologically mediated environment. In short, technology makes
tasks that require complex outcomes possible and it can make rich,
multilayered input available for achieving them. It not only enriches
learners’ opportunities for language learning but also helps to foster
electronic literacy and increase learners’ ability to handle multi-modal
communication.
By and large the model of TBLT presented in the previous sections
of this chapter is premised on a set of more or less disconnected tasks
which provide the basis for individual lessons – as, for example, in the
Communicational Language Project. Ortega (2009) suggested that
technologically driven TBLT should be reconceptualized as project-
based, where there is a series of interlocking tasks relating to the
overall goal of the project. Again, this is possible in a face-to-face
environment – in fact Skehan (1998) proposed just this – but it is
arguably easier to organize with the assistance of technology.
The increasing interest in technology-mediated TBLT is reﬂected in
the growing literature on the subject (e.g. González-Lloret and Ortega,
2015; Thomas and Reinders, 2015) and in the appearance of online
TBLT courses (e.g. Duran and Ramault, 2006). There are also prob-
lems and challenges. Learners may lack the necessary technical skills to
exploit the multi-modal resources made available to them. Teachers
often lack training in how to handle tasks in a technologically

mediated environment while the emphasis on learner-centredness can

leave them uncertain of their own role.

Task-Based Language Assessment

The development of TBLT ran in parallel with ‘a general move away
from discrete-point, indirect testing, and towards more integrated,
direct performance assessments’ (Norris et al., 1998, p. 54) based on
tasks. In fact, though, as Bachman (2002) pointed out, the use of tasks
for assessment purposes had figured in direct language testing for
some time. What was new was the idea of using tasks not as a means
of eliciting learner performances as basis for assessing learners’ general
abilities (i.e. their language proficiency) but for determining whether
they were capable of performing specific target tasks. When tasks are
used to assess L2 general proficiency, the assessor makes a judgement
of the learner’s performance of a task based on a rating scale that
specifies the different abilities being assessed and the level achieved.
Popular tests such as TOEFL (Test of English as a Foreign Language)
and IELTS (International English Language Testing System) assess
proficiency in this way. In task-based language assessment, however,
task performance is assessed in terms of task accomplishment.
The basic principle of task-based assessment was clearly stated by
Long and Norris (2000):
Task-based assessment does not simply utilize the real-world task as a means
for eliciting particular components of the language system, which are then
measured or evaluated; instead the construct of interest is performance of the
task itself. (p. 600)
For Long and Norris – in line with Long’s (1985) views about TBLT –
the tasks used for assessment should reflect target tasks (i.e. real-life
tasks). They proposed using needs analysis to identify the specific
target tasks relevant to a particular group of learners and deriving
authentic assessment tasks from these. Douglas (2000) developed a
framework for analysing target tasks as communicative events with
the aim of achieving a high level of correspondence between the target
task and the assessment task.
There are, however, problems with such an approach (see Bachman,
2002). As we have already pointed out, a needs-based approach is not
appropriate for all learners. Situational authenticity is clearly import-
ant if the purpose of the test is to assess learners’ ability to perform the
tasks in a specific target domain but it is less relevant when the purpose
is to assess the communicative abilities of general purpose learners for
whom there is no clearly defined target domain. For such learners a

more realistic aim is interactional authenticity in the assessment tasks.

However, guaranteeing interactional authenticity is not easy. The very
fact that learners know they are being assessed encourages them to
display what they know rather than to interact in a natural way.
Teachers, however, are more likely to be concerned with formative
rather than summative assessment. Formative assessment is an essential
part of TBLT and involves obtaining information about how learners
perform tasks. The information needed relates to both the product of
the task (i.e. did the students succeed in achieving the outcome of the
task?) and its actual performance (i.e. did the students engage actively
when they performed the task?). Van Gorp and Deygers (2014) pro-
vide a detailed account of a formative assessment of a reading task
designed for primary school students in Belgium. It was based on a set
of key questions that addressed whether (1) the students’ reading of the
task-based material was goal oriented, (2) they could ﬁnd the infor-
mation they were looking for, (3) the teacher could identify and address
any problems the students experienced and, more generally, whether
(4) the students demonstrated self-reliance, positive attitudes to the task
and reﬂective ability. Such a formative assessment can shed light not
just on the students’ abilities and the teacher’s contribution to their
development but also on how the task itself might be improved
for future use. There is a strong case for student self-assessment.
After completing a task, learners can be guided to self-assess their
own performance of it. Butler (2017a) was able to show that not only
are quite young children capable of this but that their self-assessment
correlates well with more objective assessment.
Task-based assessment is discussed in Chapter 9.

Evaluating TBLT
We have seen that TBLT grew out of CLT but developed into a distinct
approach to language teaching. By rejecting the premise that a language
can be taught piecemeal in linear fashion and by proposing instead an
approach catering to the learner’s natural propensity for learning a
language, TBLT can be seen as a radical alternative to traditional forms
of language teaching – what Long (1991a) called ‘focus on forms’.
There is plenty of evidence of the uptake of TBLT. Starting in
2005, there has been a biennial TBLT conference where task-based
educational ideas and research are presented and discussed.
A number of countries have ofﬁcially mandated the use of TBLT.
In 1999 the Education Department of Hong Kong launched
the Target Oriented Curriculum, which was underwritten by a
task-based approach. In Belgium task-based syllabuses and materials

were developed for teaching Dutch both as a first and second lan-
guage at the primary, secondary and adult education levels (see Van
den Branden 2006). The new English curriculum in China does not
specify any particular teaching approach but recommends the use of
task-based teaching as the means for achieving integrated skills
development, problem-solving abilities and cooperative learning
(Wang, 2007). There have also been countless small-scale implemen-
tations of TBLT in contexts where teachers are free to choose their
own approach (see, for example, Leaver and Willis, 2004 and
Edwards and Willis, 2005). TBLT has progressed well beyond theory
into actual practice but it is clearly important to evaluate to what
extent TBLT has been successfully implemented in different instruc-
tional contexts.
There have been a number of evaluations of TBLT programmes.
One of the first was Beretta and Davies’ (1985) evaluation of Prabhu’s
Communicational Teaching Project. This reported results that lent
support to the effectiveness of task-based teaching. Beretta and Davies
concluded that task-based instruction produces significantly different
learning from traditional form-focused instruction. In a follow-up
evaluation, however, Beretta (1990) questioned whether the methodo-
logical innovations required by the project were actually implemented
by the teachers involved. He concluded that the principles and meth-
odology of task-based instruction had not been fully assimilated by the
regular classroom teachers involved in the project.
Later evaluations of TBLT carried out in different teaching contexts
pointed to a number of difficulties in implementing it:
• teachers’ misunderstanding about the nature of a ‘task’
• problems with oral use of the target language in the case of teachers
for the whom the target language was also an L2
• overuse of the L1 by the students when performing tasks
• difficulty in adjusting tasks to the students’ level of proficiency
• difficulty in implementing tasks in large classes
• lack of task-based teaching resources and limited time for teachers
to develop their own resources
• uncertainty about how grammar was to be handled in TBLT
• the need to prepare students for formal examinations
• lack of training in TBLT.
This list paints a bleak picture of the viability of implementing TBLT.
However, many of the same problems are likely to arise whenever
teachers are faced with an innovation of any kind and are addressable
by ensuring that the appropriate conditions for innovation have
been established – in particular through teacher training programmes.

Also, there are cases showing the successful uptake of TBLT.

González-Lloret and Nielson (2015), for example, report a carefully
planned evaluation of a TBLT course for agents in the US Border
Patrol Academy who needed to use Spanish in their daily work. The
students in the task-based course outperformed students in a trad-
itional grammar-based course in terms of ﬂuency and also achieved an
equivalent level of grammatical accuracy. They all passed the
performance-based assessments. The students also reported ﬁnding
the course useful and relevant to their work. In Chapter 10 we exam-
ine a number of experimental studies that have compared TBLT and
other approaches, while in Chapter 11 we look at evaluation studies
that have examined how TBLT has been implemented in a range of
different instructional contexts.

Critiques of TBLT
The advocacy of TBLT has to a large extent been driven from the top
down by teacher educators with a background in applied linguistics, in
particular SLA. For this reason, perhaps, TBLT has met with consider-
able resistance and is the subject of a number of critiques (e.g. Sheen,
1994, 2006; Swan, 2005a). Many of these critiques, however, derive
from a misunderstanding of TBLT (Ellis, 2009a; Long, 2016). For
example, some critics have wrongly assumed that it necessarily involves
learners working in groups to perform speaking tasks. Often critics
have failed to recognize that TBLT is not monolithic but incorporates a
range of possibilities which share the central idea that a language is best
learned through the effort to use it communicatively. The critiques have
also been directed at TBLT for general language teaching and ignore the
obvious suitability of TBLT for specific-purpose language teaching.
However, some criticisms deserve serious consideration. One of the
main criticisms is that there is no evidence that TBLT is more effective
than a traditional focus-on-forms approach. Sheen, in particular, has
argued the need for comparative studies that investigate the relative
effectiveness of the two approaches and attempted such a study himself
(R. Sheen, 2006). Sheen is right in demanding evidence but his own study
was methodologically flawed in several ways and demonstrates the
difficulty in designing comparative method studies. In fact, though, there
is evidence from both evaluation studies and from experimental studies
(e.g. Shintani, 2015) that TBLT can deliver on its promise to foster the
development of both linguistic and communicative competence in an L2
more effectively than traditional ‘focus-on-forms’ instruction.
Another criticism worthy of serious consideration is that TBLT is
incompatible with cultures of learning that are different from those in

Western settings. Littlewood (2014), for example, argued that CLT

(including TBLT) is ill-suited to the traditional Chinese culture of
learning, where ‘education is conceived more as a process of know-
ledge accumulation than as a process of using knowledge for immedi-
ate purposes’ (p. 653) and which therefore emphasizes knowledge
transmission and teacher-centred instruction. Littlewood came out in
favour of task-supported language teaching, where tasks are used to
provide communicative practice for language items taught in accord-
ance with a traditional structural syllabus – in other words, presenta-
tion, practice, production (PPP).
This last criticism leads to an important question. To what extent
should the choice of teaching approach be determined by psycholin-
guistic or cultural factors? To a very considerable extent the advocacy
of TBLT has been based on the former. Opposition to TBLT has been
based on the need to acknowledge the cultural realities of classroom
life. If the goal is to achieve the ability to use an L2 for real-life purposes
then traditional approaches do not have a good record of success. If,
however, the alternative to these approaches – TBLT – proves difﬁcult
to implement, then, it too is unlikely to be successful. There is no easy
resolution to this conundrum except to note that a modular language
curriculum makes room for both a traditional approach and for TBLT.6

Conclusion
We have seen that TBLT grew out disillusionment with the structural
approach. It was informed by CLT and recognition of the need to
develop fluency in an L2, by theory and research in SLA that pointed
to the difficulty of intervening directly in the process of L2 acquisition,
and by educational theories that challenged traditional transmission-
style teaching and emphasized the need for holistic, experiential instruc-
tional activities. From its starting point in the 1980s fully-fledged
proposals for using tasks as the basic unit for teaching and assessment
have been developed and there are now accounts and evaluations of
complete task-based programmes. There are books that detail how
teachers can set about implementing TBLT in their classrooms. Not
surprisingly there are also critiques that have raised a number of issues
relating to both the rationale for TBLT and its implementation.
We conclude with a list of questions arising from the account of
TBLT in this chapter:
1. How should the central unit of task-based teaching – the task – be
defined?

2. What kinds of tasks are appropriate for different groups of

learners? Is a needs-based approach for identifying target tasks
appropriate for all learners?
3. How can the problems of determining the complexity of tasks be
resolved to ensure that learners of different levels of proﬁciency
are faced with tasks that pose a reasonable challenge?
4. How can task-based teaching be made to work for beginner
learners who have no or very little knowledge of the L2?
5. Is there a role for focused as well as unfocused tasks and, if so,
how should focused tasks be incorporated into a task-based
syllabus?
6. Is there merit in a modular curriculum that includes both a task-
based component and a traditional structural component? How
should such a curriculum be organized?
7. What alternatives are there for the organization of a task-based
lesson? Is the lesson format proposed by Willis (1996), which has
proved very inﬂuential, the only way?
8. How can a focus on form be best incorporated into a task-based
lesson?
9. How can teachers carry out formative assessments of task-based
lessons to gather evidence of whether learning is taking place and
what changes may be needed to the task?
10. What problems do teachers face in implementing task-based
teaching and how can these be addressed?
This chapter has offered provisional answers to these questions based
on our own views about TBLT but, as we have also pointed out, there
are alternative positions. These questions are revisited throughout the
book and in particular in the concluding chapter.

Appendix: Example of a task-based lesson plan (based on

material developed by Tom Marchand – see http://willi-elt
.co.uk/wp-content/uploads/2015/03/1StrictParents.pdf

Talking about Families—How Strict Are/Were Your

Parents?
1 Introductory questionnaire:
When you were a child:
a) Do you think your parents were strict or easy-going?
b) Did they allow you to stay out late at night?

c) Did they let you go on holiday on your own?

d) When you went out did you always have to tell them where you
were going?
e) Did you always have to do your homework before supper?
f ) Did your parents make you help about the house?
g) What jobs did they make you do?
h) Did you have to wash the car?
PREPARATION: Teacher makes sure that learners understand the
questionnaire.
TASK: Learners work in groups to answer the questions.
PLANNING: Teacher tells learners that a spokesperson from each
group will be asked to report the results of their discussion to the class
as a whole. Learners are given time to help the spokesperson plan the
report.
REPORT: Spokespersons for two or three of the groups deliver their
reports. The other groups listen and make notes comparing the report
with their own results. Teacher leads a round-up discussion which will
include contributions from groups which did not report.
2 Discussion: Whose parents were the strictest?
TASK: Learners work in groups to decide which of them had the
strictest parents.
PLANNING: Teacher tells learners that a spokesperson from each
group will be asked to report the results of their discussion to the class as
a whole. Learners are given time to help the spokesperson plan the report.
REPORT: Spokespersons for two or three of the groups deliver
their reports. The other groups listen and decide which parents were
the strictest. Teacher leads a round-up discussion which will include
contributions from groups which did not report.
3 Listening: Tim made recordings of some of his friends talking about
how strict their parents were. For example:
My Dad is a quiet man really, so he didn’t really make me do much at home. He
sometimes asked me to wash his car or cut the grass, but I was never forced
to do it, and I could usually get some pocket money for it as well. I think my
Mum was also pretty easy-going; she let me stay out late with my friends. As
long as she knew where I was, she wouldn’t mind so much what I did.
4 Language practice:
For the form-focused work, the ﬁnal stage in a task-based cycle,
activities focusing on expressions of permission and compulsion were
devised.

This section focuses on the theories and research that afford different
perspectives on task-based research. It aims to address the following
questions:
1. What theoretical view of language performance and learning
underlies each of these perspectives?
2. What key theoretical constructs inform the investigation of tasks in
the different perspectives?
3. What research methodology is used to investigate tasks in each
perspective?
4. What differences are there in the way acquisition/language use is
conceptualized and operationalized in these perspectives?
Chapter 2 presents the theory and research related to the cognitive-
interactionist perspective, which was introduced in Chapter 1. It
examines how different kinds of tasks create opportunities for inter-
action that foster the processes involved in second language (L2) and
thereby highlights the importance of social interaction for task-based
language teaching (TBLT). It addresses the role that the negotiation of
meaning and form play in the implementation of tasks and how
negotiation is achieved through interaction, especially when there is
corrective feedback (CF). This chapter also examines to what extent
and how interaction fosters acquisition when tasks are performed. It
concludes with an evaluation of this approach to investigating tasks,
pointing out both its strengths and weaknesses.
Chapter 3 presents theory and research that examine tasks in rela-
tion to the cognitive processes involved in second language (L2) pro-
duction in what we have called the psycholinguistic perspective. The
chapter explores and critiques two models of task-based performance –
the Limited Attention Capacity Hypothesis (LACH) and the Cognition
Hypothesis (CH) – which have informed a large body of research. The
chapter reviews studies that investigated how task design and

implementation variables impact on task performance in terms of the

complexity, accuracy, lexis and fluency of the learners’ production.
The chapter also considers a key issue for TBLT, namely the relation-
ship between task performance and L2 acquisition.
Chapter 4 offers a sociocultural perspective on tasks. In sociocul-
tural theory a ‘task’ is viewed an artefact for mediating learning
through interaction. It views ‘task’ as always interpreted by the par-
ticipants so that what the task is intended to achieve (i.e. the task-as-
workplan) may well not match what the task actually achieves when it
is performed (i.e. the task-as-process). Like the cognitive-interactionist
perspective, the sociocultural perspective views tasks in terms of the
interactions to which they give rise, emphasizing the importance of the
collaborative nature of the interaction for ‘learning’ (defined as other-
regulation) and for ‘development’ (defined as self-regulation). It
reviews a range of research that has investigated tasks from a socio-
cultural perspective, including studies involving ‘languaging’, dynamic
assessment and concept-based language instruction.
Individual learner factors play an important role in how a task is
performed. This requires a perspective that draws on the theory and
research that addresses the psychology of the learner – what we call
the psychological perspective on TBLT. Chapter 5 surveys the large
body of research on the role of cognitive aptitudes (including working
memory) in mediating the effects of different instructional tasks on
language performance and acquisition. The chapter will also discuss
the influence of affective factors such as motivation and language
anxiety on task performance and outcome. A key focus of this chapter
is how these psychological variables mediate the performance of a task
and the learning that results.
In Chapter 6, the final chapter in this section, we adopt and educa-
tional perspective on TBLT. The chapter begins by summarizing gen-
eral educational theories that support an approach to learning that
emphasizes experience and ‘doing’ over knowing and ‘telling’ – such
as those of Dewey (1938) and recent work on complex skill acquisi-
tion and training. It then considers research that draws on educational
accounts of the role of ‘engagement’ in task performance and learning
and the importance of investigating learners’ perceptions of the tasks
they perform as well as their actual performance.

Introduction
According to the Interaction Hypothesis (IH) ‘negotiation for mean-
ing, and especially negotiation work that triggers interactional
adjustments by the NS or more competent interlocutor, facilitates
acquisition because it connects input, internal learner capacities,
particularly selective attention, and output in productive ways’
(Long, 1996, pp. 451–2). In this way, Long captured the symbiotic
relationship between interaction and cognition; that is, interaction
activates the mental mechanisms involved in processing input and
output in ways that result in acquisition. The IH was subsequently
broadened into what Gass and Mackey (2007) called the interaction
approach, which concerns what happens ‘when learners encounter
input, are involved in interaction, receive feedback and produce
output’ (p. 176). According to this perspective, tasks will prove
effective to the extent to which they provide input and promote
interaction of the kinds that activate the mental mechanisms involved
in acquisition.
We begin this chapter with the ‘cognitive side’ by examining the
role of attention in second language (L2) acquisition and implicit/
incidental acquisition, which proponents of task-based language
teaching (TBLT) such as Long (2015) and Ellis (2003) see as central
cognitive processes. This is followed by an account of interaction
and how researchers have analysed it. The next sections, which
constitute the core of this chapter, consider how cognitive-
interactionist perspectives have informed the design and implementa-
tion of tasks, how tasks induce noticing, how the interactions
they afford result in acquisition, and how interaction takes place in
small group work. In the concluding section of the chapter we consider
the strengths and limitations of cognitive-interactionist accounts
of TBLT.

Cognitive Processes: Implicit Learning, Incidental Learning

and the Role of Attention
For Long (2015), the essential empirical problem in second language
acquisition (SLA) that needs an explanation is why children are
entirely successfully in acquiring their mother tongue implicitly while
adults are largely unsuccessful. According to Long (1990, 2015), L2
acquisition is maturationally constrained because adults no longer
have full access to the mechanisms involved in implicit learning and
thus are ‘partially disabled language learners’ (Long, 2015, p. 41).
Long maintains, however, that ‘implicit learning is still the default
learning mechanism’ (p. 43) for adults and cites as evidence studies
(e.g. Rebuschat and Williams, 2009; Leung and Williams, 2011) that
demonstrate implicit learning is still possible in adults for some aspects
of grammar such as simple form-meaning mappings and basic word
order. Long also recognizes the need for explicit learning to help
overcome the limitations of adult implicit learning. However, he
rejects the view that explicit, intentional learning constitutes an alter-
native to implicit learning as claimed by skill-learning theory and as
manifested in the language teaching approaches based on that theory.
For Long, explicit learning has a more limited role – to prompt initial
perception of L2 forms in the input and thereby to help learners to
overcome entrenched ﬁrst language (L1) routines. That is, he sees the
explicit knowledge that results from intentional learning as changing
how learners attend to input and as tuning the implicit processes
involved in acquisition (N. Ellis, 2005). In other words, TBLT, when
supported by strategies that focus learners’ attention on form, pro-
motes implicit learning.
Cognitive models of L2 acquisition also draw heavily on Schmidt’s
(1990, 1994) views about the importance of attention in L2 acquisi-
tion. Schmidt distinguishes two levels of attention – a low level which
he calls ‘noticing’ and a deeper level involving ‘understanding’. For
example, a learner may ‘notice’ the plural-s on a noun and register that
this signals ‘more than one’ or the learner may go a step further and
consciously construct a rule for expressing plurality (i.e. add ‘s’ to a
noun to signal ‘more than one’). According to this view, completely
implicit learning is not possible as ‘noticing’ involves consciousness
which is a prerequisite for acquisition to take place. However, Schmidt
(2001) later modiﬁed this position. While continuing to emphasize the
importance of ‘noticing’, he acknowledged that detection or what
Gass (1988) called apperception can take place subconsciously as
claimed by Tomlin and Villa (1994). Williams (2013, p. 39) similarly
claimed that ‘whilst attention does appear to be necessary for learning,

explicit learning explicit knowledge

(1)
interaction noticing implicit knowledge

(2)
implicit learning

Figure 2.1 Cognitive-interactionist model informing TBLT

awareness might not be’. Schmidt was ambivalent about the role of
‘understanding’, viewing it as potentially helpful for acquisition but
not necessary.
Drawing on these key constructs, we can outline the cognitive-
interactionist model that underscores TBLT (see Figure 2.1). Learners
are exposed to input through interaction. Implicit learning can occur
when learners learn without conscious attention to linguistic forms in
the input (i.e. there is an absence of ‘noticing’) and this results in
implicit knowledge – the kind of knowledge required for easy and
fluent communicative language use. However, adult L2 learners are
limited in their ability to learn in this way but, fortunately, there is
another route to implicit knowledge which can compensate for their
reduced capacity for implicit learning.1 This involves the explicit
knowledge they have gained from intentional language learning,
which serves as an activator of noticing and, in this way, facilitates
the development of implicit knowledge. In Figure 2.1 the line linking
explicit knowledge and ‘noticing’ (i.e. the conscious awareness of
linguistic features) is dotted. This is intended to show that this link is
variable, depending on both external factors such as interactive strat-
egies that attract attention to form and internal factors such as
learners’ propensity and ability to attend to form. The model posits
no direct relationship between explicit and implicit knowledge. In
other words, there is no substitute for implicit learning or noticing as
the means for achieving implicit knowledge of a language.
There remain a number of issues that require clarification. First,
implicit and incidental learning are often treated as synonymous but
they are not.2 Whereas the latter is defined as learning without con-
sciousness and involves only detection, the former involves ‘noticing’,
i.e. focal attention and conscious awareness of specific linguistic forms.
The route labelled (1) in Figure 2.1 represents incidental learning and
the route labelled (2) implicit learning. Incidental acquisition can occur
in all kinds of instruction, including explicit instruction. For example,

learners may incidentally acquire features y and z even when the

instruction is explicitly directed at feature x. This possibility has been
little investigated, however, and is perhaps unlikely given that the focal
attention directed at the target structure may inhibit attention to other
features. Loewen, Erlam and Ellis (2009), for example, found that
learners failed to acquire third person-s incidentally when the explicit
instruction directed attention at the indeﬁnite article. Incidental acquisi-
tion is more likely when learners are primarily focused on meaning-
making and take time out to attend to form when occasion calls for this.
This is what TBLT caters to through various strategies that attract
learners’ attention to form while they are communicating (as discussed
later in this chapter). TBLT aims to foster implicit learning through the
apperception that takes place as learners engage in the communication
needed to achieve the task outcome. However, for adults at least, it is
incidental rather than implicit learning that is central in TBLT.
A second issue concerns the role of output in cognitive-interactionist
theories. Long tends to emphasize input but also acknowledges the role
of pushed output in facilitating noticing. Skehan (1998), drawing on
and extending Swain’s (1985, 1995) arguments in support of compre-
hensible output, lists six roles for production: (1) It serves to generate
better input through the feedback that learners’ efforts at production
elicit. (2) It forces syntactic processing (i.e. it obliges learners to pay
attention to grammar). (3) It allows learners to test out hypotheses
about the target language grammar. (4) It helps to develop automaticity
of existing L2 knowledge. (5) It provides opportunities for learners to
develop discourse skills, for example by producing ‘long turns’. (6) It
helps learners to develop a ‘personal voice’ by steering conversations
onto topics they are interested in contributing to. We would add that
production also provides the learner with ‘auto-input’ (Schmidt and
Frota, 1986) as learners can beneﬁt from the ‘input’ that their own
output provides them with. Output is part of interaction but it can also
occur when learners are not interacting. In Chapter 3 we examine in
detail how learner production contributes to L2 development.
The third issue is controversial. It concerns the nature of the ‘explicit
learning’ in Figure 2.1. For Long (2015), explicit learning is of value
when it is embedded in the communicative interactions that result
from the performance of a task. That is, it involves only ‘brief episodes
of selective learner attention to critical segments of input (focus on
form)’ (p. 53). In other words, Long rejects providing for explicit
learning separately. Ellis (1994, 2018a), however, suggests that the
knowledge that learners gain from explicit language lessons facilitates
the ‘noticing’ of linguistic forms in the input – as indicated in
Figure 2.1 – and is also of value for monitoring output. In other

words, Ellis argues that an explicit focus on form does not always
have to be contiguous with the performance of a task. While ‘brief
episodes of selective attention’ (Long, 2015) may be desirable,3 some
explicit lessons may help learners to acquire those linguistic ‘fragile’
features which are often not learned even when focus on form accom-
panies the performance of a task. From this perspective the ideal
is a modular curriculum involving a primary task-based component
with a secondary structural component, a possibility discussed in
Chapter 7.

Analysing Interaction
Interaction occurs when two or more people engage in communica-
tion. Prototypically it occurs face to face but there is growing interest
in the interaction that arises when tasks are technologically mediated
(e.g. González-Lloret and Ortega, 2015; Granena, 2016). Interaction
can be two-way (i.e. all the participants contribute actively) or one-
way (i.e. one person does all the speaking and the other(s) just listens
as in a lecture). In two-way communication, learners have the oppor-
tunity to both receive input and produce output. In one-way commu-
nication, the speaker produces output and the other(s) receives input.
This is an important distinction for task-based instruction because it
underscores a key difference in tasks; as noted in Chapter 1, output-
based tasks aim to provide opportunities for two-way interaction
whereas input-based tasks are essentially one-way.4
However, the input that arises when an input-based task is per-
formed in an interactive situation is not fixed.5 Speakers adjust their
choice of language in accordance with their assessment of the listeners’
abilities to comprehend. In other words, input is continuously modi-
fied; often it is simplified but sometimes it can be elaborated, which
Long and Ross (1993) suggest is more facilitative of acquisition. There
is a rich literature documenting the characteristics of the ‘foreigner
talk’ that occurs when native speakers talk to L2 learners (see Gass,
1997; Ellis, 2008) and of the ‘teacher talk’ found in classrooms (Henzl,
1979). Such talk helps to provide learners with the comprehensible
input that Krashen (1985) argued is essential for acquisition. When
teachers perform input-based tasks, they naturally modify their speech
to ensure comprehension. Shintani (2012), for example, showed how
repeating the same input-based tasks with young L2 learners resulted
in changes in the teacher’s input. The teacher gradually reduced her
use of the L1 (Japanese) while increasing the length of her utterances
by elaborating the commands she gave to the children. If it is a learner
who performs a one-way task (e.g. when a learner is describing where

to locate places on a map), the resulting output is monologic. Never-

theless, when this takes place in an interactive context, learners will
still need to make efforts to ensure they are comprehensible and this
involves discourse management (i.e. deciding how much information
to provide, checking comprehension, repeating, paraphrasing etc.).
In the classroom context, discourse management is evident in the
pre-emptive strategies that both teachers and learners use to anticipate
and prevent problems arising during interaction. Teachers, for
example, pre-empt by asking questions (e.g. Do you know what
‘economy’ means?) or by warning learners to take care (e.g. Remem-
ber – you need to use the past tense when telling the story). Learners
too pre-empt by asking questions (e.g. ‘Do I need past or present tense
here?’). Some researchers (e.g. Long, 2015) have argued that what is
important is the discourse repair work that activates the internal
mechanisms involved in acquisition. That is, interaction works for
acquisition when interlocutors react to problems – communicative or
linguistic – that arise as a task is performed. Other researchers (e.g.
Ellis, 2017a), however, have argued that interaction involving pre-
emptive moves aimed at preventing problems is also facilitative of
acquisition. Arguably, both reactive and pre-emptive interactive strat-
egies can activate the cognitive processes involved in acquisition.
In accordance with the IH, research on two-way tasks has focused
on the discourse repair that occurs when communication problems
arise and negotiation takes place. Varonis and Gass (1985) developed
a model of ‘non-understanding routines’ (see Figure 2.2 and the
example below). This distinguishes the turn that triggers non-
understanding and the subsequent turns where there is an attempt to
resolve the problem. This model allows for the identification of specific
discourse strategies for resolving communication problems. Foremost
among these strategies is corrective feedback (CF).
Example:
S1: Einstein’s scientific work helped Americans make the nuclear bomb. (T)
S2: Clear bomb? (I)
S1: No nuclear, nuclear, nuclear bomb. (R)
S2: Nuclear bomb. I see. (RR)
(Aubrey, unpublished data)
Table 2.1, based on Lyster and Ranta (1997), defines the principal
types of indicator moves, along with comments on whether they are
input-providing or output-prompting and whether they are implicit or
explicit in nature, although this latter distinction is not clear cut as it
depends more on context and on how an indicator is delivered than on
the type of indicator – an important point which we consider later in

Trigger T = trigger (i.e. the utterance which causes misunderstanding)

↓
I = indicator (i.e. of misunderstanding)

Resolution R = response

RR = reaction to response

Figure 2.2 Model of non-understanding routines

Source: Varonis and Gass (1985)

this chapter. These indicators also differ in the nature of the problem
they are typically used to tackle. Some signal that there is a communi-
cation problem and thus relate to the negotiation of meaning while
others just signal that the problem is linguistic in nature and thus
involve the negotiation of form (Lyster, 2001). The indicator types
constitute the different ways of conducting CF and doing ‘focus on
form’, a major area of interest in the interaction approach. They
connect with cognitive mechanisms in different ways and have differ-
ent implications for language acquisition – see Table 2.1.
Not all indicators require a response. Input-providing indicators
such as recasts place no obligation on the addressee to respond. In
contrast, output-prompting indicators do require a response in accord-
ance with Grice’s (1975) Cooperative Principle, but even in this case
learners may sometimes opt not to respond if, for example, their
linguistic resources prevent them from doing so. The ﬁrst response
option, therefore, is no response. Other indicators – a conﬁrmation
check for example – require no more than an acknowledgement. This
often takes the form of simply saying ‘yes’ followed by a topic-
continuing move, as in this example:

S: I was in pub
(2.0).
S: I was in pub.
T: In the pub?
S: Yeah and I was drinking beer with my friend.
(Ellis, Basturkmen and Loewen, 2001).

Table 2.1 Indicator types in corrective feedback

Type Deﬁnition Example Comment

Repetition The addressee repeats the L1: I felt really chuffed with the This is implicit and output-prompting.
speaker’s utterance to signal results of my exam. Lyster (1998) pointed out that repetitions
that there is a comprehension L2: Chuffed. are ambiguous as they can also signal
problem L1: Yes, I did better than understanding. Thus, they may lead to a
I expected. resolution of the problem or they may
not.
Confirmation Any expressions ‘immediately Learner 1: Ok it’s in the it’s in Like a repetition, a confirmation check
check following an utterance by the the corner the building simply repeats the problematic utterance
interlocutor which are Learner 2: In the corner? or part of an utterance but the rising
designed to elicit confirmation Learner 1: yeah intonation more clearly signals there is a
that the utterance has been (Gilabert, Baron and Llanes, problem (i.e. makes the move more
correctly heard or understood 2009, p. 377) explicit) and thus may be more likely to
by the speaker’ (Long, 1983, lead to a successful resolution.
p. 137)
Clarification Any expression that elicits Learner 1: Go walking it’s two Clarification requests are output-
request clarification of the preceding apples further two streets prompting. They place the burden of
utterance more it looks. resolving the problem on the speaker who
Learner 2: Two what? created it. They are often viewed as
Learner 1: Two streets further. implicit as they occur naturally in
(Gilabert et al., 2009, p. 376) everyday conversation but they are in
fact quite explicit in signalling that there
is a communication problem.
Metalinguistic A move that provides a S: There are influence person Metalinguistic clues are output-prompting
clue comment or questions some who (i.e. they do not provide the learner with
aspect of the preceding T: Influential is an adjective the remedy of the problem) and they are
S: Influential person clearly very explicit as they respond to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.006
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:31, subject to the

utterance, signalling a (unintelligible) because of his form of the preceding utterance rather
linguistic problem power. than its meaning. This type of indicator is
(Sheen, 2004, p. 278). more likely to be used by a teacher than
by a learner.
Explicit A move that indicates an S1: And three pear (sounds like This type of indicator is explicit and input-
correction utterance is problematic and ‘beer’). providing. Like metalinguistic clues it is
at the same time provides the S2: Three beer. more likely to be performed by a teacher
solution to the problem T: Not beer. Pear. although learners have also been
observed to correct each other explicitly.
Elicitation A move aimed at extracting the An elicitation can take the form It is output-prompting and explicitly
correct linguistic form from a of a question (e.g. How do we corrective. It negotiates form rather than
speaker. say x in English?), a statement meaning and is used more or less
requiring completion (e.g. exclusively by teachers.
You __) or a request to
reformulate (e.g. Can you say
it another way?) – Sheen
(2004, p. 278).
Recast An utterance that rephrases an S: I stand in the first row. Recasts are input-providing and are
utterance ‘by changing one or T: You stood in the first row. generally considered implicit. However,
more of its sentence S: Yes. they can also be made more explicit,
components (subject, verb or (Y. Sheen, 2006, p. 35) especially if intonation is used to
object) while still referring to highlight the part of the utterance that
its central meanings’ – Long, has been reformulated.
1996, p. 436). Various types
and characteristics of recasts
have been identified (e.g.
partial; versus full) – see
Y. Sheen (2006).

Source: Based on Lyster and Ranta (1997).

38 Theoretical Perspectives

Learners may respond to an indicator such as a recast by echoing it or

by modifying their initial output in what is called uptake. Lyster and
Ranta (1997) pointed that uptake can be of two kinds depending on
whether the problem is or is not repaired. In uptake-with-repair, the
learner either repeats the indicator move if this is input-providing, as
in this example:

S: I got up late today morning.

T: Today morning? This morning.
S: This morning.

Or, if the indicator consists of a prompt, the learner may self-repair the
utterance that triggered the negotiation. Uptake-with-repair constitutes
one kind of modiﬁed output. However, this construct is broader as it
includes occasions when learners attempt to modify output without
being prompted to do so by feedback. Sometimes, as in the example that
follows, the learner may attempt to repair following feedback but fail to
do so, resulting in what Lyster and Ranta called uptake-needs-repair.

S: I have an ali [bi].

T: You have what?
S: an ali [bi] (i.e. S continues to mispronounce ‘alibi’).

Uptake-needs-repair can involve a complete failure to address the

problem or, in some cases, partial repair (i.e. the learner corrects part
of an erroneous utterance).
The various strategies involved in discourse management and repair
constitute the means for conducting focus on form while a task is
being performed (see Ellis et al., 2001). Focus on form is a necessary
feature of Long’s (2015) deﬁnition of TBLT. It constitutes the means
for drawing learners’ attention to linguistic features when problems
arise in the communication resulting from the performance of a task.
The linguistic problems that are addressed can be wide-ranging and
unpredictable or they can be pre-determined and therefore predictable.
This will depend on whether the task is unfocused or focused – a
distinction introduced in Chapter 1 (see p. 12). If the task is unfocused
the linguistic features addressed will be whatever happened to cause a
problem. The resulting focus on form will be extensive (i.e. many
different forms will be addressed). There is evidence to show that
problems relating to vocabulary receive more attention than grammat-
ical problems when the negotiation involves meaning (Pica, 1996)
because the wrong or mispronounced word is more likely to lead to
communication breakdown than a missing grammatical morpheme.

However, when it is form (rather than meaning) that is negotiated all

forms – phonological, lexical and grammatical – are likely to receive
attention. If the task is focused, the focus on form will be directed at
whatever is the target feature of the task with the same feature
addressed repeatedly (i.e. it will be intensive). As we noted in
Chapter 1, some proponents of TBLT favour unfocused tasks. How-
ever, many researchers investigating the effects of CF (e.g. Lyster,
2004; Ellis, Loewen and Erlam, 2006) have elected to investigate
focused tasks because these make pre-testing and post-testing of the
targeted feature possible and thus allow for the effect that focus on
form has on acquisition to be investigated.
Early research in the cognitive-interactionist paradigm focused on the
negotiation of meaning and reactive focus on form but increasingly
researchers have broadened the frame of reference to examine how
learners and their interlocutors attend to form when purely linguistic
problems arise. In particular, language-related episodes (LREs),6
defined as ‘any part of dialogue where the students talk about the
language they are producing, question their language use, or correct
themselves or others’ (Swain and Lapkin, 1998, p. 326), have proved a
popular way of investigating task-based interaction. Plonsky and Kim
(2016), for example, found that 25 per cent of all the interactional
features examined in the studies included in their meta-analysis were
LREs. Studies examining tasks from a sociocultural perspective, which
have increased in number over time, favour LREs. These studies will be
considered in detail in Chapter 4. There have also been attempts to
investigate how tasks impact on other aspects of interaction. Gilabert
and Baron (2013), for example, examined how task type and task
complexity affected learners’ use of pragmatic features (requests and
suggestions). Sato (2017) focused on ‘collaborative sentence comple-
tion’, where one learner helps another learner to complete an utterance.
Task-based interaction studies are of two basic kinds. There are
what Plonsky and Kim (2016) call ‘task-based learner performance
studies’ (p. 74). These are studies that manipulate various dimensions
of task design and implementation conditions in order to investigate
what effect they have on interaction. This type of study was dominant
in early research and is viewed by Plonsky and Gass (2011) as the first
phase of interactionist research. The second kind is ‘task-as-treatment
studies’, where the aim is to investigate what effect a particular task
design or implementation condition has on either the processes
involved in acquisition (e.g. noticing) or on L2 acquisition. These
studies figure in the second phase of interactionist task-based research.
We will begin by looking at task-based performance studies and then
move on to task-as-treatment studies.

Task-Based Learner Performance Studies

A major research strand in the Interaction Approach has focused on
what kinds of tasks are most likely to result in the types of interaction
that foster language acquisition. Much of the early research, reviewed
in Ellis (2003, 2012) examined how different design features and
implementation strategies impacted on the negotiation of meaning.
Table 2.2 shows the variables investigated.
This research typically involved laboratory-based studies so the
findings may not be applicable to classroom contexts (see Ellis, 2012)
but it provides a clear indication that task design and implementation
influence the level and kinds of interaction that take place. Regarding
task design variables, Ellis concluded that negotiation of meaning was
more likely to occur with tasks that required information exchange (i.e.
information-gap tasks) than with tasks where information exchange
was optional (i.e. opinion-gap tasks ) (Long, 1980; Foster, 1998) and
also more likely when a task was two-way (i.e. the information to be
exchanged was split among participants) than when it was one-way
(i.e. one learner held all the information to be exchanged) (Long,
1989). Closed tasks also led to more negotiation of meaning than
open tasks (Long, 1989). Learners also negotiate more in tasks with
unfamiliar topics than in those with familiar topics (Gass and Varonis,
1984) and more in tasks with human-ethical content than in tasks
with objective-spatial content (Berwick, 1990). Regarding task
implementation variables, the participant role, task repetition, and
interlocutor proficiency were all found to impact on the level of

Table 2.2 Task design and implementation variables investigated in

interaction studies

Design variables Implementation variables

1. required vs. optional information 1. participant role (e.g. listener vs. active
exchange participant)
2. information gap: one-way vs. two- 2. task repetition
way 3. interlocutor familiarity (i.e.
3. task outcome: open vs. participants familiar with each other
closed tasks vs. not familiar)
4. topic (e.g. topic familiarity) 4. interlocutor proﬁciency
5. discourse mode (e.g. narrative vs.
description)
6. cognitive complexity (e.g. context-
embedded vs. context-reduced)

negotiation. Learners negotiate more when their role requires them to

speak as well as listen (Gass and Varonis, 1994) in the first perform-
ance of a task than in a repeated performance (Gass and Varonis,
1985), and when the less proficient learner is put in charge of the
interaction (Yule and McDonald, 1990). Pica, Kanagy and Faludan
(1993) attempted to consolidate the findings of these studies into a
general framework that could account for the differential effect that
tasks have on interaction.
More recently, researchers have drawn on Robinson’s (2007b) Cogni-
tion Hypothesis (CH) to investigate how task complexity affects inter-
action, measured in terms of LREs. Robinson argued that more complex
tasks result in more acquisition-rich interaction than less complex tasks.
[7] Task complexity was operationalized in terms of the presence or
absence of specific task features (e.g. +/ there and now; +/ reasoning;
+/ few elements) with simpler tasks being those with - features. A number
of studies (e.g. Gilabert et al., 2009; Kim, 2009; Révész, 2011; Baralt,
2014; Kim and Taguchi, 2015) lend support to Robinson’s claim. For
example, more complex tasks lead to more LREs than simple tasks.
Solon, Long and Gurzynska-Weiss (2017) is a good example of this
kind of study but it also illustrates the danger of over-generalizing the
effect of task complexity on interaction. They investigated seventeen
dyads of intermediate L2 Spanish proficiency performing focused tasks
that were designed to contextualize the pronunciation of Spanish
phonemes built into street names (e.g. Calle Copa vs. Calle Capa).
The tasks were designed to differ in complexity according to the number
of elements they contained. To establish whether this task feature did in
fact distinguish the complexity of the tasks, the actual time and the
learners’ retrospective estimate of the time they had spent on the tasks
were recorded. The learners performed the map tasks in pairs; one
learner held a version of the map showing the route but minus place
names and the other a version with the place names but minus the route.
LREs were identified in all the interactions. Overall, there were more
LREs in the complex task, as predicted. However, pronunciation-
focused LREs were highest in the simple task. Solon et al. (2017)
suggested that whereas learners may be used to consciously reflecting
on grammar they lack experience in reflecting on pronunciation features
and thus are less likely to engage in LREs focused on pronunciation. In
other words, the extent to which task design affects LREs may differ
according to the linguistic features being investigated.
Somewhat surprisingly, there has been very little research that has
investigated tasks in relation to pragmatic aspects of language use.
Gilabert and Baron (2013) compared university-level English as a
Foreign Language (EFL) learners’ performance of two tasks which

differed in type (one was a two-way split information task with a

convergent goal and the other a two-way shared information task
with a divergent goal) and in complexity (there was a simple and
complex form of both tasks). They found that the divergent task
produced more overall moves and also a greater variety of moves.
The complexity of the tasks affected the number of overall moves but
had no effect on the variety of moves. Given that pragmatic
competence is a key feature of language proficiency, there is an obvi-
ous need for more studies that shed light on how task selection and
task design impact pragmatic aspects of language production.
There has been increasing interest in how computer-mediated (CM)
tasks affect interaction (see, for example, Ziegler, 2016). Studies have
sought to compare the similarities and differences between tasks per-
formed face to face and in a CM context. Ziegler, drawing in particu-
lar on Smith (2003), points out that one difference is that whereas the
response to a communication problem usually occurs immediately
after the trigger in face-to-face interaction, it is often delayed in CM
tasks, especially if these involve text chat. There are also differences in
the effect that different CM modalities (i.e. text chat, audio, or video)
have on negotiation. For example, Jepson (2015) found that repair
moves were more frequent when tasks were performed in voice than in
text chat. Researchers have also been interested in whether the type of
task affects negotiation in CM interaction. Blake (2000) reported that
one-way tasks where the information was split resulted in more nego-
tiation than decision-making tasks where the information was shared,
thus replicating one of the main findings for face-to-face tasks. Ziegler
calls for more research investigating all the issues.
These studies afford a general picture of how task variables can
impact on interaction. But they need to be viewed with circumspection.
Each task involves a cluster of variables that are likely to interact in the
effect they have on interaction. Studies have typically manipulated task
types in terms of pairs of variables (e.g. one-way versus two-way;
familiar versus unfamiliar topics; closed versus open; +/ reasoning;
+/ here and now; +/ number of elements) but the tasks used in these
studies inevitably involved a cluster of variables. This makes it difficult
to generalize the research findings. We cannot be sure, for example, that
unfamiliar tasks will lead to more negotiation of meaning than familiar
tasks in all cases as other variables – such as the number of elements
involved in the task – may counteract this general tendency. Further-
more, how a task is implemented will also affect the interaction that
results. A one-way task may not result in less negotiation than a two-
way task if the learners are instructed to interact actively and if
the information to be exchanged is held by the less proficient learner.

The combinations of design and implementation variables are multitu-

dinous. Thus, while it may be possible to design studies that investigate
how two or three variables impact jointly on interaction, taking into
account all the potentially inﬂuential variables that make up a task will
prove impossible. At best, then, the research provides only clues as to
how task design and implementation can affect interaction.

Task-as-Treatment Studies
When tasks are used as the means for intervening in the process of L2
acquisition researchers have addressed three key questions: (1) What
do learners pay attention to (i.e. notice) when they perform a task?
(2) What is the relationship between noticing and learning? (3) What
learning occurs when interactive tasks are performed? We will con-
sider studies that have investigated these three questions.

Tasks and Noticing

Researchers interested in whether noticing occurs when a task is
performed have investigated both pre-modified input (i.e. input that
has been specially designed to facilitate comprehension) and interac-
tionally modified input (i.e. input that is modified when a learner
signals a comprehension problem).

NOTICING IN PRE-MODIFIED INPUT

In the case of pre-modiﬁed input, efforts are made to draw learners’

attention to specific linguistic features either through (1) ‘flooding’ the
input with exemplars of a specific feature, or (2) highlighting a specific
form in the input. By and large input enhancement of this kind has
been undertaken with written input in reading tasks but it is also
possible with oral tasks. One obvious way in which this can happen
is by repeating the key lexical or grammatical items, using intonation
and stress to highlight them, or by providing additional clues, such as
gesture, to help learners decode their meanings. The input modifica-
tions found in pre-modified oral input are very similar to features
found in teacher-talk (Ellis, 2015b). Overall, the research shows that
input enhancement does facilitate noticing but, as Lee and Huang’s
(2008) meta-analysis showed, its effect is often quite limited. Han,
Park and Combs (2008, p. 600) noted that ‘there are numerous
methodological idiosyncrasies characterizing the individual studies’ –
such as the number of times a specific feature was highlighted, the
number of texts involved, and whether learners received explicit

instruction on the targeted feature(s) prior to exposure to the input –

all of which can impact on noticing.

NOTICING IN INTERACTIONALLY MODIFIED INPUT

Noticing is arguably more likely to occur in the pre-emptive and

reactive focus on form episodes that arise when interactive tasks are
performed than in pre-modiﬁed input. One measure of whether
noticing occurs in such episodes is whether there is learner uptake.
Ellis et al. (2001) investigated the frequency of focus-on-form episodes
(FFEs) when experienced teachers performed communicative tasks
with adult English as a second language (ESL) learners, reporting that
there was an FFE approximately every one and a half minutes. They
investigated whether the different kinds of FFEs resulted in uptake and
whether this uptake was ‘successful’ (e.g. resulted in learners repairing
their own errors).8 Learner-initiated pre-emptive focus on form – for
example, when a learner posed an explicit question about a linguistic
form – resulted in a high level of successful uptake. Reactive focus on
form (where the teacher responded to a learner utterance containing
an error) also regularly led to successful uptake. In contrast, teacher-
initiated pre-emptive focus on form , where it was the teacher who
made a linguistic form the topic of the interaction, was much less likely
to result in successful uptake. This study suggests that noticing can
occur frequently in task-based interactions but that this may depend
on what kind of focus on form learners experience.
Other noticing studies have focused solely on reactive focus on
form – in particular, recasts. Recasts, which juxtapose a learner
utterance containing an error with a target-like reformulation of the
utterance, provide what Long (1996) considered the ideal context for
the learner to not just notice the target form but also to compare their
erroneous form with it – what Schmidt and Frota (1986), called
noticing-the-gap. In the exchange that follows, for example, the learner
overgeneralizes the use of the regular past tense (doed) – a common
error – and the teacher immediately recasts the erroneous part of the
utterance. The learner uptakes the correction, repairing the error.
The learner then continues, making another tense error, but this time
he corrects himself without the teacher’s intervention.

L: When he 18 years old he m– if he doed it.

T: Did it.
L: Uh did it.
T: Yeah.
L: Must go to the prison? went to the prison.

Prompts can also result in noticing by inducing learners to attend to

the linguistic problems in their utterances. Prompts in fact are more
likely to result in uptake-with-repair than recasts as they require a
response from the learner (Lyster and Ranta, 1997).
To investigate whether noticing occurs following CF researchers
have either investigated uptake-with-repair as in Ellis, Basturkmen
and Loewen (2001) or elicited retrospective self-report from learners,
for example using stimulated recall (Gass and Mackey, 2000). An
example of the ﬁrst approach is Ellis and Mifka-Provozic (2013). They
reported that 84.5 per cent of the recasts were followed by uptake-
with-repair, suggesting that in the context of a foreign language class-
room recasts were highly salient to the learners and that the teacher’s
corrections were consistently noticed. However, as Sheen (2004)
showed, the level of uptake-with-repair following recasts varies con-
siderably depending on the instructional context. It is much more
likely to occur in contexts that encourage a focus on form (e.g. foreign
language classrooms) than in contexts where meaning remains pri-
mary (e.g. immersion classrooms).
Egi’s (2007) study is a good example of the use of stimulated recall.
She investigated whether learners paid attention to speciﬁc linguistic
forms in the recasts they received by replaying extracts from their
conversations that contained recasts and inviting them to comment on
them. She distinguished comments where (1) they showed awareness
that an error had been made but no awareness of the target-like form in
the recast, (2) awareness of the target-like model but no awareness that
their original utterance was problematic and (3) awareness of both the
error and the target-like model in the recast. No noticing was reported
most of the time for morphosyntactical features. However, Egi reported
that 18.7 per cent of the learners’ comments demonstrated awareness in
terms of (2) and a further 26.05 per cent awareness in terms of (3).

NOTICING IN ORAL, WRITTEN AND CM INPUT

Learners’ ability to engage in noticing and subsequent form-meaning

mapping when the task involves processing oral input may be limited.
The ephemeral nature of input does not allow them time for internal
processing. In contrast, written input ‘may encourage learners to move
beyond simple registration of new forms items and engage in intake
processing’ (Gilabert, Manchon and Vasylets, 2016, p. 125). These
authors point out that many tasks typically combine oral and written
input in an ‘interweaving of modes’ (p. 129). Many decision-making
tasks, for example, include substantial written input to prompt group
discussion. There is a need to investigate how hybrid tasks involving

both oral and written modes affect key processes such as noticing but,
as Gilabert et al. noted, little has been done to date.
This mingling of oral and written modes is also evident in tasks
involving synchronous text-based communication where interaction
unfolds in real time but affords a visible record in writing (Smith,
2003). CM tasks performed via text chat seem ideal for promoting
noticing as learners can inspect a written record of their interactions.
Text chat also affords an opportunity to use eye-tracking technology
to identify when learners attend to specific forms – an opportunity
that is not possible for face-to-face interaction. Smith (2012) used
this method, reporting that learners regularly attended to the specific
words that had been reformulated in the recasts that followed utter-
ances containing an error. Yuksal and Inan (2014) used stimulated
recall to compare noticing in CM and face-to-face interaction; they
found that while negotiation occurred more frequently in face-to-face
contexts, noticing was more likely to occur in synchronous CM
communication. Other studies (e.g. Gurzynski-Weiss and Baralt,
2014), however, failed to find any clear advantage for noticing in
CM interaction in comparison to face-to-face interaction.

Noticing and Acquisition in Task-Based Interaction

Cognitive-interactionist theories distinguish intake (i.e. the initial regis-
tration of linguistic forms in working memory) from acquisition (i.e. the
modification of the learner’s interlanguage system in long-term
memory) – see, for example, Leow (2015). Intake is likely to involve
noticing (i.e. the conscious registration of linguistic forms). But noticing
does not guarantee acquisition. Evidence for this comes from Mackey
(2006). She found that the level of noticing following recasts or clarifi-
cation requests varied according to target structure, with higher levels
evident for question forms, much lower levels for past tense and inter-
mediate levels for plurals. Eighty-three per cent of the learners who
reported noticing question forms demonstrated acquisition (i.e. they
improved in their ability to form questions in a post-test). However, the
relationship between noticing and the other two target features was not
established. Other studies (Mackey and Philp, 1998; Ellis and Mifka-
Provozic, 2013) also cast doubt on whether uptake-with-repair (the
clearest sign that noticing has occurred) is important for acquisition.
These studies all involved focused tasks. In contrast, Loewen (2005)
investigated the pre-emptive and reactive FFEs that arose incidentally in
lessons based on unfocused tasks. As in Ellis et al. (2001), he first
identified the FFEs and then devised tailor-made tests to assess the
linguistic features addressed in these episodes. He administered the tests

to the particular learners who had participated in the episodes one day
later and also two weeks later to see if participation in the form-focused
episode had led to learning. Out of the 473 FFEs that were tested, 47.6
per cent of the responses were correct in the immediate test and 39.3 per
cent in the delayed test. Loewen noted that these results were roughly
comparable to other studies (e.g. Williams, 1999; Nabei and Swain,
2001) and felt they were ‘encouraging, given the incidental and gener-
ally brief nature of the FFEs’ (Loewen, 2005, p. 381). Loewen also
investigated whether particular features of the FFEs were more likely to
result in correct test responses. The feature most likely to predict
learners’ correct responses to tests items was successful uptake, suggest-
ing that when noticing had taken place learning also occurred. How-
ever, successful uptake was only predictive of correct responses to the
grammar and vocabulary items in the test. For pronunciation the key
features of the FFEs were complexity (i.e. long FFEs were associated
with more correct responses than brief ones) and source (i.e. FFEs
involving the negotiation of meaning led to more correct responses than
FFEs involving the negotiation of form). However, while successful
uptake can be seen as evidence of noticing, its absence does not preclude
the possibility of noticing having occurred. Arguably, long FFEs and the
negotiation of meaning are exactly those features of interaction likely to
induce noticing. Loewen acknowledged a limitation of his study,
namely that the tests he used most probably measured the learners’
declarative rather than their procedural knowledge.

Tasks and Acquisition

The bulk of the task-based research based on cognitive-interactionist
theories has been entirely product-oriented. That is, it has investigated
whether focus on form results in acquisition without also investigating
what happens when learners perform tasks. A typical study involves
the use of focused tasks. Learners are ﬁrst tested on the linguistic
feature (typically grammatical) that is targeted by the task, perform
one or more tasks during which they receive some type of focus on
form, and then complete an immediate post-test and, some time later,
a delayed post-test. We begin by considering studies that have investi-
gated input-based tasks followed by a detailed look at research that
has investigated output-based tasks and CF.

INPUT-BASED TASKS

TBLT has generally been seen as involving production-based tasks.

However, as noted in Chapter 1, tasks can also be ‘input-based’.

An input-based task aims to promote interlanguage development by

directing learners’ attention to L2 input through listening or reading.
One of the key definitional features of a task is that learners are
required to use their own linguistic and non-linguistic resources to
communicate. In the case of input-based tasks, this means that they
have to use their knowledge of the L2 in conjunction with contextual
clues provided by the task to process the input they are exposed to
where ‘process’ refers both to comprehending the meaning of the input
and, potentially, attending to linguistic form (i.e. noticing). It is
important that the outcome of the task can only be achieved if the
learners are successful in comprehending the input. Although input-
based tasks do not require leaners to produce in the L2, learners may
elect to respond to the input they receive using their L1 or, if they are
able, their L2. In other words, production is not ruled out and in fact
has been shown to occur (e.g. Shintani, 2015).
An input-based task often takes the form of a listen-and-do task,
which requires learners to listen to commands or descriptions and then
perform actions (e.g. a physical action or pointing to a picture) to show
they have understood. Early studies of listen-and-do tasks (e.g. Ellis,
Tanaka, and Yamazaki, 1994; Loschky, 1994; Ellis and He, 1999; Ellis
and Heimbach, 1997) showed that they can lead to learning vocabu-
lary. These studies, motivated by the IH, investigated the effect of
exposure to different kinds of input (unmodified, pre-modified and
interactionally modified) on acquisition. They reported somewhat
mixed results, reflecting differences in the design of the studies. In
Loschky’s (1994) study the target items were first presented to the
learners before they listened which may explain why he found no
differential effect for types of input. In Ellis et al. (1994) there was no
prior presentation of the target words. Learners in the interactionally
modified condition were encouraged to signal their non-understanding
if they did not understand the input. The study reported a clear advan-
tage for modified input, especially interactionally modified input. How-
ever, the interactionally modified input typically took longer to
complete than tasks involving pre-modified input. In Ellis and He
(1999), the time taken for the pre-modified and interactionally modified
conditions was the same and in this study there was no difference in
their effectiveness. These studies suggest one reason why interactionally
modified input has been found to assist acquisition – it gives learners
additional time to process the input. If pre-modified input allows
adequate processing time it can be just as effective.
These early studies investigated the effect of performing input-based
tasks on vocabulary acquisition. Later studies included grammatical
features as well as vocabulary in their design. The tasks were designed

to force attention to the target grammatical features. For example, in

Shintani and Ellis (2010) and Shintani (2015), the learners (six-year-
old Japanese children) needed to attend closely to the input to distin-
guish English singular and plural nouns and could only complete the
tasks successfully if they did so. In Erlam and Ellis (2018) the learners
were required to distinguish L2 French singular and plural forms.
These studies, like the earlier ones, showed that performing input-
based tasks results in the acquisition of vocabulary. They also showed
that they can help grammar acquisition. They were clearly effective in
developing learners’ receptive knowledge of the target grammatical
features but less so in developing productive knowledge, possibly
because the tasks did not provide the amount of exposure needed to
develop productive use of the grammatical forms.[9]
For input-based tasks to work for acquisition they must create a
functional need for learners to attend to the target items and to engage
in form-meaning mapping. It is relatively easy to ensure that this
happens in the case of lexical items as learners will only be able to
do the tasks if they understand the vocabulary in the input. It is more
difficult in the case of grammar. Some grammatical features, for
example, convey no meaning. Shintani (2015) investigated whether
the input-based tasks she designed helped learners attend to English
copula-be as well as plural-s. The input exposed the learners to
numerous exemplars of copula-be but they did not acquire it, which
Shintani suggested was because it is a non-meaning bearing feature so
they could understand the directions without having to process it. This
is, of course, a limitation of focused production-based tasks as well
and, in fact, as Loschky and Bley-Vroman (1993) pointed out, it is a
lot easier to design input-based tasks that make the processing of
specific grammatical features ‘necessary’ (as opposed to ‘natural’ or
‘useful’) than production-based tasks.
Considerable skill on the part of the teacher is required to ensure that
input tasks are effective. Learners need instant feedback on whether
their efforts to understand have been successful. The teacher needs to
utilize a variety of strategies to help learners to understand – using
repetition, highlighting key items intonationally, utilizing contextual
clues etc. The extract from Shintani (2016) that follows illustrates the
skills involved. The teacher is instructing learners to find the picture
cards representing different animals and to take them to the ‘zoo’
(depicted in a frieze on the wall of the classroom). To select the right
card the learners need to be able to distinguish ‘squirrels’ from ‘squirrel’
(for which there was a separate card). The teacher first ensures the
learners’ attention, then gives the direction, repeating the key item
(‘squirrels’) three times. When a learner uses Japanese (his L1) to check

understanding, the teacher immediately conﬁrms. When the students

mention colours to help them identify the correct animal, she corrects
them and provides a contextual clue by pointing at a brown object in
the classroom. The learners performed the task successfully, demon-
strating their ability to process plural-s.

T: Okay the next. Okay, listen. please take the squirrels, squirrels to the zoo.
Squirrels.
S2: Doubutsuen [the zoo]?
T: Zoo, that’s right.
S5: Green? Blue?
S1: White?
T: No, no, no, not white. Not green. Not blue. Brown (pointing to a brown
item in the classroom)
S2: Brown.
S3: (Showing ‘two’ with his ﬁngers) two?
T: Two, yes. Three (.) two (.) one (.) go.
Ss: (All the students showing the correct cards)
T: Yes everyone is correct.

Input-based tasks have an important role to play in TBLT. In the

case of beginner-level learners they are essential as learners cannot
be expected to produce in the L2 until they have built up a linguistic
repertoire receptively. But they also have an important role to
play in later stages of L2 development. Apart from their contribution
to acquisition, they facilitate what might be called ‘input-processing
ﬂuency’ – the ability to rapidly and effortlessly segment input
and derive meaning from it. This is an essential element of L2
proﬁciency.

OUTPUT-BASED TASKS AND CORRECTIVE FEEDBACK

The bulk of the research investigating the effect that performing

output-based tasks has on acquisition has focused on the implementa-
tion rather than the design of the tasks – in particular, the role of CF.
The research has been experimental in design and has almost invari-
ably involved focused tasks and reactive focus on form (i.e. CF). Sato
and Loewen (2018) refer to ‘the burgeoning body of research investi-
gating variables that moderate the effectiveness of corrective feedback’
(p. 536). These variables include:
• Feedback versus no feedback
• Extensive versus intensive feedback

• Type of CF (e.g. input-providing versus output-prompting; explicit

versus implicit)
• Immediate versus delayed
• The target structure
• Effect of pre-teaching the target structure
• Moderating role of learner factors (e.g. working memory).
We will brieﬂy consider each of these variables but will leave the
moderating role of learner factors to Chapter 5.

Feedback versus No Feedback

Several of the studies involved a task-control group – i.e. a group that
performed the task(s) but without receiving any CF. These studies
enable us to address whether CF has any additional value for acquisi-
tion. This is an important question and one clearly relevant to theory-
building. According to theories that reject a role of negative evidence
in acquisition and view acquisition as driven entirely by positive
evidence (see, for example, Schwartz, 1993), performing tasks without
CF will be as effective as performing them with it. In contrast,
cognitive-interactionist theories predict that CF will enhance the
effectiveness of tasks by facilitating noticing and modified output. As
already noted, for Long (2015) reactive focus on form is necessary to
help adult learners overcome their limited capacity for incidental/
implicit language learning especially if the grammatical features lack
saliency and/or acquisition is blocked by their L1.
The research indicates that acquisition can sometimes take place
simply as a result of performing tasks. As we will see in Chapter 3, this
is an underlying assumption of the cognitive theories of Skehan (1998)
and Robinson (2007b). High-proficiency learners in particular may
benefit just as much from just performing tasks as performing them with
feedback (e.g. Ammar and Spada, 2006). However, some studies (e.g.
Lyster, 2004) have found that performing tasks without CF does not lead
to acquisition. The most general finding is that CF results in greater
acquisition than no CF. Meta-analyses of CF studies, for example, report
a clear effect for CF. Lyster and Saito’s (2010) meta-analysis – arguably
the most relevant here as it was restricted to classroom-based studies –
reported a medium-sized mean effect size (d = 0.74) for comparisons
involving groups that received and did not receive feedback. The
research, then, lends clear support for CF when learners perform tasks.

Extensive versus Intensive Corrective Feedback

As we have noted, intensive CF is possible when the task is a focused
one because it makes it possible for the teacher to focus correction on

the features targeted by the task. In contrast, CF is extensive when the

task is an unfocused one and the teacher corrects whatever errors
learners happen to make as they perform the task. Researchers have
opted to investigate focused tasks and intensive CF. Pre-testing is
possible if there is a pre-selected target but problematic if there is not.
In an interesting study, Nassaji (2017) compared the relative
effectiveness of intensive and extensive CF. Three groups of
intermediate-level ESL learners completed two tasks. One group
received intensive recasts directed at article errors, a second group
received extensive recasts directed at a range of errors, and a third
group performed the tasks with no CF. The main ﬁnding was, some-
what surprisingly, that the extensive recasts were more effective than
the intensive recasts in enabling acquisition of English articles. It
should be noted, however, that both experimental groups received
almost the same number of corrections of their article errors. The
difference lay in the fact that the extensive group received an add-
itional 122 recasts directed at other errors. Nassaji suggested that
the advantage found for extensive CF might have been because its
sheer frequency oriented the learners to pay greater attention to the
corrections. This study, then, is not a good test of the relative effect-
iveness of extensive and intensive CF. For a grammatical feature such
as articles, extensive feedback is likely to be effective because it is in
fact intensive. But this will not be the case for grammatical structures
that occur infrequently with unfocused tasks. There is an obvious
need to investigate whether extensive feedback is effective and this
will call for longitudinal studies.

Type of Corrective Feedback

We have noted that CF strategies can be distinguished in terms of two
dimensions – input-providing versus output-prompting and implicit
versus explicit. We also noted that in fact strategies can vary in terms
of how implicit/explicit they are. Figure 2.3 shows how the two
dimensions intersect. Researchers initially focused on these two dimen-
sions but more recently have begun to examine how varied execution
of the same strategy can affect acquisition.
Both input-providing (typically recasts) and output-prompting CF
have been found to be effective but overall output-prompting CF has a
stronger impact on acquisition than input-providing (Lyster, 2004;
Ammar and Spada, 2006; Ellis et al., 2006; Mackey, 2006; Loewen
and Nabei, 2007; Yang and Lyster, 2010; Sato and Loewen, 2018).
Lyster and Saito’s meta-analysis of classroom-based CF studies
reported a large effect size for comparison of prompts and control

OUPUT-PROMPTING

Clarification Repetition Elicitation Metalinguistic

Requests Clues

IMPLICIT EXPLICIT

Recasts Explicit Correction

INPUT-PROVIDING

Figure 2.3 The explicit/implicit continuum

Source: Slightly modiﬁed from Lyster and Saito, 2010, p. 278.

group (d = 0.83) and a medium effect size for the same comparison for
recasts (d = 0.53). However, Lyster and Saito also pointed out that the
standard deviations and confidence levels for these contrasts varied
widely, which they suggested was because of the difficulty of imple-
menting these CF strategies consistently in a classroom setting. In fact,
the comparison between recasts and prompts is conflated with that
between implicit and explicit CF. While recasts might be considered
generally implicit (but see later in this section), prompts vary consider-
ably, with some (e.g. clarification requests) implicit and others (e.g.
elicitation) much more explicit. Thus the superiority of prompts may
be due to their explicitness rather than to the fact that they lead to
learners modifying their output. Some recent studies, however, have
attempted to address these design problems by comparing a single,
implicit input providing strategy with a single output-prompting strat-
egy. Mifka-Profozic (2013) compared recasts with an implicit type of
prompt (clarification requests) and found the former more effective in
enabling high school learners of L2 French to improve accuracy in the
use of passé composé and imparfait. Sato and Loewen (2018) investi-
gated the same two implicit corrective strategies on Chilean university
students’ acquisition of English third person-s and possessive deter-
miners (his/her), reporting that the output-prompting strategy proved
more effective but only for possessive determiners. To my mind the
jury is still out regarding the relative efficacy of input-providing and
output-prompting CF and is unlikely to be resolved quickly given the
multitude of learner and contextual factors that can impact on how
these two types of CF are implemented and how they are perceived
and responded to by learners.

The distinction between implicit and explicit types of CF is also

problematic. Li (2013a) noted that it is important to distinguish the
perspective of the instructor/researcher from that of the learner. From
the perspective of instructor, the difference depends on whether the
feedback directs or just attracts attention to the linguistic form that is
the focus of the feedback. From the perspective of the learner, it
depends on whether or not the learner perceives the feedback as
corrective. Li argued that the explicit/implicit distinction should be
viewed from the perspective of instruction. CF is implicit if it is
implemented in such a way that it does not aim to make learners
aware they are being corrected (irrespective of whether they do in fact
become aware); it is explicit if it is implemented in a way that overtly
signals a correction is being made. In other words, the difference lies in
how the CF is implemented.
There are two ways of investigating the relative effectiveness of
implicit and explicit CF. One way is to compare different types of CF
that are generally accepted to be either implicit or explicit. The other
way is to compare the same CF type when it implemented in an
implicit and explicit way. Most of the studies to date have adopted
the first approach. Ellis et al. (2006) reviewed a number of early
comparative studies of explicit and implicit feedback. The implicit
feedback typically consisted of recasts in these studies. Explicit
feedback was operationalized by means of explicit correction or
metalinguistic clues. Ellis et al. concluded that overall the studies
pointed to an advantage for explicit feedback. Their own study also
reported that explicit CF was superior. Ellis (2019) reviewed a
number of later studies (e.g. Varnosfadrani and Basturkmen, 2009;
Li, 2010; Yilmaz, 2013a, 2013b) and confirmed the superiority of
explicit CF, which is evident in both immediate and delayed post-
tests and in also in tests that require controlled and more automatic
processing. Not all the studies reported in favour of explicit CF. For
example, Goo (2012) found no difference between the effects of
metalinguistic feedback and recasts. Li (2014) also reported that
implicit CF (recasts) were more beneficial for low-proficiency
learners of a complex Chinese structure.
Arguably, the more interesting comparisons of implicit and explicit
CF are those where the comparison involved the same CF strategy
implemented in different ways. Nakatsukasa (2016), compared a
group that received recasts with a group that received recasts accom-
panied by gestures. Nakatsukasa was interested in whether gestures
enhanced the effect of recasts but her study can also be seen as investi-
gating the relative effects of implicit-type recasts and recasts made
more explicit through gestures. She reported that only the group that

was exposed to recasts with gestures outperformed the control group

on the delayed post-test. Nassaji (2009) distinguished implicit and
explicit variants of both recasts and prompts in the interactions that
arose in a task-based lesson. Learning was measured in terms of the
learners’ ability to correct their errors in a written task that they had
completed prior to interacting with the teacher. The more explicit
forms of recasts and prompts led to higher rates of correction than
the more implicit forms of both strategy types. Both these studies
suggest that explicit CF is more effective than implicit. Zhao (2015),
however, reported that implicit and explicit recasts were equally
effective for Chinese university students’ acquisition of English third
person-s and argued that this may have been because the sheer fre-
quency of the implicit recasts, along with the students’ predisposition
to attend to form, made them perceptually salient to the students.
It is important to recognize that all these studies reported that CF
helped acquisition. But it is not easy to come to clear conclusions
about the relative effectiveness of different types of CF or of different
ways of implementing the same CF strategy. This is not so surprising
given differences in the participants, the context and the design of the
studies. Attempts to compare their results runs into the apples-and-
oranges problem. Recasts may be less effective than prompts in
immersion classrooms, for example, but very effective in classrooms
where learners are focused on form (Lyster and Mori, 2006; Zhao,
2015). It may be desirable from a theoretical perspective to tease out
the relative contributions of different types of CF, but from the peda-
gogic perspective (which is must be the primary concern in a book
about TBLT) it is sufﬁcient to acknowledge that all types can be
effective and that the best advice that can be given to teachers is to
opt for a variety of strategies as Lyster and Ranta (2013) proposed.

Immediate versus Delayed CF

Some proponents of TBLT (e.g. Willis and Willis, 2007) recommend
delaying correction until learners have completed a task. Cognitive
theories of L2 acquisition suggest that CF will work best when it is
offered in a ‘window of opportunity’ (Doughty, 2001) (i.e. immedi-
ately after an error has been committed). However, there are grounds
for believing that delayed CF can foster metalinguistic understanding
by encouraging learners to reﬂect on the corrections they receive. One
clear advantage of delayed feedback is that makes it much easier for
the teacher to select which type of CF to implement. Delayed feedback
is, of course, inherently explicit even if it involves recasting learners’
errors.

Despite the theoretical and pedagogic relevance of this issue, there has
been relatively little attention paid to it by researchers. Rolin-Ianzati’s
(2010) study of delayed CF distinguished two different approaches
corresponding to the input-providing and output-prompting types of
feedback found in immediate CF. In one approach the teacher provided
the corrections while in the other the teacher elicited corrections from the
students. Drawing on sociocultural theory, Rolin-Ianzati suggested that
eliciting correction will be more effective but she did not investigate this.
In a laboratory-based study (Quinn, 2014), ninety intermediate-level
adult ESL learners were randomly assigned to immediate, delayed or
no CF conditions. The grammatical target was English passive construc-
tions. The immediate and delayed feedback consisted of a prompt that
pushed the learners to self-correct followed by a recast if needed. There
were statistically signiﬁcant improvements resulting from both feedback
conditions but no differences between them. The task-only condition (no
CF) was just as effective as the CF conditions. Li, Zhu and Ellis (2016)
compared the effects of immediate and delayed CF involving prompts
followed by recasts on Chinese high school learners’ acquisition of past
passive constructions. Both types of CF resulted in gains on a grammat-
icality judgement test (GJT) but no effect for either type of CF was found
on an elicited imitation test (EIT). A slight advantage was found for
immediate feedback on the GJT, which was explained in terms of the
learners using the feedback progressively in the production of new past
passive sentences as they performed the tasks. These studies do not allow
any clear conclusion to be reached about the relative effects of immediate
and delayed CF and point to the need for further research.

Target Structure
Another variable potentially affecting the learning that results from CF
when learners perform tasks is the linguistic feature(s) targeted by the
CF. In the case of grammatical targets, it is quite likely that CF will
vary in how effective it is. Studies that have investigated this have
compared the effect that CF has on grammatical structures hypothe-
sized to differ in terms of their difficulty.
Several studies suggest that the difficulty of the target structure is a
factor influencing how effective CF is. Ellis (2007) compared the
effects of two types of CF (recasts and metalinguistic comments)
on two grammatical structures – comparative adjectives and past tense
(-ed). He evaluated the learning difficulty of these two structures using
a variety of criteria (e.g. input frequency, processability, reliability of
the explicit rule) and concluded that overall the comparative was likely
to pose a greater learning burden than past tense (-ed). The effect of

the recasts on the acquisition of these two structures did not differ but
the effect of metalinguistic comments was more evident on the com-
parative structure. Li (2014) also compared the effect of recasts and
metalinguistic corrections. The target structures were Chinese classi-
fiers and perfective aspect markers with the former deemed more
salient and therefore the more easily learned. Results for the two
structures only differed for recasts. In the immediate post-test recasts
had stronger effect than metalinguistic comments on the more salient
structure in the delayed post-test for low-proficiency learners. In the
post-test, however, recasts had a greater effect on the less salient
structure for high-proficiency learners. Yang and Lyster (2010) com-
pared the effects of recasts and prompts on regular and irregular past
tense forms in English. Prompts were more effective for regular -ed but
there was no difference for irregular forms. Finally, Sato and Loewen
(2018) compared the effects of two implicit types of CF (recasts and
clarification requests) on the acquisition of third person-s (considered
non-salient) and possessive determiners (considered salient). They
reported that the group that received clarification requests outper-
formed the group receiving recasts but only for possessive determiners.
Perhaps the only clear conclusion that can be reached from these
studies is that the nature of the grammatical structure does indeed
mediate the effect of CF as Long (2007) claimed would be the case.
However, the results of these studies do not allow for any firm conclu-
sions about the interaction between CF type and grammatical struc-
ture. Again, this is perhaps not surprising given that the grammatical
targets in these studies varied greatly and there is a lack of an agreed
set of criteria for evaluating their difficulty. Li’s (2014) study also
suggests that the learners’ proficiency level moderates the interaction
between CF types and grammatical structure. From a practical point
of view, it is difficult to see how teachers should take account of the
grammatical target in deciding what type of CF to provide. At best, all
they can do in intensive CF is select a structure that they deem is within
the developmental level of their students, e.g. a structure that students
have started to deploy but often erroneously.

Effect of Pre-teaching the Grammatical Target

An important issue for TBLT is whether teachers should explicitly
teach the grammatical target of a task in the pre-task phase of a lesson
(see Chapter 1). Long (2015) is adamantly opposed to this on the
grounds that it constitutes a return to focus on forms. However, in
many CF studies (e.g. Lyster, 2004; Sato and Loewen, 2018) explicit
instruction was included in the pre-task phase.[10] Two questions

arise. First, does pre-teaching the grammatical structure have an

impact on how the task is performed? Second, does pre-teaching the
grammatical structure enhance the effect of the CF?
The first question is of interest because of the claims made for how
task performance assists acquisition. This is dealt with in Chapter 3,
where the effect of task design and conditions on the complexity,
accuracy and fluency (CAF) of learners’ production is considered.
Does pre-teaching a grammatical structure affect CAF? Again, there
is little research that has addressed this. In Mochizuki and Ortega
(2008) there were three treatment conditions – a no-planning condi-
tion, an unguided planning condition and the condition of greatest
interest here – a guided planning condition, where the participants
(first-year high school students in Japan) were given a handout about
English relative clauses, listened to a pre-recording of the task per-
formance, and then had five minutes to plan before they performed
the task. The performance of the task, which was designed to elicit
use of relative clauses, was analysed in terms of amount of relative
clause use, the accuracy of relative clause use, and the global com-
plexity and global fluency of the language produced. The guided
planning group produced more than twice as many relative clauses
as the other two groups and their relative clauses were also more
target-like. However, there was no difference in global complexity
and fluency in the three conditions. This study, then, suggests that
pre-teaching does not impact negatively on how a task is performed
and in fact can have a positive effect on production of the target
feature.
Another study, however, suggests otherwise. In Ellis, Li and Zhu
(2018) one group of Chinese high school learners received a brief
grammar lesson on the English passive voice followed by five minutes
of practice activities prior to performing two dictogloss tasks while
another performed the same task but with no pre-task instruction or
practice. The learners’ task performance was coded in terms of their
production of the target structure and on global measures of CAF. The
results showed that the pre-task instruction led to more frequent but
not more accurate use of the target structure, but that it had detrimen-
tal effects on CAF.
The second question was addressed by Li et al. (2016). Using
the same learners as in Ellis et al. (2018), they compared the effects
of performing the tasks (1) without pre-task explicit instruction,
(2) with pre-task explicit instruction, (3) with CF but no explicit
instruction and (4) with both explicit instruction and CF on acquisi-
tion of past passive constructions, as measured by means of a GJT
and an EIT. On the GJT, conditions (2), (3) and (4) led to significant

gains with explicit instruction plus feedback showing the largest

effects. On the EIT, there was no effect for any of the three treatment
groups when the data were analysed for the whole cohort. However,
when the learners were subdivided into those with zero and some
prior knowledge based on their pre-test EIT scores, condition (4) was
more effective than the others.
It would be premature to advise teachers about whether to include
explicit instruction in the pre-task phase of a lesson but the evidence
suggests that it is likely to result in more attempts to use the target
structure and may sometimes be beneﬁcial for learning. It may,
however, interfere with the general performance of a task. There
are, of course, other ways of providing explicit instruction – during
the performance of the task if learners fail to use the target structure
(as in Samuda, 2001) or after the task has been completed in the post-
task stage (as suggested by Ellis, 2003). These are considered in
Chapter 8.

Task-Based Interaction in Small Group Work

All the research we have considered to date has examined interaction
in tasks performed in a teacher–class participatory structure. In some
ways this runs contrary to a general understanding of what TBLT
entails. Mainstream accounts of TBLT assume that tasks will be
largely carried out in small group work. We need to ask, therefore,
whether the interactions that arise in small groups when tasks are
performed manifest focus on form similar to that found in teacher-
led lessons and whether learning results.
Group work is generally seen as advantageous for language learn-
ing. In an early study, Long et al. (1976) reported that students
working in small groups produced a greater quantity of language
and also better quality language than students in a teacher-fronted,
lockstep classroom setting. Small group work provided more oppor-
tunities for language production and greater variety of language use in
initiating discussion, asking for clariﬁcation, interrupting, competing
for the ﬂoor and joking.
Group work has been found to result in more interactional adjust-
ments than in teacher–class interaction but only if the task is of the
required information exchange type (Pica and Doughty, 1985a,
1985b). In an often-cited article, Long and Porter (1985) pointed to
a number of advantages of group work:
1. Quantity of practice (i.e. there is more opportunity for language
practice in group work than in lockstep lessons).

2. Variety of practice (i.e. in group work learners can perform a wide

range of language functions).
3. Accuracy of student production (i.e. learners have been shown to
use the L2 just as accurately in group work as in lockstep lessons).
4. Correction (i.e. students engage in self- and other corrections to a
greater extent in group work than in lockstep teaching).
5. Negotiation (i.e. students engage in more negotiation of meaning
sequences when performing communicative tasks in group work
than in teacher-led lessons).
6. Task (i.e. group work lends itself to the performance of two-way
tasks that elicit negotiation of meaning sequences).
Nassaji (2009) also found a clear advantage for group work. He
compared the interactions occurring in three participatory structures –
whole class, group work, one-on-one (all involving a teacher). The
study involved fifty-four hours of communicative lessons in seven intact
classes. Overall 1,986 FFEs were identified – 1,325 in whole class,
511 in group work, 150 in one-on-one interactions. He found that
reactive FFEs are more likely to occur in individual and small group
interactions than in the whole class context but pre-emptive FFEs,
which were more frequent overall, were more likely in a whole class
context. FFEs in small group work and in one-on-one interactions are
more likely to lead to correct responses in tailor-made tests administered
after class than FFEs in a whole class context (69% and 66% vs. 48%).
Nassaji concluded that focus on form is a socially mediated process.
However, not all researchers provide such a favourable account of
group work. Researchers have noted that learners sometimes overuse
their L1 when performing tasks in groups (Carless, 2004). Adams,
Nuevo and Egi (2011) observed that the studies vary in how fre-
quently CF occurs in learner–learner interactions, with some showing
that it is very infrequent. Toth (2008) noted that learners in groups
tend to use a limited range of feedback strategies and, in contrast to
teachers, they tend to focus on a wide range of linguistic features.
Few studies have actually investigated whether group work inter-
action results in acquisition. Adams et al. (2011) investigated whether
the implicit and explicit feedback in learner–learner interactions was
related to acquisition. Learners in high-intermediate classes in an adult
ESL school in the United States worked in pairs to complete tasks
designed to elicit the use of past tense and locatives. They were not
instructed to provide CF. The interactions were coded for all instances
of feedback, whether the feedback was implicit or explicit, and whether
the learners modified their output when they were corrected. Acquisi-
tion was measured by means of a GJT and an oral production test.

Adams et al. reported that a third of the feedback was non-target-like

and that there was relatively little output-prompting feedback. There
was no evidence that either implicit or explicit corrections promoted
acquisition of past tense but recasts were related to scores for the
locative in the delayed post-test. They concluded ‘it seems likely that
feedback may not play as important role in learner–learner interactions
as it plays in NS–learner interactions [native speaker–learner inter-
actions]’ (p. 56). They suggested that the widely focused nature of the
feedback may have made it less salient and that the learners may have
been hesitant in accepting that the feedback they received was correct.
The difference in the accounts of group work interaction found in
these studies is not so surprising as its effectiveness must surely depend
on the particular learners involved. What may be crucial is the mindset
of the learners. Sato (2017) deﬁned mindset as ‘a disposition toward
the task and/or interlocutor prior to and/or during the interaction’
(p. 255). His study indicated that learners with a positive mindset
engaged in more correction, language-related collaboration and col-
laborative sentence completion than learners with a more negative
mindset. One way of inducing a positive mindset is through training.
Sato and Lyster (2012) reported that training Japanese university
students to make use of CF during group work led to them providing
both more and more effective CF (prompts) and also to more repair
work. For group work to deliver on its promises in TBLT, then,
learners may require guidance in the behaviours that are needed to
ensure that it is effective for acquisition.

Conclusion
Cognitive-interactionist theories support TBLT by emphasizing that
(1) acquisition of an L2 occurs incidentally/implicitly when learners
are focused on meaning as they perform tasks, but that (2) focus on
form is needed to ensure that learners attend to the linguistic forms
they are exposed to in the input. Interaction facilitates learning when it
promotes noticing and noticing-the-gap. To examine whether and to
what extent this happens, researchers have utilized discourse analysis
to identify those features of interaction that theory predicts will facili-
tate acquisition. Key overlapping constructs are negotiation of
meaning, negotiation of form, and pre-emptive and reactive focus on
form. Interaction works for acquisition when it enables learners to
map form onto meaning in a context where they are communicating
purposively as they perform tasks.
We distinguished two strands of research – (1) task-based learner per-
formance studies and (2) task-as-treatment studies. The former sheds light

on how task design and implementation features affect the kinds of inter-
action that result from a task. The latter shows how performing tasks can
induce noticing and facilitate acquisition. A major focus of research that has
drawn on cognitive-interactionist theories is CF. We discussed various
aspects of this research – whether CF has add-on value to performing a
task, extensive versus intensive feedback, type of feedback, immediate
versus delayed, choice of target structure and the effect of explicit instruction
in the pre-task phase. We also considered interaction in group work.
What then do we know?
• Closed tasks of the required information exchange type are best for
promoting negotiation of meaning.
• Interactionally modified input (including in text chat) is more likely
to induce noticing of linguistic forms than pre-modified input but
this may be because it affords learners more time to process input.
• Participating in FFEs helps acquisition especially if learners have
opportunities to repair their errors.
• Input-based tasks work for acquisition providing that they create a
functional need for learners to map forms onto their meanings. This
is more likely to occur if learners have an opportunity to interact
when they do not understand the input.
• Intensive CF facilitates acquisition.
• On balance, explicit feedback is more effective than implicit
feedback and, in the eyes of some researchers, output-prompting
feedback is more effective than input-providing.
• CF can work for acquisition even when it is delayed until the post-
task stage of a lesson.
• The effectiveness of feedback depends on the grammatical targets to
which the correction is directed.
• Explicit instruction in the pre-task phase may help to elicit use of the
target structure when the task is performed and assist acquisition
but may also impact on the overall quality of learners’ production.
• While group work may be generally beneficial, learners may not
always engage in much correction unless they are trained to do so.
Many of these conclusions are necessarily tentative – partly because in
some cases the research is still very limited and partly because the
available research findings are not always consistent. Also, as we have
noted, the complexity of interactional phenomena such as CF makes it
difficult to arrive at clear conclusions based on studies that investigate
just one or two variables at a time and cannot take account of the
intertwined relationships among a host of variables.
There are some obvious limitations in the research we have
reported. One concerns how learning is measured. Uptake-with-repair

may serve as a measure of noticing (intake) but it cannot be used as a

measure of learning, which requires administering tests. A great var-
iety of tests have been used. In some studies (e.g. Ellis et al., 2006; Li,
2013a) care was taken to include tests that measured both declarative
and procedural knowledge of the target structures. This is important
as the aim of TBLT is to develop learners’ procedural L2 ability so
GJTs or tests that tap into controlled language cannot be used to assess
whether tasks have worked as they are intended to. We are still a long
way off knowing whether TBLT enables learners to develop the impli-
cit knowledge of an L2 that Long (2015) saw as the goal.
Interaction is a complex phenomenon with many facets. In this
chapter we have focused quite narrowly on those constructs that
SLA researchers have deemed theoretically important for understand-
ing how interaction ‘connects input (what learners hear and read);
internal learner capacities, particularly selective attention; and output
(what learners produce) in productive ways’ (Long, 1996, pp. 451–2).
It should be noted, however, that the acquisition potential of inter-
action does not rest solely in the fairly narrow set of constructs that we
have considered. Other aspects of interaction – for example, the extent
to which learners have the opportunity to play an initiating as well as a
responding role in interaction (see Ellis, 1999) or the extent to which
learners have the opportunity to produce long turns – are also import-
ant. Some researchers are moving away from the narrow approach we
have adopted to consider how ‘engagement’ is evident in interaction.
This more encompassing approach is considered in Chapter 6.
Another limitation of the research that the cognitive-interactionist
perspective has spawned is that it has focused almost entirely on the
short-term effects of performing interactive tasks. Thus we know
almost nothing about how interaction feeds acquisition over time.
There is an urgent need for longitudinal studies that investigate how
patterns of interaction change and what the accumulative effects of
engaging in interaction over time are. More studies, such as those of
Shintani’s (2016) ﬁve-week classroom-based study of TBLT and Saito
and Akiyama’s (2017) one-semester study of video-based interaction,
are needed. These studies revealed that task-based interaction does
help acquisition in the long term but also point to possible limitations.
Shintani’s young learners demonstrated little productive ability. Saito
and Akiyama’s adult learners demonstrated no improvement in
aspects of language that only develop slowly and gradually (e.g.
pronunciation), leading these researchers to suggest that explicit
instruction may be needed to complement interaction.

Introduction
The focus in this chapter is on the work that has been done, both
theoretical and empirical, to understand what happens in task perform-
ance from a psycholinguistic perspective. This does not, in any way,
conflict with the more socially oriented chapters in Part II of the book,
but it does reflect the considerable work with a more purely cognitive
perspective that has been done. First, typical methods of measuring task
performance are described, essentially because these become the touch-
stone for the more substantive discussions which follow. Then two
approaches are covered which account for a considerable amount of
the recent research: the Limited Attentional Capacity (LAC) approach
(Skehan, 2014c) and the Cognition Hypothesis (CH)/Stabilize, Sim-
plify, Automatize, Restructure, Complexify (SSARC) model (Robinson,
2015), focusing on tasks themselves – task design, task characteristics
and so on. The discussion attempts to address six questions:
• What is the main focus of the approach, and correspondingly, what
is de-emphasized?
• What is the role of acquisition?
• How important are performance issues?
• What theoretical accounts are provided?
• What is the research base?
• What research methods are typical for the approach?
The following section compares the two approaches and explores
strengths and weakness with each. Then the final section of the chapter
focuses on current issues with psycholinguistic approaches.

Measuring Task-Based Performance

It may seem odd to start the chapter with a concern for the rather
‘technical’ area of measurement, but there are issues to be resolved in

this area, and these have an impact on the more substantive discus-
sions which follow. So the section is necessary, but only to prepare the
ground for what comes later.
Cognitively oriented task research has been remarkably focused on
a limited number of performance areas. Initially (Ellis, 1987; Crookes,
1989) these were complexity, accuracy and fluency (CAF). More
recently the area of lexis has been added to this list, and so, strictly
speaking, we should now be concerned with structural complexity and
lexical complexity separately, although the acronyms CAF (Housen
and Kuiken, 2009) and CALF (Skehan, 2009a) are common in refer-
ring to this approach to measurement. There are several justifications
for this approach:
• Statistically, the four areas have distinctness as shown by factor
analyses of datasets that generate separate factors (Skehan and
Foster, 1997; Tavakoli and Skehan, 2005). In other words, it is
quite possible that someone will obtain high scores in one dimension
of performance and not in others: there is no guaranteed proficiency
effect leading to even performance in each area.
• The four areas, as will be shown in more detail throughout this
chapter, can be influenced by different variables, so that what raises
one (e.g. complexity), may not generalize to other areas, and may
even lower it (e.g. fluency).
• It has been argued (Skehan, 1998, 2014c) that there is an acquisi-
tional sequence consistent with the four areas, with complexity
(structural or lexical) coming first, as an interlanguage system is
destabilized and grows, followed by greater control, first through
the reduction and even elimination of error and then followed in
turn by the development of fluency, as not only is accuracy
increased, but this is done at reasonable speed of production.
• There is some evidence of the four areas reflecting different prior-
ities, of personal styles, with some learners emphasizing accuracy or
fluency, and others complexity (Skehan and Shum, 2017).
Given this background, it is worth exploring each of the areas, and the
progress that has been made in measurement in recent years.

Structural Complexity
Early research in this area focused on two approaches. These were to
explore the range of structures that were used in a task, on the one hand,
and to compute a measure of subordination, on the other. The former
emphasizes structural variety, assuming that this reﬂects a greater under-
lying structural repertoire. The latter takes subordination to be an

effective surrogate for general complexity, on the assumption that the

more speakers can pack more information into what they say or write
through this linguistic device the greater the underlying structural system
that they have available. Although the former, range, has generated
considerable interest, it has not been used nearly as widely as subordina-
tion-linked measures. These have become fairly routine in task-based
research. There is even a journal article (Foster, Tonkyn and Wiggles-
worth, 2000) which has been very influential in the field in the attempt to
standardize the measures used to facilitate cross-study comparisons.
As operationalizations of complexity, the subordination-based
measures have been very useful in task research. More recently an
alternative approach has been proposed. Norris and Ortega (2009)
suggest that, for more advanced levels, the subordination-based meas-
ures do not discriminate so well, and they propose such measures
should be supplemented by indices based on the average number of
words per clause. Since this proposal such measures have been used in
many research studies, so it is interesting to explore what the relation-
ship is between these two types of measure. Inoue (2013), Pang and
Skehan (2014), Skehan and Shum (2017), and Wang and Skehan
(2014) all report correlations on this issue: the typical correlation from
this range of studies, over different proficiency levels and with different
research designs, is less than 0.20. In other words, the two measures
are clearly concerned with different constructs, and so research studies
need to include both. Reviewing this area, Skehan (2018) also pro-
poses that there are indications of systematic influences. Narrative (vs.
interactive) tasks and planned conditions raise both subordination and
words-per-clause. Structured tasks and there-and-then conditions raise
subordination only. Low proficiency (possibly surprisingly), native-
speaker status, a here-and-now condition and less structure raise
words-per-clause only. This is clearly an area where more research is
needed to account for these emerging generalizations. But for now, it is
clear that both types of measure are independently essential.

Accuracy
As with complexity, there are also some alternative choices with
regard to measures of accuracy. Prominent amongst these are: (1)
the proportion of error-free clauses, (2) errors per (usually) 100 words
and, more recently, (3) error gravity (Foster and Wigglesworth, 2016).
The use of each of these can be defended on theoretical and/or prac-
tical grounds. Perhaps the first has accumulated most findings in
existing research, and so one can have confidence in it on that basis.
The second has been advocated as more appropriate for some

languages, such as German (Mehnert, 1998). The third has been

proposed as having more construct validity, since it is argued that
treating all errors as equal is inappropriate since some impact on
communication more than others (Foster and Wigglesworth, 2016).
In an examination of the functioning of all of these measures, through
the examination of a number of datasets, Skehan (2018) proposes
that, empirically, it does not matter which is used: unlike the structural
complexity measures, which do not intercorrelate highly, all the
accuracy measures show very high correlational levels – almost always
above 0.80, and often clearly above that. So in this case, we appear to
have simplicity of measurement – the choice that is made does not
seem to have a severe impact on results.

Lexical Complexity
Two measures of lexical complexity have been widely used in task
research. Lexical diversity measures are based on type-token ratios.
Given the well-established and strong relationship between text length
and type-token ratios (correlations of 0.70 are typical – Foster and
Skehan (2012) – demonstrating that the longer the text, other things
being equal, the lower the type-token ratio), there needs to be a
correction made to compensate for text length. Typical, but by no
means the only alternatives, would be the mean segmental type-token
ratio, and D, computed by the Child Language Analysis (CLAN) suite
of programmes (Macwhinney, 2000). Lexical sophistication aims
more at the construct of lexical richness, and is based on the propor-
tion of words that are used in a spoken or written performance which
are deemed difficult. Difficulty is usually defined in terms of frequency,
and so the claim is that ‘penetration’ of a text by more difficult words
is reflective of a more extensive mental lexicon.
As with the structural complexity measures, it is interesting that the
different lexical measures do not intercorrelate highly (Skehan,
2009b). Each appears to be doing something different. High lexical
diversity reflects speakers or writers who do not reuse the same words
so much in a text, and this seems distinct from second language (L2)
users who draw upon less ‘obvious’ words, the target of lexical
sophistication. There is no literature yet on what influences each of
these measures selectively (in contrast, as we have seen, to the different
structural complexity measures). Lexical diversity, though, does dis-
tinguish very clearly between native and non-native speakers (Skehan
and Shum, 2017), whereas lexical sophistication does not. The former
seems to be a capacity of the speaker, whereas the latter seems more
task-influenced (Skehan and Shum, 2017).

Fluency
In some ways this is the most complex sub-dimension of performance.
It has been argued that fluency can be subdivided into breakdown-
linked fluency, repair-linked fluency and speed (Tavakoli and Skehan,
2005). The first is typically measured through pausing, the second
through behaviours such as reformulation, replacement, repetition
and false starts. Speed is typically measured through words or syllables
per minute, but De Jong et al. (2013) advocated the use of mean
syllable duration, i.e. inverse articulation rate, since it is more nor-
mally distributed. There are also composite measures such as length of
run and phonation time, as well as, possibly, double occurrences, e.g.
the number of times that pauses and repair coincide (Kahng, 2014).
The sub-dimensions of fluency intercorrelate at a level between the
correlations reported for accuracy and structural complexity. In other
words, the relationships are positive, but not necessarily strong. In
addition, there are several emerging issues in fluency measurement.
First, there is the issue of pause or repair location. Skehan (2009b,
2018) argued that end-of-clause dysfluencies should be considered to
be distinct from mid-clause dysfluencies, and indeed it may be the case
that mid-clause pausing is similar to repair, more generally. Second,
there are also concerns about surrogate measures, as, for example,
with mid-clause filled pauses being taken as a surrogate (easier to
measure) for unfilled mid-clause pauses (notoriously more difficult to
measure). Lambert, Kormos and Minn (2016) use them in this way
whereas Skehan (2018) reports quite low intercorrelations between
filled and unfilled mid-clause pauses. Most challenging and exciting of
all, linkages have been proposed between different types of dysfluency
and the detail of psycholinguistic speech production processes (Lam-
bert et al., 2016) from Levelt’s model of speaking, and we will return
to this in the section on the Limited Attentional Capacity Approach.

General vs. Speciﬁc Measures

All the measures we have considered so far have been generalized in
nature, taking an entire speech sample and then measuring the various
sub-dimensions of performance while drawing on the entire sample.
Speciﬁc error types, e.g. of aspect or agreement, do not ﬁgure in this
analysis. Such generalized approaches have the advantage that they
are based on the largest amount of data possible. They also enable the
sub-dimensions of performance to be measured separately. As a result
they are, perhaps, the most effective way of detecting differences
between experimental conditions that affect task performance.

But they are also crude, and so there is a strong case for using more
specific measures. R. Ellis (1987), for example, used the different
forms of the English past tense (regular, irregular, copula) to try to
capture differences between rule-based and lexical forms. Using spe-
cific measures of this type connect more naturally with acquisitional
processes, and, perhaps, patterns of development of different interlan-
guage subsystems. They also enable more precise hypotheses to be
framed. Robinson (2015), for example, makes linkages between spe-
cific measures which derive from cognitive linguistic analyses of per-
formance, linked to different experimental conditions. But the major
disadvantage here is that the use of specific measures risks reducing the
amount of data that can be analysed. Specific hypotheses, while desir-
able, mean that particular tokens have to be generated, and in suffi-
cient quantities to confirm or disconfirm the hypotheses. Engineering
research designs which do this are not at all easy and there is the
danger that such a design may compromise the ‘taskness’ or natural-
ness of the data collection by constraining too narrowly what needs to
be done. As a result, the researcher has a considerable dilemma. Both
types of measure, generalized and specific, have their uses, and so
where possible, using both is desirable. It is simply that, for much of
the time, specialized measures will not be feasible to enable effective
statistical testing. Where it can be done, however, it is highly desirable.

Models of Task-Based Performance

This, the main section of the chapter, will consider ﬁrst Skehan’s LAC
approach, followed by Robinson’s CH, and in its most recent form the
SSARC model.

Part One: The LAC Approach

Skehan’s LAC approach arose out of a series of studies conducted with
Pauline Foster (Foster and Skehan, 1996, 1999; Skehan and Foster,
1997, 1999). The results of these studies suggested (bearing mind that
they used measures of structural CAF) that there was often a trade-off
between the performance areas, particularly between accuracy and
complexity. For example, in one study (Foster and Skehan, 1996)
exploring planning, a group simply given planning time produced
the highest level of accuracy, whereas a group given planning time
and instructions which emphasized ideas to be expressed, produced
the highest levels of structural complexity and lower accuracy.
A factor analysis conducted in Skehan and Foster (1997) conﬁrmed
this separation between complexity and accuracy. These ﬁndings

contributed to the development of the importance of trade-off, against

the background of limited attentional resources and working memory.
In other words, given such limitations, if one is performing close to
one’s limit (as these tasks were designed to provoke) there is the strong
possibility that prioritizing one performance area may well be at the
expense of another. Since this start, a wide range of studies that are
consistent with this interpretation have appeared (Foster and Skehan,
1999, 2013; Skehan and Foster, 1999, 2005, 2007).
This general approach is still actively pursued, as in Tavakoli and
Foster (2008), Foster and Tavakoli (2009), Foster and Skehan (2012,
2013) and the chapters in Skehan (2014a). But it has changed over the
years, not least since the emphasis has moved away from trade-off,
since this term implies inevitability, and that there will always be a
tension between accuracy and complexity. Now it is more appropriate
to refer to the LAC approach, since this emphasizes the importance of
attention and working memory constraints, but not that these con-
straints cannot be overcome. The approach is characterized through a
series of principles (Skehan, 2015, 2018). These are:

Principle 1: Working memory and attention are limited. The working

memory literature is huge, and demonstrates quite clearly that this
aspect of memory, effectively current consciousness, has limited size
and that this limitation has an impact on attention (Wen, 2015;
Skehan, 2016). In working memory research there may be dispute
about the exact size of the memory, but the disagreements are so
slight that this does not really change the impact this limitation has
on the field of L2 acquisition and performance (Cowan, 2015).
LAC also assumes limited attentional availability, and while it is
accepted there may be variation in how much attention is available
(e.g. for motivational reasons, more, or less, attention may be
mobilized at a particular time), there is still a maximum and this
maximum is assumed to represent a significant functional con-
straint for the L2 user.
Principle 2: The CALF framework is useful. Essentially this is a
restatement of the first section of the chapter. The claim is simply
that viewing performance in this way, with these sub-dimensions, is
revealing about the different influences, e.g. task characteristics,
task conditions and the effects they have on performance. The areas
are particularly important in capturing the effects of different inde-
pendent variables, and of tensions between them.
Principle 3: Tasks are analysable, but difficult to work with. Much
research with tasks explores broad task types (e.g. personal infor-
mation exchange, narrative, interactive) or more specific

characteristics (structured or not; requiring information transform-

ation versus simple retrieval). The aim is to establish generalizations
about performance on such tasks which would be useful to predict
future performance or to understand development and acquisition.
These analytic schemes have been productive, and have contained
considerable promise, delivering a number of very useful results,
practically and theoretically. But tasks are also difficult to work
with, since there may be a fundamental distinction between the
intended task and the actual task (Breen, 1984; see also the discus-
sion in Chapter 1). In other words, individuals doing tasks can take
the task in different directions to those anticipated, and so one of
the fundamental problems in working with tasks is that analytic
schemes do not always transfer into actual task performance in
straightforward ways. This is discussed more extensively in the
Issues part of the chapter, particularly the section on Task Condi-
tions, where it is also argued that task conditions are more depend-
able sources of influence (Skehan, 2016).
Principle 4: Linking task performance (as well as notions of atten-
tional limitations, CALF-measured performance and task
characteristics) to the Levelt model of speaking is productive and
a potential basis for effective predictions. Levelt (1989, 1999) dis-
tinguishes between three major stages of speaking. These are
intended to account for the case of first language (L1) speaking,
but they are generalizable, with modification, to the L2 case (De
Bot, 1992; Kormos, 2006; Skehan, 2014c). The three stages are
conceptualization, formulation and articulation. The first is con-
cerned with developing the ideas to be expressed, working out, in a
conversational context, what needs and is appropriate to be said.
This stage ends with the development of the pre-verbal message.
Formulation is concerned with clothing the pre-verbal message
(propositional rather than linguistic in nature) with language. The
first sub-stage is to retrieve appropriate lemmas from the second
language mental lexicon (SLML), the second is to use the rich
information contained in the lemmas as the basis for syntax-
building, and the third is to take the semi-assembled message and
to access relevant phonological information, again from the
lemmas. This information is then fed to the final stage, Articulation,
which converts the phonological outline into actual speech. In
addition, the process of speaking is also accompanied by the cap-
acity to monitor (and therefore modify) what is being assembled,
although according to the Levelt model this can only be done at
particular points in the speech production process. It is also import-
ant to clarify that the different components in this model,

conceptualization, formulation, articulation, are encapsulated and

modular. What this means is that each gets on with its job, operat-
ing simultaneously to the other modules within the system. Each
module is working on something different. The Conceptualizer
delivers the pre-verbal message to the Formulator, and then gets
on with the next communicative cycle at the very same time as the
Formulator is operating upon the input it has received from the pre-
verbal message. The same applies to the Articulator, which also gets
on with its job in parallel fashion. This simultaneous operation of
each stage is vital for the smooth capacity we have to use our L1.
The LAC approach, though, is concerned with L2 speaking.
Crucial for this is the SLML. This is smaller, slower, less well organ-
ized and, importantly, often not as rich in the information that is held
in a lemma (information such as appropriate syntactic frames, mul-
tiple potential meanings, phonological information, collocates, dis-
course functioning and so on). As a result, during speaking (or
writing, come to that), the Conceptualizer may make demands upon
the SLML to underpin Formulation, but these demands cannot be
met at all, or only partially, or cannot be met quickly enough. As a
result, parallel functioning, the norm in L1 language, cannot be
sustained, and effortful, attention-consuming serial-processing
results, as problems at one stage of the speech production process
divert attention from the others, until the problem is solved. Then the
thread of discourse has to be (painfully) retrieved.
A final point in this regard is that there are connections between CALF
and the stages within the Levelt model. Conceptualization does map
reasonably on to structural and possibly lexical complexity, and
Formulation has stronger links with accuracy and fluency. These
mappings are not, by any means, exact. Conceptualization, for
example, is likely to link with fluency also, when it underpins a more
macro approach to discourse, spanning several clauses or sentences
(Skehan, 2018). It may also, as when Conceptualization is assisted by
planning opportunities (Pang and Skehan, 2014), help accuracy by
avoiding more difficult language. But in the main, there is something
of a connection between CALF and the Leveltian stages.
Principle 5: Task characteristics and task conditions influence per-
formance separately and in combination. This may seem simply a
statement of the obvious, but in the LAC approach it does have
some theoretical importance! Recall the claim that attentional
limitations are a constraint, not an inevitability. The point is that
allocating attention to one area may have a negative impact on
other dimensions of performance (a statement, in itself, consistent
with a trade-off interpretation). But essentially this represents a

challenge – the need to overcome attentional limitations by judi-

cious task design/choice and implementation through task
conditions. A task ‘event’ is a bundle of things: a combination of
task characteristics (because tasks are not characterized by one
feature only) and combinations of task conditions as well. Task
and task condition research has delivered a range of generaliza-
tions. Briefly and incompletely, but relevant to the present argu-
ment, these are:
• structured tasks raise accuracy
• tasks requiring information transformation or integration raise
complexity
• tasks based on familiar and concrete information increase fluency
• pre-task or strategic planning raises complexity and fluency and
to a lesser extent, accuracy
• online planning, if there is room for conceptualizer use, raises
accuracy
• repetition raises CAF
• post-task conditions raise accuracy.
These generalizations relate to the point, just made, that task
characteristics and conditions influence performance separately. In
themselves, they could be consistent with a trade-off interpretation
(with the exception of task repetition, to which we return later in
this section). But the important point is to consider combinations of
task characteristics and conditions. Careful combinations can lead
to increases in more than one dimension of task performance. For
example, Tavakoli and Skehan (2005) and Tavakoli and Foster
(2008) report that structured tasks which also require information
integration (in both these cases linking background and foreground
information) raise accuracy and complexity. The push to accuracy
comes from the structured nature of the task while complexity is
induced by information organization. Foster and Skehan (2013)
report that a complex (interactive decision-making) task allied to a
post-task condition raises complexity and accuracy. The decision-
making nature of the task raises complexity, and the post-task is
important for accuracy. Such results suggest that attention alloca-
tion can be manipulated, and that, within the constraints of the total
amount of attention available, more than one performance area can
be raised. The central claim (and this is important in relation to the
CH discussion below) is that the LAC approach contends that
careful combinations of tasks and conditions are all that is needed
to produce the joint raising of accuracy and complexity, and that it
is the separate but interacting influences that produce this effect.

Repetition is an interesting part of the task literature (Bygate,

2001). Wang (2014) shows how it raises all aspects of performance.
Lambert et al. (2016), in a study with multiple repetitions, also
report positive effects. It appears that repetition is a technique which
enables more effective conceptualization and formulation and that
these translate into raised structural CAF. Lambert et al. (2016) link
successive repetitions with different Leveltian stages. Wang (2014),
in slight contrast, suggests that one repetition, provided that it is
immediate, allows all performance areas to benefit. Her argument is
that the first performance engages the Conceptualizer, the Formula-
tor and the Articulator, and this underpins raised CAF. So this task
condition, in itself, seems to contain all the ingredients needed to
raise three performance areas.
Principle 6: Task difficulty needs to be analysed distinctly for the
Conceptualizer and the Formulator. The LAC approach does use
the concept of difficulty (cf. the different approach taken by the CH
below), and regards this as inherent within the task. Obviously
people will respond to this difficulty differently, but it is contended
that it is useful to regard tasks as more or less difficult than one
another. But the major complication is that what makes a task
difficult in terms of the Conceptualizer may not be the same as
what makes a task difficult in terms of the Formulator. The first is
likely to emphasize the ideas within a task, their accessibility, their
need for manipulation and so on. The second is concerned with
how the SLML is accessed, is adequate and can respond to the
demands that are made upon it by the Conceptualizer. Obviously
this also links with the conditions under which a task is done, e.g.
time pressure, so it is not simply the task itself. But the task itself is
central and influences how the speaker or writer has the SLML
resources to respond to the demands of the task. Potentially, there-
fore, there are separate variations in difficulty for the Conceptuali-
zer and the Formulator. This is developed further in Chapter 9,
when task-based assessment is considered.

Acquisition
LAC addresses this in two ways (Skehan, 2007, 2012, 2013). First, it
is assumed that the CALF categories can represent an acquisitional
sequence, and so tasks which promote greater complexity are pushing
for new language, while tasks which promote accuracy or fluency
are supporting control of an existing interlanguage level. In this view,
first there is destabilization, and then there is a concern for control
(eliminating inaccuracy first, and then achieving fluency second).

But second, and more fundamentally, LAC regards the task itself as
having the important function of making some aspect (or aspects) of
language salient. It is assumed that the teacher records what language
has been made salient in this way, as when some language has been
noticed (Schmidt, 1994), or a gap has been noticed through the
creation of ‘a need to mean’ (Samuda, 2001). Then it is assumed that
important acquisitional work takes place at a post-task stage, where
the teacher can react to the language which has emerged in this way,
and use pedagogic techniques to bring understanding, or extension, or
integration or consolidation, as appropriate. The important point is
that this language is what has emerged when the learner has transacted
a task. The language is not pre-selected, but comes into focus because
of the needs of the learner.

The Research Base

The research base for the LAC approach is not enormous and its
claims focus on consistency with the underlying principles, rather than
on predictions followed by confirmation through controlled experi-
mentation. The initial impetus for the approach, as mentioned, came
from a series of studies done by Skehan and collaborators. A first set of
studies explored issues of task type (personal information exchange,
narrative retelling, decision-making[interactive] tasks) and task
characteristics (e.g. task structure, information organization), as
well as task conditions such as planning and post-task conditions.
The results of the studies were interpreted as suggestive of limited
attentional capacities. In addition, various generalizations were appar-
ent, such as familiarity of information and task structure being associ-
ated with raised accuracy; interactive tasks associated with raised
complexity; information distribution being associated with
raised complexity also; and a series of influences (familiar information
and structure again) with greater fluency. But beyond these general-
izations, there seemed to be tensions such that different performance
areas might be associated with different influences, and if one was
raised, others were often lowered. A major competition for resources,
evident in these studies, was between accuracy and complexity (Foster
and Skehan, 1996; Skehan and Foster, 1997), but competition
between fluency and other areas, usually accuracy, also occurred.
(Lexis did not come into sharp focus in this earlier set of studies, and
so does not figure in any of these claims.) Hence the notion of trade-off
at that time, and more recently the notion of LAC.
The same research team also reported occasions when accuracy and
complexity were jointly raised, which might appear to challenge a

limited attention account. Two studies by Foster and Skehan (1999,

2013) showed this pattern. Skehan (2014c) provides an interpretation
of these results, in effect what is covered as Principle 5: that separate
influences on accuracy and complexity combine to raise each of these
areas. In Foster and Skehan (1999), with several planning conditions,
it was the teacher-led planning which produced the joint raising, and it
was argued that teachers pushed learners to engage with both
accuracy and complexity. In Foster and Skehan (2013) it was the
conjoint influence of a decision-making format (which led to raised
complexity) linked with a post-task condition (which raised accuracy).
There is one important additional point to make here. Skehan (2014c)
argues that to establish that accuracy and complexity are jointly raised
in a way that questions the relevance of limited attention being rele-
vant, one has to report the correlation coefficient between accuracy
and complexity and this needs to be at least reasonably high. This
would indicate that the same individuals can sustain higher accuracy
and complexity. In the studies in question, despite the joint accuracy-
complexity effect evident through group mean scores and inferential
statistics, the accuracy-complexity correlations were either very low or
low. In other words, it appears as though some participants prioritized
accuracy while others prioritized complexity, but not often both. In a
way, this is indirect evidence of a trade-off effect.
More recently there have begun to be studies which explore the
LAC approach from a more prediction-oriented standpoint, exempli-
fying Principle 4 – the connection with the Levelt model. Typical here
is Wang and Skehan (2014). They manipulated the two variables of
task structure and time perspective. The former was operationalized
and interpreted as a general problem-solution structure in a video
narrative retelling compared to non-structured narrative stories. The
latter was operationalized broadly as in the Cognition Hypothesis
(CH), through video-based narrative retellings, simultaneously or
delayed, to achieve time perspective difference. But the interpretation
of the two time perspective conditions was very different to the CH
proposals. The here-and-now condition was interpreted as low in
longer-term memory demands but higher in working memory and
general pressuring demands through less negotiability of content.
The there-and-then condition was interpreted as higher in long-term
memory demands but lower in working memory demands. It was also
regarded as much more negotiable. Predictions were therefore made
that structure and time perspective (there-and-then) would have posi-
tive influences on accuracy and complexity and that they would inter-
act. These predictions were clearly upheld for complexity. Accuracy
was more mixed in that main effects were not evident but there was an

interaction, with the least error in the there-and-then structured con-

dition. So in this case, the LAC approach led to testable predictions,
and reasonable, if not total conﬁrmation.

An Evaluation of the LAC Approach

The LAC approach is essentially bottom-up in nature. It attempts to
account for task characteristic and task condition effects as they
emerge from research studies. Tasks are not seen as the result of
overarching conceptual categories but rather accumulate from a range
of sources. Foster and Skehan (1996), for example, based the three
task types they used (personal information exchange, narrative retell-
ing, problem-linked decision-making) on a survey of language teach-
ing materials. The previous literature is also important in identifying
researchable tasks – as in the impact of R. Ellis (1987) and Crookes
(1989), whose findings on planning were suggestive of the fertility of
this research area. Similarly, post hoc analyses of research results may
be suggestive of important patterns which need to be confirmed and
extended. Typical examples here would be a focus on structured tasks,
as well as the importance of information integration, since these
variables were apparent only after reflection on research results. Any
general analysis of task types then is based on the pattern of studies
and results which accumulate, interpreted through the six principles
that have described.
Inevitably, as results accumulate, and confirm or disconfirm or
modify, so generalizations may change. This, clearly, is a post hoc
approach, which can be criticized as such (Robinson, 2007b). In other
words, the LAC has not been strong on predictions, and accordingly is
weakened in face of the criterion of falsifiability. Indeed, one of the
central concepts, limited attention leading to trade-offs in perform-
ance, is rather elastic. If a particular set of results occurs indicating
trade-offs in performance, a ready-made account is available, just as, if
they do not occur, one can say that attentional limitations were not
relevant and trade-offs did not occur. At the very least, therefore,
replication becomes more important to at least demonstrate that any
effects of trade-off are consistent, even if they are not strongly motiv-
ated by any particular theory.
It is important, therefore, to consider whether the LAC approach
could make predictions and be falsifiable. There are, perhaps, two
responses that can be made in this regard. First, there is the possibility
of what might be termed mini-theories. The range of generalizations
that has emerged from task research may be disappointingly small (see
discussion below), but we do now have several. The various

generalizations can then be the basis for predictions. For example,

structure has been reported as raising accuracy and fluency. Skehan
(2018) offers an analysis as to why this is so, involving the connection
between wider discourse functioning (macro processes) and detailed
focus on the surface of language at the clause level (micro processes).
All of this is based on a particular interpretation, specifically problem-
solution, of task structure (Winter, 1976; Hoey, 1983). But this is only
one way of structuring a task. The prediction could then be that other
forms of structure will produce a similar impact on performance. The
same approach could be taken with other generalizations which have
emerged in the task literature, such as the role of information integra-
tion and transformation. Other tasks could be analysed through these
concepts, exploring other forms of information integration and trans-
formation, and the predictions made for the new contexts which are
based on existing findings.
This is rather opportunistic and particular, though. More relevant
would be a wider framework within which to locate the different sorts
of tasks and conditions which have been used. It is here, once again
(see Principle 4), that the use of the Levelt model (1989), applied to the
L2 case (Kormos, 2006), is vital. The model proposes three stages in
speech production, as we have seen, and it suggests a structure for the
Formulator which makes the SLML central. Already this suggests
ways in which predictions can be made regarding tasks and task
conditions. The mini-theories mentioned, regarding structure and
information organization, fit in nicely. The former, task structure,
unites Conceptualizer work (to develop a macrostructure to the task)
with Formulator work (as attention is more available for SLML
operations, given the easing effect on attention of a broader
macrostructure). The latter, information organization and transform-
ation, makes it clear that Conceptualizer work is going to be more
intensive, as information is manipulated, leading to higher structural
complexity.
Drawing on the Levelt model in this way enables predictions
regarding more than task characteristics and task conditions. The
SLML is, as we have seen, central to the LAC. One implication of this
is that, as proficiency develops, it is likely that the SLML will also
grow, in size, organization, richness and speed. This development has
the potential to recast the relationship between the Conceptualizer and
the Formulator for the L2 speaker. A more effective SLML means that
Conceptualizer demands can be greater and still be met, and Formula-
tion can proceed more effectively. This, in turn, allows an interesting
connection with the CH, which, amongst other predictions, argues
that complexity and accuracy of performance can be jointly raised.

The LAC approach, too, could predict that accuracy and complexity
could be more easily jointly raised – but only at higher proﬁciency
levels. Indeed, this possibility has conﬁrmation in Malicka and Lev-
kina (2012), who report that the accuracy–complexity relationship is
higher in precisely this way.

Part Two: The CH and the SSARC Model

The CH was developed by Peter Robinson to account for task per-
formance, again broadly using a CAF framework. In addition, the CH
makes proposals regarding pedagogic issues, such as task sequencing
and syllabus design, as well as the role of tasks in catalysing feedback
and acquisitional processes (Robinson, 2011). The basic CH approach
underwent significant change with the development of the SSARC
model (Robinson, 2015), and so the present account will first deal
with the earlier phase, which only implicated the CH, before going on
to explicate the newer approach, incorporating SSARC.
The CH has a more prominent foundation in linguistic analysis than
the LAC. In earlier formulations the main influence was Givon (1985)
and functional linguistics. More recently there has been greater
emphasis on cognitive linguistics (Ellis and Robinson, 2008). With
both, the assumption is that communication is meaning-driven and that
linguistic exponents respond to the functional/cognitive demands placed
upon them. In addition, Robinson (2011) takes a radically different
view of attentional functioning to the LAC. While limitations in
working memory are accepted, the same is not so true for attentional
resources. Following Wickens (2007) and Sanders (1998), Robinson
proposes that attention can expand to meet the demands placed upon it
and that it is more appropriate to speak of ‘resource pools’ which can be
drawn on provided that they do not compete for exactly the same
resources. Hence, for example, separate demands on different modal-
ities should not compete for resources. Thus, the trade-off aspect of the
LAC does not function as a constraint in the same way.
As one moves to detail, Robinson proposes a triadic componential
framework, subsuming task complexity, task conditions and task
difficulty. The first of these focuses on the task itself, and its
complexity: this is the central feature of the CH (and see below). Then
task conditions are concerned with participation variables, while task
difficulty involves what the speaker (or writer) brings to the task, a set
of characteristics which may change how the task is approached, given
that some people may have more relevant abilities than others. Each of
these will now be described in more detail. Chapter 5 focuses narrowly
on the learner factors involved in task difficulty.

Task complexity is concerned with the cognitive demands of a task

and how these connect with actual performance. First of all, there is a
major distinction between resource-directing variables and resource-
dispersing variables. Resource-directing variables are hypothesized to
push the speaker towards engagement with language itself. The con-
sequences of this are that accuracy and complexity are both raised,
with this linking with the way attention expands to meet functional
needs. One performance area does not have a negative influence on
the other because each is a reflection of the engagement with lang-
uage. Another consequence is that it is hypothesized that there is
greater likelihood of noticing, and of generating, interactional feed-
back and negotiation for meaning. Resource-directing variables are
exemplified by:
1. time perspective (here-and-now vs. there-and-then)
2. intentional reasoning
3. spatial reasoning
4. causal reasoning
5. number of elements
6. perspective-taking.
In contrast, there are resource-dispersing variables. These do not push
the speaker towards language engagement, but they do affect the
general dispersal of resources. They do not connect with predictions
such as the accuracy–complexity relationship. Their impact on per-
formance is consequently not linked to language itself. Typical
resource-dispersing variables are:
1. planning time, or not
2. single vs. dual task
3. task structure, or not
4. number of steps
5. independence of steps
6. prior knowledge.
The discussion so far is, then, in terms of task complexity and its
impact on language performance. Task conditions, in Robinson’s
model, in contrast, are concerned with the interactional demands of
tasks. These are of two general sorts: participation variables and
participant variables. The former are concerned with the nature of
the task outcome, the number and relationship of the participants in
the task, and the scope for negotiation within the task. The latter are
concerned with how the participants relate to one another, in terms of
things like proficiency level, gender, degree of familiarity with one
another and the relevant world knowledge that each of them has.

Task difﬁculty, ﬁnally, is concerned with learner factors. Note here

that difficulty is not seen as a quality of a task (since it is complexity
that is emphasized in that regard). Instead task difficulty involves the
qualities (both ability and affective) that the participant brings to the
task. Ability factors concern things like working memory, capacity to
switch attention, reasoning ability, mind/intention reading, aptitude
and field independence. Affective variables consist of task motivation,
anxiety, willingness to communicate, openness to experience, control
of emotion and self-efficacy. The difficulty of an actual task is then a
result of the interaction between task complexity and task difficulty
factors.
The heart of the CH is task complexity, but it is clear that the
enumeration of task condition and task difficulty variables contains
massive potential for a research programme. Each of the condition or
difficulty variables represents a hypothesis as to what influence they will
have, e.g. the effect on performance of a task with an open solution, or a
convergent solution and so on (and see discussion in Chapter 2).
Largely, this potential research programme is embryonic rather than
realized, but there are many interesting possibilities here. In the SSARC
model, the task difficulty variables are discussed a little more extensively
than before, and linked to different ways that task complexity resource-
directing variables might function. Several of the resource-directing
variables involve reasoning, and the task difficulty reasoning-ability
variable is related to these, with the suggestion that tasks which require
different sorts of reasoning might function more effectively with partici-
pants higher in reasoning ability. This is, as yet, untested.
The CH model has generated a considerable amount of research (see
below). It has lent itself to predictions and to research designs which
probe, particularly, the impact of resource-directing variables. But
there are pedagogic implications of the CH also and, more recently
(Robinson, 2015), the CH has been extended in this direction. Two
principles pave the way for this, and each is concerned with task
sequencing:
Task Sequencing Principle 1: Only the cognitive demands of tasks
relating to intrinsic conceptual and cognitive-processing complexity
(i.e. resource-directing and resource-dispersing variables) are
involved in task sequencing. Task condition and task difficulty
variables, while important, do not influence sequencing itself.
Task Sequencing Principle 2: In sequencing tasks, resource-dispersing
variables should be increased first, and only then should resource-
directing variables be increased. The intention here is to guide
learners from the known, through the development of automaticity,

to the need to develop new form-function mappings and to

restructure-complexify.
This leads to the SSARC model, expressed in three equations, as in
Robinson (2015, p. 94):
Step 1, SS (stabilize, simplify) = i x e [(‘s’rdisp) + (‘s’rdir)]n
Step 2, A (automatize) = i x e [(‘c’rdisp) + (‘s’rdir)]n
Step 3, RC (restructure, complexify) = i x e [(‘c’rdisp) + (‘c’rdir)]n
where i = current interlanguage state, e = mental effort, ‘s’= simple task
demands, ‘c’ = complex task demands, rdisp = resource-dispersing
tasks, rdir = resource-directing tasks and n = potential number of
practice opportunities on tasks.
To expand on these equations:
SS (Simple, stable): This level involves the use of current interlanguage.
Demands from the task are kept low so there are low resource-
dispersing demands and low resource-directing demands.
A (Automatization): The point here is to increase resource-dispersing
demands. This in turn is seen as promoting speedier access to
resources and also automatization.
RC (Restructuring, complexifying): At this stage, the increase in
resource-directing demands is hypothesized to promote restructur-
ing and lead to new form-function mappings. Increasing complexity
then is intended to destabilize interlanguage.
The development of the SSARC model is, then, intended to expand the
CH and make it relevant to pedagogic decision-making regarding
syllabus design.

Acquisition
The CH offers two basic inﬂuences on acquisition. First, there is a
performance-supporting role. Task complexity pushes for raised
complexity and accuracy as attention responds to functional needs.
In this way the learner is pushed to develop abilities to use language,
and possibly, through the greater complexity of language, to
restructure and become a more effective communicator. Second, tasks
of greater complexity are seen as more likely to generate negotiation of
meaning and noticing. In other words, they are seen as nurturing the
sort of personalized, timely feedback that the Interaction Hypothesis
(IH) advocates, as described in Chapter 2.
The development of the SSARC model extends this picture. Recall
that the sequence embodied in this acronym is:

Table 3.1 Potential sequences in the SSARC model

Possible Possible Possible

sequence 1 sequence 2 sequence 3
Low res. dis., low Structured, here- Planned, few Few steps, no
res. dir. and-now elements reasoning
High res. dis., Unstructured, Unplanned, few More steps, no
high res. dir. here-and-now elements reasoning
High res. dis., Unstructured, Unplanned, more More steps,
high res. dir. there-and-then elements reasoning

1. stabilize, simplify
2. automatize
3. restructure, complexify.
These stages, portrayed through equations, also have terms in the
equations for the amount of potential effort, and the number of
practice opportunities at each stage. This system is then consistent
with the sorts of sequences exempliﬁed in Table 3.1.

The Research Base

It is too early to draw on a research base for the SSARC model so the
focus here is on the CH. The CH is more straightforward to deal with
in this regard. This is because it makes some predictions which are
much clearer than the LAC approach, principally that task complexity
will raise both accuracy and complexity.
There are two major strands to the research base: individual
research studies and meta-analyses. Regarding individual studies, the
CH has generated a very large number of examples. Typically the
research design is to manipulate one or more of the resource-directing
variables to explore whether the more complex task (judged on this
basis) generates raised accuracy and complexity. Sometimes a
resource-directing variable is combined with a resource-dispersing
variable, although this can be a difficult design to manage because
the two types of variables can make conflicting predictions (Inoue,
2016). Typical studies of this sort (and there are many more, but
essentially more of the same) are Niwa (2000: cited in D. Ellis,
2011), where task complexity only influenced lexical density; Robin-
son (2001), where accuracy and fluency were affected by task com-
plexity, but not structural complexity; Kuiken and Vedder (2008),
where accuracy was affected but structural and lexical complexity

were not; Gilabert, Barron and Levkina (2011), where again accuracy
was affected but complexity was not; and Michel (2011), where no
effects were found. The generalization which emerges from these (and
many other studies) is that there is a link between task complexity and
accuracy, but that is all, and finding studies which report increases in
both is very difficult to do, although Ishikawa (2007) does report such
results, albeit for a writing-based study. This consistency of results is
important, since a central and innovative claim of the CH is that
accuracy and complexity can be raised simultaneously.
There have been two major meta-analyses of the CH. These will be
developed at greater length in the first of the Issues sections (on Tasks
and Conditions), and will only be summarized here. Jackson and
Suethanapornkul (2013) report results which essentially focus on
resource-directing variables, and even within these, largely the variable
of time perspective. They report average effect sizes of 0.16 for
fluency, 0.28 for accuracy and 0.02 for complexity as well as 0.03
for lexis (all Cohen’s d). This confirms the pattern given. There is no
joint raising and the one effect size of note has to be regarded as small.
A much larger meta-analysis, with more studies and a wider range of
variables, is provided by Malicka and Sasayama (2017). They ana-
lysed enough studies to be able to provide average effect sizes for two
resource-directing variables. With time perspective they report values
of 0.03 for fluency, 0.15 for accuracy and 0.41 for complexity. Lexis
gives an average effect size of 0.12. With reasoning demands, the
values are 0.12, fluency; 0.12, accuracy; 0.09, complexity; and
0.34, lexis. Again the typical result is a close-to-zero effect size, with
the exceptions of 0.41 for complexity, time perspective, and 0.34 for
lexis, reasoning demands. There is also the point that the pattern here
does not totally agree with Jackson and Suethanapornkul’s (2013)
findings: they found their higher value for accuracy (principally based
on time perspective) whereas Malicka and Sasayama (2017) report
their highest value for complexity, also for time perspective. The
general conclusion has to be that resource-directing variables do not
generate high values, that there is inconsistency in the results. Basically
there is little evidence of joint raising of accuracy and complexity.

An Evaluation of the CH/SSARC Model

The ﬁrst point to make is that the CH has been extremely successful in
generating research. The number of CH-oriented studies is now more
than a hundred. Connected to this, a considerable strength of the
hypothesis is that it makes predictions, particularly that of task
complexity raising accuracy and complexity together. The main

hypothesis also has interesting supporting hypotheses, such as the

functioning of attentional resources. All of these features have to be
evaluated positively.
But there are also some problems with the CH, both logical and
empirical. First, the distinction made between resource-directing and
resource-dispersing variables, central to the hypothesis, is not entirely
clear. There are resource-directing variables which do not seem
to sit comfortably with the others. Time perspective, for example,
which contrasts the (less complex) here-and-now condition with the
(more complex) there-and-then condition, on the basis of memory
demands, could be viewed quite differently to the CH interpretation.
The former condition may also have greater and less avoidable input-
processing problems, through the presence of visual material, with all
its detail. The latter, although requiring memory, enables considerable
negotiability in how to select and organize material. Different per-
formances might be down to these non-resource-directing influences.
More generally, there is scope to clarify the unity that is intended to
link the various resource-directing variables. There is an obvious
connection between the different reasoning-based variables (inten-
tional, causal, spatial). But the relationship of these to perspective-
taking and number of other elements would benefit from greater
clarification, as would how all these variables impact upon language
involvement. Conversely, it could be argued (Skehan, 2015, 2018)
that the resource-dispersing variables of planning and task structure
could lead to language involvement akin to resource-directing vari-
ables. Qualitative research with planning (Pang and Skehan, 2014)
suggests that some people use the time to engage with ideas (which
have language implications) or even language itself. Structure may
push speakers to engage with the language required to do justice to
this design feature. The problem-solution structure, a key feature of
LACH research studies (Tavakoli and Skehan, 2005) is fundamentally
concerned with causal reasoning. So in these cases it is difficult to
argue that a resource-dispersing variable is only concerned with pro-
cessing. One has to conclude that there is potential for realignment of
what should be considered resource-directing and resource-dispersing.
There is also the issue of evidence. As we saw in the section ‘The
Research Base’, the clear majority of studies report a resource-
directing variable impact on either complexity or accuracy, usually
accuracy. This applies when one examines individual studies. It also
applies to the meta-analyses which are available; there is a slight
contradiction between Jackson and Suethanapornkul (2013) and
Malicka and Sasayama (2017) in that the former reports a
small accuracy effect but no complexity effect, while the latter larger

meta-analysis reports the reverse. This is damaging for the CH since

the hypothesis predicts joint raising as central to the hypothesis. To
discover that only one area is raised suggests little more than the
proposition that tasks and task condition can influence performance,
which is, really, the starting point for all task research.
It is too early to offer much of an evaluation of the extension to the
SSARC model. The model seems largely a set of proposals which need
fleshing out. For example, the formulae which are used are currently
rather opaque, without any detail regarding terms such as ‘e’ for
mental effort and ‘n’ for number of practice opportunities. In addition,
while the model clarifies how one might sequence particular combin-
ations of resource-directing and resource-dispersing variables (e.g.
time perspective and planning), it does not clarify how particular
combinations like this would fit into a longer sequence as one would
expect from a practical syllabus. Finally, the sequence of stabilize –
simplify – automatize – restructure – complexify seems unclear, and
possibly conflicts with some theories of learning. It seems to propose
automatization before restructuring-complexifying, rather than the
other way around. Perhaps connected with this, it lacks clarity as to
what happens after the restructuring-complexification has taken place.
One wonders if there is need for consolidation or practice at this point.

Comparing the Two Approaches

As a first view of the comparison between the two approaches,
Table 3.2 presents a series of categories drawn from questions raised
in the introduction to this chapter, and then a capsule description of
the LAC approach and the CH/SSARC model in each case. The table
will be the basis for the subsequent discussion.
The first two categories in the table bring out clear differences
between the two approaches. What comes across is that the LAC
approach is firmly rooted in psycholinguistic approaches to L1 pro-
duction applied to the L2 case while the CH emphasizes more linguis-
tic approaches of a functional/cognitive persuasion, and then does not
elaborate any particular model of speaking as relevant (though see
Kormos, 2011). The hinterland for each approach, in other words, is a
long way apart.
The different performance influences similarly show marked differ-
ences. The LAC approach is somewhat opportunistic and accumulates
influences through research findings. There are some more general
categories (complexifying, pressuring, easing, focusing: see Skehan,
2009c) and these are relatable to the extension of the Levelt model
but there is no subsystem of influences. In contrast the CH has, as

Table 3.2 A comparison between the LAC approach and the CH/SSARC model

Area of similarity/
difference LAC CH/SSARC
Linguistic underpinning/ • Levelt model is central • Cognitive linguistics
model of speaking • No specific linguistic analysis • Little emphasis on a psycholinguistic model of
speaking (but see Kormos, 2011)
Analysis of memory and • Both working memory and attention are limited • Expandable attention
attention • They are a constraint that has to be worked • Resource pools
around
Performance and • Task and task condition are individual • Fundamental role for resource-directing and
influences upon it influences which then combine resource-dispersing variables
• Influences largely emerge through research • Prediction of accuracy –complexity relationship
• The constraint of limited attention pervasive influenced by task complexity
Measuring performance • Generalized measures of CALF • General and specific measures, usually of CAF
• Some importance for noticing, feedback and
interactional moves
Acquisition • Transacting tasks makes language salient. This • Resource-directing tasks push for greater
language needs to be recorded and worked on interaction and uptake of task-relevant input.
at the post-task stage SSARC sequence to promote interlanguage
development
Pedagogy • Methodology influenced by above view on • Clear SSARC sequence advocated in focused
acquisition areas within syllabus design, but little on a
• No suggestions on syllabus design general syllabus
88 Theoretical Perspectives

central, the contrast between resource-directing and resource-

dispersing variables. There are three aspects to this. First, there is the
different functioning of the two classes of variable, one claimed to
have language implications, and the other claimed to have processing
implications. Second, there is the issue that the different variables
within the resource-directing heading are seen as similar, just as the
variables within the resource-dispersing heading are seen as similar to
one another, but obviously different from the resource-directing vari-
ables. The third CH aspect in relation to performance concerns general
vs. specific measures. LAC, as we have seen, focuses on general meas-
ures. These are an important component for the CH/SSARC model,
but, in addition, there are proposals that particular measures should
also be used. So it is proposed that a resource-directing variable like
complex reasoning links with cognitive state terms, or spatial
reasoning with expression of motion events, or time perspective to
tense and aspect in the present (here-and-now) compared to events
happening elsewhere in time and space (there-and-then).
The two approaches have fairly distinct approaches to the process
of acquisition. LAC addresses this in two ways (Skehan, 2007, 2012,
2013). First, it is assumed that the CALF categories can represent an
acquisitional sequence, and so tasks which promote greater complex-
ity are pushing for new language, while tasks which promote accuracy
or fluency are supporting control of an existing interlanguage level. In
this view, first there is destabilization, and then there is a concern for
control (eliminating inaccuracy first, and then achieving fluency
second). But second, and more fundamentally, LAC regards the task
itself as having the important function of making some aspect (or
aspects) of language salient (Willis and Willis, 2007). It is assumed
that there will be a record of what language has been made salient in
this way, as some language has been noticed (Schmidt, 1994) or a gap
has been noticed through the creation of a need to mean (Swain, 1995;
Samuda, 2001). Then it is assumed that important acquisitional work
takes place at a post-task stage, where the teacher can react to the
language which has emerged in this way and use pedagogic techniques
to bring understanding, or extension, or integration or consolidation,
as appropriate. The important point is that this language is what has
emerged when the learner has transacted a task. The language is not
pre-selected but has emerged because of the needs of the learner. In
contrast, the major emphasis within the CH/SSARC is towards the use
of interaction leading to feedback and noticing. It is assumed that
greater task complexity through resource-directing tasks leads to a
greater engagement with language and that part of this will mean that
learners will be more likely to negotiate, more likely to provoke and

process relevant feedback, more likely to notice and, therefore,

following the IH (see Chapter 2), more likely to make progress.
Linked to this is what can be said about syllabus, and the
sequencing of pedagogic activities. Essentially, the LAC has very little
to say here. There is a broad commitment to tasks varying in difficulty
(note: not Robinson’s analysis of difficulty – more his analysis of
complexity), and the suggestion that simple tasks should come first
and lead subsequently to greater difficulty. In this respect, the range of
task characteristics and task conditions would come into play here.
One interesting point is that LAC proposes that difficulty is different
for the Conceptualizer and Formulator, and this would then have an
impact on how pedagogic tasks might be sequenced. Complexification
and destabilization would come first, and this would be followed by
the development of control. In contrast, the CH/SSARC has more to
say about sequencing. This is particularly so in the SSARC model, as
described, with the move from low resource-directing and dispersing
to low resource-directing but high resource-dispersing, to high
resource-directing and dispersing. This is linked to the engagement
of existing interlanguage at the first stage, then the development of
automatization, followed by a push for restructuring and complex-
ification and a challenge to the existing interlanguage system.
There are, then, many points of difference between the two
approaches. They differ clearly in their connections with linguistics,
psycholinguistics and attentional functioning. They also contrast in
their views of acquisition (provoking and using feedback compared to
counterpunching at the post-task stage). But there are other areas
where, although differences are clear, there might be scope to argue
that the differences may not be as great as first sight would suggest.
Both, for example, are committed to basing claims about tasks on
research evidence. Both, also, are committed to understanding how
tasks have an impact on performance, and also how they might be
used in pedagogic contexts.
An interesting way of bringing them closer together is to relate them
to proficiency level, as we have seen, and also modality. LAC essen-
tially offers a view of language production where Conceptualization
makes demands upon the Formulator which are met, or not, through
the operation of the SLML. The impact of attentional limitations is
clearest at this point. If the SLML is not large enough, or rich enough,
or fast enough, processing difficulties will occur. When this happens,
parallel performance is derailed, and the speaker (or writer) has to deal
with the problem which has arisen (need to find alternative expression
because a needed lemma does not exist; repair because an existing
lemma does not contain enough information, e.g. syntactic frame,

collocation; or repair because the Formulator does not handle its work
in time). These problems, first of all, are not so acute when writing is
involved. One of the central claims of the CH is that accuracy and
complexity are jointly raised, in response to (resource-directing) lan-
guage demands. When less communicative pressure is involved, as in
writing, or online planning, this may be more achievable. Exactly the
same would apply at higher proficiency levels. As the SLML grows (as
proficiency grows) the sorts of problems that the speaker has when
Conceptualizer demands create pressure for the Formulator, will be
lessened and so parallel function (and perhaps jointly raised complex-
ity and accuracy) is more likely. One could even add to this a task
difficulty factor from the CH: if working memory is greater, this too
might ease the functioning of Formulator use during communication.
Indirect evidence for this comes from studies (reviewed in Skehan,
2018) that working memory has an impact with online planning but
not with strategic planning. So one can conclude that the two
approaches may not be so far apart on the accuracy-complexity issue,
depending on the proficiency level, modality and relevant individual
differences.

Issues in Psycholinguistic Analyses of Tasks

So far the discussion has been very much through the lenses of the
LAC approach and the CH/SSARC model, considered separately and
also together. The two approaches have been inﬂuential and generated
a lot of research, generally from one perspective or another. But a
good deal of research has been conducted from more neutral stand-
points, and interesting issues have emerged. Five such issues will be
considered in this last section.

Tasks and Conditions: Perspectives from Meta-Analysis

It is useful, at the outset, to explore the general evidence relating to the
effects of tasks on L2 performance. Malicka and Sasayama (2017)
report on a meta-analysis of task-related studies. The motivation for
this research was to gain greater understanding of task complexity,
the central part of the CH. But the analysis makes important contri-
butions, beyond its speciﬁc focus on the CH, because it includes a
wide range of variables, including resource-dispersing variables
(in contrast to Jackson and Suethanapornkul (2013) who focused
only on resource-directing variables in their meta-analysis). Malicka
and Sasayama’s (2017) results are shown in Table 3.3. The table
shows the results arranged by CALF area, in columns, and then

resource-directing and resource-dispersing variables, in rows. The

values shown in the table are effect sizes, using Cohen’s d. The data
represents the most comprehensive analysis presently available of
general task effects on performance. By way of clarification, with
resource-directing variables, a negative sign means that the variable
in question reduces the level of performance. With resource-dispersing
variables, a negative value indicates that the ‘simpler’ condition
(planning, a repeated task, a structured task, a task with familiar
information, a task with support) produced the higher level of
performance.
Several generalizations emerge from Table 3.3. The analysis which
follows takes effect sizes greater than 0.30 as worth commenting on.
1. Resource-dispersing variables clearly have stronger and more con-
sistent relationships with performance than do resource-directing
variables.
2. None of the variables, resource-directing or resource-dispersing,
raise all performance areas. Resource-dispersing variables, in gen-
eral, raise three areas (CAF) as do the specific resource-dispersing
variables of repetition (CAF) and task structure (C, L, F).
3. No resource-directing variables have large effect sizes for accuracy
whereas two resource-dispersing variables (planning, repetition) do.
4. Resource-directing variables are rather inconsistent in their effects,
with reasoning raising lexis, and time perspective (i.e. the there-
and-then condition) raising structural complexity.
5. In general, modality (i.e. oral vs. written performances) suggests that
written performances respond more positively to the range of vari-
ables for complexity; that accuracy shows lower response, and that
fluency is more affected by resource-dispersing variables. (However,
it is important to point out that the measurement of fluency is most
altered when one moves from oral to written modes.)
The results broadly agree with the existing literature but with some
points where additional claims might be made. After all, Malicka and
Sasayama’s (2017) work derives from a meta-analysis and this means
that a sufficiently large number of studies on a particular task feature
have to be available to enable any sort of effect-size-based generaliza-
tion. There is room, in other words, to draw on other studies as well,
which probe variables not investigated sufficiently widely to be
included in the meta-analyses. So, based on such a wider reading,
and on a synthetic basis, one might add:
• tasks based on more concrete information tend to raise fluency
• tasks which require integration of information tend to raise
complexity

Table 3.3 Malicka and Sasayama’s analysis of CH-linked variables

Task feature Complexity Accuracy Lexis Fluency

Resource-directing 0.13 0.13 0.28 0.09
Reasoning 0.09 0.12 0.34 0.12
HnN/TnT 0.41 0.15 0.12 0.03
Referential demand 0.08
Resource-dispersing 0.77 0.73 0.27 0.34
Planning 0.88 0.87 0.21 0.25
Repetition 0.57 0.61 0.11 0.59
Structure 0.53 0.16 0.5 0.62
Familiarity 0.12 0.19 0.22
Support 0.56 0.66
All Oral All Written All Oral All Written All Oral All Written All Oral All Written
Tasks Tasks Tasks Tasks Tasks Tasks Tasks Tasks
Resource-directing .02 .26 .17 .06 .34 .19 .10 .07
(all combined)
Resource-dispersing .75 .91 .80 .47 .24 .33 .27 .83
(all combined)

Note: HnN/TnT = Here-and-now/There-and-then.

Psycholinguistic Perspectives 93

• tasks which require transformation of material also tend to raise

complexity
• post-task conditions raise accuracy, and sometimes complexity.
In addition, there are some inconsistencies between other research and
the findings reported in Table 3.3. Jackson and Suethanapornkul
(2013, p. 15), in their meta-analysis, also motivated by the CH, and
based only on resource-directing variables, focused mainly on time
perspective. They reported mean effect sizes of complexity: 0.02;
accuracy: 0.28; lexis: 0.03; and fluency: 0.16. The accuracy effect
here may not be large but does contrast with the smaller finding in
Malicka and Sasayama. In a reverse direction this is also true for
complexity, but here Malicka and Sasayama (2017) report the higher
value, at 0.41. Even so, the prevailing impression from Jackson and
Suethanapornkul (2013) is of quite low mean effect-size values.
The most important conclusion one can draw from Table 3.3 is that
the resource-dispersing variables generally are associated with consist-
ently higher effect sizes: planning (strategic and online) and repetition
are very clearly involved here, as are task support and task structure.
In fact, though, three of these resource-dispersing variables are task
conditions, and so it may be more appropriate to think of them as
more clearly defined by this aspect. This leads to the possible claim
that task conditions are more consistent influences on performance – a
point developed in Skehan (2016). Correspondingly, and more con-
tentiously, task characteristics themselves, with exceptions, such as
structure, have been something of a disappointment. It is to examining
why that might be so that we now turn. First we will re-analyse the
notions of tasks and task characteristics. Then we will explore the
nature of task conditions more generally.

The Feasibility of Analysing Tasks

In a study of the impact of the resource-directing variable concerning
the number of participants, Sasayama (2015) used narrative picture
series to compare performance with the four values of one, two, four
and nine participants. She argued that most studies with this variable
typically only use two values, and with possibly no great difference in
the number of participants. Using a range from one to nine was then a
response to this – a greater overall range plus the inclusion of inter-
mediate numbers of participants. Although she did ﬁnd a broad effect
for the number of participants, this general ﬁnding was complicated by
the issue of clarity. For the manipulation to work, the number of
participants in a picture series has to be interpreted in the way that
was intended. This was not the case in her study, with the result that

there was confusion as to the roles and relationships involved in some

of the picture series. The largest number of participants did make a
difference, but the intermediate values presented much less clear a
picture, since misinterpretation (given the time available for the task)
occurred.
Another issue emerges from work by Inoue (2016). She explored, in
a testing context, whether two picture narrative retellings could be
established to be equivalent. Despite careful attention to task design,
she found that although in many ways the picture series did seem to be
equivalent, the performances that they elicited differed in signiﬁcant
ways (and this would have implications for their satisfactoriness in a
testing context). It turned out that in one case the amount of inference
was greater, while in the other, in addition to greater clarity of how the
story developed (cf. Sasayama, 2015), one could analyse the story as
being structured (i.e. there was a disguised variable involved). These
unforeseen aspects of the two narratives led to different methods of
using complex language. The ‘structured’ narrative led to more
subordination while the higher-inference narrative elicited higher
words-per-clause measures. So once again, two initially well-analysed
narratives produced different results through unforeseen design
factors, and these were only revealed clearly by careful data analysis.
The studies by Sasayama (2015, 2016) and Inoue (2013, 2016)
provide particular examples of how tasks may not produce the per-
formances that were intended. The Sasayama (2015) study provides
another insight because she also used qualitative techniques to probe
why her participants did not respond in the predicted way to the one,
two, four and nine participants in the picture-based narratives they
were given. This data was very revealing. In general, the participants,
if they saw clarity in the narrative, got on with telling the story. If they
did not, they simply made the best of things, and developed hypotheses
about the characters in the stories, their relationships and their
motivations. If something did not seem to ﬁt in to their interpretation,
the participant ignored it. The question then arises as to whether these
two studies, each very careful investigations of task functioning, were
simply untypical in the unforeseen results they obtained, or whether
they connect with more general and fundamental issues in the use
of tasks.
There is a link here with a major generalization that comes out of
contemporary cognitive psychology. Kahneman (2011) offers a gen-
eral analysis of problem-solving and typical cognitive behaviour.
He distinguishes between two systems. System 1 is intuitive, fast
and does not require much effort. System 2 is logical, slower and
requires conscious effort and thought. System 2 is, of course,

fundamentally superior (if slower), but human beings are lazy. The
result is that we default to System 1 even when this is not appropri-
ate. Worse, when we are confronted by a difficult problem, a typical
response is not to engage System 2 and solve the difficult problem in
a logical way. Instead we change the problem to make it easier, and
best of all, easy enough for System 1 to (appear to) work. Kahne-
man (2011) reviews a massive literature in cognitive psychology
which supports these claims. In L2 task-based performance, we
need to take this literature seriously. Its implication is clear, and
entirely consistent with Inoue’s (2013) and Sasayama’s (2016) find-
ings – the more a research design introduces more difficult tasks,
especially ones based on cognitive analysis and rather subtle differ-
ences between conditions, the more likely it is that their will not be
a focus on the full range of the details of the task (since this would
require System 2), and instead System 1 will be brought into play.
This is not to say, remotely, that these designs are wrong. It is,
fundamentally, the problem that the predictability of tasks is by no
means an exact science. This difficulty can be overcome, to some
extent. Piloting may give some confidence if it can demonstrate that
the intended task and the actual task coincide sufficiently. Gathering
qualitative data can also help in this regard, and even give indica-
tions as to how a task can be modified to achieve the intended goals.
But for now, the possible lack of convergence of intended and actual
tasks remains a nagging worry.
There are still two troublesome aspects of task research that need to
be considered. The first of these concerns the importance of negoti-
ability. There is always the paradox that the essence of taskness is that
participants are expressing meanings, and if this is the case, there is the
likelihood that they will want to express their own meanings! But in
doing this, they may take a task away from what the task designer
intended. But also, if a task participant is able to exert some personal
control and direction when they are doing a task, in addition to
introducing a lack of standardization in the language the task elicits,
there is also the major factor that the speaker can nudge the task into
areas where they feel more comfortable, or more knowledgeable, or
have relevant linguistic resources. They may not be so ‘imprisoned’,
that is, within the tight designs of a particular task (e.g. a narrative
picture series). This may not be so important within a study based
on only one task type, but where different task types are compared,
if one task type has potential for negotiability while the other does
not (or less so), this has a problematic effect on comparisons.
Consider, for example, a here-and-now task, where the stimuli are
clearly present, and have to be attended to, with a there-and-then task,

where the stimuli are no longer present, but there is the potential to
shape the discourse more freely. The contrast may then not simply
be time perspective, but also the speaker’s capacity to shape the
discourse.
But it gets worse! A sociocultural account of tasks (see Chapter 4)
would argue that a task, through the interactions of the participants,
has a life of its own, and can be reinterpreted in whatever direction the
participants want to develop. We are back to the point that if a task is
an opportunity to express meanings, then the meanings which are
expressed cannot be preordained or constrained. As just discussed,
the point about negotiation concerned freedom to use whatever lan-
guage elements are preferred. But it was assumed that the broad
parameters of the task were accepted. Here, in contrast, we are con-
cerned with ways in which the task itself is modified, transformed,
ignored or subverted (Coughlan and Duff, 1994). Participants might
simply redefine the task. It can be argued that, if a task is worthwhile,
this is inevitable!
In summary, we see that:
• tasks may be unclear;
• tasks may contain hidden variables other than those under
investigation;
• participants may approach a task from a very different perspective
to the task designer;
• tasks which enable participants to shape the language that they use
may be easier to do, and more richly done, than tasks which
constrain tightly;
• tasks themselves may be changed, subverted and even ignored.
If we reflect upon the findings reported in Table 3.3, one can come up
with the generalization that tasks are a perilous area within which to
research! Effects have been found, with some consistency, but perhaps
less often than one would wish, given the amount of theorizing that
has gone into task design linked to task performance. This is a major
challenge for the future of task research.

Task Conditions
In contrast to task characteristics, the task condition variables in
Table 3.3 (planning, repetition, support) showed appreciable effect
sizes with more than one aspect of performance (taking appreciable
to mean an effect size of 0.30 or more). The exception with task
characteristics, which we will return to at the end of this section, is
task structure, which showed medium effect sizes with three areas

(complexity, lexis and ﬂuency). This raises the interesting question as

to why these conditions-linked variables should have such consistent
and appreciable effects.
Skehan (2016, 2018) makes two general suggestions as to why this
is so. The first has already been touched upon: the flexibility that is
associated with conditions regarding the content of what is said or
written. Generally, with planning or repetition, the task concerned is
likely to allow the speaker (or writer) to make choices and select the
emphasis or stance to be used, as well as the content to be included. In
turn, there is the possibility that where language itself is concerned,
there is greater opportunity to shape what is said so that it fits areas of
strength, whether these be organized and possibly previously used
ideas, or the specifics of language, of lexis and so on. The result is
that planning or repetition provide greater opportunity to polish what
is said and to achieve higher levels of performance.
The second point is that condition-type research is more likely
to link naturally with models of speaking. Given the stages of Concep-
tualization, Formulation (syntactic encoding, lemma retrieval) and
Articulation, and the associated process of monitoring (Levelt, 1989,
1999), one can quickly sketch out possible linkages. Strategic planning
enables more effective Conceptualization, as the ideas in the task are
wrestled with and developed. It also facilitates processes of retrieval
and rehearsal (Ortega, 2005), which, if retained, will advantage the
lemma retrieval and syntactic encoding components of the Formula-
tor, as well, possibly, as the Articulator. Online planning can then ease
the operations of the Formulator, as lemma retrieval has a little more
time to function as a deeper and more extensive process, and also,
possibly, as the speaker has time to adapt the message to avoid diffi-
culties. Repetition has all the advantages of strategic planning (organ-
ization of ideas and rehearsal). But in addition the first performance
(or earlier performances if they are multiple) pushes the speaker to
confront the limitations of shallower lemma retrieval, and possibly
leave traces which are undeveloped in the first performance but can be
exploited subsequently. The repeated performance can then be based
more firmly on more native-like lemma retrieval and more efficient
Formulator operation. In addition, and very distinct from strategic
planning, the first performance will have engaged the Articulator and
sensitized the speaker to problems that occur at that stage of speech
production (Wang, 2009).
One other condition in task-based research, not covered in Malicka
and Sasayama (2017) because of the small number of studies, fits this
analysis quite neatly. It is to explore the effects on task performance
itself of anticipation of a post-task. In a series of relevant studies

Skehan and Foster (1997), Foster and Skehan (2013) and Li (2014)
have shown that if speakers doing a task know that there will be a
post-task (needing to engage in a public performance; being expected
to transcribe a recording of their own performance doing the task)
they are likely to be more accurate in the actual task performance, and
with decision-making, interactive tasks, more complex also. Again, in
Leveltian terms, one can relate the higher accuracy (and complexity) to
the process of monitoring where participants have been induced to
focus on form to a greater degree than they otherwise would have
done. Anticipation of what is to come changes the focus of attention
during the actual task performance. Once more, a standard model of
speaking illuminates the way a task condition can have an effect on
performance.
Interestingly, the task characteristic highlighted here as being con-
sistently successful – task structure – is analysable in Leveltian terms.
The very fact of structure pushes the speaker to be concerned with the
connection between elements (indicating relationships, expressing
causality), essentially describable in resource-directing terms, as
argued earlier in this chapter. This is not particularly Leveltian. But
also important is the complementary inﬂuence on macro and micro
processes. The broad macrostructure of a task provides an organizing
framework for the speaker which guides overall structure. This then
allows the speaker to see a clearer relationship between the overall
message and the details of what are being said at any particular time.
This enables more attention to be directed at the surface of language –
essentially a Formulator operation. It also enables the speaker to
recover from any glitches in speaking (e.g. because of lemma retrieval
difﬁculties) and make links between Conceptualizer and Formulator
(and regain parallel processing after it has been knocked off course).
So a large part of the advantage of structured tasks can be related to a
model of speaking also.

The Challenge of Calibrating Task Complexity

Both the LAC and the CH indicate important roles for task com-
plexity. The LAC is concerned with difficulty, and distinguishes
between Conceptualizer-linked difficulty and Formulator-linked dif-
ficulty. But in each case, task demands are seen as having an impact
on the structural and lexical complexity of the language which is
induced by a task. The CH/SSARC approach makes task complexity
central and the driving force for language development and syllabus
design, with resource-directing influences central for interlanguage
change.

It has to be admitted, though, that in both cases the approach has

been to analyse tasks from what might be euphemistically called a
‘logical’ perspective and then make inferences about their complexity.
The optimistic hope has been that research results would confirm the
analyses which were the starting point for research studies. Clearly the
pattern of research results in the literature has been rather disappointing
in this regard. There are successes but there are a fairly extensive list of
failures, as the meta-analyses and more narrative accounts of research
are beginning to show. This, in turn, raises the obvious question as
to why the starting point for these enquiries might not have been better
if, first, task complexity had been researched, and then, once tasks of
different complexity have been empirically established, more theory-
based research studies could have been conducted more effectively.
More recently a range of research studies have indeed taken this
more foundational approach. Central here is the need to have some
method of establishing task difficulty/complexity that is independent
of initial theorizing, and a key component of this is the concept of
cognitive load (Sweller, 1988, 1994). This concept originated in cog-
nitive psychology, and proposes that different tasks can exert a differ-
ent load on the cognitive and psycholinguistic processes that underpin
speech. If greater cognitive load can be identified for a task, then we
have an independent means of making claims about task difficulty/
complexity. In that respect a range of secondary measures have been
proposed which could function as indices of cognitive load. These
include:
• self-report of, for example, mental effort required by a task, or of
the difficulty of the task
• expert analysis
• time estimation
○ where participants are told ahead of time they will be asked to
estimate how much time a task has taken
○ where participants are not warned they will be asked to estimate
time, but are indeed asked to do so after task completion
• secondary tasks, where participants need to monitor some add-
itional task, such as detecting some sort of colour change. Reaction
time and accuracy in doing so are the major indices
• eye-tracking
• pupillary response, for example where the amount of pupillary
change is measured.
Such techniques are being used more frequently in psycholinguistically
based task research. The approach is little more than ten years old, but
now an increasing number of studies is beginning to convert the

original promise of such measures into more dependable results. So it

may be possible to make substantial claims of task difficulty/
complexity other than general theory or inferences made only after
research results are obtained. The result of this is that it may be
possible to evaluate competing claims, e.g. the LAC approach, or the
CH/SSARC model, in a more defensible manner.
At present it is too early to offer any major conclusions because,
although research using such techniques is growing, we are not yet to
be able to offer wide-ranging generalizations. A few points are worth
making, though. First, there is broad congruence, and therefore
encouragement, in the way different measures do show some conver-
gent validity. Révész, Michel and Gilabert (2016), for example, report
that accuracy in a colour-change secondary task, self-ratings and
expert ratings give similar results across a series of tasks. We need
more studies to demonstrate such congruence of results. Second, even
so, there are some less consistent results. Révész et al. (2016) report
that accuracy in colour-change detection worked as intended, but
reaction time in the same secondary task did not. Interestingly,
Sasayama (2015) also used a colour-change secondary task (which
differed in some details) and she found that reaction times did reflect
proposed task complexity, but accuracy did not. There is some way to
go here. Third, many researchers have focused on one secondary
measure, and then compared it to general task analysis and self-rating
measures, of mental effort and/or task difficulty. But there is also scope
to explore how the different secondary measures interrelate, and
whether, for example, time estimation, secondary tasks and physio-
logical measures such as eye-tracking, deliver consistent results. There
is little evidence so far on this point. Fourth, it is possible, at present,
that the intervals that have been used between simple and complex are
rather large. It may be more challenging to explore whether smaller
intervals, or stepped intervals, can generate consistent results also.
Sasayama (2015), for example, found that when she used a narrative
retelling with one, two, four or nine participants, the secondary meas-
ures confirmed the largest difference but were not so effective with the
intermediate values. Finally, there may be scope to broaden the types
of task which have been investigated. So far, entirely reasonably, the
impetus for tasks that are researched has been the CH. But we also
saw that Skehan (2018) proposes that one needs to distinguish
between difficulty (his term) as perceived in the Conceptualizer and
in the Formulator stages. It might be useful therefore to explore
whether the range of secondary measures is equally effective when
tasks independently vary these two speaking stages.

Native-Speaker Baseline Data

One final area is worth mentioning in terms of psycholinguistically
motivated task research. It is to argue for the relevance of native-
speaker baseline data (D. Ellis, 2011). In general, it is not typical of
task studies to gather such data. This is understandable because to
gather such data would introduce a significant complication in any
research study. Researchers want to get on with investigating the
variables, the tasks, the conditions that are important to them, rather
than find ways of complicating life. But nonetheless to have some
baseline data is very important if we are to make solid judgements
about tasks.
The chapter has provided many examples of variables which have
been studied and linked to different aspects of performance. Similarly,
theoretical accounts also propose that certain tasks, or categories of
task, such as resource-directing or resource-dispersing, will have con-
sistent influence on performance. Claims are then made that some
tasks are more likely to foster acquisition, or to help L2 learners to
achieve higher levels of performance. But one has to consider the
question as to whether the same tasks or the same conditions would
produce the same effects with native speakers. Accuracy is a bit of a
complication here. The point of using native-speaker baseline data is
that native speakers, by definition, have complete language systems.
As a result, one can expect that they will simply be accurate, and so
comparing them with non-native speakers for accuracy will simply
demonstrate that the non-native speakers are non-native speakers. But
the situation is more instructive with structural or lexical complexity
and different aspects of fluency. If a task does not produce change in
performance with native speakers in these areas, then we have to
question what we are doing if the task is used with non-native
speakers. There may be good reasons for using the task, but they are
certainly not so obvious in these cases. Equally, if a task or a task
condition does produce a different level of performance with native
speakers, then we have a touchstone against which to judge the
performance of non-native speakers. Is performance affected in the
same direction? To the same extent? Or even, possibly, to a greater
extent? Without this information, it is difficult to disentangle the
effects of tasks and task conditions, on the one hand, and the effects
of speaker status, on the other.
There are some studies comparing native and non-native perform-
ance. Foster (2001) showed that native speakers use formulaic lan-
guage differently to non-native speakers. They rely on such language
more in unplanned speech, and less in planned speech. Being given

planning opportunities seems to enable them to produce more creative

language. Non-native speakers are the reverse and use planning
opportunities to access formulaic speech more effectively. Skehan
(2009b) reports that there are very large lexical differences, for lexical
sophistication and for lexical diversity, between these two groups,
with the difference being largest for lexical sophistication with a
narrative task and for lexical diversity with an interactive, decision-
making task. He also reports that planning has little effect for native or
non-native speakers regarding lexical diversity, whereas it does have
an effect for lexical sophistication, for a personal information
exchange task and a decision-making task for non-native speakers,
but a narrative for native speakers. Skehan and Shum (2017), based
on four video-based narrative retellings, which vary in degree of
structure, confirm these results of a native to non-native difference
for lexical diversity but less so for lexical sophistication. There are
slight differences between the two groups for structural complexity,
with native speakers producing higher subordination scores on two
tasks, but the non-natives having a significantly higher value for
words-per-clause on one of the four tasks (and arithmetically higher
scores on the other three). There are also clear differences between the
two groups for fluency, with the native speakers consistently pausing
less, repairing less and speaking faster. As a final point, there is also
interesting evidence of prevailing styles of speaking. Both groups show
strong cross-task consistency for lexical diversity, pausing (both clause
boundary and mid-clause) and repetition. Lexical sophistication
has clear style involvement for the native speakers but not for the
non-natives. Finally, structural complexity scores show much less
evidence of style.
It would be helpful if more research is reported which shows native
and non-native speaker performance of this sort. This should give us a
clearer understanding of how tasks have a contribution to make,
specifically in influencing non-native-speaker performance.

Introduction
As Zuengler and Miller (2006) noted, the ﬁrst twenty years (i.e.
1970–1980) of second language acquisition (SLA) research were dom-
inated by a cognitive view of how a second language (L2) is acquired
(i.e. the mental processes involved in the conversion of input into
intake and the role of L2 production in acquisition). This view
informed the previous two chapters. Chapter 2 examined how task-
based language teaching (TBLT) has drawn heavily on the cognitive-
interactionist theories of Long, Gass and Mackey (among others).
Chapter 3 drew on cognitive models of speaking to show how
task design and implementation features impact on the complexity,
accuracy, lexis and ﬂuency (CALF) of learners’ production. Zuengler
and Miller went on to point out that, starting around 1990, an
alternative paradigm emerged in SLA – one that emphasized the social
nature of L2 acquisition. They reviewed a number of social theories –
sociocultural theory (SCT), language socialization, community of
practice, Bakhtin’s dialogic perspective and critical theory – all of
which emphasize the importance of the social context and the central-
ity of participation in explaining how learners use and acquire – or fail
to acquire – an L2. From this perspective, the development of an L2
was not a question of taking possession of knowledge but of taking
part in social activity (Sfard, 1998). It rejects the distinction between
‘use’ and ‘acquisition’, which lies at the heart of cognitive theories of
L2 acquisition, and prefers the metaphor of ‘appropriation’ over that
of ‘acquisition’. Of the social theories mentioned by Zuengler and
Miller the one that has been the most fully developed and has the
greatest relevance to TBLT is SCT.
The advent of this ‘social turn’ (Block, 2003) in SLA was not
welcomed by some advocates of cognitive SLA (e.g. Gass, 1998; Long,
1998) and, arguably, its impact on TBLT has been much less than that

103

of cognitive SLA. Ellis (2003) included a chapter on sociocultural SLA

and tasks in his book Task-Based Language Learning and Teaching
but Long (2015) gives it no space at all in his Second Language
Acquisition and Task-Based Language Teaching. Neither ‘social’ not
‘sociocultural’ appear in the index of his book. Clearly, though, there
is a social dimension to tasks. Also, as proponents of sociocultural
SLA (e.g. Lantolf, 2000; Swain, 2000) take pains to point out, SCT is
not just a social theory; it aims to explain how the social use of a
language serves as both the source and context of the development of
higher-order abilities. In other words, there is an important cognitive
element to sociocultural SLA. Nevertheless, as we will see, not all
proponents see sociocultural SLA as providing a theoretical basis for
TBLT. Lantolf (see Lantolf and Thorne, 2006; Lantolf and Poehner,
2014), for example, sees SCT as supporting the development of
declarative knowledge through the explicit teaching of linguistic forms
in an approach that Long would doubtlessly dismiss as ‘focus on
forms’ and therefore as antithetical to TBLT.
This chapter begins with a brief account of sociocultural SLA and of
activity theory. It then reconsiders the incidental-intentional learning
distinction (so central to an understanding of TBLT) from the perspec-
tive of sociocultural SLA. The next two sections address what insights
SCT has to offer for the design and the implementation of tasks. The
chapter ends with a reconsideration of the role of explicit instruction
in L2 acquisition in the light of the evidence provided by research
based on SCT.

Sociocultural SLA
Sociocultural SLA draws heavily on the work of Vygotsky (1978,
1986), Leontiev (1981) and Wertsch (1985), among others. There
are now a number of accounts of the theory, as applied to SLA (e.g.
Lantolf, 2000; Lantolf and Thorne, 2006; Lantolf and Poehner, 2014;
Storch, 2017). We will draw on the account provided by Swain,
Kinnear and Steinman (2014). Their unique book offers readers a set
of narratives told by L2 learners and their teachers and uses episodes
from these narratives to illustrate and discuss the key concepts. For
readers interested in developing an understanding of sociocultural
SLA, this book is highly recommended.
Swain et al. explain that the most basic concept of Vygotsky’s work
is that ‘the individual cannot be understood in isolation but only as
part of a history, a culture and a society’ (2014, p. x). It follows that in
order to understand how an individual’s mental development takes
place it is necessary to examine how this individual engages with

people in social activities. They point to Vygotsky’s basic premise,

namely that individuals’ minds develop through the guidance of a
more experienced and knowledgeable person. Vygotsky’s method for
investigating development involved observing individuals performing
tasks that they could not do by themselves,1 providing them with help
and noting how these individuals made use of the guidance provided.

Basic Concepts
Swain et al. acknowledge that it is not easy to explain SCT because the
interconnectedness of its concepts make it difficult to know where to
start. Their approach is to distinguish a set of basic concepts that
constitute the core of the theory and a number of related concepts
that figure strongly in sociocultural SLA. Table 4.1 provides a brief
description of these basic concepts along with examples taken from
Swain, Kinnear and Steinman (2011).
Sociocultural SLA explains L2 development in terms of the inter-
relatedness of these concepts. It views learning an L2 as like any other
kind of learning. That is, it is mediated; learning occurs when a learner
has the chance to interact with cultural artefacts, with social inter-
action (one type of cultural artefact) serving as the primary means of
mediation. Learning commences within an interaction between an
expert (a teacher or a more advanced learner) and a learner, resulting
in the co-construction of a ZPD. It is through participating in ZPDs
the learner comes to understand scientific concepts, which are crucial
for higher-order thinking. Non-ZPD interactions also contribute to
learning, but only of everyday concepts. What the learner manifests in
a ZPD may be internalized, allowing the learner to achieve self-
regulation. In other words, there is a progression from intermental
behaviour to an intramental state. The routines and patterns that
figure in social interaction are also internalized and can re-emerge
in private speech (talking to oneself ), which serves as means for
self-regulating when a problem cannot be immediately solved.
Mediation and the construction of ZPDs generally occur when the
learner experiences positive emotions.

Scaffolding, Languaging and Imitation

An issue of both theoretical interest and of obvious relevance to TBLT
is how social interaction facilitates the construction of a ZPD. In other
words, what does an expert and a learner have to do when interacting
to create a ZPD? This is analogous to the question we addressed in
Chapter 2 when we considered cognitive-interactionist perspectives on

Table 4.1 Basic constructs in sociocultural SLA

Concept Description Example

Mediation The material and symbolic tools that organize or Mona tells how her father gave her a grammar
regulate our behaviour. Generally speaking book and how Mona’s use of the grammar
material tools (e.g. a hammer) are directed book was mediated through interaction with
towards the environment whereas symbolic her father.
tools (e.g. language) are directed towards
changing our psychological selves (Swain et al.,
2011, p. 152).
Zone of proximal An interaction during mediation enables an A student (Brock) accidentally runs into his
development (ZPD) individual to achieve more than he/she could teacher, who expects him to apologize. At ﬁrst
have achieved working alone. The ZPD is co- Brock does not do so but when prompted he
constructed by the learner working with an apologizes in English (his L1) rather than
expert.a French (the L2). He is reprimanded by his
teacher. Another student (Sarah) then helps
Brock out by whispering Je m’excuse, which
Brock repeats.
Private speech Speech that is social (intermental) in origin and Jody was about to catch a bus when a stranger
form but psychological (intramental) in asked her which direction the bus was going to
function. It is speech addressed to oneself and is in Chinese. She replied sei (west). On the bus she
used by individuals to mediate their own begins to question herself in English to try to
behaviour when they experience a cognitively decide whether sei was correct. She rehearses
complex problem. L2 learners typically use their the ingrained sequence dong, lam, sei, bach to
ﬁrst language (L1) for private speech. herself until she realizes that it corresponds to
east, south, west, north and that she had given
the man the wrong information. Sei actually
means ‘west’ and the bus was going east.
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.008
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:31, subject to the

Intermental; intramental; Intermental refers to processes that occur between Sarah helped Brock with je m’excuse when Brock
other-regulation; self- individuals. Intramental refers to processes that was unable to say it himself. This constitutes an
regulation; occur within one individual. Other-regulation is example of intermental, other-regulated
internalization behaviour that is regulated (i.e. mediated) by behaviour in a ZPD. Brock was then able to
another person; self-regulation is behaviour apologize in French to his teacher. However, we
where an individual exercises control over him/ do not know whether he internalized je
herself. Internalization refers to how a social (i.e. m’excuse and thus can use it to apologize in
intermental) process is transformed into a French in the future. Further evidence would be
psychological (i.e. intramental) process. It needed to demonstrate self-regulation
captures the progression from other-regulation to (internalization).
self-regulation.
Everyday concepts; Everyday concepts are ‘understandings individuals Thaya read a story he had translated from Tamil
scientific concepts develop from their experiences to solve various (his L1) into English to a group of postgraduate
cognitive/emotional problems’ (p. 150). They students who commented on the story. In their
do not constitute part of a system, are context- discussion of Thaya’s story, the students drew
dependent and are applied without on a variety of scientific concepts (i.e. rhetorical
consciousness. Scientific concepts are devices and literary strategies) which they had
systematic, hierarchical, context-free and appropriated from their readings and
subject to conscious manipulation. They are presentations in the postgraduate course they
acquired through mediation. were taking.
Cognition and emotion Emotions, like cognition, are socially constructed. Grace recounted her experience as a bilingual
The emotional experience arising from a (Greek-English) when living in Greece and in
situation determines what effect it has on a Canada. She reported that her sense of a lack of
person’s mental development. Thus cognition competence in Greek (a negative emotion) when
and emotion are interrelated and cannot be living in Greece led her to keep silent. In
considered separately. contrast, feeling embarrassed in elementary
school in Canada when she inadvertently used a
Greek word led her to learn the equivalent
English word.
a.
The Zone of Proximal Development (ZPD) is not the same as Krashen’s (1981a) i+ 1 concept, which tied to the notion of a fixed
order of acquisition. See Dunn and Lantolf (1998).
108 Theoretical Perspectives

TBLT, namely what kinds of interaction trigger the internal processes

(such as noticing) involved in L2 acquisition? SCT emphasizes the
joint interactional behaviour of the expert and the learner (or between
learners) as a ZPD is always co-constructed. It defines interactional
behaviour much more broadly than in the cognitive-interactionist
approach. A key construct is scaffolding.
Scaffolding refers to the interactional work by which one speaker
(usually the expert) assists another speaker (usually the novice) to
perform a skill or a linguistic feature that he/she cannot perform
by him/herself. Storch (2017), in her summary of L2 classroom
research based on SCT, emphasizes the importance of scaffolding.
She traces interest in it to Aljaafreh and Lantolf’s (1994) study of the
feedback that a tutor provided on a L2 learner’s writing. The tutor
systematically (and without training) utilized a variety of feedback
strategies, commencing with very implicit strategies and then moving
to more explicit to help the student self-correct his/her errors. The
importance of this study is that it established the necessity – from
a sociocultural perspective – of providing the minimal amount of
scaffolding needed to construct a ZPD. It also indicated that learning
could be measured not just in terms of whether a learner had achieved
self-regulation but also in terms of whether there was a change in
the amount and level of scaffolding needed to mediate the correct
production of a specific feature. The role played by an expert has
been further developed in dynamic assessment, which integrates
assessment and instruction by identifying a learner’s potential for
learning in scaffolded conversations. This is considered in the section
‘Dynamic Assessment’. Other researchers have focused on how L2
learners scaffold each other when performing a task. Storch’s work
on effective collaboration is a notable contribution here and will also
be discussed later.
A second concept is related to scaffolding but offers a broader
perspective on how the use of language can mediate the learning of a
language. Swain (2006) coined the term languaging, which she defined
as ‘the process of making meaning and shaping knowledge and experi-
ence through language’ (p. 98). Languaging can occur in social inter-
action or in the individual learner when it takes the form of private
speech. In a whole series of studies – see Ellis (2012) for a review –
Swain and her co-researchers explored how L2 learners used language
(both their L1 and the L2) to address linguistic problems that arose
when they were performing tasks. Learners engaged in language-
related episodes (LREs), defined as any part of a dialogue where they
‘talk about the language they are producing, question their language
use, or correct themselves or others’ (Swain and Lapkin, 1998, p. 326).

In LREs linguistic features become the topic of talk and the focus of
explicit attention. In this respect, SCT differs from the psycholinguistic
and cognitive-interactionist perspectives by placing a premium on
explicit and intentional learning – a point we will pick up later in this
chapter.
Vygotsky (1986) saw one way in which everyday concepts can be
transformed into scientific concepts as through imitation – the means
by which socially constructed forms of mediation are internalized.
Imitation is not the same as the ‘copying’ that occurs when an individ-
ual mimics a stimulus, as in behaviourist learning theories. Rather it is
a conscious, reflective activity on the part of the learner. A distinction
can be made between ‘simple imitation’ and ‘persistent imitation’
(Lantolf and Thorne, 2006). The former involves the unreflective
attempt to reproduce a model – as, for example, when a learner just
repeats a recast that repairs an error. The latter is intentional and is
related to the goal the learner has for performing an utterance. It is
also cyclical, with each attempt at imitation based not on the original
model but on the previous imitation. Crucially, too, imitation is trans-
formative; it involves modification of the model. For example, to
qualify as imitation a learner would have to not just repeat a recast
but to build on it. Imitation can occur in both social interaction and in
private speech.
These three concepts all relate to the central concept of mediation.
Learners participate in social interaction, which scaffolds the produc-
tion of utterances they are incapable of producing independently,
offers opportunities for learners to ‘language’, and creates contexts
where learners can modify and extend the models they are exposed to
by means of imitation. For mediation to be effective, however, it has to
be goal-directed. This takes us to Activity Theory.

Activity Theory
Activity Theory was a development of Vygotsky’s ideas by Leontiev
(1978). It was extended further by Engeström (1999). The core of the
theory is represented in Figure 4.1. For our purposes the ‘subject’ is
a language learner. The ‘goal’ can be construed narrowly as the
performance of a single utterance designed to achieve some purpose
(e.g. politely refuse an invitation or deploy a grammatical feature
accurately) or more broadly as the achievement of the outcome of
a task (in the sense of this term in TBLT). The ‘mediational means’
refer to the various artefacts (material and symbolic) that mediate
an activity. The interior of the triangle represents the activity that
takes place. Engeström’s extended model incorporates three

Figure 4.1 Activity theory model

Source: Engeström (1999)

contextual elements that impact on the activity that takes place: (1)
rules (i.e. the norms that govern the use of language in a particular
context, (2) community (i.e. the social group that a learner belongs to)
and (3) division of labour (i.e. how the work done to achieve the goal
is shared out among the participants in the activity). These compon-
ents of an activity system interact among themselves in complex ways.
Neither the components nor the relationships between them are
static. It is this that gives Activity Theory its ‘messiness and power’
(Swain et al., 2011).
An activity has three levels: (1) motive, (2) action and (3) conditions
(or operations). That is, a learner has a motive for performing an
action that is directed at achieving the object (goal) and utilizes appro-
priate operations to achieve this. The motive determines the goal and
also the operations that are used to achieve it. A good example is
Wertsch, Minick and Arns’ (1984) study. They compared how urban
schoolteachers and rural mothers in Brazil mediated children’s per-
formance of a puzzle copying task. Although the task was the same for
both the teachers and the mothers, the activity that resulted was very
different. The teachers’ motive for performing the task was educa-
tional (i.e. the children needed to learn how to function independently)
and the goal was to help the children to carry out the actions them-
selves. Accordingly, they offered the children clues about what parts of
the puzzle they needed to attend to. The rural mother’s motive was to
complete the task as quickly as possible and the goal was to prevent

their children from making errors.2 To this end they gave explicit
instructions about how to build the puzzle. The fact that the same
task can result in very different activities, involving different oper-
ations, has important implications for TBLT, which we will consider
later in this chapter.

Incidental and Intentional Language Learning Revisited

One of the main differences between the cognitive theories that inform
TBLT and sociocultural SLA lies in how they envisage the contribu-
tions of incidental and intentional language learning. Central to
Long’s (2015) cognitive-interactionist theory of instructed language
acquisition is the notion of implicit/incidental learning. He cites
Doughty (2003):
[T]he findings of a pervasive implicit mode of learning, and the limited role of
explicit learning in improving performance in complex control tasks, point to
a default mode for SLA that is fundamentally implicit, and to the need to
avoid declarative knowledge when designing L2 pedagogical procedures.
(p. 298)
However, Long also acknowledged adults’ reduced power for implicit
language learning and thus accepted that there might be a need for
‘facilitating intentional initial perception of new forms and form-
meaning connections’ (Long, 2015, p. 49; italics in original). For Long
(and in mainstream views about TBLT) this is not to be achieved by
explicit language teaching but through focus on form as learners
perform tasks (see Chapter 2). Long rejects the explicit teaching of
predetermined linguistic features (i.e. focus on forms). The justification
for Long’s position lies in the attested orders and sequences of acqui-
sition of grammatical structures, which arise because of the way in
which implicit L2 knowledge is acquired and which cannot be altered
through instruction. The goal of instruction is to speed up progression
through the natural route of acquisition by fostering incidental
acquisition.
Sociocultural SLA takes a radically different stance on the role of
incidental and intentional language learning. Social mediation is
directed in helping learners form scientific concepts through both
external guidance, imitation and private speech. In Swain’s research,
when learners perform tasks they engage in LREs, many of which
involve quite explicit attention to linguistic forms and intentional
language learning. Sociocultural theorists such as Lantolf and Swain
focus on how learners develop their conscious understanding of the
meanings encoded by specific linguistic forms and how mediation
helps them to achieve this. The goal of instruction is not implicit

knowledge (which is rarely referred to in the sociocultural literature)

but declarative knowledge. In other words, they see instruction as
helping learners to go well beyond the incidental perception of new
forms, which result only in everyday concepts, to the conscious for-
mation of scientific concepts about language through intentional
learning.
In taking the position that explicit instruction and intentional lan-
guage learning are not just desirable but necessary for developing an
L2, Lantolf (2009) disputes the empirical basis for Long’s position,
namely the existence of a natural route of acquisition. His view is that
any grammatical structure can be learned at any time provided that it
is appropriately mediated. This claim, however, runs up against
research that indicates instruction is powerless to alter the route
learners follow. To counter this evidence Zhang and Lantolf (2015)
conducted a study designed to test Pienemann’s (1985) Teachability
Hypothesis, according to which instruction directed at a specific gram-
matical feature will only be successful if that feature is next in line to
be acquired in terms of the natural sequence of acquisition. Using
an approach called concept-based language teaching, based on
Vygotskian principles (discussed later in this chapter), they taught
adult learners of Chinese an L2 grammatical structure that lay well
ahead of their current developmental stage and concluded that ‘it is
possible to artificially construct a developmental route different from
the one predicted by natural developmental sequences, in agreement
with the claims of Vygotsky’s developmental education’ (Zhang and
Lantolf, 2015, p. 152). Other studies also provide evidence of variable
developmental routes (see Ellis, 2015b).
A key issue here, however, is whether instruction results in implicit
knowledge, on which claims about developmental sequences are
based. While Lantolf does not deny the existence of implicit know-
ledge, which he correlates with everyday concepts, he argues that the
goal of instruction – for adults at least – is automatized declarative
knowledge, i.e. explicit knowledge that is available for use without the
need for conscious control. This is what Zhang and Lantolf claimed
the instruction achieved in their study. In this way the objection that
Long and others raise to focus on forms is circumvented. There is,
however, a further problem. DeKeyser (1998) questions the viability
of teaching complex grammatical rules. To overcome this objection
to explicit instruction, Lantolf and Thorne (2006, pp. 298–302)
argue that concept-based language teaching, where the focus is
on mediating learners’ conscious understanding of meaning-form
mappings, can result in automatized explicit knowledge of even very
complex structures.

In short, whereas mainstream TBLT – as reﬂected in Long (2015) –

is founded on cognitive-interactionist theories that prioritize incidental
acquisition and focus on form, sociocultural SLA supports explicit
instruction – as long as it conforms to Vygotskian principles – and
intentional language learning. In effect, sociocultural theorists such as
Lantolf lend support to task-supported language instruction, where
tasks provide opportunities to apply conscious linguistic knowledge
that has been explicitly taught in meaning-focused activity.

Sociocultural SLA: Task Selection and Design

As we have seen in Chapters 2 and 3, both cognitive-interactionist and
psycholinguistic perspectives afford insights about the selection and
design of tasks in a task-based course. Skehan (2001), for example,
draws on his earlier research to suggest how a number of design
features (e.g. the familiarity of information, the degree of structure
and the complexity of the outcome) affect the complexity, accuracy
and fluency (CAF) of learners’ production. He proposed that such
information is useful because it enables course designers to select a
variety of tasks to ensure balanced L2 development. Robinson (2001)
also identified a number of factors (e.g. whether a task requires
reasoning) that affect the complexity of a task. Researchers working
in the psycholinguistic paradigm (e.g. Sasayama, 2016) have shown
that task factors impact directly on the cognitive load imposed by the
task and indirectly on how the task is performed. Cognitive-
interactionist researchers (e.g. Pica, Kanagy and Falodun, 1993) have
investigated how task design features influence the extent to which
negotiation for meaning takes place. Designers of task-based courses
are able to draw on this research to make informed decisions about the
design of task-based syllabuses, although, as we have seen, neither the
cognitive-interactionist nor the psycholinguistic approach has solved
all the problems associated with task selection and sequencing.
SCT, however, offers no assistance to the course designer. According
to Activity Theory, the same task can result in very different activities
as a result of differences in the motives of the task participants and the
goals they set for the performance of the task. Differences in how
the same task is performed can also arise as a result in differences
in the rules, community and division of labour. Several studies testify
to the fact that the same task can be performed in very different ways.
The most often cited is Coughlan and Duff (1994). They showed
that the same task can be performed differently by different learners
(Hungarian school students and an adult English as a Second
Language [ESL] learner) and by the same learner (the adult ESL

learner) at different times. Donato (2000) also provided evidence to

show that ‘tasks do not manipulate learners to act in certain ways
because participants invest their own goals, actions, cultural back-
ground, and beliefs (i.e. their agency) into tasks, and thus transform
them’ (p. 44). A similar position can be found in Dynamic Systems
Theory (de Bot and Larsen-Freeman, 2011). From the perspective of
this theory, how learners perform a task cannot be predicted because
of the interconnectedness of all the variables involved in the task itself,
in the learner and in the situation and psychological context. The
activity system created by the interactions among these variables is
inherently dynamic and task performance unsystematic.
It would seem then that, from the perspective of sociocultural SLA
and Dynamic Systems Theory, ‘task’ does not constitute a viable unit
for designing a task-based course or for conducting research because
how it is performed is unpredictable – a view argued strongly by
Seedhouse (2005a). Nevertheless, teachers and researchers do have
to make decisions about what tasks to select. There has to be a starting
point and, in fact, some SCT researchers have made suggestions about
what kinds of tasks are needed. Storch (2017), for example, com-
mented ‘what SCT implies for L2 learning/instruction is the need for
two key ingredients: challenge and effective support’ (p. 77; italics
added). In other words, (1) there is no sense in asking learners to
perform easy tasks because learners will draw on everyday concepts
or the scientiﬁc concepts they have already internalized and thus do
not require mediation, and (2) the co-construction of ZPDs requires
tasks that challenge but are not so far beyond the learners’ current
abilities that they cannot be successfully performed through
mediation. Bygate (2018) also suggested that predictability is a matter
of degree and pointed out that Dynamic Systems Theory acknow-
ledges the emergence of ‘patterns’ of behaviour. He suggested that
tasks manifest ‘pragmatic predictability’ and that it would be possible
to select tasks in terms of the trajectories of behaviour they elicit. He
illustrated this idea by pointing out that there were predictable phases
in the performance of a picture story task where learners work in
groups, each holding one of the six pictures that make up the story,
and pass through phases (e.g. description, comparison, interpretation,
narration) to achieve the task outcome.
Research on dynamic assessment (not to be confused with Dynamic
Systems Theory) implicitly acknowledges that there are inherent dif-
ferences in the complexity of tasks. A key issue when investigating
mediated assistance is the extent to which learners are able to transfer
the learning that takes place in the performance of the mediated task
to a new, more complex task (Poehner, 2008). Van Compernolle,

Weber and Gomez-Laich (2016), for example, investigated the effect

of instruction motivated by SCT on learners’ acquisition of the prag-
matic uses of Spanish tú and usted. They deliberately included items of
differing levels of complexity in an appropriate judgement task and a
written discourse completion task in order to see whether the learners’
understanding of these two linguistic forms was generalizable beyond
the specific contexts that figured in the mediated test. Ableeva and
Lantolf (2011) also reported a study in which they sought to demon-
strate transfer from the mediated development of listening comprehen-
sion to a new, more difficult listening task. Notions of task complexity
are inherent in ‘transfer’ and clearly sociocultural researchers do
employ some intuitive sense of what constitutes complexity in a task.
By and large, however, SCT addresses the issue of what constitute a
‘challenging’ or ‘complex’ task in terms of the teacher’s (or researcher’s)
experience of working with tasks with particular learners. In other
words, what tasks to incorporate into a course are not decided a priori
but as the course progresses through an inspection of how learners
perform particular tasks and the teacher’s developing understanding of
what degree of difficulty is needed for the ongoing construction
of ZPDs.
A good example of this approach can be found in Mochizuki’s
(2017) proposal for ‘contingent needs analysis’, which she suggested
can help to identify the gap between the task-as-workplan and the
activity that arises from it. Clearly, if there is no close correspond-
ence between ‘task’ and ‘activity’ as SCT claims, a traditional
needs analysis of the kind that Long (2005, 2015) recommends
makes little sense. Mochizuki’s idea is that an analysis of how
learners perform a task can help to identify the contradictions and
tensions that arise and this information can be fed back into what
teachers need to do to ensure the effective mediation of subsequent
tasks. Mochizuki’s study used Activity Theory to examine how two
different groups of learners participated in feedback sessions on
doctoral students’ writing. The ‘needs’ she identified included the
importance of the facilitators equipping students with the strategies
required for giving and receiving feedback and the need to address
the power relations that suppress some students’ participation in
giving and receiving feedback.
It would seem then, a task-based course informed by SCT will, at
best, just have a provisional syllabus, which is redeveloped as the
course continues. The actual syllabus is the one the teacher ends up
with. We accept that this is always the case but we will argue that
in many instructional contexts there is a clear need for an a priori
task-based syllabus.

Sociocultural SLA: Task Implementation

SCT has much more to say about how tasks should be implemented
than about task design or selection. Indeed, the focus of SCT is more
or less entirely on how tasks can be implemented in order to mediate
learning effectively. It is in this respect, then, that SCT has the most to
offer TBLT. For reasons of space, we will illustrate what SCT-inspired
research has shown by focusing on three major areas of enquiry –
graduated feedback, collaborative dialogue and dynamic assessment.

Graduated Feedback
In Chapter 2 we saw that cognitive-interactionist views of L2 acquisi-
tion have motivated a number of studies that have investigated
whether one type of corrective feedback (CF) (e.g. explicit feedback)
is more effective in promoting acquisition than another type (e.g.
implicit feedback). We noted that the results of these studies do not
enable a single type of feedback to be identified as the most effective.
This is because there are differences in how learners react to feedback
depending on individual learner factors (such as working memory)
and contextual factors. Such differences are to be expected from the
perspective of SCT and Activity Theory. SCT has taken a radically
different approach to investigating CF, which is conceived of as a form
of mediation aimed at learner development.
Aljaafreh and Lantolf (1994) (see also Lantolf, Kurtz and Kisselev,
2016), examined writing conferences where a tutor provided oral
feedback on students’ written work. They developed a ‘regulatory
scale’ to reflect the extent to which the tutor’s oral feedback was
implicit or explicit. For example, asking learners to find and correct
their own errors constitutes an implicit strategy while providing
examples of the correct pattern is a highly explicit strategy. An inter-
mediate level occurs when the tutor indicates the nature of an error
without identifying it for the learner. This scale reflects a central claim
of SCT, namely that for CF to be effective if must be fine-tuned to the
learner’s development (i.e. provide the minimal assistance needed to
induce a self-correction).
In a study based on Aljaafreh and Lantolf’s regulatory scale, Nassaji
and Swain (2000) investigated two Korean learners of English. One
learner was provided with graduated assistance (i.e. the tutor system-
atically worked through the regulatory scale to tailor the feedback
supplied) while the other learner was given only random help (i.e. the
tutor was supplied with a random list of correcting feedback strat-
egies). Nassaji and Swain reported that systematic graduated feedback

was more effective in assisting development than the random feed-

back. However, a limitation of this study is that random feedback is
highly unnatural and unlikely to occur in actual teaching. A more
interesting comparison would be one that compared graduated feed-
back with a speciﬁc type of feedback that cognitive-interactionist
research has found to be effective.
Erlam, Ellis and Batstone (2013) conducted such a study. They
compared the effects of graduated feedback and explicit correction
on two grammatical structures – English past tense and articles. In
contrast to Aljaafreh and Lantolf (1994), they found no evidence of
any shift in the quality of graduated feedback over the two occasions
that correction was provided on the students’ writing. Evidence from a
post-test showed that graduated feedback resulted in greater gains in
accuracy than explicit correction for articles but not for past tense. It is
possible, therefore, that when the form-meaning mapping is transpar-
ent (as is the case for English past tense), graduated feedback is not
necessary. Responding to Erlam et al.’s (2013) study, Lantolf et al.
(2016) argued that development, as measured by a graduated feed-
back index, is not linear, that unidirectional change from one week to
the next is not to be expected and therefore investigating development
by means of a pre- and post-test as in Erlam et al. is not appropriate.
A strength of the SCT research on CF is that it recognizes that
feedback is contingent on the learner’s response to it. In research based
on cognitive-interactionist SLA, CF is typically construed as of
the one-shot kind. That is, every time an error occurs it should be
corrected by a pre-determined corrective strategy (e.g. recasts or
prompts). In other words, CF is something done to a learner. In
contrast, in SCT, CF is seen as co-constructed between an expert and
novice and as continuing over several turns and times in the search for
a ZPD.

Collaborative Dialogue
Following Swain (2000), we have elected to use the term ‘collaborative
dialogue’ rather than ‘scaffolding’ because it more accurately captures
how social interaction mediates development. The scaffolding
metaphor implies a pre-planned architecture but, according to
SCT, mediation is a jointly constructed activity and is thus ﬂexible
and collaborative. The research on graduated CF is one example of
collaborative dialogue. In this section, we look at other ways in which
it has been investigated in research involving tasks. We will not
attempt an extensive review of the research but rather focus on a few
representative studies.

Intersubjectivity
Swain et al. (2011) pointed out that for the co-construction of a ZPD
‘there needs to be some level of intersubjectivity’ (p. 24) in the sense of a
shared understanding of the goal of performing a task. Ellis (2003)
illustrated the importance of intersubjectivity in an exchange between
a teacher and a beginning L2 learner. To begin with the teacher and
learner have different goals. The teacher’s goal was to help the learner
to describe what is wrong in a picture of a bicycle with no pedals. This
goal was beyond the learner’s linguistic ability, who therefore estab-
lished a simpler, different goal – identifying the colours of objects in the
picture. As a result, in the early part of this sequence, the participants
are functioning at cross-purposes. Eventually, intersubjectivity is
achieved in turn (7) when the teacher accepts the learner’s goal. As a
result, a ZPD is constructed in turn (8) when the learner builds on the
teacher’s preceding utterance to produce what Ellis claimed was the ﬁrst
instance of a two-word utterance (black taes) in his data for this learner.
1. T. I want you to tell me what you can see
in the picture or what’s wrong with
the picture
2. L. A /paik/ (= bike)
3. T. A cycle, yes. But what’s wrong?
4. L. /ret/ (= red)
5. T. It’s red yes. What’s wrong with it?
6. L. Black.
7. T. Black. Good. Black what?
8. L. Black /taes/ (= tyres).
(From Ellis, 2003, p. 181)

The importance of shared goals is obvious. It should be noted,

however, that the rubric for a task cannot guarantee shared goals –
as was the case in this sequence – and as SCT predicts. When learners
work in pairs or small groups, they may need to agree on the goal for
a task before they start to perform it. This can involve meta-talk
about the task, which in monolingual groups may well be carried out
in the L1. Brooks and Donato (1994), for example, described how even
though the teacher carefully explained the task goals, the learners often
felt the need to discuss these between themselves. They argued that this
constitutes a legitimate use of the L1 in task-based teaching.3

Talk in Group Work

According to SCT, development occurs when an expert mediates the
novice’s performance of a task. Accordingly, much of the research has

investigated how ZPDs are constructed in teacher–learner interaction.

But, in fact, learners can also successfully mediate each other’s devel-
opment. Thus, as in research in the cognitive-interactionist paradigm,
many SCT-motivated studies have investigated how learners perform
tasks in pairs or small groups.
In an early study, Donato (1994) investigated groups of university
students of French performing an oral activity. In a detailed analysis of
an exchange involving the negotiation of the form tu t’es souvenu,
Donato showed how the students jointly managed components of this
structure, compared what they produced with what they perceived as
the ideal solution, and used their collective resources to minimize
frustration and risk. This collaborative scaffolding enabled the joint
construction of the correct form of the verb even though no single
learner had demonstrated knowledge of it prior to the task. This study,
therefore, demonstrated the central claim of SCT, namely that ‘higher
mental functioning is situated in the dialectal processes embedded
in the social context’ (p. 46). Donato also provided evidence of intern-
alization by showing that the joint performance of new structures
on one occasion was frequently followed by individual learners’
self-regulated use of them on a later occasion.
One of the most complete studies of group work from a SCT
perspective is Ohta’s (2001) account of beginner learners in a Japanese
foreign language classroom. Through the detailed analysis of
sequences of talk by these learners, she identified the various
scaffolding techniques they used to help construct ZPDs. For example,
when a listener observed a partner struggling to produce an utterance
he/she would wait to give the partner time to complete it, prompt him/
her by repeating a syllable, co-construct the utterance by providing a
syllable, word or phrase that contributed towards its completion, or
sometimes, provide an explanation in the L1 (English). Ohta empha-
sized the reciprocal nature of assisted performance: ‘This is the key to
peer assistance – that both peers benefit, the one receiving assistance
and the one who reaches out to provide it’ (p. 125).
The quality of collaboration in group work varies, however, so a
key question is what constitutes effective collaboration. Storch (2002)
investigated this by analysing the patterns of dyadic interaction
found in ESL students’ performance of a range of tasks. She identified
four basic patterns based on two intersecting dimensions involving (1)
mutuality (i.e. ‘the level of engagement with each other’s contribu-
tion’) and (2) equality (i.e. ‘the degree of control or authority over a
task’) (p. 127). Storch investigated the extent to which ‘learning’,
as evidenced in the interactions, led to ‘development’, as shown in
the performance of subsequent tasks. She reported that the most

collaborative dyad (i.e. the dyad manifesting high mutuality and high
equality) achieved the most instances of transfer of knowledge.
SCT prioritizes social interaction as the primary means of
mediation. However, SCT, also acknowledges that a learner can
mediate him or herself by means of private speech. A question of some
interest, therefore, is whether there is any advantage of a task being
performed socially as opposed to individually. Storch (2007) com-
pared ESL students completing a text-editing task in pairs and
individually. She found no difference in the accuracy of their edited
texts. Nevertheless, she argued that performing the task in pairs was
advantageous because it afforded opportunities for using the L2 for a
range of functions that would promote language learning. Other
studies (e.g. Swain and Lapkin, 2007), have shown that interaction
with the self (i.e. private speech) is effective in mediating development.

Languaging
The languaging that occurs when learners focus on linguistic problems
as they perform tasks has been investigated in a series of studies by
Swain and her co-researchers (e.g. Swain and Lapkin 1998, 2001,
2002; Watanabe and Swain, 2007; Swain et al., 2009). The typical
design of these studies involved: (1) transcribing a recording of
learners performing a task, (2) identifying LREs and coding them as
successfully resolved, unsuccessfully resolved or unresolved, and (3)
investigating whether the learners were subsequently able to use the
features they had targeted in the LREs accurately. Swain often chose
tasks that were likely to result in linguistic problems. Swain (1995), for
example, reported a study by La Pierre involving a dictogloss task that
resulted in 140 LREs being identiﬁed.
Languaging in Swain’s studies is clearly a collaborative activity. The
analyses of the talk generated by learners reveal the mental processes
that mediated L2 learning (e.g. generating alternatives, assessing alter-
natives through hypothesis testing and applying rules to new L2
contexts). In some of the studies, Swain included pre- and post-tests
in order to see whether the dialogic activity that occurred as learners
performed the tasks enabled them to move from incorrect to correct
responses. There was clear evidence of this happening (see, for
example Swain and Lapkin, 1998).
However, the studies also demonstrated considerable variability in
learners’ ability or preparedness to ‘language’. Watanabe and Swain
(2007), drawing on Storch’s (2002) research on collaboration in small
group work, investigated the patterns of interaction that took place in
pairs of learners and the relationship of these to learning. The pairs of

learners differed in their L2 (English) proﬁciency. The learners wrote

an essay in pairs (the pre-test), were given a reformulated version
which together they compared with their own text, and then individu-
ally rewrote their essay (the post-test). Learning was operationalized in
terms of whether the changes that the learners made to their original
text were correct. Proﬁciency differences in the pairings made little
difference in post-test performance but the extent to which the pairs
engaged in collaborative patterns of interaction did. Swain et al.
(2009) reported that ‘high languagers’ produced more and better
quality LREs, demonstrated greater depth of understanding and had
higher scores in tests than ‘low languagers’.
These studies provide clear evidence of the power of languaging as a
mediating tool. All of the studies involved learners grappling with
language problems, working towards a conceptual understanding in
order to solve them, and thereby learning to use linguistic forms in a
target-like way.
The strength of Swain’s research is that ‘development’ is not investi-
gated solely in terms of the LREs that arise in social interaction or in
the ‘language units’ observed in interaction with the self but also in
transfer to subsequent tasks or tests. There are, however, some limita-
tions. In general, the studies did not convincingly demonstrate full
transfer of learning as they did not show that the learners were able
to generalize their learning to new tasks and new contexts. Nor did
they provide convincing evidence that the learners had automatized
their linguistic knowledge. Many of the studies involved writing rather
than speaking.

Dynamic Assessment
Dynamic assessment also involves mediation through collaborative
talk. It aims to achieve ‘the dialectic integration of instruction and
assessment’ (Lantolf, 2009), thereby overcoming the dualism evident
in much of the applied linguistics literature. Drawing on the idea of
graduated feedback, the tester aims to show both what learners can do
independently and what they can achieve with assistance, the aim
being to measure learners’ potential for future learning as well as their
actual learning.
Dynamic assessment has now become one of the major lines of
research in sociocultural SLA (Poehner and Lantolf, 2005; Poehner,
2008; Lantolf, 2009; Poehner and Infante, 2017). It can be carried out
in two ways – the interventionist and the interactionist. In the former
the training provided by the tester is pre-planned. This makes it well
suited to assessing large numbers of learners and also to the

computerized delivery of mediation. In interactionist dynamic assess-

ment the mediation provided is highly ﬂexible and tailored to the
individual learner and thus corresponds more closely to the original
idea of graduated feedback. ‘In interactionist DA, the priority that
trumps all others is learner development’ (Poehner, 2008, p. 66).

Interventionist Dynamic Assessment

Poehner (2008) discussed a number of interventionist models of
dynamic assessment. Budoff’s Learning Potential Measurement
Approach pioneered the sandwich format (i.e. there was a pre-test
and a post-test in order to identify to what extent different learners
beneﬁted from the training). Carlson and Widl’s Testing-the-Limits
Approach asked examinees to give reasons for both their correct and
incorrect choices during the training. The resulting learner proﬁles
included information about the learners’ ability to verbalize their
linguistic choices. In Brown’s Graduated Prompt Approach, transfer
tasks are included to see whether the improvement resulting from
mediation transfers to both similar and dissimilar tasks from those
used in the treatment. These possibilities have been incorporated into
SLA studies involving dynamic assessment.
A good example of an interventionist study is van Compernolle and
Zhang (2014).4 They administered a version of the oral elicited imita-
tion test (EIT) (Erlam, 2006). The test consisted of six sets of sentences
with three pairs of sentences in each, one grammatical and one
ungrammatical. Each sentence contained two exemplars of each target
feature (plural -s; past-tense -ed; third person -s). If the learners were
able to imitate a sentence correctly without assistance they scored four
points. The researchers suggested that this indicated that they had
implicit knowledge of the target structure. If they failed to imitate it
correctly, they were given three clues ordered from implicit to explicit
and awarded marks on a declining scale (three, two or one) depending
on the level of assistance they needed to imitate a sentence correctly.
Van Compernolle and Zhang suggested that these clues prompted
learners to use their metalinguistic knowledge. If a learner ultimately
failed to produce a sentence they scored zero. They reported detailed
results for one learner. Interestingly, the unsupported accuracy level
for the three structures was the same as for the natural order of
acquisition (Krashen, 1981a). With assistance, however, the learner
was able to produce all three structures accurately and needed less
assistance as he moved through the sets of sentences in the test.
This study, then, lends support to the central claim of sociocultural
SLA. It demonstrated that the mediation enabled the learner to

produce grammatical structures accurately when he was unable to do

so independently.
Commenting on their study, Van Compernolle and Zhang noted
‘one of the issues . . . is whether, and to what extent, mediation in DA
promotes greater control and speed of access to metalinguistic know-
ledge during performance, or if it supports the development of a
learner’s implicit (unconscious, procedural) competence’ (p. 401).
We will take up this important issue in the conclusion to this chapter.

Interactionist Dynamic Assessment

As Poehner (2008) pointed out, interactionist dynamic assessment
owes much to Feuerstein’s ideas. Feuerstein was concerned with
helping children who were ‘retarded performers’ as a result of their
impoverished social experiences. His work was premised on the
assumption that ‘the more a child is subjected to mediated learning
experiences, the greater will be his capacity to benefit from direct
exposure to learning’ (Feuerstein, Falik and Rynders, 1988, p. 58).
In other words, Feuerstein was not just concerned with helping chil-
dren perform tasks but, crucially, with assisting their general cognitive
development (i.e. their ability to learn). The value of Feuerstein’s work
for dynamic assessment with language learners lies, in particular, in
the very detailed account he provides of how mediated language
experiences can be distinguished from other types of interaction.
He identified eleven attributes of effective mediation. These include
the importance of reciprocity (i.e. the collaborative nature of the
mediation), transcendence (i.e. true development manifests itself in a
child’s ability to perform increasingly complex tasks) and the
mediation of meaning (i.e. the importance of engaging the child in
cause-and-effect and inferential thinking).
Lantolf (2009) provides an example of interactionist dynamic
assessment involving two interactions between a mediator and an
advanced L2 learner of French. In the first interaction where the
learner narrates a scene from a Hollywood movie, the learner experi-
ences problems in deciding whether to use passé composé or imparfait.
She initially opts for imparfait but is challenged by the mediator,
which prompts the learner to try to justify her choice. As she launches
into an explanation, she talks herself into the more appropriate
option and settles on passé composé. In the second interaction a
problem emerges with a complex negative construction. This time
the mediator has to engage in more extensive assistance involving
hints, explicit explanation and finally recasting the learner’s attempt
to produce the structure. Lantolf argued that whereas the choice

of verb aspect lay well within this learner’s ZPD, necessitating

minimal assistance from the mediator, the complex negative construc-
tion was, at best, only in the very early stages of development and thus
necessitated more overt mediation.

The Relevance of Dynamic Assessment to TBLT

Dynamic assessment studies provide some of the richest examples of
how ZPDs can be scaffolded. They illustrate the kinds of strategies
that teachers can use when implementing tasks. These studies also
shed light on assessment involving tasks. Advocates of TBLT are clear
that assessment, like teaching, must be task-based (Norris, 2009a).
The kinds of assessment that have been proposed (see Chapter 9),
however, are based on the unmediated performance of tasks. That is,
the focus is on learners’ solo performances. SCT questions the validity
of such assessments. It proposes that a more valid assessment can be
derived by examining learners’ mediated performance of tasks in order
to demonstrate their potential.
There are problems, however. Where teaching is concerned, inter-
actionist dynamic assessment is not practical in instructional contexts
involving large classes although computer-delivered interventionist
dynamic assessment may be. Where assessment is involved, there
needs to be a way of deriving scores from a dynamic assessment.
Qin and van Compernolle (in press) suggest three scores are possible:
(1) an actual score, (2) a mediated score and (3) a learning potential
score. This, however, is only feasible in interventionist DA where there
is a pre-determined number of mediational clues.

Concept-Based Language Instruction

We have seen that SCT emphasizes the importance of developing
scientific concepts in the learner. Lantolf and Zhang (2017) argued
that ‘discovery learning’ – of the kind that TBLT aims to foster -
cannot ensure that L2 learners will develop the necessary scientific
concepts. They therefore argued for the explicit teaching of grammat-
ical concepts. However, Lantolf maintained that this cannot be
achieved by teaching learners rules of thumb that figure in traditional
grammar teaching. He commented ‘rules of thumb are not necessarily
wrong, but they generally describe concrete empirical occurrences of
the relevant phenomenon in a fairly unsystematic fashion and, as a
result, fail to reveal deeper systematic principles’ (Lantolf, 2007,
p. 36). Accordingly, he set out the case for presenting learners
with conceptually organized grammatical knowledge where the

links between semantic/functional concepts and linguistic form are

specified in detail. To this end, he saw cognitive linguistics and
systematic-functional grammar as providing the necessary bases for
a ‘developmental education’ involving concept-based language
instruction.
Concept-based language instruction is based on Gal’perin’s (1989)
proposal for systemic-theoretical instruction. Lantolf and Zhang
(2017) identified five phases in this kind of instruction. In the orienting
phase, a clear goal is specified and the means for achieving it estab-
lished. The aim is to create a dissonance between the learner’s existing
everyday knowledge and the new scientific knowledge. In the second
phase, the scientific knowledge is given a material or visual instanti-
ation in the form of Schema for the Orienting Basis of Action
(SCOBA). This can take the form of a diagram, a chart, a picture or
physical objects (e.g. Lego pieces). The aim here is to discourage rote
memorization of the new information and to facilitate its application
in practical activity. The third phase connects the SCOBA to practical
activity. This is where tasks can come in. In the fourth phase learners
engage in overt verbalization by explaining their understanding of the
new information to someone else or aloud to themselves. This is when
‘languaging’ occurs. In the final phase learners demonstrate fluent
control of the new knowledge by performing additional communi-
cative activities (i.e. tasks).
The study that provides the clearest account of concept-based lan-
guage instruction is Negueruela and Lantolf (2006). This study investi-
gated twelve students in a university Spanish as a foreign language
class, which met three times a week for fifteen weeks. The explicit
instruction involved a SCOBA for presenting grammatical aspect. It
consisted of a flow chart that led the learners through a series of
questions to help them understand when to use the preterite and
imperfect tenses in Spanish. The students were asked to verbalize the
Schema six times while carrying out a number of oral and written
communicative activities. Finally, they completed a written communi-
cative task. The students’ verbal explanations of the grammatical
structures were collected at the beginning and end of the course.
Initially these were simplistic and incomplete, reflecting the rules of
thumb in student textbooks with which the students were familiar.
Their explanations at the end of the course, although not always
complete, were generally more coherent and accurate, which Neguer-
uela and Lantolf suggested demonstrated that internalization of the
concepts was taking place. The study also provided evidence to show
that the learners’ improved conceptual understanding was reflected in
improved accuracy in new production tasks.

By and large, concept-based language instruction studies have

focused on the learning of grammatical features. In an interesting
study, Kim and Lantolf (2018) investigated whether this kind of
instruction was effective in enabling L2 learners to comprehend sar-
casm. They noted that previous research has demonstrated that
learners often fail to detect sarcasm and argued that to overcome this
problem learners needed a full, scientiﬁc account of how sarcasm
functions in English. They provided learners with a list of the linguis-
tic, paralinguistic and contextual cues that signal sarcasm and asked
them to view of series of video clips to decide whether they contained
sarcastic utterances and if they did to verbalize the clues they had used
to detect them. Eight out of nine learners increased their scores in the
immediate post-test, maintained improvement in the delayed test and
also were better able to identify relevant cues.
Concept-based language instruction can be seen as a version of task-
supported language teaching. For this reason it will probably be
rejected by proponents of task-based language instruction. Clearly it
involves intentional learning, not the implicit/incidental learning that
TBLT is primarily directed at fostering. The studies to date do not
provide convincing evidence of full automatization of the target struc-
tures as they did not include free oral production tasks, but they do
indicate that the learners were able to use them in activities that allow
for controlled processing. The studies have also typically involved
university-level students, so the suitability of this kind of instruction
for younger learners or even adult learners with low language
analytical ability, who might be less able to handle the very detailed
information provided in SCOBAs, is doubtful.

Conclusion
SCT made a relatively late entry into SLA but since the 1990s it has
become increasingly inﬂuential. There are numerous books devoted to
it. Readers looking for a comprehensive account can refer to Lantolf
and Thorne (2006) or Lantolf and Poehner (2014). There is also a
journal (Language and Sociocultural Theory) devoted to the applica-
tion of the theory to language, including L2 acquisition.
By claiming that learning originates within social activity, SCT
offers an explanation for L2 development that is radically different
from that of cognitive or cognitive-interactionist SLA, which views L2
acquisition as an essentially mental phenomenon. Not surprisingly,
SCT has been largely ignored by cognitive SLA. Long (2015), for
example, dismissed SCT (along with Piaget) on the grounds that
the core constructs of SCT – inner speech, appropriation, mediation,

self-regulation and the ZPD – are ‘nebulous’ and inadequate both

theoretically and experimentally. While it is perhaps true that these
concepts do not interconnect into a tightly woven theory, as Swain
et al. (2011) acknowledged, Long’s dismissal of sociocultural SLA is
unwarranted. The central concept – mediation – provides a basis for
investigating how participation in the social uses of language does not
just facilitate learning – the cognitive perspective – but where it
happens on the fly.
In what ways, then, does SCT constitute a theoretical framework for
TBLT? We have argued that SCT has little to say about the design of
task-based courses. It offers no obvious basis for selecting which tasks
to use or how to sequence tasks to ensure they offer the right level of
challenge to L2 learners. SCT proponents assume that teachers will use
their experience of their students to gauge what tasks to use. In many
instructional contexts such as foreign language classes in state schools,
however, teachers need the support of a syllabus.
In contrast, we have shown that SCT has much to offer teachers
when it comes to how tasks can be implemented to foster L2 learning.
Research that has investigated graduated feedback, collaborative talk
and dynamic assessment provides rich accounts of how learning can
be mediated in both teacher–learner and in learner–learner inter-
actions. It shows how learners can be talked and can talk themselves
into using linguistic features that lie outside their independent control.
This research, perhaps more clearly than any other, shows how par-
ticipation is learning and thus feeds directly into our understanding of
task-based language instruction. For example, it demonstrates convin-
cingly that learners’ use of their L1 has a positive role to play in task
performance and as such supports the growing recognition that the L1
is a valuable resource in the L2 classroom (Hall and Cook, 2012).
SCT, then, is of value when it comes to deciding how tasks can be
effectively implemented. In Chapter 3 we noted that it is task
implementation rather than task design that had emerged as important
for creating the types of language use deemed important for acquisi-
tion. From this perspective, SCT has much to offer TBLT.
SCT also provides a theoretical basis for task-supported language
teaching. By rejecting the existence of a universal route for L2 acquisi-
tion, proponents of SCT dismiss the principal objection to ‘focus on
forms’. Research on dynamic assessment has focused on how to
mediate the development of specific grammatical features. Concept-
based language instruction is based on key principles of SCT (i.e. the
importance of scientific concepts and of verbal mediation). It has
been found to help learners develop declarative knowledge of highly
complex grammatical and socio-pragmatic features. However, to date

both dynamic assessment and concept-based language instruction

have only been tried on adult, university-level learners, who are adept
at explicit, analytical language learning. Whether they are appropriate
for younger, school-based learners, who are better equipped to engage
in the implicit/incidental learning that task-based instruction caters to,
remains to be shown. Also – potentially the main caveat – research
based on SCT has not convincingly shown that learners can utilize
the scientiﬁc concepts that instruction helps them to develop in the
kind of language use that TBLT prioritizes – spontaneous, naturally
occurring speech.5
The cognitive and sociocultural perspectives are often seen as
incompatible and therefore incommensurate. It is unfortunate that
proponents of each remain hostile to each other. The position adopted
by Ellis (2000) was that pedagogy involving tasks can beneﬁt from
both perspectives. He cited van Lier’s (1996) plea for a ‘dual vision’ –
the need for teachers to keep in mind ‘a long-term sense of direction
and the need to make on-line decisions that take account of the
exigencies of the moment’ (p. 215). Cognitive perspectives arguably
provide a long-term sense of direction for TBLT as they address head-
on the need for a syllabus about which SCT has virtually nothing to
say. SCT, however, offers rich insights as to how teachers can best
handle online decision-making as they implement tasks.

The psychology of second language (L2) task performance refers to

factors relating to ‘the mental experiences, processes, thoughts, feel-
ings, motives, and behaviours of individuals involved in language
learning’ (Mercer, Ryan and Williams, 2012, p. 2). The psychological
dimension of task-based language teaching (TBLT) thus defined
includes the learner characteristics that are, in Snow’s (1991) terms,
‘propaedeutic’ (‘required as preparation for a learning condition’)
(p. 205) to a learning goal, including affective (feelings and emotions),
conative (motivation) and cognitive (reasoning and memory) vari-
ables. In this chapter, we discuss the role of these variables in affecting
task performance and the effects of task-based instruction by elabor-
ating the theoretical underpinnings for the role of psychological
factors and synthesizing the research on these factors.
The factors are broadly divided into cognitive and affective factors,
following Robinson (2011). In line with the current mainstream second
language acquisition (SLA) literature (Dörnyei, 2005; Mercer et al.,
2012; Ellis and Shintani, 2014) these variables are collectively referred
to as individual difference variables. Special attention is given to lan-
guage aptitude and working memory in the cognitive domain and
motivation and anxiety in the affective domain, due to their importance
for TBLT, the relatively clear definitions and operationalizations of the
constructs, and the availability of a body of relevant empirical research.
The synthetic review of each of these variables starts with an overview
of the construct in L2 research in general, including, but not limited to,
the conceptualization, operationalization and methods of the research,
followed by a more specific discussion of the TBLT research.

Theoretical Issues
The theoretical basis of the role of individual differences in language
learning can be found in Robinson’s Cognition Hypothesis (CH)

129

(2001, 2011; see Chapter 2). Robinson’s theory posits a three-

component framework: task complexity, task conditions and task
difﬁculty, which concern the conceptual/cognitive, interactional and
perceptual demands of tasks, respectively. Among the three dimen-
sions, task difﬁculty relates to individual difference variables, includ-
ing affective (e.g. motivation and anxiety) and cognitive (e.g. language
aptitude and working memory) variables, which account for inter-
learner variation in task performance. Individual differences constitute
a key component of Robinson’s triadic framework that interacts with
the other two components in affecting learners’ task performance.
The CH makes the following predictions about how individual
difference variables interact with the other two groups of variables.
First, individual differences in the affective domain are implicated
when tasks are performed under different conditions leading to emo-
tions and interpersonal relationships coming into play. For example,
monologic tasks (e.g. narratives) that require public reporting may
lead to more anxiety than dialogic tasks performed in pairs or small
groups; tasks that require equal contribution from all participants
are likely to be more motivating than those that only require some
participants to contribute. Second, individual differences in cognitive
abilities are related to performance along different dimensions of task
complexity. For example, reasoning ability is important for successful
performance in tasks that are complex along the resource-directing
dimension which pose greater processing demands; attention control is
predictive of performance in tasks manipulated along the resource-
dispersing dimension. Third, the role of individual differences is
more evident in complex tasks that are more demanding of cognitive
abilities than simple tasks that require less mental effort and fewer
cognitive resources.
It can be seen that the CH affords a theoretical basis for the role of
individual difference factors in task performance involving learners’
existing knowledge and skills, but it does not spell out how these
variables impact L2 development or learning (gains in new knowledge
and skills). However, as we will see, empirical studies have investi-
gated both performance (e.g. Ahmadian, 2012) and learning (e.g. Li,
2013a, 2013b). Another caveat is that the matching of different types
of individual difference variables on the one hand and variables relat-
ing to task condition and task complexity on the other may not be as
transparent as predicted by the CH. For example, anxiety is an affect-
ive variable that is postulated to be drawn upon when tasks are
performed under different conditions, but it is also possible to posit
a logical link between anxiety and the procedural aspects of tasks,
such as ‘with or without planning time’ (Mak, 2011) or ‘with or

without task structure’ (Trebits, 2014). Nevertheless, the CH is the

only theory that has attempted to map the complicated relationships
among the three groups of task variables and foreground the import-
ance of learner factors in accounting for variation of task performance
in the triadic framework.

Language Aptitude
Overview
According to Carroll (1981), language aptitude is a componential
construct that consists of three cognitive abilities, namely phonetic
coding ability, language analytic ability (which entails grammatical
sensitivity and inductive learning) and rote memory, which correspond
to the learning of pronunciation, grammar and vocabulary respect-
ively. Language aptitude is considered to be (1) domain specific in the
sense that it is only important for learning a foreign language, (2)
distinct from other individual difference variables such as motivation
and anxiety, and (3) not subject to change. While some of these
characteristics have been empirically confirmed, others remain contro-
versial. Gardner and Lambert (1965) found that foreign language
learners’ scores on the subtests of the MLAT (Modern Language
Aptitude Test) (Carroll and Sapon, 1959) loaded on different factors
from their scores on the subtests of the PMA (Primary Mental Abil-
ities) – a test of academic intelligence – suggesting that language
aptitude involves distinct abilities for other academic subjects. How-
ever, in a meta-analysis on the construct validity of language aptitude
(Li, 2016), aptitude was found to overlap with intelligence. The meta-
analysis also found that aptitude was unrelated to motivation and
negatively correlated with anxiety. Regarding whether aptitude is
subject to change, there is no clear answer. While some studies found
that learners with more language learning experience had higher
aptitude scores than those with less experience (e.g. Einstein, 1980),
it is possible that those learners with more experience may have had
high aptitude to begin with. Therefore, there is a need for research to
show (1) the higher scores of those with more experience are not due
to their higher aptitude, and (2) the improvement in the same learners’
aptitude scores is only attributable to study experience instead of
maturation effects.
Language aptitude has been measured via test batteries consisting of
multiple subtests that tap the three components, and the most influen-
tial test is the MLAT. The MLAT was validated with more than 5,000
foreign language learners and therefore has strong predictive validity,

but it has been criticized on a number of accounts. First, it was

validated in the 1950s using traditional audiolingual classes charac-
terized by rote learning and mechanical practice. Thus, whether it is
relevant to current meaning-oriented approaches that emphasize the
importance of exposure to authentic linguistic materials and incidental
learning is questionable. Second, it was developed based on observa-
tions of what happened in language classes, not on SLA theories, and
therefore it lacks theoretical basis. Third, the five subtests do not
correspond with the three hypothesized aptitude components, making
it difficult to interpret the related findings. Fourth, these abilities are
only important for learning the formal aspects of language and for
learning language as discrete items and they do not account for how
the pragmatic and contextual aspects of a language are learned
(Skehan, 2002). Despite the criticisms levelled against the MLAT, it
is still the most dominant aptitude test in current research. Recent
developments include the Hi-LAB (High-Level Language Aptitude
Battery Test) (Linck et al., 2014), which targets high-level learners,
and tests of implicit aptitude (see Wen et al., in press).

Aptitude and TBLT

Aptitude research falls into two major categories: predictive and inter-
actionist, and TBLT falls into the latter. Predictive research aims to
investigate the associations between aptitude and learning rate regard-
less of learning conditions. Interactional studies, which are mainly
based on Robinson’s triadic framework (2011) and his Aptitude Com-
plexes Hypothesis (2002), seek to ascertain whether the role of apti-
tude or different aptitude components varies as a function of different
learning conditions such as the following four (see Robinson, 2002):
(1) short-term classroom treatments developed on the basis of a set of
pedagogical constructs, such as deductive vs. inductive instruction
(Hwu et al., 2014);
(2) laboratory-based treatments deﬁned and operationalized in terms
of the degree of explicitness such as explicit, implicit and inciden-
tal (de Graaf, 1997);
(3) instructional treatments involving interactional corrective feed-
back (CF) (Sheen, 2007);
(4) speciﬁc instructional approaches such as communicative teaching
(Ranta, 2002) or immersion (Harley and Hart, 1997).
Among the four streams of research, (1) and (2) are not entirely conducted
with instructional tasks; (3) concerns how focused tasks containing CF
facilitates L2 development; and (4) caters to some of the fundamental

principles of TBLT (Ellis, 2003): the primary focus is on meaning, the

tasks relate to the real world and task outcomes are non-linguistic.
The focus of this section is the studies in the third and fourth categories.
Aptitude and CF. As discussed in Chapter 2, there has been exten-
sive research on the role of CF in SLA because feedback embedded
within meaning-oriented tasks caters to an important principle of
TBLT – focus on form (Ellis, 2003; Spada et al., 2014; Long, 2015).
One line of feedback research concerns whether the effectiveness of
feedback is constrained by individual differences in language aptitude.
For example, Sheen (2007) conducted a classroom study where Eng-
lish as a Second Language (ESL) learners received recasts and metalin-
guistic feedback in learning English indefinite articles a/an. She
reported that language analytic ability only predicted the effects of
metalinguistic feedback, not those of recasts. Yilmaz (2013a) reported
that explicit correction was more effective than recasts only when
learners had high analytic ability. These two studies suggest that
aptitude is more clearly relevant in explicit learning conditions.
However, two studies (Trofimovich, Ammar and Gatbonton, 2007;
Sachs, 2010) that investigated computerized feedback reported that
aptitude was also important in implicit learning conditions such as
when no feedback or implicit feedback (recasts) was provided. How-
ever, the recasts in Trofimovich et al.’s study, which included a correct
model regardless of whether the utterance was correct, are not really
implicit. An explanation for Sachs’ finding is that the instructional
treatment required the learners to process the linguistic target to
complete the task, which drew on their analytic ability.
The influence of aptitude in CF also depends on the nature of
the linguistic target. Li (2013a, 2013b) investigated the three-way
interaction between feedback type, language aptitude and the linguis-
tic target. He found that in the learning of Chinese classifiers, analytic
ability was correlated with the effects of recasts but not metalinguistic
correction. In contrast, the data for perfective -le showed that
the reverse was true: analytic ability predicted the effects of metalin-
guistic correction, but not recasts. Li attributed this discrepancy to the
different linguistic properties of the two structures: the classifier is
syntactically and semantically simple, so the provision of metalinguis-
tic explanation levelled off the role of analytic ability. However, in the
recast condition where metalinguistic explanation was unavailable,
analytic ability came into play. The perfective -le is an opaque
structure that involves complicated form-meaning mapping, and
understanding the metalinguistic explanation poses challenges for
analytic ability. When metalinguistic information was absent, the
learners were unable to learn this complicated structure despite

support in the form of recasts (which were not effective in learning the
aspect marker). In this case, analytic ability did not play a role because
learners were unable to learn the linguistic target by relying on their
analytic and they were unable to benefit from the instruction.
Aptitude and meaning-focused language teaching. As mentioned,
the MLAT – the most influential aptitude test – was validated in
traditional audiolingual classes involving rote learning and mechan-
ical drills, which led to questions regarding whether it is relevant in
more meaning-oriented approaches such as communicative language
teaching (CLT) or immersion. Ehrman and Oxford (1995) stated that
the suspicion was unfounded because their study showed that aptitude
was the strongest predictor of learning in foreign language classes
which were ‘heavily influenced by the communicative teaching trends’
(p. 77). However, the classes that contributed the data were from
state-funded intensive programmes that, as the researchers admitted,
partly relied on drilling, and therefore the extent to which they were
communicative is uncertain. Stronger support for the relevance of
aptitude in CLT comes from Ranta’s (2002) study, which showed that
aptitude was significantly correlated with learning outcomes on mul-
tiple measures in classes judged to be communicative based on obser-
vations and interviews with the teachers. One caveat about Ranta’s
study is that aptitude was measured by means of a metalinguistic test,
not a validated measure such as the MLAT, although first language
(L1) metalinguistic knowledge has been shown to be related to lan-
guage analytic ability (Alderson, Clapham and Stee, 1997).
Harley and Hart’s studies (1997, 2002) show that aptitude was
implicated in French immersion classes for young learners where the
L2 was learned through exposure to the language. However, these
studies reported an interaction between age and aptitude components,
that is, the learners whose initial age of exposure was younger relied
on memory and the later starters on analytic ability. Harley and Hart’s
findings demonstrate that (1) aptitude is not only important in form-
based instruction but also in meaning-based instruction, (2) aptitude is
drawn on by young learners (10th and 11th graders), and (3) learners
of different age groups or at different stages of learning may draw on
different aptitude components.

Summary
The three types of studies discussed allow us to reach the following
tentative conclusions. First, the feedback research indicates that apti-
tude is more likely to be drawn on in tasks with an explicit focus on
form, which disadvantages low-aptitude learners. However, because

overall explicit feedback has proven more effective than implicit

feedback (Ellis, Loewen and Erlam, 2006; Li, 2010), at least in the
short term, it is advisable to make the corrective intention known to
the learner when CF is used as a form-focusing device in TBLT. (See
Chapter 2 for more detailed discussion of CF in TBLT.) Second, despite
the need for more research, the ﬁndings to date suggest that aptitude is
relevant in meaning-based instruction such as CLT and immersion
classes. Furthermore, the ﬁnding that younger and older learners draw
on memory and language analytic ability respectively suggests that a
heavy dose of form-focused instruction is not ideal for young learners.

Working Memory
Overview
Working memory refers to the ability to simultaneously store and
process incoming information. Baddeley (2007) proposed a compon-
ential model where working memory consists of a central executive
and three slave systems – a phonological loop, a visuospatial sketch-
pad and an episodic buffer. The central executive coordinates different
components, controls attentional shifts between meaning and form
and between information retrieval and task performance, and inhibits
irrelevant information (Miyake and Friedman, 1998; Juffs and Har-
rington, 2012). The phonological loop is responsible for storing and
rehearsing verbal information. The visuospatial sketchpad deals with
visuospatial information such as images, shapes and locations. The
episodic buffer integrates information from the slave systems and
long-term memory. Although working memory has been argued to
be a component of language aptitude, research has shown that it is
separate from aptitude (Li, 2017), probably because working memory
is a domain-general cognitive device that is essential for learning in
general, not just language learning.
Working memory has been measured in two ways – by using simple
tasks that only tap the storage component and complex tasks that
gauge both the storage and processing components (Conway et al.,
2005). Simple tasks include the word span or digit span tests that
require learners to repeat series of unrelated words, non-words or
digits. A complex task typically consists of two parts: one that requires
the learner to conduct some sort of information processing and one
that requires the learner to recall an element of the item in question.
For example, in a typical reading or listening span test, the learner
reads or hears sentences divided into sets of two to seven sentences
(called span sizes), judges their semantic or syntactic plausibility

(e.g. ‘The man standing in his ofﬁce was bitten by a wall’), and at the
end of each set, recalls the ﬁnal word of each item in that set. In
addition to listening or reading span tests, other measures of complex
working memory that have been used in the literature include oper-
ation span tests that ask the learner to perform some mathematical
computation and remember the letter or word that follows the equa-
tion in the item (e.g. ‘10/2 – 2 = 5 Q’) and backward digit span tests
where the learner is presented with sets of unrelated digits and asked
to recall the digits in the reverse order. Forward digit span is con-
sidered a simple task and backward digit span a complex task.

Working Memory and TBLT

We will now consider the research that has investigated how working
memory is implicated in task-based instruction. This research falls into
two broad categories: studies examining the effects of working
memory on task performance under different conditions and studies
exploring how working memory mediates the learning that results
from interactional feedback. These studies are based on three theoret-
ical models of TBLT and SLA: the Limited Attention Capacity
Hypothesis (LACH) (Chapter 3), the CH (Robinson, 2011) and the
Interaction Hypothesis (IH) (Long, 1996, 2015) – see Chapters 2 and
3. The LACH posits a central role for working memory – a limited
capacity device – in affecting learners’ task performance, which is
often assessed through measures of complexity, accuracy and ﬂuency
(CAF). The LACH draws on Levelt’s (1989) theory of speech produc-
tion, which holds that the production of spoken language undergoes
three stages: conceptualizing the message, formulating the language
representation (selecting the linguistic forms for the message) and
articulating the message. In Levelt’s model, the role of working
memory is restricted to message conceptualization, and formulation
and articulation are ‘underground processes’ (p. 22) that happen
without awareness and that are beyond attention control. However,
while it is perhaps true that formulation and articulation are auto-
matic in L1 oral production, L2 oral production often relies heavily on
attention control in all three phases, not only during message concep-
tualization. This suggests a more crucial role for working memory in
L2 production.
The CH states that complex tasks involving resource-directing
variables divert the learner’s working memory resources to the ‘input
that complex tasks promote’ (p. 19), and therefore the role of
working memory should be more evident in complex tasks than
simple tasks.

The theoretical justiﬁcation for a mediating effect of working

memory on the effects of interactional feedback can be found in the
IH (Long, 1996, 2015). This emphasizes the role of selective attention
when focusing on linguistic forms in meaning-oriented tasks as
learners switch attention between form and meaning, necessitating a
heavy reliance on working memory. In the case of CF, the learner must
attend to and temporarily hold the information contained in the
feedback and retrieve information from long-term memory in order
to process the available negative and/or positive evidence. At the same
time, the learner needs to maintain the continuation of the ongoing
discourse. Some feedback types such as output-prompting feedback
(Ellis, 2010) push the learner to modify their output, and this also
requires working memory resources.
Working memory and task performance. Studies investigating the
role of working memory in performing tasks have examined how it
interacts with planning, learner proficiency, +/ task structure and
task complexity. With regard to planning, researchers (R. Ellis, 2005)
distinguish pre-task or strategic planning and within-task planning
(i.e. whether learners are pressured to perform the task rapidly or
are given time to think about the information to be communicated
and the language needed).
Within-task planning, then, can be studied by determining the time
learners are given to perform a task, as in Ellis and Yuan (2004).
However, in many task-based studies within-task planning was either
not controlled or there is a lack of information about whether or not it
was controlled. Ahmadian (2012) is one of the few studies investi-
gating the role of working memory in careful online planning. The
study showed that working memory as measured through a listening
span test was significantly correlated with accuracy and fluency but
not complexity. Guará Tavares (2011) examined the effect of working
memory in pre-task planning and found that in the planning condi-
tion, the learners with high working memory outperformed those with
low working memory in terms of complexity and fluency but not
accuracy. Working memory did not affect the task performance of
the no-planning group. One problem with this study, however, is
that the working memory was measured using a speaking span test
where the learners were asked to create grammatically and semantic-
ally acceptable sentences with given words, which might be considered
more like a speaking test than a memory test. Also, there was no
information about whether and how online planning was restricted.
These findings are derived from separate studies conducted in dif-
ferent settings and using varying methods. One study that investigated
both pre-task and unpressured within-task planning with learners

from the same instructional context is Li and Fu (2018). The study

found significant correlations between working memory (measured
via an operation span test) and accuracy and fluency in the within-
task planning condition, but no significant correlations were found for
the pre-task planning condition. Also, the majority of the correlations
for the within-task planners were positive while those for the pre-task
planners were mostly negative, suggesting that larger working
memory capacity did not help in pressured performance.
The role of working memory may also be constrained by learners’
general L2 proficiency, as found by Gilabert and Munoz (2010). In
this research, adult L2 English learners at a Spanish university were
divided into high- and low-proficiency groups based on their scores on
the Oxford Placement Test. They performed a video narrative task
where they watched a video twice and retold the story with no pre-task
planning and no time limit for task performance. Learners’ working
memory scores on a reading span test were found to be only predictive
of lexical complexity for the high-proficiency learners’ oral perform-
ance and there were no significant results for the low-proficiency
group. One possible explanation is that at the lower-proficiency level
it was the learners’ linguistic proficiency rather than their working
memory capacity that affected task performance.
Kormos and Trebits (2011) reported a study examining whether
working memory had differential effects on structured vs. unstruc-
tured tasks. In the structured task, learners were asked to tell a story
based on a set of cartoon pictures sequenced in the correct order. In
the unstructured task, they had to invent a story based on a set of
unrelated pictures. In both tasks, learners were allowed two minutes to
plan before starting the narratives, but it is not clear whether there was
a time limit for task performance. Significant effects for working
memory were found for the structured task but not the unstructured
task, but the relationship between working memory and task perform-
ance was non-linear. For example, learners with high working
memory capacity outperformed those with lower memory abilities in
terms of clause length, but the latter performed better in terms of
subordination. The researchers speculated that although the struc-
tured task was assumed to be simpler, it may have turned out to be
more cognitively demanding because the learners had no choice but to
follow the provided storyline, which posed a greater challenge than in
the unstructured task where they had more freedom to draw on their
own linguistic resources. This might explain why working memory
was only predictive of the performance under the structured task. The
researchers further pointed out that working memory may not always

be beneﬁcial because those with higher working memory may try to

attend to too many aspects of performance, which can have a detri-
mental effect on their performance.
Finally, there is one study (Crespo, 2011) that investigated the
interaction between working memory and task complexity operation-
alized as +/ reasoning. Adult L1 Spanish EFL (English as a Foreign
Language) learners performed two versions of the same decision-
making task, the more complex version requiring learners to figure
out the relationships between more elements, consider more factors
when making decisions and have access to fewer resources. The study
included measures of three aspects of working memory: phonological
short-term memory, attention control – one function of the central
executive – and working memory as a global construct. Surprisingly,
neither working memory nor attention control showed strong correl-
ations with performance in either of the two task conditions. Phono-
logical short-term memory, however, was significantly correlated with
a number of outcome measures for both the simple and complex tasks.
The study failed to confirm Robinson’s prediction that complex tasks
are more likely to draw on working memory. It suggests that increas-
ing task complexity along the resource-directing dimension may not
increase the processing load. The study also suggests that despite the
putative links between working memory and L2 task performance,
phonological short-term memory, which has received little attention in
task-based research, may prove to be critical in speech production.
To sum up, the studies on the impact of working memory on L2
task performance showed the following:

(1) Working memory seems to be implicated in unpressured perform-

ance during within-task planning, while its role during pressured
performance after pre-task planning is inconclusive;
(2) the role of working memory is greater for advanced learners;
(3) tasks that provide a clear structure for performance may tax
learners’ working memory resources to a greater extent than tasks
without a clear structure (contrary to what is commonly assumed);
(4) there may be a non-linear relationship between working memory
and task performance and greater working memory capacity may
have adverse effects in some task conditions;
(5) complex tasks along the resource-directing dimensions do not
necessarily draw more on working memory than simple tasks;
(6) the role of phonological short-term memory in oral task perform-
ance may be of particular signiﬁcance – see the section ‘Working
Memory and CF’.

Working Memory and CF

Mackey et al. (2002) was the first study to explore the role of working
memory in noticing interactional feedback embedded in communica-
tive tasks and facilitating L2 development. Thirty ESL learners whose
L1 was Japanese were paired with native speakers of English and
performed three communicative tasks during which the learners
received recasts on errors relating to English question formation. The
results showed that learners with high working memory capacities
reported more noticing of feedback but that those with low working
memory scores manifested greater development initially. However,
those learners with high scores did better in the delayed post-test.
Kim et al. (2015) confirmed Mackey et al.’s findings. Working
memory was predictive of ESL learners’ noticing of recasts and the
effects of recasts on question formation in dyadic interaction. Kim
et al. also examined the +/ reasoning variable, reporting that task
complexity was not a significant predictor of the noticing of feedback
or learning gains. However, more learners with high working memory
in the complex task advanced to higher stages of question formation
than in the simple task.
Révész (2012) reported a complex interface between working
memory and outcome measures in a study of the effects of recasts on
learning. The study included two measures of phonological short-term
memory (digit span and non-word repetition) and one measure of
working memory (reading span). The effects of the recasts were meas-
ured by means of an oral description task, a written production task
and a grammaticality judgement task. It was found that phonological
short-term memory was correlated with gains in accuracy in oral
production and working memory with gains on the written tests.
Révész argued that phonological short-term memory facilitates the
acquisition of proceduralized/implicit knowledge whereas working
memory is more useful for the development of declarative/explicit
knowledge. The hypothesis about the differential roles of different
types of working memory in facilitating the acquisition of different
types of knowledge is important and needs to be investigated further.
Goo (2012) and Yilmaz (2013a) probed the interface between
working memory and feedback type, both studies investigating both
explicit feedback – metalinguistic feedback in Goo’s study and explicit
correction in Yilmaz’s – and implicit feedback – recasts in both studies.
However, they obtained different results. Goo found working memory
to be a significant predictor of the effects of implicit feedback while
Yilmaz reported it was correlated with the effects of the explicit
feedback. Goo explained that the learners who received recasts utilized

their working memory to notice the linguistic target, whereas the

metalinguistic feedback did not pose attentional demands. Yilmaz also
resorted to the concept of noticing when interpreting his results,
arguing that the recasts in his study were not explicit enough to trigger
the learners’ working memory for conscious learning. However, an
explanation is still needed for the conflicting findings for the explicit
feedback types. One possibility is that the explicit feedback in Goo’s
study contained rule explanation, which does not require heavy use of
the storage function, whereas the explicit feedback in Yilmaz’s study
took the form of explicit correction, necessitating storage of the cor-
rections in order to induce the general rule.
Li (2013a, 2013b) investigated the interaction between working
memory, feedback type and the nature of the linguistic target. Li
reported that working memory was drawn upon when learners received
metalinguistic feedback in learning both Chinese classifiers – a simple
structure and the perfective -le – a complex structure, but not when they
received recasts. However, one striking finding was that working
memory was a positive predictor of the effects of the explicit feedback
in the learning of classifiers but a negative predictor for the perfective
-le. In other words, learners with high working memory capacity
benefited less from metalinguistic feedback when learning a complex
linguistic structure. Li referred to Newport’s Less Is More Hypothesis
(1990) when interpreting the results, that is, with high working memory
capacities tend to store linguistic input as large chunks and ignore the
detail, while those with smaller working capacities engage in deeper
processing of the linguistic input. Li’s studies again show the contingent
relationship between cognitive resources and task conditions.

Summary
It seems that learners make heavy use of their memory resources when
planning their speech during unpressured performance. Allowing pre-
task planning may alleviate the burden on working memory. Further-
more, given the positive effect of unpressured online planning on task
performance, allowing both pre-task planning and unpressured
within-task planning can be expected to have even greater effects on
task performance. One important implication from this line of
research is that tasks that are assumed to be simple may turn out to
be complex and consequently be more taxing on working memory
resources. A further ﬁnding is that working memory is implicated in
processing the online feedback embedded in communicative tasks. To
date, however, there has been no study investigating the role of
working memory in delayed, ofﬂine feedback.

Motivation
Overview
Motivation is considered a primary determinant of L2 success, which
explains why it has been one of the most extensively studied individual
difference factors. Dörnyei (2005) explained that the importance of
motivation lies in that fact that it ‘provides the primary impetus to
initiate L2 learning and later the driving force to sustain the long and
often tedious learning process’ (p. 65). Motivation, according to Ellis’
(2015b) synthesis, is a complex construct consisting of three compon-
ents: (1) the reasons why a learner wants to learn an L2, (2) the effort
one invests in the learning process and how it is influenced by the
immediate context, and (3) the impact of the evaluation of the
outcome and progress of learning on subsequent behaviour.
Of these three aspects of motivation, those in (1) constitute general-
ized, macro motives that relate to the general goal to be achieved and
the general orientation towards the language, culture and speech
community. These include the traditional integrative and instrumental
motivation in Gardner’s (1985) model, with the former referring to
motives arising out of positive attitudes towards the speakers of the
target language and the desire to integrate and identify with the
community, and the latter to pragmatic motives such as getting a job
or promotion. Those involved in (2) and (3) can be regarded as the
specific, micro aspects of motivation that relate to the process of
learning or the ongoing learning tasks.
The tripartite framework also incorporates the dynamic, situated
model of motivation proposed by Dörnyei and his associates (Dörnyei
and Ottó, 1998; Kormos and Dörnyei, 2004; Dörnyei, 2005; Dörnyei
and Ushioda, 2009), which differs from the traditional static model,
where motivation is viewed as a trait that correlates with the ultimate
learning outcomes. In this model, motivation is (1) subject to temporal
variation and (2) influenced by multiple contextual factors such as the
school, the course, the class and the target language. Dörnyei’s ideas
are well represented in the so-called ‘process model’ (2005, p. 84)
where different conglomerates of motives are drawn on at different
stages of learning. At the pre-actional stage prior to the start of the
learning process, learners’ motivation is generated and the goal is set.
This is called choice motivation, and it relates to learners’ general
dispositions or the macro factors in Gardner’s model. During the
actional stage or the learning process, the general motivation is influ-
enced by various supporting factors as well as factors that inhibit
distractions and supress unfavourable behaviours (e.g. off-task

behaviours). These factors inﬂuence executive motivation. In the post-

actional stage when the learning process/task is completed, the learner
makes retrospective evaluations of what transpired during the learning
process, and the results of the evaluation in turn affect subsequent
actions. This concerns the attributional dimension of motivation.
One motivation theory proposed by Dörnyei (2005) that has had a
profound inﬂuence on recent research is the L2 self system. This entails
three dimensions of motivation: Ideal L2 self, ought-to L2 self and L2
learning experience. The Ideal L2 self refers to the motivation driven
by the desire to reduce the discrepancy between one’s current state and
the future state to be reached (e.g. ‘I often imagine myself speaking
English ﬂuently’). The ought-to L2 self concerns the motivation that
prompts the learner to study an L2 to avoid the negative consequences
(e.g. ‘If I don’t study English, others will be disappointed’). L2 Learning
Experience relates to the motives associated with the immediate con-
texts of learning as determined by the curriculum, the course materials
and the teacher (e.g. ‘I always look forward to English classes’).
Whereas the two types of self-related motives are associated with the
macro dimensions of motivation drawn upon in the pre-actional stage,
the motives relating to learning experience concern the micro aspects of
motivation and are involved in the actional stage and the post-actional
stage (although Dörnyei was not clear about which of the three types
of motivation are important in the post-actional stage).

Motivation in TBLT
Task motivation encompasses all motives that may affect task per-
formance or engagement at any of the three stages of a task cycle (pre-
task, main task and post-task). It will be influenced by both general
motives such as the Ideal Self and the Ought-to Self but in particular
by the more specific motives relating to the task-as-workplan and task-
as-process (see Chapter 1), such as attitudes towards the task, percep-
tions about the difficulty or complexity, how the task is implemented
and the other participants, all of which are subsumed under the L2
Learning Experience component of the L2 Self System. In this section,
we discuss what the little empirical research has shown about task
motivation and how the concept of motivation has been investigated
in relation to TBLT.
Dörnyei (2002) was one of the first to explore the multifaceted
nature of task motivation in dyadic interaction. The study included
measures of different levels of motivation, including generalized dis-
positions such as integrative and instrumental motivation, as well as

motivation relating to the immediate contexts such as attitudes

towards the course and the task. The outcome measure was task
engagement, which was operationalized as the number of words and
turns produced by the learners (forty-four Hungarian secondary
school EFL students) in an argumentative task. It was found that all
motivation variables except for integrative motivation were signifi-
cantly correlated with task engagement. However, a different picture
emerged when the learners were divided into two groups based on
their task attitudes. Whereas both generalized and task-specific
motives were predictive of high-task attitudes and learners’
engagement, the engagement of learners with low-task attitudes was
only predicted by course attitudes. This suggests that their lack of
interest in the task was compensated for by their positive attitudes
towards the course. The study also found that the task engagement of
the learners with low-task attitudes was affected by their interlocutors’
motivation. Taken together these results indicate that both task-
specific and general motives contribute to task motivation and that
task motivation is co-constructed.
Drawing on the same data as Dörnyei (2002), Kormos and Dörnyei
(2004) found that although the learners’ task attitudes predicted the
quantity of their oral production (number of words and turns), this
factor did not have a positive effect on the quality (the linguistic
accuracy and complexity) of their production. Course attitudes were
found to be significantly correlated with both quantitative and quali-
tative aspects of production. In a study by Al Khalil (2011), learners’
general motivation measured by Gardner’s Attitude/Motivation Test
Battery (AMTB) (1985) was significantly correlated with the CAF of
the oral production of forty-four L2 Arabic learners in the United
States when engaged in dyadic communication with a native speaker.
Integrative motivation was significantly correlated with the noticing of
recasts, an index of task engagement. These two studies seem to
indicate that task-specific motivation affects task engagement, and
that more general types of motivation may affect both task
engagement and the linguistic aspects of task performance. Following
Dörnyei and Ushioda (2009), we can suggest motivation is best con-
ceived as not directly related to achievement but as the antecedent of
action that may have indirect effects on achievement. It would follow
that measures of motivated behaviours such as task engagement are
better measures of motivation.
Dembovskaya (2009) examined whether pre-task motivation-
enhancing strategies lead to more positive perceptions of learners’ task
experience and whether pre-task cognitive strategy training improves
students’ task performance in terms of the CAF of their oral

production. L2 French learners at a US university performed an

information-gap task where they received a list of objects and clues
found in the apartment of a crime suspect and were asked to work in
groups to come up with a description of the suspect. To motivate one
group of learners the teacher informed them about the value of the
task (e.g. by telling them it would help them improve their communi-
cative competence), enhanced their interest (e.g. by telling students
that other students had performed the task and found it enjoyable) and
promoted their self-confidence (e.g. by telling students that they could
do it well). The students in the other group were equipped with the
linguistic and strategic tools needed to complete the task, such as
activating their schematic knowledge and informing them what infor-
mation to look for to identify a suspect. The results showed that the
group that received the pre-task motivational instruction perceived the
task to be more interesting and valuable and themselves as more
autonomous than the groups that did not receive motivational
training, but the finding was only true of Year 2 students, not Year
3 students. Also, there was no difference in the two groups’ perform-
ance of the task. Noteworthy is the fact that although the pre-task
phase can be thought of as corresponding to the pre-actional stage in
Dörnyei’s process model as it relates to so-called ‘choice motivation’,
the pre-actional stage in Dörnyei’s conceptualization concerns L2
learning in general and therefore is seen as involving generalized
motives. In this study, the pre-actional stage concerns a task and,
consequently, the choice motivation that the learners received training
for relates more clearly to task-specific motivation.
Jauregi et al. (2012) investigated whether learners’ motivation can
be improved through authentic, video-web communicative tasks.
A group of L2 Dutch learners in Czech attended three 30-minute
virtual interaction sessions with some Dutch pre-service language
teachers. A thirteen-item questionnaire was utilized to measure differ-
ent dimensions of motivation – attitudes towards interacting with
native speakers, attitudes towards the course and attitudes towards
the L2 culture. Significant effects were found and the effects were more
prominent for the beginner group than the more advanced learners.
What is unique about this study is that motivation was examined as a
dependent variable that served as the ‘effect’ rather than ‘cause’, and it
(both integrative motivation relating to the culture and speech com-
munity and more specific motivation relating to the course) was found
to be improved through task-based interaction.
One commonality between these two studies is that learners’ motiv-
ation seems to more easily influenced at the beginning stage of L2
learning than at more advanced stages. One possible explanation is

that the attitudes and perceptions of more advanced learners about the
learning process or tasks are entrenched and not easily swayed. It can
also be speculated that learners that choose to proceed to higher stages
of learning are more motivated to begin with. In any case, it would
seem more important to enhance learners’ motivation at the beginning
stages of learning, so the interface between motivation and learner
proﬁciency seems to be a promising area of research.

Summary
The few studies that have investigated task motivation show that it is a
dynamic, complex and multi-componential construct. It consists of
motives relating to different facets of the learning task, including
general learning goals, course motivation and motives to do with the
performance of the task per se. These motives have been found to be
signiﬁcant predictors of learners’ task engagement and performance.
Pre-task motivational strategies can enhance learners’ motivation to
perform a task while participation in meaning-oriented tasks can
improve learners’ course and integrative motivation. Finally, initia-
tives to increase learners’ motivation seem to work better for low-level
learners than high-level learners, suggesting that practitioners should
make a special effort to stimulate and maintain beginning L2 learners’
motivation when implementing TBLT.
Motivation is one of the most promising areas of TBLT research
and Robinson’s (2011) triadic framework provides a useful frame-
work for so doing. For example, while increasing the cognitive
demands of a task may enhance learners’ task performance, increasing
task complexity beyond a certain threshold may have a harmful effect
on learners’ motivation, which may in turn have adverse effects on
their task performance and engagement. Second, with regard to the
variables relating to task condition, research on whether and how
factors pertaining to participatory structure and participant character-
istics affect motivation may provide valuable insights for the imple-
mentation of TBLT. For example, Dörnyei’s (2002) ﬁnding that
learners’ motivation was affected by their partners’ motivation sug-
gests that it is advisable to pair up learners’ with different levels of
motivation. However, Dörnyei and Kormos (2000) found that the
relationship between task participants was predictive of task
engagement when the task was performed in their L1 but not when
the task was performed in the L2. This suggests that participants’
relationships may not be as important as teachers have assumed but
clearly more research is needed. Third, it is surely important to exam-
ine how motivation in contrast to and in combination with other

individual difference factors affects task performance but to the best of

our knowledge, there has been no research in this regard.

Anxiety
Overview
Horwitz, Horwitz and Cope (1986) defined anxiety as ‘the subjective
feeling of tension, apprehension, nervousness, and worry associated
with an arousal of the autonomic nervous system’ (p. 125). Three
types of anxiety have been identified in the literature: trait anxiety,
state anxiety and situation anxiety (Ellis, 2015b). Trait anxiety is a
personality variable that refers to the general disposition, state anxiety
relates to one’s emotional condition at a particular moment and
situation anxiety is associated with what one experiences in particular
contexts. Trait anxiety accounts for interpersonal variation, that is,
certain individuals are inherently more anxious than others. State and
situation anxiety can be considered as intrapersonal variables in the
sense that the same individual may experience different levels of
anxiety at varying moments in a particular situation and in different
situations. Language learning anxiety is a type of situation anxiety,
and it occurs when a leaner produces or comprehends an L2. Horwitz
et al. (1986) argued that language learning anxiety is principally
derived from three sources: spontaneous communication, fear of nega-
tive evaluation and test anxiety. In L2 research, anxiety has been by
default been associated with speaking, and measures of anxiety –
typically questionnaires – primarily consist of speaking-related items
(Phillips, 1992; Aida, 1994), although anxieties for other skills such as
listening (Elkhafaifi, 2005), writing (Cheng, Horwitz and Schallert,
1999) or reading (Saito, Horwitz and Garza, 1999) have also been
investigated. Furthermore, anxiety can be debilitative or facilitative,
that is, while too much anxiety may have a negative influence on task
performance or learning outcomes, a certain amount of anxiety may
play a positive role. However, the distinction has not received much
attention in empirical research.
In general, anxiety has been found to have negative effects on
language learning (e.g. Ewald, 2007), which is in line with the harmful
effect of anxiety on general academic performance (r = 0.25),
according to a meta-analysis of 126 studies (Seipp, 1991). Tobias
(1985) attributed the adverse effect of anxiety to cognitive interfer-
ence, that is, anxiety-prone learners have to split their cognitive
resources between task-relevant and task-irrelevant processes, thereby
affecting their task performance. Drawing on Tobias’s model,

MacIntyre and Gardner (1994) posited that anxiety causes interfer-

ence at all three stages of learning: (1) when learners receive linguistic
input, (2) when learners try to organize and store input, and (3) when
learners are required to produce previously learned material.
MacIntyre and Gardner developed the Input, Processing and Output
Anxiety (IPOA) scale to measure the types of anxiety for the three
stages of learning, with six items for each of the three stages. Theoret-
ically sound as it might be, some of the items in the questionnaire do
not seem to be clear measures of the three types of anxiety. For
example, the items for input and processing are not clearly distinguish-
able, and one item in the output section is about test anxiety. Finally,
while the mainstream view is that anxiety is a cause for low achieve-
ment, Sparks and Patton (2013) contended that it is the consequence
of learning difﬁculties or lack of aptitude.

Anxiety and TBLT

The research on the role of anxiety in TBLT has revolved around three
themes. One is the correlation between anxiety and task complexity.
As discussed, the CH (Robinson, 2011) predicted that stronger correl-
ations between anxiety and task performance can be expected when
task complexity is increased along either resource-directing or
resource-dispersing dimensions. Put in another way, as the processing
demands of tasks increase, the negative impact of anxiety becomes
more evident. A second line of research has focused on whether
learners’ anxiety levels vary as a function of the modality of inter-
action. For example, computer-mediated (CM) task-based instruction
may alleviate learners’ anxiety in the absence of the pressure and
interaction demands that characterize face-to-face communication.
A third stream of research focuses on the mediating effects of anxiety
on the learning that results from CF provided in communicative tasks.
In the following, we discuss these studies in more detail.
Task complexity. There have been two studies investigating the
correlations between anxiety and the resource-directing variable of
‘with or without reasoning demand’. In Robinson’s (2007c) study,
forty-two L1 Japanese university EFL students formed twenty-one
dyads, each performing three narrative tasks at different levels of
reasoning demand. Each dyad was given a set of jumbled pictures.
One learner narrated the story based on the sequence he/she decided
on and the other put the pictures in the sequence based on the
speaker’s narrative. Anxiety was measured using MacIntyre and
Gardner’s (1994) IPOA scale. Output anxiety was found to be nega-
tively correlated with syntactic complexity, and with the increase of

task complexity the correlations became stronger. However, one inter-

esting finding that was not discussed in detail is that processing anxiety
was significantly and positively correlated with the accuracy of the
production under the simple task condition.
Kim and Tracy-Ventura (2011) examined the mediating role of
anxiety in affecting the learning of the English past tense morphology
under three task conditions that differed in terms of the presence of
reasoning demand (simple vs. complex) and number of elements
(complex vs. more complex). The study involved 128 Korean EFL
learners who performed four dyadic interaction tasks within a two-
week period. The researchers found that the learners with low anxiety
(measured through a six-item questionnaire) outperformed their high-
anxiety peers in all task conditions. They concluded that there was no
interaction between anxiety and task complexity because the role of
anxiety did not vary across task conditions.
Whereas these two studies concern a resource-directing variable that
involves information manipulation, Trebits (2014) examined a
resource-dispersing variable relating to the procedural dimension of
task complexity: single vs. dual task. In this study, which is based on
the same data as Kormos and Trebits (2011) (discussed in the section
on ‘Working Memory and TBLT’), the learners performed a cartoon
description task where they told a story following a given sequence,
and a picture description task where they told a story based on unre-
lated pictures that must be sequenced logically during the narrative.
The cartoon task was easier in terms of content organization than the
picture task – a dual task condition where the learners had to attend to
both content organization and language formulation (selection of
linguistic forms). However, the author argued that the cartoon task
was more challenging in terms of language formulation than the
picture task because the former required the learners to select linguistic
forms to match the prescribed content while the latter allowed learners
the flexibility of tailoring the content to match their linguistic reper-
toire. As in Robinson (2007c), anxiety was measured via the IPOA
battery. It was found that: (1) in the cartoon task, processing anxiety
correlated positively with lexical and syntactic complexity, and (2) in
the picture task, output anxiety correlated negatively with accuracy.
As can be seen, it is difficult to draw unequivocal conclusions about
the interface between anxiety and task complexity due to the conflict-
ing findings, which in turn may have resulted from the methodological
differences between the studies. However, these studies did show some
interesting patterns. First, anxiety does have some negative effects on
task performance and L2 development. As Robinson (2007a) and
Trebits (2014) showed, it had a negative impact on task performance,

particularly in complex tasks. Kim and Tracy-Ventura (2011) showed

that low-anxiety learners consistently outperformed high-anxiety
learners while learning the English past tense. Second, it would seem
that anxiety may be facilitative of some aspects of oral production in
simple tasks where the burden of content organization is alleviated
and where anxiety may serve as an impetus for more attention to the
linguistic aspects of their task performance. This is evidenced by the
finding in both Robinson and Trebits’s studies that significant, positive
correlations were found for anxiety in the simple task conditions.
Third, Kim and Tracy-Ventura found that low-anxiety learners out-
performed high-anxiety learners in both simple and complex task
conditions. However, the data also showed an advantage for complex
tasks in comparison with simple tasks in enhancing L2 development
regardless of anxiety – a finding that was not discussed due to the
focus of the study. This raises the question of whether priority should
be given to the effectiveness of instruction or the concern over the
possible negative consequence of incurring more anxiety by using
complex tasks.
Task modality. Computer-mediated (CM) communication in the
form of text chat has been assumed to be effective in easing learners’
anxiety in comparison with face-to-face communication because of the
opportunity for more online planning and lack of requirement for
public performance in the former mode. However, Baralt and
Gurzynski-Weiss (2011) showed that there was no difference between
the two modes of communication in terms of the amount of anxiety the
learners experienced. What is unique about this study is that it exam-
ined state anxiety, that is, whether there was any change in learners’
anxiety measured during and after task performance. Twenty-five
fourth-semester Spanish learners from a large public university in the
United States performed an information-gap task where they engaged in
dyadic interaction with a native-speaker interlocutor in CM and face-
to-face communication. The results revealed no difference in the
learners’ anxiety between the two modes of communication either
during or after task performance. The researchers explained that this
might be because the novelty of engaging in online chat in the foreign
language led to anxiety levels comparable to face-to-face communica-
tion. While this stands to reason, an alternative speculation is that the
study happened in a laboratory setting that did not require public
performance, which is often the direct cause of anxiety.
In another study, Satar and Ozdener (2008) compared the effects of
text and voice chat on the development of Turkish EFL learners’ oral
proficiency and on learners’ anxiety. Ninety high school students
were divided into three groups: text chat, voice chat and control.

The learners in the experimental groups were paired up and performed

two tasks in each of four treatment sessions. The task types included
information gap, problem-solving, jigsaw and decision-making.
The control group followed the normal curriculum and did not receive
any treatment. The learners also answered the Foreign Language
Classroom Anxiety Scale (FLCAS) anxiety questionnaire (Horwitz
et al., 1986) before and after performing the tasks. The results showed
that (1) both chat groups outperformed the control group in their
speaking performance as a result of the treatment, and (2) the text
chat group showed significantly lower anxiety scores after the treat-
ment, although the anxiety level of the voice chat group was also
reduced (the change was non-significant). The researchers explained
that online chat led to lower anxiety because it was a secure environ-
ment without peer pressure. However, surprisingly, the control group
showed consistently higher anxiety scores than both experimental
groups both before and after the study even though they did not
participant in any chat sessions.
In these two studies, interaction happened either between a native-
speaker interlocutor and a learner or between two learners. A recent
study by Côté and Gaffney (2018) examined the impact of anxiety on
learners’ public performance in front of a large group. The researchers
claimed that this setting is more representative of a real classroom and
therefore the findings have higher ecological validity. The study
involved sixty-one beginning learners of French at a Canadian univer-
sity who completed a grammar and vocabulary lesson followed by two
production activities in a face-to-face mode and a CMC (computer-
mediated communication) mode. The learners also answered a revised
version of the FLCAS questionnaire. The study revealed that the
learners experienced significantly less anxiety and produced signifi-
cantly more turns and words in the CMC mode than the face-to-
face mode.
Overall, the limited research on the effects of modality of instruction
delivery on learner anxiety show that learners indeed suffer from
lower anxiety in the CMC mode compared with the face-to-face mode.
However, the outcomes investigated in the studies relate primarily to
the process aspects of task-based instruction, such as the number of
words produced, and more research is necessary whether anxiety
affects the product dimensions or learning gains in the two modes of
instruction in different ways.
Task-based feedback. There have been two studies on the mediating
role of anxiety in affecting the effectiveness of CF embedded in com-
municative tasks (Sheen, 2008; Rassaei, 2015). In Sheen’s study, sixty-
one learners from a large ESL programme in a US community college

were divided into two groups: recast and control, each subdivided into
high and low based on their anxiety, which was measured through a
six-item questionnaire. The recast group performed two narrative
tasks, each including a practice stage where they worked in groups,
followed by a reporting stage where each student produced a few
sentences before passing the speaker role to another group member.
The teacher corrected their errors on English articles a/the by
using recasts. The results revealed that low-anxiety learners not only
benefited more from recasts in learning the target structure but also
produced more modified output (responses after feedback) than high-
anxiety learners.
Rassaei carried out a study with Iranian EFL learners from a private
language teaching institute following Sheen’s procedures, but the
researcher included two types of feedback: recasts and metalinguistic
correction, aiming to see whether anxiety has differential impacts on
the effects of the two types of feedback. The study found that high-
anxiety learners benefited more from recasts, and low-anxiety learners
benefited from both recasts and metalinguistic feedback. However, a
closer inspection of the results showed that in the recast group, high-
and low-anxiety learners were similar in their scores across all three
tests of treatment effects, but in the metalinguistic group, low-anxiety
learners performed consistently better than high-anxiety learners. It
would seem that anxiety did not play a role when recasts were pro-
vided – a finding that is different from Sheen’s finding that anxiety
played a negative role in affecting the effects of recasts. However, one
piece of useful information to take away from this study seems to be
that metalinguistic feedback – an explicit form of correction – does
have an adverse effect on learning outcomes.

Summary
To conclude this section, we would like to point out that the amount
of research on the role of anxiety is in disproportion to its importance
given its putative connections with TBLT. The limited research seems
to suggest that anxiety is unfavourable for speech performance under
complex task conditions, and that there is a possibility that simple
tasks favour anxious learners by freeing up their cognitive resources
and diverting their attention to the linguistic aspects of their perform-
ance. In terms of its role in affecting the learning of new linguistic
knowledge, anxiety seems to have a harmful effect, regardless of task
complexity. The research on task modality suggests that in laboratory
settings text-based CM interaction seems to lessen learners’ anxiety
compared with face-to-face interaction, and that within CMC, text

chat has a positive effect in reducing anxiety compared with voice

chat. The research on CF indicates that explicit feedback in the form of
metalinguistic correction disadvantages highly anxious learners while
mixed ﬁndings are obtained for whether anxiety affects the effects of
recasts.

Final Comments
In general, the research on aptitude was conducted to ascertain
whether aptitude mediates the effects of different instructional treat-
ments or whether it is implicated in different learning conditions. The
research has shown that traditional aptitude is a set of cognitive
abilities that are most likely drawn on in tasks with an explicit focus
on form. One promising area of research that has been recently
initiated is identifying those abilities that are important in implicit or
unconscious learning (Granena, 2013, 2015), which TBLT is claimed
to facilitate. It would be interesting to ascertain whether a task-based
approach draws more on implicit aptitude than traditional aptitude,
which has been found to be relevant to explicit learning.
The research on working memory explored whether it affected
learners’ performance under different task conditions and how it
mediated the effectiveness of interactional feedback. One general find-
ing supported by several studies is that the role of working memory is
more evident in unpressured performance where learners have oppor-
tunities to plan the content and language of their speech. The role of
working memory in mediating the effects of CF has been attributed to
its function in noticing the corrective force of feedback. One objective
of this stream of research should be to identify tasks that facilitate task
performance or L2 learning but do not pose heavy processing
demands on working memory resources. For tasks that enhance task
performance or learning outcomes but are taxing on working memory
resources, it is important to find ways to support learners by alleviat-
ing the processing load.
The few studies of task motivation investigated the predictive power
of generalized and task-specific motives for task engagement and
performance and ways to improve task motivation. Unlike the
research on the two cognitive variables, studies on motivation are less
uniform and the findings are less robust, reflecting in part the lack of
theorization about its role in L2 acquisition (Dörnyei, 2005; Ellis,
2015b) and the uncertainty over the nature, composition and
measurement of the construct. However, as pointed out, the investi-
gation of task motivation may provide important insights, given that
students’ lack of motivation has been considered a major hindrance to

the implementation of TBLT in some contexts – for example, in

foreign language as opposed to L2 settings (Ellis, 2003).
The research on anxiety affords a complex, interesting picture that
to some extent undermines the commonly held beliefs about this
affective factor. For example, while it did have some adverse effects
under some task conditions, there was also evidence for a possible
positive effect for anxiety. One principle that should be adhered to is
to (1) prioritize task or instruction type when facing the choice
between tasks that enhance learning and those that cause less anxiety,
but (2) ﬁnd ways to adapt the aspects of the task or instruction that
disadvantages highly anxious learners. For example, if complex tasks
consistently show larger effects than simple tasks in improving
learners’ L2 knowledge for both high- and low-anxiety learners, as
Kim and Tracy-Ventura’s (2011) data showed, then it would be
advisable to use complex rather than simple tasks even though com-
plex tasks may cause more anxiety. However, teachers then should
take steps to reduce learners’ anxiety by manipulating the procedural
aspects of task, such as by allowing learners to plan before performing
a task. However, to date, there has been no research on the role of
anxiety in different planning conditions.
Finally, we would like to call for more research into the role of
individual difference variables in affecting task performance and learn-
ing gains. It is a mistake in tacitly (and mostly) assuming that tasks
work in the same way for all people, and this concern has been borne
out by the limited research we synthesized in this chapter.

Introduction
This chapter provides an overview of some key aspects of twentieth-
century educational philosophy in relationship to contemporary
task-based language teaching (TBLT) principles and practices. It
constitutes a very different approach to the preceding chapters in
Part II, which drew on theory and research in second language acqui-
sition (SLA). As we noted in Chapter 1, the initial impetus for TBLT
came from SLA but educational perspectives have increasingly
informed developments, drawing attention to how general educational
principles can shape TBLT and reinforce the perspectives offered
by SLA.
Theories of experiential learning will be the focus of the chapter as
these are particularly relevant to TBLT. The chapter begins by outlin-
ing key principles of TBLT as reference points for discussing experien-
tial learning. Some key theories are then outlined and TBLT principles
and practices are weighed against them to determine the extent to
which the essential aspects of experiential learning theory have been
addressed in the TBLT literature to provide an educational rationale
for TBLT and point to possible directions for future work.
The ﬁrst three principles form the foundation of TBLT in relation to
other approaches to second language (L2) instruction:
1. Learning by doing: In TBLT, language learning takes place
through task performance rather than for task performance. Lan-
guage learning is seen primarily as an incidental process that takes
place in line with learners’ communicative needs.
2. Individual development: Learning takes place in line with learners’
internal syllabuses. Tasks provide the space for learners to integrate
task content and their own language resources. In TBLT, learners
use their own language to successfully accomplish tasks in their
own ways.

155

3. Relevance: Tasks present learners with real communicative

demands. The abilities that are developed in connection with tasks
are thus directly relevant to learners’ communicative needs and
their conceptions of what being proficient in a language involves.
In TBLT, the learner is thus the agent in the learning process, and
teachers or course designers serve to facilitate this process through
planning and implementation of learning opportunities. This process
of facilitation typically involves additional principles by which
learners’ performance of tasks is optimized:
4. Interaction: Although not all target tasks involve interaction, TBLT
provides ample opportunities for learners to negotiate meaning and
develop shared understanding based on their own L2 resources.
Interaction is a driving force in language acquisition (Long, 2015),
and learners must be able to negotiate solutions to the range of
tasks that they face effectively and appropriately (Ellis, 2003).
5. Input-based tasks (listening, reading): In addition to opportunities
to perform interactive tasks, learners at different ages and profi-
ciency levels can benefit from performing receptive tasks in which
they are not required to produce language (Ellis, 2001, 2018a:
Long, 2015; Shintani, 2016).
6. Output-based tasks (speaking, writing): Likewise, in addition to
opportunities to perform interactive tasks, learners must have
opportunities to produce extended discourse (Yule, 1997), test
hypotheses and notice the aspects of language that connected dis-
course requires (Swain, 1995).
7. Focus of form: There is not a one-to-one relationship between task
performance and language learning. Learners’ attention must be
drawn to forms that are difficult to acquire incidentally during the
performance of communicative tasks (e.g. Long, 1991b, 2015;
Ellis, Basturkmen and Loewen, 2002).
8. Supporting performance: Learners need opportunities to optimize
their own task performance by being provided with time to plan
and reflect on their performances as well as repeated opportunities
to complete tasks and improve their performances (e.g. Skehan
et al., 2012).
9. Individual differences: Learners have different aptitude,
motivation and anxiety profiles. Optimizing the performance of
learners with different learner profiles on tasks requires adjust-
ments to task-based instruction (TBI) (Chapter 5; Robinson, 2011).
These nine principles illustrate the focus on the learner and the learning
process in TBLT. These principles are consonant with twentieth-century

educational philosophy and will be the focus of the present chapter.The

chapter will begin with Dewey’s (1913, 1938) theory of interest and
effort in education and the ways in which TBLT is consistent with this
theory. However, the roots of this theory stretch back to the eighteenth
century (see Samuda and Bygate, 2008; Long, 2015, for reviews). Subse-
quent developments in educational theory related to Dewey’s work will
then be outlined and applied to contemporary TBLT practices. Recent
empirical work will then be summarized that points to was of expanding
TBLT practice in line with the educational theories discussed.

Learning by Doing
Experiential learning involves learners acting on and refining what
they know in order to achieve specific outcomes and objectives (TBLT
Principle 1). In particular, the aim of instruction is to develop the
learner’s ability to independently observe and evaluate immediate
conditions and organize their own means to achieve their purposes
within a given context (Dewey, 1938, p. 28) (TBLT Principle 2). In
designing instructional materials, it is thus essential that educators tap
into the possibilities inherent in ordinary experience (p. 89) and make
learning relevant to learners (TBLT Principle 3). This section will
provide an overview of Dewey’s theory of developing experience. In
this theory, affect and cognition are integrated and inseparable. As
Swain (2013) points out, this is also the case in Vygotsky’s sociocul-
tural theory of mind (1978, 1987, 2000; Chapter 4). Subsequent work
has tended to address one dimension or the other, however, and the
cognitive dimension of learning has been the dominant concern
in TBLT.
Dewey (1938) argues that for meaningful growth to occur, the
learner must be the agent in the learning process. This requires intelli-
gent effort on the part of the learner (p. 69). Intelligent effort,
according to Dewey, involves learners initiating activity, taking con-
trol of it, and understanding the consequences of different behavioural
alternatives in achieving different ends. Dewey argues that to generate
intelligent effort, learners must feel personally involved in the purposes
which direct their actions (1938, p. 67) and have a sense of responsi-
bility for the outcome of learning activities (pp. 53–61). According to
Dewey, a genuine purpose always begins with an impulse or a desire
on the part of the learner, and the intensity of this drive will ultimately
determine the strength of the effort that the learner puts forth in
transforming this impulse or desire into a plan of action and a method
for achieving its fulfilment. Impulse and desire give impetus to action,
but intelligent effort directs it towards an end (p. 69).

Dewey (1913) elaborates the role of interest and effort in the learn-
ing process. In doing so, he distinguishes interest from feeling. Dewey
defines feelings as mental states that do not lead to further action on
the part of the learner, and he argues that it is a mistake to orient
instruction towards learners’ feelings. For example, activities might
amuse learners for a while (e.g. hearing anecdotes or jokes, watching
movies, listening to popular songs), and learners may enjoy the diver-
sions that these activities provide, but this type of feeling of enjoyment
quickly dissipates and learners immediately require new stimuli from
the teacher to maintain their attention. According to Dewey, instruc-
tional activities aimed only at pleasing or amusing learners have no
real educational value. In contrast, Dewey argues that the essential
characteristic of educational activities is that they generate interest on
the part of learners. Dewey uses interest in a technical rather than
colloquial sense. According to Dewey, it has two essential characteris-
tics. First, interest, in contrast to feeling, relates to an object external to
the learner and generates action on the part of the learner aimed at
achieving that object. In other words, we are interested in something,
and this object of interest channels our attention and our subsequent
behaviours. Second, interest, in Dewey’s sense, is always personal.
Any object of interest, by definition, must be perceived by the learner
to be connected with his or her sense of self or well-being in some way.
In other words, we have some personal investment in anything that we
are genuinely interested in. For Dewey, interest is the driving force
behind intelligent effort on the part of the learner, and effective
instruction must tap into learners’ current interests and build on
learners’ current knowledge and abilities if meaningful development
is to result.
According to Dewey, interest, which is embodied and personal,
generates intelligent effort in the form of unified activity (i.e. activity
in which means and ends are suffused and transform one another). In
a unified activity, learners allocate attention unreservedly and are
absorbed in the activity which drives them. There is no distinction
between means and end. Rather than having to push themselves to
continue, they have difficulty pulling themselves away. Thus, in
Dewey’s theory, interest is primary and intelligent effort in the form
of unified activity on the part of the learner follows naturally. Two key
questions remain. The first is how interest results in the development
of learners’ abilities, and the second is how interest might be generated
in the types of learning tasks used in TBLT.
Regarding how interest relates to ongoing development, Dewey
(1913) argues that interests become more complex and involve more
factors as they take on a longer time span (e.g. an interest in causal

debates in high school can lead to interest in dialectic methods in

Plato and eventually to the study of the Ancient Greek language at
university). For Dewey, the development of interest involves two
processes: (1) a direct interest generates a range of indirect interests,
and (2) these indirect interests eventually supplant or at least margin-
alize the original interest. In learning an L2, for example, a romantic
attachment during an undergraduate study abroad programme
(direct personal interest) might lead to the need to communicate in a
foreign language (an indirect interest). This language ability might
then lead to a subsequent internship in a company abroad after
graduation, and the internship might in turn lead to a specialized
degree in international business and formal study of the language,
which has become an interest in itself. These qualifications might
then lead to personal interest in the specialized target tasks required
to secure a permanent position and promotion in a company abroad.
This example illustrates Dewey’s (1913) argument that learners’
experience develops from an initial state of direct embodied interest
which is driven by the learner’s current impulses and desires. It is
also associated with the learner’s sense of self and well-being.
The specialized interests associated with the learner’s ultimate target
language needs would not have existed without the learner acting on
an initial personal drive.
This brings us to the question of how such embodied and personal
interest might be engendered in the design of pedagogic tasks for
TBLT. Dewey addresses this issue at the abstract level of the condi-
tions that must be engendered in materials design. Concrete ways in
which these principles might be operationalized in both the design and
implementation of tasks in TBLT will be addressed in the section
‘Personal Investment and TBLT’. Dewey (1913) argues that the subject
matter of instructional activities must provide learners with a means of
self-identification and generate self-initiated action. Other people play
an important role in most learners’ sense of self. Humans generally
have a natural desire for intimate contact in the form of sympathy,
approval and shared understanding when they interact with one
another. Learners’ social needs are also deeply suffused in the objects
of their interests and often play a key role in the genesis of these
interests. Dewey argues that learners are naturally interested in topics
and content that create sympathy, approval and shared understanding
with others. Social and emotional needs motivate learners’ interest.
And when the learner’s social or emotional sense of self is invested in
classroom work, the likelihood increases of their engaging in unified
activity and putting forth intelligent effort in completing classroom
activities.

The difﬁculty for the course designer is in tapping into the back-
ground experience of individuals and determining how this experience
might be incorporated to drive the learning of new subject matter
(Dewey, 1938, p. 75). Instructional materials must create conditions
that incorporate previous experience and current interests and build
on these experiences and interests to lead to new experiences (p. 80).
However, current deﬁnitions of tasks in TBLT typically focus on the
cognitive demands of performance and learners’ allocation of
attention during performance (TBLT Principles 4–9). Little work has
been done to systematically address the issue of learners’ interests in
the sense the term is used in Dewey’s theory of experience. Some
researchers argue that if instruction is relevant to learners’ L2 needs
that interest will take care of itself (e.g. Long, 2015, p. 65). Others
argue that the issue can be dealt with by selecting tasks and topics that
are generally interesting to the population of learners in question
(Prabhu, 1987; Yule, 1997; Ellis, 2003). Only recently has attention
been devoted to ways of designing and implementing tasks to engage
the learner socially and emotionally and generate personal investment
in L2 task performance. Recent work on task engagement (e.g.
Lambert, 2004, 2017; Aubrey, 2017a, 2017b; Butler, 2017b;
Lambert, Philp and Nakamura, 2017; Stroud, 2017) begins to provide
a systematic basis for addressing Dewey’s notion of interest in TBLT.
This work will be discussed later in this chapter, together with con-
crete ideas for how tasks might be designed and implemented to
develop a personal investment in learning and engage learners socially
and emotionally in the learning process.

The Affective Dimension of Learning by Doing

Maehr’s theory of personal investment provides a basis for modelling
the role that learners play in performing tasks in the TBLT classroom.
Maehr’s theory suggests that ﬁve classes of variables determine the
meaning that classroom activities have for learners. Learners bring a
range of motivational traits to the classroom that result from their
previous experiences and the sociocultural context in which the pro-
gramme is situated. While these traits may be amended over timescales
of weeks, months or years, teachers are typically more concerned
with how learners’ motivational state can vary within the time period
of a given lesson as a result of the instruction that they provide.
In terms of the subconstructs of L2 motivation (see Kormos,
Kiddle and Csizer, 2011 for a data-based discussion), relatively ﬁxed
motivational traits might include ideal L2 self, international orienta-
tion, intrinsic/instrumental goals for learning, peer pressure, parental

Meaning
1. Sense of competence or qualification
2. Perceived behavioral options
3. Standards of success
4. Goals -Orientation
a. Task goals
b. Competitive goals
c. Social solidarity goals
d. Extrinsic goals

Investment of Personal Resources

1. Talent
2. Time
3. Energy

Performance Effects

1. Direction 2. Persistence 3. Continued motivation 4. General activity level 5. Performance level

Figure 6.1 Maehr’s theory of personal investment

Source: From Lambert (1998)

encouragement, technology, anxiety. Motivational state variables will

depend on the immediate context but, on the other hand, might
include learners’ motivational intensity, their satiation control, their
resourcefulness and their sense of self-efficacy.
The first two classes of variables at the top of Figure 6.1 might thus
function as resources that can be drawn on in planning instruction
rather than independent factors that can be manipulated in instruc-
tional design. The other three classes of variables at the top of
Figure 6.1, however, are within the teacher’s immediate control in
planning and implementing instruction. The design of the tasks used
in instruction, the social expectations established in the classroom and
the information provided to learners can be modified, with direct
effects on learners’ performance during specific lessons.
All five classes of variables at the top of Figure 6.1 contribute to the
meaning that instructional activities have for learners. Maehr defines
meaning as a technical term, comprised of learners’ (1) sense
of competence, (2) perceived behavioural options, (3) standards of
success, and (4) goal orientation (i.e. whether learners are trying to
improve their ability to perform the activity, compete with one
another, develop social solidarity or achieve some extrinsic goal such
as a test score or a promotion).
The meaning that an activity has for learners will, in turn, determine
learners’ willingness to invest their personal resources into the

performance of it. If an activity is meaningful for learners, they will be

more willing to become involved in its performance in terms of volun-
tarily devoting their personal talents, time and energy into completing
it. Thus, the learner will be more likely to complete the activity well
rather than well enough.
Finally, personal investment in instructional activities on the part of
the learner will have concrete performance effects. According to
Maehr, the ﬁrst three types in the bottom of Figure 6.1 are the direct
result of learners’ level of personal investment, and they are thus the
best indications of it. These are: (1) learners’ direction or their decision
to work on the activity rather than to devote their attention to some-
thing else (e.g. their mobile phone or what their partner did over the
weekend), (2) learners’ persistence or their willingness to continue
working on the activity for a longer period of time without the need
for continual encouragement and support from the teacher (cf.
Crookes and Schmidt, 1991), and (3) learners’ continued motivation
or the extent to which they are willing to revisit the activity in subse-
quent lessons.

Personal Investment and TBLT

In Lambert’s (1998, 2004, 2017) adaptation of Maehr’s theory of
personal investment for TBLT, tasks are designed to incorporate
learners’ interests and sense of self and to use this experience to drive
L2 use and build new experiences in the classroom. Lambert’s work
focuses on the respective benefits of tasks based on learner-generated
content (LGC) and those based on teacher-generated content (TGC) in
TBLT (see Lambert, 2017; Lambert and Zhang, 2019). In LGC tasks,
the immediate classroom context creates the physical, social or emo-
tional need that drives the language that learners use. Instead of
exclusively relying on fixed content supplied by instructors and mater-
ials writers based on what teachers think learners will be interested in,
tasks can also be designed to allow learners to generate context-
specific content that is based on real experience. Specifically, Lambert
argues that for learners to have a personal investment in L2 task
performance, they must choose and discuss content that: (1) they think
is genuinely interesting, (2) they genuinely want to share with the
specific interlocutors they are working with on the task, and (3) they
genuinely think these interlocutors will be interested in hearing. As
argued later in this chapter, these three criteria might be argued to
define the conditions for any authentic language use whether it takes
place inside or outside of the classroom.

By designing tasks to tap into the real-life thoughts, experiences,

activities and talents that interest learners and occupy their attention
on a daily basis, all learners (even those who are not good at languages)
have the opportunity to make a personal investment in classroom
interaction. An example for lower-intermediate learners might be
asking them to recommend local restaurants or other venues that they
like, want to share, and think their partners might actually
visit (Lambert, Gong and Zhang, in press). Another example for
higher-intermediate learners might be choosing an anecdote about
something that happened to them that they think is funny or interest-
ing, want to share with the speciﬁc person they are working with, and
think that this person will enjoy hearing (Lambert et al., 2017). Finally,
an example for advanced-level learners might involve discussing unfor-
tunate situations they have experienced and trying to arrive at an
agreement on who is responsible (Lambert and Minn, 2007; Lambert
and Zhang, 2019). Consistent effects for these LGC tasks on L2 use
have been documented across task types and target languages. Before
discussing these effects, however, the construct of personal investment
must be disambiguated from related constructs in the TBLT literature.
Partially overlapping proposals for task selection and implementa-
tion include the relevance of tasks for learners in terms of the current
or future needs (Long, 2015), the provision of background knowledge
in conjunction with task performance (Robinson, 2011), learners’
familiarity with task content (Prabhu, 1987; Ellis, 2003) and learners’
control over task content (Foster and Skehan, 1996). Personal invest-
ment in the sense described implies a degree of each of the variables.
However, it is a distinct construct in that none of these variables,
individually or in combination, necessitate personal investment. For
example, personal investment in the form of LGC might be argued to
imply a degree of relevance to learners as their language use is con-
nected with their own experience. However, something may be rele-
vant to learners’ current or future external needs for the language, but
meet no urgent need (physical, social or emotional) within the imme-
diate context. An example might be the need to translate user manuals
for new machinery or summarize L2 economic news articles in the ﬁrst
language (L1). These were determined by a needs analysis in Japan to
be tasks that Japanese English majors at a public university are likely
to have to complete if they enter trading companies after graduation
(Lambert, 2010). However, tasks connected with these needs gener-
ated very little interest or effort on the part of learners. It is possible,
of course, that learners’ experience would grow to include these indir-
ect interests eventually but learner external needs as determined by

Lambert’s needs analysis may have been too distant from learners’
current interests and social and emotional needs to generate personal
investment in task performance.
Personal investment through LGC will also imply a degree of back-
ground knowledge related to task content (Robinson, 2011). Again,
however, background knowledge does not imply personal investment.
For example, it is possible for a learner to have grown up with a father
who was very interested in golf or in fishing. This might provide her
with a degree of background knowledge about these activities and
what is involved. This knowledge may in turn facilitate her develop-
ment of a conceptual framework for completing tasks associated with
these activities in the classroom. As useful as this background know-
ledge might be in facilitating her language use on tasks in other ways,
however, it does not imply that the learner would have any more
interest in completing a task connected with the topic of fishing or
golf than she would in completing a task that she knew nothing at all
about. In fact, she may be more interested in the new topic.
Finally, familiarity with and control over task content represent
similar cases. In a typical map task which requires learners to describe
routes and draw them based on the descriptions (see Prabhu, 1987;
Yule, 1997; Long, 2015 for examples), a materials writer might ask
students to describe the real routes that they take to and from school
every day rather than narrate a random route on a generic map (e.g.
Foster and Skehan, 1996). The task will certainly be more familiar to
learners, and it may also allow them more freedom to structure task
content and encode language in line with their current L2 resources.
However, this familiarity and control is unlikely to impact learners’
interest or personal investment in the performance of the task in the
sense outlined. Familiarity in itself does not create any interest or
urgency in the outcome of the task. Learners are unlikely to have
any more interest in telling someone about their route to school than
they would in discussing a random route on a generic map – nor is
their interlocutor any more interested in hearing about it. As Paradis
(2004) points out, in such pedagogic tasks learners are more motivated
by the fact that it is their turn to speak rather than by any physical,
social or emotional need for their partner to successfully understand
the route that they are describing. In contrast, if learners are asked to
think of a restaurant or club in the city that they frequent and really
like, want to introduce to the specific interlocutor(s) they were
working with for a task and think that these interlocutor(s) may
actually visit, they may be more likely to have a physical, social and
emotional need to ensure the semantic, pragmatic and interpersonal
effectiveness of the conversation.

Personal Investment and L2 Performance

Initial empirical research on the effect of personal investment on L2
use during task performance has produced promising results. Lam-
bert’s early work on the topic (1997, 2002, 2004) suggested that
content produced by learners in earlier tasks in a given sequence
may function as learner-generated input to subsequent tasks and thus
promote learners’ personal investment in the learning process. Lam-
bert and Minn (2007) subsequently investigated how a range of task
types in different discourse genres (instruction, narrative and opinion),
each designed to contrast learner-generated and teacher-generated
task content, affected the number of utterances that served to expand
on the semantic content of the tasks as well as the range of lexical
items the learners employed. This study was replicated with L2
speakers of Chinese (Lambert and Zhang, 2019) using a broader range
of performance variables, including engagement in L2 use, speech
processing and clause combination strategies. It was found LGC tasks
resulted in more engagement, more ﬂuent speech processing and the
use of devices which create pragmatic meaning (see Tables 6.1 to 6.3),
but that TGC tasks resulted in more clause combination and focus on
the language code. Finally, Lambert et al. (2017) compared thirty-two
learners’ performances on narrative tasks in which the solution to a
culturally relevant and familiar problem was explained based on

Table 6.1 Instruction task discourse based on LGC

Original Chinese English translation

啊。你不用粉底吗？ Ah. Don’t you use foundation?
对对对。 Correct, correct.
哇! Wow!
如果放粉底，我觉得心里有点儿别扭。 If I put foundation, I feel a bit awkward.
不太习惯。 Not used to do that.
啊，皮肤不能呼吸的感觉？ Ah, the feeling that the skin can’t
breathe?
*** ***
大概用多长时间？ How long will it take?
大概用二十分钟就够了。 About twenty minutes is enough.
哇，很快。 Wow, very quick.
嗯，很快。我不知道怎么化妆。（笑） Well, pretty fast. I don’t know how to
make up. (laughter)
没关系。很美。（笑） It doesn’t matter. You are very
beautiful. (laughter)
不如你呀。（笑） Not as good as you. (laughter)

Table 6.2 Narrative task discourse based on LGC

Original Chinese English translation

欸，口水掉在那个馅儿里边？ Eh, your saliva fell into the stuffing?
对对对对对。然后我妈妈说，哇, 你的口 Yes, yes, yes, yes, right. My mother
水在里面！然后我妈妈说，啊, 一点点口 said, ‘Wow, your saliva is inside!’
水没问题。（笑） Then my mother also said, ‘Ah, a
little bit of saliva, no problem.’
(Laughs)
嗯，然后拌完水饺，就煮嘛。煮完，先煮 Well, having finished mixing the
的给姥爷。姥爷，我帮忙, 这个水饺好不 dumpling stuffing, we boiled
好吃? 姥爷说，比一般的更好吃.（笑） dumplings. My grandpa was
（笑）有那个口水的效果。 invited to eat first. I asked him,
可能。 ‘Grandpa, I helped with making
dumplings, do they taste good?’
He answered, ‘They are more
delicious than the usual ones.’
(Laughs)
(Laughs) That’s the effect of the
dribble.
Possible.

Table 6.3 Opinion task discourse based on LGC

Original Chinese English translation

留学的时候很孤单嘛, 不是吗？ It’s very lonely when one studies
abroad, isn’t it?
留学就可以 Do you mean when one studies
abroad, they then can . . .
然后，有了帅哥就 And there’s this handsome guy, then. . .
可以吗？ Is that ok?
不是，我没有说可以，可是，我觉得没 No, I didn’t say that it’s okay to have
办法。 an affair, but I feel it is simply one of
those things.

picture strips supplied by the researcher with those in which they were
asked to select a problem that: (1) they had actually experienced and
solved in the past, (2) they felt was genuinely funny or interesting and
wanted to share, and (3) they thought their partner would be genu-
inely interested in hearing. Results showed that learners tended to
speak more, elaborate more, clarify more and support one another

more through back-channelling behaviour when discussing tasks that

created a personal investment in the learning process through LGC.
The examples in Tables 6.1, 6.2 and 6.3 serve to illustrate the type
of personally invested discourse that occurs as a result of LGC across
three discourse genres (instruction, narration, opinion). The samples
are taken from Lambert and Zhang (2019). They were produced by
L1 Japanese students speaking Chinese as an L2. The two speakers are
twenty-one-year-old undergraduate Japanese females majoring in Chi-
nese at a large public university in Japan.
In the first sample (Table 6.1), the learners are completing an
instruction task based on LGC. The speaker has just finished explain-
ing a method that she uses for putting on her cosmetics that saves her
time. The listener expresses her interest.
In terms of the amount of language and interaction produced, the
conversations on LGC tasks were roughly comparable to TGC tasks
(see Lambert and Zhang, 2019, for details). As the excerpt in Table 6.1
illustrates, however, there was a personal dimension to the learners’
conversations on the LGC tasks that was not present on the TGC tasks.
The speaker elaborates on her own feelings, and the listener shares her
own experiences to demonstrate her understanding before further elab-
orating in line with her own interests. Furthermore, the closing
sequence demonstrates the pragmatic device of self-deprecation in
response to praise, and the learners then proceed to personally comple-
ment one another. Their emotional investment is reinforced through
laughter. Use of such devices to create pragmatic meaning was absent
on comparable TGC instruction tasks (Lambert and Zhang, 2019).
In the second sample (Table 6.2), the learners are completing a
narrative task based on LGC. The speaker has just finished a story
about an event that took place while she was preparing dinner with
her mother, and the listener expresses her surprise.
The excerpt in Table 6.2 demonstrates how the rapport between
learners was deepened on this LGC tasks. The elaborative details of
the conversation related more closely to the learners personally and
emotionally. The listener encourages the speaker with laughter and
humour, and the speaker plays along by elaborating the details of her
story. The interaction was quick and animated, and the atmosphere was
humorous and light-hearted. This contributed to a rapport between the
learners that is frequently absent on pedagogic tasks in the classroom.
In the final sample (Table 6.3), the learners are completing an
opinion task based on LGC. The speaker has explained a complicated
affair that took place while her friend was studying abroad, and the
students are trying to decide who was responsible. The excerpt begins
with a comment by the listener.

What is apparent in the three excerpts in Tables 6.1 – 6.3 is that

pragmatic language use takes on a role in learners’ task-based
interaction. The learners’ role in the content makes strategic lan-
guage use necessary. In the excerpt in Table 6.3, learners use prag-
matic devices such as tag questioning to elicit agreement,
uncompleted sentences to deliver veiled challenge, mitigated ques-
tioning, unarticulated confrontation and softened tone. Such devices
are what Sabet and Zhang (2015) refer to as vague language, which
functions to create pragmatic meaning such as tentativeness, self-
protection, collaboration and cooperation. There is simply no need
for the use of such strategic devices in many of the pedagogic tasks
that are based on content supplied by the teacher or materials
designer. While such tasks may aid learners in developing inter-
active competence or gaining control of unfamiliar language associ-
ated with speciﬁc task content, tasks based on LGC may help to
engage the learner personally and ensure the need for pragmatic
language use. As TBLT relies primarily on incidental learning
(TBLT Principle 1), learners cannot be expected to acquire prag-
matic devices such as those in Tables 6.1 to 6.3 if the tasks
they complete do not involve the level of personal investment that
necessitates their use.

Strengths of LGC and TGC Tasks within the L2 Curriculum

Although this chapter has pointed out several advantages of incorpor-
ating tasks based on LGC in TBI, there are many aspects of L2
instruction that may be better suited to TGC tasks. The contrasts in
Table 6.4 illustrate the need for a balance of both LGC and TGC
in TBI.
Recent research on the relative effects of LGC and TGC (e.g. Lam-
bert, 2017; Lambert and Zhang, 2019) indicates that LGC tasks often
tend to result in ﬂuent use of known language, with most negotiation
of meaning being limited to negotiation of content. Furthermore, as

Table 6.4 Strengths of LGC and TGC tasks within the L2 curriculum

Learner-generated content Teacher-generated content

Fluency Complexity
Control of known language Experimentation with new language
Negotiation of content Negotiation of form
Pragmatic competence Interactive competence

illustrated in this chapter, LGC tasks might engender the type of

pragmatically authentic interaction that is often lacking in pedagogic
tasks supplied by teachers and materials writers.
In contrast, TGC tasks may result in learners experimenting with
more syntactically complex discourse and the negotiation of form
(Lambert and Zhang, 2019). TGC tasks might thus be best suited
for promoting noticing and the ongoing development of learners’
linguistic competence during meaning-focused communication.

Personal Investment and Task Implementation

This chapter has so far focused on task design. However, it may also
be possible to engage the learner personally in achieving task outcome-
s through task implementation. One possibility is in grouping learners
to perform tasks. Aubrey (2017a, 2017b), for example, found that
learners in a homogenous EFL context who worked with intercultural
interlocutors reported being more absorbed in the performance of
oral tasks in the classroom and that this intercultural contact also
resulted in improved turn-taking during task performance. Phung
(2017) reports similar findings in a heterogenous English as a Second
Language (ESL) context for learners’ preferences for tasks, arguing
that grouping learners with those different to themselves creates a
greater need to communicate and that this positively affects learners’
preference for tasks as well as increasing the negotiation of form and
content that takes place during task performance. In short, language
learners come into the classroom with a range of individual differences
(L1, background knowledge, proficiency level, gender, etc.) and
grouping learners to create gaps in these differences during the per-
formance of tasks may be one means of generating interest in task
performance.
Another means of engendering personal interest during the imple-
mentation of tasks might be the use of goal-setting devices over
repeated task performances (e.g. Reese and Wells, 2007; Butler,
2017b; Stroud, 2017). Reese and Wells (2007) and Stroud (2017),
for example, report positive effects over semester-long studies in
English as a Foreign Language (EFL) conversation classes for
implementing a goal-tracking system based on conversation cards
weighted for different point values for varying behaviours during task
performance (e.g. giving reasons to support opinion, disagreeing with
a partner’s opinions, clarifying unknown meanings, confirming under-
standing, interrupting, correcting misunderstandings, etc.). In both
studies, these goal-tracking systems were found to have positive effects
on learners’ L2 use as well as their affective response to instruction as

measured by questionnaires. Similarly, Butler (2017b) points out the

popularity of digital instructional games employing avatars as motiv-
ational devices to keep students playing and reinforce positive learning
behaviours. These devices allow learners to earn credits for positive
behaviours which they can use to maintain and develop their avatar’s
appearance in the game. Implementing tasks with goal-tracking
systems of this kind may offer a means of engendering interest in task
performance. In most approaches to goal-tracking, the assumption,
based on self-determination theory, is that the use of such extrinsic
motivators will ultimately lead to intrinsic motivation on the part of
learners when a certain level of proﬁciency is reached (e.g. Reeve and
Lee, 2014; Oga-Baldwin and Nakata, 2017). In short, some learners
may beneﬁt from being provided with a task-extrinsic means of con-
trol over their own progress in the form of game elements they can
identify with at a personal level.

Personal Investment and SLA

The crucial issue, of course, is how personal investment, or interest in
Dewey’s sense, might be related to SLA. Neither Dewey’s experiential
learning theory nor Maehr’s theory of personal investment speciﬁcally
address language. Work in neurolinguistics (e.g. Lamendella, 1977;
Paradis 1994, 2004; Schumann, 2001), however, provides a basis for
conceptualizing the relationship between personal investment and
language learning on tasks. The limbic system is a group of intercon-
nected structures dedicated to linking visceral states and emotion to
cognition and behaviour (Mesulam, 2000). Recent models of
the limbic system (e.g. Catani, Dell’Acqua and de Schotten, 2013;
Lovblad, Schaller and Vargas, 2013) identify components that work
in unison to assess and encode new memories. Lovblad et al. (2013)
argue that the hippocampus deals with the factual content of memory,
whereas the amygdala assigns emotional content to memories and
acts to enhance memory formation by increasing the importance of
events. Personal investment in task-based language use could thus
result in better registration and encoding of both form and meaning
during incidental language learning (Paradis, 1994, 2004). Memory
for the language used during task performance may be one way in
which personal investment impacts L2 learning. Empirical work by
Keenan, MacWhinney and Mayhew (1977) and MacWhinney,
Keenan and Reinke (1982) has demonstrated effects on memory for
both form and meaning when L1 conversations involve personal as
opposed to factual content. Furthermore, recent empirical work with
beginning-level L2 learners by Lambert, Gong and Zhang (in press)

has found that both immediate and delayed lexical recall is better for
LGC than TGC lexical items.
A second possibility for measuring the impact of personal invest-
ment in TBLT is through indicators of engagement in language use
during task performance. Engagement has been a mercurial construct
in L2 research and has tended to piggyback on trends in research
on information processing and cognitive-interactionist theories of
SLA. Since the late 1990s, aspects of L2 performance that have
been investigated in connection with the notion of motivation and
engagement have ranged from the number of words and turns
used (Dörnyei and Kormos, 2000) to the amount of intake and pushed
output (Batstone, 2002), learners’ use of corrective feedback (CF)
(Hyland, 2003), language-related episodes (Storch, 2008; Baralt,
Gurzynski-Weiss and Kim, 2016), language awareness (Svalberg,
2009), and most recently a combined set of measures from interac-
tionist research (Lambert et al. 2017) matched to a model of perform-
ance engagement from general education (Fredricks, Blumenfeld and
Paris, 2004; Philp and Duchesne, 2016). In all cases, engagement has
been measured in relative terms of learners doing more of what
other theories of SLA have argued to be relevant for learning. The
construct of engagement itself and how it relates to SLA is currently
undertheorized.
In addition to its impact on memory and L2 performance, however,
personal investment may also affect learners’ sense of absorption or
‘flow’ during task performance (Csikszentmihalyi, 1975, 1990;
Egbert, 2003; Aubrey, 2017a, 2017b). Flow might be an important
indicator of learners’ emotional engagement on tasks (Aubrey, 2017a,
2017b). Aubrey (2017a) argues, based on Egbert (2003), that one
factor contributing to learners’ experience of flow during task per-
formance is that the content of the tasks is perceived as important,
urgent or meaningful. A second factor is learners’ control when they
perceive themselves as enacting a choice over task topic, content or
procedures. Aubrey (2017b), in a follow-up analysis of learners’ post-
task reflections, makes the case that such flow experiences can lead to
a sense of accomplishment and self-confidence, which can facilitate the
attainment of flow states in subsequent tasks. If this is the case,
questionnaires eliciting learners’ self-reported levels of flow during
L2 task performance (see Egbert, 2003; Aubrey, 2017a, for examples)
or post-task diaries documenting learners’ retrospective thought
processes and perceptions (Aubrey, 2017b) may provide an important
and feasible means of gaging the extent to which materials designed
to engender interest on the part of learners have worked in the
classroom.

Implications for TBLT Principles

Although TBLT is generally consistent with Dewey’s theory of experi-
ential learning (TBLT Principles 1–3), the focus has been primarily on
the cognitive dimension of learning (TBLT Principles 4–9). Current
approaches to TBLT conceptualize tasks in two primary ways. The first
is in terms of learners’ real-world needs and the relevance that this has
for them (Long, 2015; Robinson, 2011). The second is in terms of
learners’ cognitive and interactive needs for incidental SLA in instructed
settings (e.g. Prabhu, 1987; Yule, 1997; Ellis, 2003). In both
approaches the role of learners’ personal interest in completing tasks
has been a secondary consideration which has been assumed to be a by-
product of relevance to future needs (Long, 2015) or has been left to
teachers and materials writers to decide based on what they feel will
appeal to learners (Prabhu, 1987; Yule, 1997; Ellis, 2003).
In Dewey’s theory of experience as well as Vygotsky’s theory of
mind, however, the affective dimension of learning is central to subse-
quent cognition. As we have seen in this chapter, Dewey argues that
interest is always embodied and personal, and that it provides the
driving force behind both unified activity and intelligent effort in
achieving new objectives. Learning is thus a personal process in which
learners act on their current impulses and experiences to achieve
objects of interest that are connected with their sense of self and
well-being. These objectives might be shared personal understanding
in social interactions in the classroom or achieving a much-needed job
or promotion to support one’s family. One way or the other, learners
who achieve their best are typically personally invested in their learn-
ing and take an agentive role in what they do. The role for learners is
too often missing in the types of interactive pedagogic tasks used in
TBLT research and instruction. We might thus add a tenth TBLT
principle to the list at the beginning of this chapter.
10. Personal investment: To optimize learning opportunities on tasks,
learners must take an agentive role in the learning process. TBLT
must incorporate learners personally so that task interactions and
outcomes reflect their sense of self and well-being.
Until recently, the affective dimension of task-based learning has
received little attention in the TBLT literature. By incorporating ample
opportunities for learners to tailor task content based on their per-
sonal drives and experiences to the specific classroom context in which
they are working, TBLT researchers and practitioners might generate
more opportunities for intelligent effort, unified activity and pragmat-
ically authentic communication in classrooms.

Conclusion
This chapter has presented an educational perspective on many TBLT
practices based on Dewey’s (1913, 1938) theory of experience and
subsequent work on the affective dimensions of learning by Maehr
(1984). It has been argued that while the cognitive dimension of L2
performance has received considerable theoretical and empirical atten-
tion in TBLT, the affective dimension is only beginning to be theorized
and researched. This chapter suggested directions for addressing this
gap in TBLT theory and practice by summarizing some initial empir-
ical evidence on the different approaches to engendering personal
investment and interest through task design and the implementation
of tasks in the classroom. The chapter has also suggested some direc-
tions for evaluating the effects of these approaches on memory, lan-
guage use and learners’ experiences during task performance.

The chapters in Part II had a primary focus on the theory and research
that informs task-based language teaching (TBLT) but they also con-
sidered pedagogical issues. The primary focus of this part is on the
pedagogy of TBLT but we will continue to draw on research to
support the proposals we advance. The three chapters in Part III
address the following questions:
• What principles inform the selection and sequencing of tasks in a
task-based course?
• What does a task-based lesson consist of?
• What methodological principles underlie proposals for implement-
ing a task in the classroom?
• What kind of assessment is compatible with TBLT?
Chapter 7 addresses what tasks to include in a task-based course and
how the tasks selected can be sequenced to assist learning. The chapter
examines four different proposals for designing a task-based course. In
Prabhu’s Communicational Language Teaching Project (CLTP), the
syllabus serves as an operational construct. That is, it has low internal
structure and leaves implementation issues to be decided by teachers
based on their experience of what works in their instructional context.
Prabhu aimed to select tasks on familiar topics that would motivate
students to engage mentally in using language to achieve meaningful
outcomes. The tasks were sequenced intuitively by drawing broadly
on a set of general principles such as cognitive demand and type of gap
involved (i.e. information, reasoning, opinion). In contrast, Long’s
syllabus functions as an illuminative construct, aiming to bring what
is learned in line with what is taught. That is, tasks in his syllabus have
a training function. Target tasks are identiﬁed ﬁrst by a needs analysis
and then restructured into pedagogic tasks. They are sequenced based
on their frequency and criticality as revealed by the needs analysis and
also in terms of the cognitive demands they impose. Robinson also

175

views the syllabus as serving an illuminative purpose. His is the most

ambitious proposal as he proposes a syllabus that takes into account
the cognitive complexity of tasks, their propensity for promoting the
kinds of interaction that facilitate acquisition, and the cognitive abil-
ities and affective dispositions of individual learners. Central to
sequencing tasks, however, is their cognitive complexity. Finally, Ellis,
like Prabhu, treats the syllabus as an operational construct. He identi-
fies a range of factors that influence task complexity (e.g. the linguistic
level of the input and whether the task outcome is closed or open) but
suggests that sequencing tasks is largely a matter of intuition that can
be guided only roughly by such factors. In Ellis’ proposal – unlike the
other proposal in Chapter 7 – there is room for a more traditional,
structural module to fit alongside a task-based module in a complete
course.
Chapter 8 considers the other side of language pedagogy – the
methodology for implementing tasks. It identifies a range of options
relating to each of the three phases of a task-based lesson – the pre-
task phase, the main task phase and the post-task phase. The viability
and effectiveness of the different options is considered in relation to
both research findings and the recommendations found in popular
teacher guides. Pre-task options have three purposes: (1) to motivate
students to perform the task, (2) to prepare them to perform it, and (3)
to encourage the use of strategies that will help them. Various options
are considered with special attention given to pre-task planning and
the various ways in which this can be carried out (e.g. in terms of its
focus and the time allocated). The key option in the main task phase is
within-task focus on form. Various ways of accomplishing this are
considered – in particular corrective feedback (CF). Another main task
option that research has shown impacts on task performance concerns
whether learners perform a task with or without time pressure. Of the
post-task options, asking learners to repeat a task has attracted the
most attention of researchers. Various types of repetition (exact, con-
tent and procedural) are considered along with the relative strengths of
each. The chapter does not seek to be prescriptive but it does point to
particular options that research has shown to be effective.
Chapter 9 discusses how tasks can be used for purposes of assess-
ment. The chapter begins by reviewing some general issues in testing
to create a foundation for considering task-based assessment – the
different functions of tests (proficiency, achievement, diagnostic),
summative and formative evaluation; norm versus criterion referen-
cing; washback. It then outlines the theory of testing that underlies
task-based assessment, arguing that it must take account of both
competences (e.g. grammatical, sociolinguistic and strategic) and the

ability to use language. The interactive approach to testing is then

contrasted with the real-life approach. The challenge facing task-based
assessment is how tasks can provide evidence of learners’ ability to use
an L2 that is generalizable (i.e. not limited to the particular task used
in a test). Drawing on research discussed in Chapter 4, a proposal is
advanced for designing assessment tasks that can provide evidence of
learners’ overall proﬁciency by showing their ability to use language
that is complex, accurate and ﬂuent. The chapter concludes by looking
at the development of task-based tests in three case studies.

Introduction
The perceived importance of allowing learners opportunities to
employ their linguistic repertoire under relatively natural conditions
has generated discussion on the use of tasks in L2 instructional design
since the 1980s, and the role that tasks might play within the language
syllabus has been an issue of considerable debate (e.g. Brown and
Yule, 1983; Brown et al., 1984; Candlin, 1987; Prabhu, 1987; Nunan,
1989; Yule, 1997; Ellis, 2003; Willis and Willis, 2007; Robinson,
2010, 2011; Skehan, 2014a; Long, 2015). In task-based syllabuses,
tasks function as the primary unit in selecting and sequencing course
content. There is considerable theoretical support for using tasks in
this way in syllabus design (Long, 1985, 2015; Skehan, 1998, 2014;
Robinson, 2001, 2011; Ellis, 2003, 2018a), and initial attempts have
been made to develop courses in line with these theoretical rationales
(e.g. Kelly and Kelly, 1996; Lambert and Hailes, 2002; Benevides and
Valvona, 2008).
Tasks have also been used to support syllabuses organized around
other units of analysis (e.g. grammatical structures, lexis, topics,
situations). These task-supported syllabuses have gained support in
English as a Foreign Language (EFL) contexts in different parts of
the world (see Shehadeh, 2005; R. Sheen, 2006; Littlewood, 2014),
and they are also supported by well-established learning theories
(DeKeyser, 2007; Ellis, 2018a). Published sets of teaching material
have also appeared which are in line with these theories (e.g. Kelly
and Kelly, 1996; Cutrone and Beh, 2014; Harris and Leeming, 2016).
Some researchers argue that the theories of learning associated with
task-based and task-supported syllabuses are incompatible (e.g. Long,
2015). Others argue that when certain conditions are met they can be

179

complementary in facilitating instructed L2 learning (Ellis, 2018a).

The present chapter will compare key proposals for designing
task-based syllabuses as well as some ideas on how task-based sylla-
buses might be introduced in programmes where a more traditional
structural syllabus is in use.
Like any approach to syllabus design, proposals for task-based
syllabuses must address procedures for task selection (how the
content to be included in the course is to be determined) and task
sequencing (how course content is to be ordered). In addition, how-
ever, approaches to syllabus design differ in the aims of the syllabus
and consequently in the scope of the syllabus within the broader
curriculum. The present chapter will discuss four proposals for
designing syllabuses based on tasks. Each of these approaches will be
compared in terms of: (1) the aims and scope of the syllabus within the
L2 curriculum, (2) the procedures by which tasks are selected, and (3)
the principles on which tasks are sequenced. In the ﬁnal sections of the
chapter, examples of how these approaches have been put into prac-
tice in designing actual L2 syllabuses used in EFL instruction in Japan
will be considered, and the pros and cons of each approach will be
summarized.

Prabhu’s Approach to Syllabus Design

The Communicative Language Teaching Project (CLTP, also known
as the Bangalore Project) was devoted to developing a L2 curriculum
for use in general English classes in public secondary schools in India
between 1979 and 1982. The syllabus used in the project was organ-
ized according to topic, and a range of tasks, loosely sequenced based
on their procedural demands, were suggested to help teachers cover
each of these topics with their learners. The distinctions that Prabhu
(1987) makes regarding the role of this syllabus in instructional design
are helpful in comparing subsequent developments.
One key distinction that Prabhu (1987) makes is between syllabuses
which are intended to function as operational constructs in the cur-
riculum and those that are intended to function as illuminative con-
structs. When a syllabus is designed to function as an operational
construct in the L2 curriculum, the aim is to provide course content
as a resource for teachers to construct their lesson plans and reach
curricular goals with different learners in the varying classroom con-
texts in which they teach. This type of syllabus has low internal
structure and leaves methodological and implementational issues to
be determined by teachers based on their experience of what works at
the local level and the problems that different learners face in the

classroom. In short, the syllabus specifies only what will be taught, not
how it will be taught. The content of syllabus is fixed, but how the
teacher uses this content is flexible.
By contrast, when a syllabus is designed to function as an illumina-
tive construct in the L2 curriculum, the aim is accountability for both
what will be taught and for what will be learned as a result. The focus
at the planning stage is on ensuring accurate prediction so steps are
taken to bring what is taught and what is learned into careful align-
ment. In achieving this end, the line between syllabus and method-
ology naturally blurs, and the syllabus takes on a much broader role
within the curriculum. Illuminative syllabuses are frequently used in
subjects which are based on lists of content items that must be taught
together with discrete-point questions that assess learners’ mastery of
this content. Illuminative syllabuses are also used in workplace
training. Employees are briefly trained in how to perform key tasks
(e.g. cleaning bathrooms at an airport, preparing a hotel room before
check-in, detailing an automobile at the dealer before it is picked up,
preparing a jet for take-off, etc.). Trainees’ performances on tasks are
then evaluated according to a checklist of the points on which they
have been trained to ensure that they can do the tasks properly (e.g.
mirrors have been cleaned, soap dispenser refilled, floor mopped, bins
emptied, air freshener used, etc.). In both content instruction and
workplace training, what is learned can be brought closely in line with
what is taught. Syllabus content is divided into discrete points that
all learners can be expected to master with little or no variation.
Methodology is reduced to explanation and feedback.

Scope
Prabhu’s syllabus was designed to function as an operational con-
struct. It speciﬁed what would be taught at a low level of internal
structure and allowed teachers to make intuitive decisions and adjust-
ments to ensure adequate mastery of the syllabus content. The
motivation for this was in line with the aim of the project’s curriculum.
The curriculum aimed to develop a general capacity in learners that
would allow them to acquire the skills necessary for mastering any
number of tasks that they face in the future. Prabhu refers to this
capacity as grammatical competence and he used communicative tasks
as tools to develop this competence rather than to train learners to
perform these tasks as ends in themselves. He points out that develop-
ing competence in a language is a variable process and that a syllabus
cannot be expected to anticipate all sources of challenge for different
learners in different classrooms.

Selection
In Prabhu’s approach, syllabus content was organized in terms of
topic areas (e.g. monthly calendars, maps, school timetables) rather
than in terms of tasks. Prabhu’s primary concern in selecting topics
was generating the interest and effort necessary to engage learners
mentally in the process of using language to achieve meaningful
outcomes. To achieve this end, content was selected for the project
with a view to allowing learners to draw on their background know-
ledge and experiences. As we have seen in Chapter 6, there is a strong
educational rationale for selecting tasks that draw on established areas
of knowledge. Lambert, Philp and Nakamura (2017), for example,
found that tasks which draw on content which learners found inter-
esting, wanted to share and thought that their partner would enjoy
hearing about resulted in increased engagement in language use in the
classroom. There is also good theoretical and empirical evidence to
suggest that it might result in improved memory for both content and
language (see Lambert, 2017; Lambert, Gong and Zhang, in press;
Chapter 6 of the present volume).
Prabhu defines the pedagogic tasks used in the CLTP as pieces of
logical thinking. According to Prabhu, successful tasks had five char-
acteristics: (1) a clear outcome, (2) clear criteria for success, (3) a
balance of predictability and unpredictability, (4) content which
allowed learners to draw on their background knowledge, and (5)
content that generated interest and sustained engagement. Pedagogic
tasks meeting these five criteria were selected to support each topic
area specified for the syllabus (e.g. listening to stories of the ‘whodunit’
kind and completing them with appropriate solutions, see Prabhu,
1987, pp. 138–43, for more examples). However, it is worth noting
that the tasks used in the CLTP also seem to have met the criteria of
‘tasks’ established by Ellis (see Chapter 1) in that they typically
required a focus on meaning, involved a gap that necessitated lan-
guage processing, required learners to draw on their own resources in
completing them and resulted in a communicative outcome beyond
practising language for its own sake.

Sequencing
The tasks selected for each topic in the syllabus were then sequenced.
The primary criterion for this sequencing seems to have been the
mental procedures that learners might have to engage in arriving at
the task outcome. Prabhu distinguishes three levels of procedural
demand that ﬁlling the gaps required to arrive at task outcomes might
entail (see Chapter 1):

1. Information gap: Transfer of given information related to a topic.

2. Reasoning gap: Derivation of new information from given infor-
mation through processes of inference, deduction, calculation, col-
lation, application or identification of relationships/patterns.
3. Opinion gap: Identification and articulation of personal prefer-
ences, feelings or attitudes in response to a given situation.
Under the topic of ‘maps’ (Prabhu, 1987, p. 139), for example, we find
a sequence of six versions of a map task sequenced to increase in
difficulty according to these three levels of procedural demand.
Learners initially (1) identify and (2) describe locations on a map.
They then (3) draw maps based on instructions and (4) draw routes
on maps based on instructions. Up to this point, only information
transfer (the lowest level of procedural demand) is involved in the
sequence. However, the next task requires learners to reason based on
available information by (5) deciding forms of transport when given
information on bus routes, fares, etc. (i.e. deriving new information
from existing information). Finally, the last task requires learners to
(6) make a decision about the best place for new facilities such as a
hospital or a school. The last task in the sequence thus requires the
identification and articulation of a personal preference in response to
the situation which, according to Prabhu, represents the highest level
of procedural demand.
On careful examination, however, other factors used in sequencing
the tasks in the project can be identified (see Table 7.1). Procedural
demand is thus one factor among others. This illustrates Ellis’ (2003,
2018a) point that tasks are invariably conglomerations of features that
work together to impact the demands placed on learners (see
Chapter 3).
Prabhu’s approach to task sequencing was thus an expedient one,
but it seems to have met with some success in the context in which it
was used.

Table 7.1 Prabhu’s criteria for task sequencing

1 Information provided Amount, type, variety, circumstance

2 Reasoning needed Number of deductive, inferential or calculative
steps
3 Precision needed Level of interpretation and expression required
4 Experience assumed Experience of task purpose and constraints
5 Abstractness involved Reference to concepts vs. objects and actions
6 Language mode Receptive vs. productive; oral/aural vs. written
required
7 Cognitive demands Information transfer, reasoning, opinion

As noted, the purpose of the syllabus in the CLTP was simply to

provide teachers with a set of landmarks and suggestions for working
on a topic in the classroom. Teachers were free to adjust or eliminate
specific tasks based on each group’s needs and problems. In terms of
the level of specification of an operational syllabus of this type, Prabhu
(1987, pp. 87–8) argues that the syllabus needs to include four types of
information to provide teachers with the resources that they need: (1)
the topics to be covered (monthly calendars, maps, school timetables,
itineraries, etc.), (2) information connected with each topic (schedule
contents, prices, distances, etc.), (3) the procedures to be used in
addressing each topic (inference, calculation, collation, application,
etc.), and (4) the outcomes to be reached (the shortest time, the lowest
cost, the most symmetrical pattern, etc.).
Thus, Prabhu believes a language syllabus should function as an
operational construct in the L2 curriculum and specify course content
at a low level of internal structure. His fear is that an over-specified
syllabus will reduce what learners and teachers can bring to the
learning process and that the outcome will be unified levels of learn-
ing associated with task training rather than individual development
and the general problem-solving capacity associated with language
education. He argues that training involves mastery of a series of
fixed stages to criterion-level performance on specific real-world
tasks, and education involves mastery of a general capacity that
allows learners to engage in adaptable behaviour and ongoing learn-
ing on any future tasks. Prabhu (1987, p. 91) claims that the aim of a
language syllabus is to enable the learner to employ language for
meaning exchange and, in the process, to achieve conformity to
linguistic norms. However, his focus throughout is the subconscious
acquisition of language through rational thinking (reasoning, infer-
encing etc.).
Prabhu’s work on the Bangalore Project has made a seminal contri-
bution to the field, but subsequent theories of task-based instructional
design are in agreement that something more is needed at the syllabus
level and during implementation if learners are to acquire the ability
to complete tasks and achieve conformity with linguistic norms in
doing so.

Long’s Task-Based Syllabus

Long (1985, 2015) argues that the syllabus should be conceptualized,
organized and assessed in terms of tasks. In contrast to Prabhu’s
approach to syllabus design, where the focus was on using tasks as
tools to develop a general language capacity, the focus in Long’s

approach is on the ability to perform speciﬁc tasks. This change of

focus affords several practical advantages for the syllabus designer.
Syllabuses aimed at language competence of any kind must face the
fact that language development is a holistic, organic and variable
process that is difficult to assess in terms of fixed stages which are
the same for all learners. Syllabuses aimed at the ability to complete
specific real-world tasks, on the other hand, can potentially avoid this
issue by subordinating language to what it is used to accomplish,
namely tasks. In fact, this is how abilities are typically evaluated
outside of the language classroom. As we have seen earlier in the
chapter, in workplace training there is a high degree of alignment
between what is taught and what is learned, and instructional content
can be divided into discrete points that all learners can be expected to
master with little or no variation.
To take an example, consider the situation of non-native speakers
of English who work on the floor of furniture stores or car dealer-
ships in English-speaking countries. One of their most important
duties will naturally be achieving sales. Their manager knows that
customers come in ready to buy or not. If they are ready to buy, the
job of the showroom staff is to facilitate rather than inhibit that
process. The manager will know through experience, at least tacitly,
what this involves. It might involve greeting the customers at the
right time, shaking hands, maintaining a smile, incorporating chil-
dren in the decision-making process (What do you think about this
car? Which one do you like?), comparing products in a given range,
discussing prices and options, completing the appropriate paper-
work to close transactions etc. Initial and on-the-job training may
involve spending some time going through these things with
trainees, asking them to observe experienced sales staff for a period,
bringing certain things to their attention that went well or badly,
and then observing them as they do these things themselves to
ensure that they are done adequately, intervening whenever neces-
sary to bring performance up to standard. All employees would be
expected to do the things that are taught, and a checklist, explicit or
tacit, might be used to ensure that they are performing well. The
employees’ linguistic competence is subordinated completely to the
tasks that need to be performed. The training process and evalu-
ation process themselves are not aimed at grammatical competence
as in Prabhu’s project. The focus is on successful task performance.
Language is seen as a by-product or artefact of successful task
performance. Task performance in Long’s sense is thus trainable
and measurable in terms of objective criteria that apply to all
learners.

Scope
By subordinating language to task performance as it is in the real
world, Long’s syllabus can function as an illuminative construct
within the curriculum. It can bring what is learned in line with what
is taught and thus become accountable for what is learned as a result
of instruction. Long defines tasks broadly as the events that learners
participate in ‘in everyday life, at work, at play, and in between’
(Long, 2015, p. 108). He does not attempt to define tasks at the
syllabus level in terms of conditions that connect them to language
acquisition. In fact, at this level, he does not attempt to connect them
to the use of language at all. In Long’s approach, the pedagogic tasks
that learners work on in the classroom are direct reflections of real-
world tasks. In addition to tasks in which language plays a central role
such as ‘buying a pair of shoes’ or ‘making a hotel reservation’, the
examples of tasks Long (2015) provides include those in which lan-
guage plays only a distal role if it is necessary at all (e.g. ‘painting a
fence’, ‘dressing a child’ or ‘weighing a patient’).

Selection
In Long’s approach, the tasks which constitute the content of the
syllabus are identified based on a needs analysis (Long, 2005, 2015).
Long argues that this provides a rational and empirical means of
content selection in contrast to approaches such as Prabhu’s, which
rely on teacher or course designer’s intuitions about what is interesting
and challenging for their learners. Second, it is essential for
accountability in the case of learners who have educational and career
goals involving the target culture or who have to survive socially in
target language communities (Long, 2015, pp. 63–83).
The process by which tasks are selected for the syllabus involves
three stages of analysis (Long, 2015, pp. 223–7). The first is identify-
ing the things that learners will need to do in everyday life (e.g.
making or changing a plane, train, hotel, restaurant or theatre reser-
vation). The outcome of this analysis is a list of target tasks that
learners need to be able to complete. This list, however, provides
only the ‘raw input’ for a task-based syllabus. Following this, the
second stage of analysis involves classifying these target tasks into
task types or more abstract superordinate categories based on their
common features (e.g. making reservations, changing reservations) in
order to meet the needs of heterogeneous groups of learners effi-
ciently. When several target tasks overlap, such as in the example

of different types of reservations, it would be uneconomical to treat

them all separately in the syllabus so this intermediate step of identi-
fying task types, according to Long, provides a more efficient basis
for initially organizing course content. Long also points out that
target tasks that are unique to specific subgroups of learners might
be eliminated at this stage and only task types which all learners need
might be selected for inclusion in the syllabus. Finally, when the list
of task types to be included in the syllabus has been determined, the
final stage of analysis in Long’s approach to task selection is to create
pedagogic tasks connected with each task type. These pedagogic
tasks serve as the actual content of the syllabus (e.g. filling out
or changing a reservation form while listening to a telephone call;
role-playing customers and clerks who are making/changing reserva-
tions, etc.). The ultimate goal in selecting pedagogic tasks for a
syllabus, according to Long, is to provide adequate coverage of each
task type by ensuring that learners develop the skills necessary to
complete the target tasks they represent.
As an example of how Long’s approach to selecting tasks for a
syllabus was put into practice, Lambert (2010) conducted an analysis
of the English language use of graduates over a twenty-five-year period
from an English programme at a Japanese university. The analysis
generated a broad range of tasks which overlapped in varying degrees.
In order to incorporate a representative number of the target tasks
specified by informants in key areas of job placement, task types
representing a broad range of specific tasks were initially established
(e.g. locating information on the Internet, locating information in
newspapers, translating users’ manuals, translating e-mail messages,
editing teaching materials, editing contracts, etc.). Only after this
initial classification was it possible to create a reasonable number of
yet more abstract task types in the sense that Long suggests (e.g.
locating information, translating from English to Japanese, editing
English documents, etc.). The process was a dynamic rather than a
linear one, which required considerable intuition and experience with
the cultural milieu and the educational context in which the syllabus
would be implemented.

Sequencing
Before developing sequences of pedagogic tasks connected with each
task type, Long (2015) recommends initially arranging the various
task types that will be included in the syllabus in relationship to one
another based on ‘the relative frequency and or criticality of the task as

determined by the needs analysis’ (p. 233). In other words, the task
types which learners can be expected to encounter often or that are the
most essential for their success in reaching their goals will be addressed
earlier in the syllabus, whereas those that they will need to complete
less frequently or which are not as critical to their success will be
addressed later. Selecting tasks based on a needs analysis thus has
the added advantage of providing a principled basis for the initial
arrangement of task types to be included in the syllabus content.
Following this initial sequencing of task types, pedagogic tasks
representing simplified versions of each task type are sequenced in
relationship to one another, in line with learners developing capacity
to complete them. The goal is that the demands of the versions of these
pedagogic tasks will increase until they match the demands of the
target tasks represented by the task type. The last task in each module
should thus be a full ‘proxy’ or ‘virtual’ version of a target task that
learners need to complete outside of the classroom, and this task will
be used as an ‘exit task’ to assess learners’ abilities to deal successfully
with this target task on the basis of which the task type and associated
pedagogic tasks were created.
In terms of the specific factors used to grade the demands of versions
of a given task, Long advocates using factors related to the intrinsic
complexity of the targeted task. According to Long, task intrinsic
complexity relates to the inherent, unchanging qualities that make
tasks more or less challenging. The examples that he provides are
those identified in early proposals for L2 task design by Brown and
Yule (1983) and Brown et al. (1984). These factors include: (1) the
number of components, elements and steps in a task, (2) the distinct-
iveness of task’s components, elements and steps, and (3) the disloca-
tion in time and space of the task content in relation to the speaker.
Long also acknowledges the importance of Robinson’s (2001, 2010)
SSARC (simplify, stabilize, automatize, restructure, complexify) model
of task sequencing which will be discussed in the section ‘Robinson’s
Task-Based Syllabus’.
To take a concrete example, in comparison to the task sequence
related to ‘maps’, which was summarized in the section ‘Prabhu’s
Approach to Syllabus Design’, Long provides a sequence of eight
pedagogic tasks to prepare learners to ‘obtain and follow street direc-
tions’: (1) learners listen to recordings of real examples of a native
speaker giving directions to orient them to the demands of the target
task; (2) they follow short fragments of oral directions on a map; (3)
they follow longer directions to more distant locations on the map
followed by questions asking them to confirm their location (street
name, nearby buildings, etc.); (4) they work in pairs to read out

fragments/questions and follow/answer them; (5) they listen to direc-

tions and follow routes marked on an authentic map as they listen; (6)
they listen to directions from a starting point on the map, trace the
route as they listen and answer questions about where they are at the
end; (7) they complete the same task more rapidly with no breaks; (8)
they complete a simulation of the target task ‘obtain and follow street
directions’ which acts a form of assessment for the module. Thus,
Long’s syllabus is more illuminative in scope than Prabhu’s syllabus.
Long’s approach provides a more detailed level of description which
includes not only the content to be taught, but a considerable amount
of information about how it is to be taught (4, 7) and what is to be
learned as a result (1, 8). In other words, the traditional distinction
between syllabus and methodology begins to blur.
Thus, as we have seen, Prabhu (1987) believed that the cognitive
demands of tasks could drive the development of L2 learning in terms
of grammatical competence and allow flexible behaviour and the
capacity for ongoing learning on the part of learners. As this compe-
tence, and the processes by which it develops at the individual level,
are variable, Prabhu opted to leave the specifics of how these learning
processes would be engendered for teachers and learners to negotiate
in the classroom based on their intuition and experience. In contrast,
Long’s approach (1985, 2015) avoids the problem of variability in L2
acquisition at the syllabus level altogether by focusing on task learning
and subordinating language to what it is used to accomplish. In this
way, Long is able to propose a syllabus which aims to account for
unified, criterion-based learning outcomes for all learners based on the
things that they need to accomplish in their lives and careers. Long’s
task-based syllabus functions as an illuminative construct with respect
to task learning. This change in focus from language learning to task
learning has had a seminal effect on the field of instructed second
language acquisition (ISLA) in terms of both theory and practice.
The problem, of course, is with the accountability of the syllabus as
a language syllabus as well as a task syllabus. The goals of stakehold-
ers (learners, parents, teachers and educational administrators) are
often not only to accomplish tasks but to develop the linguistic
resources to accomplish them in comparable ways to proficient
speakers of the target language. While language learning and task
learning overlap at times, there is not a one-to-one relationship
between them. If a syllabus is to be accountable as a language syllabus,
evidence of how tasks relate to the specific processes of L2 learning is
required. Long’s syllabus does not deal with this directly. It is left to be
dealt with through focus on form as a set of procedures to address
language problems as they arise in situ (see Chapter 2).

Robinson’s Task-Based Syllabus

Scope
Robinson (2001, 2010, 2011) provides a theoretical foundation for
connecting task performance to specific processes of L2 performance
and acquisition. He outlines what Prahbu would term an illuminative
task-based language syllabus par excellence. To account for the spe-
cific language learning processes that students will engage in in the
classroom, Robinson (2010, 2011) proposes a syllabus that takes
into account factors relating to the situational and cognitive demands
of tasks as well as factors that affect the difficulty that learners
will experience with tasks based on their individual differences. See
Chapter 3 for a detailed account of the Cognition Hypothesis (CH)
that underlies Robinson’s ideas about a task-based syllabus. The
proposed syllabus involves a relatively high level of specification and
takes on a very broad scope within the language curriculum.

Selection
Robinson adopts Long’s (1985, 2015) definition of tasks as the things
that learners have to do outside of the classroom for the purpose of
initially identifying and determining the communicative conditions of
the tasks that learners will complete in the classroom. This requires the
type of behavioural needs analysis that Long advocates. However,
Robinson (2011) also argues that it is necessary to understand the
cognitive, ability and affective demands that tasks place on learners if
predictable gains in L2 performance and learning in the classroom are
to be ensured. The selection of tasks in Robinson’s approach thus
involves analyses related to three aspects of task demand: (1) an initial
behavioural analysis of tasks in situ to identify target tasks and deter-
mine their essential communicative conditions, (2) an information-
theoretic analysis of the cognitive demands that tasks place on all
learners to determine how they should be graded and sequenced into
syllabuses to promote balanced L2 learning, and (3) an ability analysis
to determine the difficulties that learners of different aptitudes and
motivational profiles will experience in completing them (Robinson,
2011, pp. 5–8) so that learners can be matched with tasks suiting their
individual needs.
The first stage of task analysis involves a behavioural needs analysis
of the type discussed in the section ‘Long’s Task-Based Syllabus’, in the
subsection ‘Selection’ to identify the real-world tasks learners need to

complete as a basis for selecting tasks for the syllabus. Robinson

claims that six types of information about target tasks should be
collected at this stage: (1) the list of target tasks, (2) the interactive
demands of these tasks and the roles that they imply for learners, (3)
the subtasks that make up these tasks, (4) the steps needed to complete
the tasks and their subtasks, (5) discourse samples of proficient
speakers completing the tasks, and (6) criterion measures of success
in completing these tasks. This information will be used to set up the
essential structure of the syllabus and establish the situational
and interactive demands that govern language use in the classroom.
This involves the goal orientation of tasks (open/closed, convergent/
divergent), the division of task-essential information between partici-
pants (shared, one-way, two-way), the number of learners involved in
a task, the contributions required (productive, receptive), and whether
the outcome of the task is subject to negotiation of meaning or not.
The information acquired during this first stage of task selection
can also be used to inform the grouping of the learners in terms of
proficiency levels, gender, familiarity, content knowledge, status and
cultural background. According to Robinson, all of these aspects of
pedagogic tasks fall into the category of task conditions. They are part
of the essential structure of the target tasks that motivated
their inclusion in the syllabus, and they are fixed based on an initial
behaviour analysis. They are not subsequently manipulated in
sequencing tasks within the syllabus.
The second stage in selecting tasks for the syllabus is classifying
them based on the information-processing stages involved between
input, output and interaction. According to Robinson, this will consist
of analyses related to component task achievement and the sequencing
of sub-task learning (2011, p. 7). Robinson does not discuss exactly
what these forms of analysis entail or the specific procedures by which
they might be conducted, but he notes that Skehan’s (1998, 2009a)
theory of how the cognitive demands of tasks can be classified and
manipulated during instruction has been the most influential example
of this type of analysis in the field of second language acquisition
(SLA). Presumably, Robinson’s approach to sequencing tasks
demands in terms of whether they direct or disperse learners’ cognitive
resources during task performance (to be discussed in the section
‘Robinson's Task-Based Syllabus’, in the subsection ‘Sequencing’) is
another. Briefly, Skehan argues that tasks allowing exemplar-based
processing, by drawing on learners’ current language resources, pro-
mote the development of fluency and accuracy, and that tasks requir-
ing rule-based processing push learners to experiment with new and

partially mastered language and will result in increased complexity

(see Chapter 3). For instructional purposes, Skehan argues that tasks
which force learners to alternate between both types of processing be
balanced across the syllabus so that learners develop the ability for
dual-mode processing. In this way, fluency does not dominate at the
expense of ongoing development nor does ongoing development dom-
inate at the expensive of fluency. In the first case, learners would chat
fluently using broken language. In the second case, they would know a
lot of language that they cannot use in real time. Robinson’s (2011)
argument is that an information-theoretic analysis of this type will
provide a basis for determining how the cognitive demands of tasks
can be controlled for all learners through the manipulation of factors
in the design of tasks.
The first two stages of task selection in Robinson’s approach thus
make a clear distinction between the situational conditions of real-
world tasks, that were the focus of Long’s approach, and the range of
cognitive demands and psycholinguistic processes that in rudimentary
form was the focus of Prabhu’s approach. This distinction is very
important as it potentially provides Robinson with the conceptual
means of bridging the gap between broad definitions of tasks as the
things that learners need to accomplish outside of the classroom
and narrower definitions of tasks as tools for promoting cognitive
processes related to L2 use and acquisition in the classroom. To
achieve this end predictably, however, Robinson’s approach necessi-
tates a third dimension of task analysis which involves identifying
factors affecting how learners of different aptitude and affect profiles
will respond to them.
Robinson acknowledges that individual differences in learners affect
how they perform tasks and that specific task characteristics will make
some tasks better suited to individual learners than others (2011,
pp. 23–6). For Robinson, the third stage in the selection of tasks is
thus the analysis of the abilities (aptitudes) and affective dispositions
(motivation) that tasks favour so that tasks can be adjusted to accom-
modate the individual differences of learners. Robinson identifies four
types of aptitude that can be argued to affect learners’ performance
and learning on tasks: (1) working memory capacity (tasks without
visual support during performance could favour learners with higher
working memory capacity); (2) executive control (tasks that require
dual-tasking could favour learners who can switch quickly from one
task to another); (3) causal reasoning ability (tasks requiring complex
arguments to be structured could favour learners with high causal
reasoning ability); (4) sensitivity to the mental states of others (tasks
that require learners to speculate on why different people behaved the

way that they did could demand a high level of sensitivity to the
mental states of others). Likewise, Robinson argues that factors
related to output anxiety, self-efﬁcacy, self-regulation, openness to
experience or tolerance of ambiguity could also affect learners differ-
entially. For example, performance on tasks with an open solution as
opposed to a single correct answer may cause problems for students
with a low tolerance for ambiguity and result in unpredicted variation
in the language used. However, in terms of the implications of ability
and affect analyses for task selection, Robinson claims that to optimize
and unify learning gains on tasks, learners of differing dispositions
will need to be matched with tasks which suit their affect and
ability proﬁles and reduce variability due to individual differences in
aptitude and affect. He does not discuss how tasks catering to individ-
ual differences might be incorporated in the design of task-based
syllabuses, however. We will return to this point in the section ‘Ellis’
Task-Based Syllabus’.

Sequencing
Robinson’s approach to sequencing tasks is based only on the cogni-
tive factors in their design. Factors relating to the interactive and
situation demands of tasks remain fixed throughout the syllabus, as
explained. Ability and affect demands of tasks are likewise not used in
sequencing tasks in the syllabus but to match tasks to learners of
different individual difference profiles. Robinson’s approach to
sequencing involves two sets of cognitive demands, which could be
argued to relate directly to the complementary cognitive processes of
analysis and control (Bialystok, 1994). Briefly, Bialystok argues that
analysis is related to the destabilization, restructuring and develop-
ment of the language system, whereas control is related to the stabil-
ization of the language system in terms of fluent access to current
resources. She argues that these two processes account for the devel-
opment of symbolic systems and complex skill acquisition over the
human lifespan. In Robinson’s model, task design factors that direct
learners’ cognitive resources (e.g. attention, working memory) to spe-
cific features of task content and the language needed to express it
could be argued to relate to processes of analysis, whereas task factors
which disperse learners’ cognitive resources over several aspects of the
larger performance context could be argued to relate to processes of
control. The examples he provides of resource-directing factors are:
(1) the number of elements in the task, (2) the temporal and spatial
displacement of these elements, (3) the need to make the reasoning
(spatial, causal, intentional) explicit, and (4) the need to make

perspectives on elements explicit. All of these tend to be explicitly

grammaticalized in English and they direct learners to specific task
content and require them to grammaticalize that content. Examples
that he mentions of resource-dispersing task factors, on the other
hand, are: (1) the number of steps in tasks, (2) the relationships
between these steps, (3) the inherent structure of the task, (4) the
need to multitask, (5) time to plan, and (6) familiarity with the task
(Robinson, 2011, p. 7). In contrast to Skehan’s (1998, 2009a)
model of cognition and L2 performance summarized in ‘Robinson's
Task-Based Syllabus’, in the subsection ‘Selection’, Robinson
argues, based on work by Neumann (1987) and Sanders (1998) that
increasing the resource-directing demands of tasks will promote
greater effort towards production and more vigilant monitoring of
output, impacting language production in terms of both complexity
and accuracy as well as increasing uptake from input.
Robinson (2010) proposes a three-stage procedure for manipulating
these two types of cognitive task design factors in sequencing tasks
within a syllabus. He refers to this as the SSARC model of L2 task
sequencing. This model was briefly introduced in Chapter 3. The first
stage involves simplifying the new task in terms of both the resource-
directing and resource-dispersing factors in its design. This provides
initial support for performing the new tasks. It also allows learners
to access their current L2 resources and stabilize these resources in
relationship to the new task. The second stage is increasing resource-
dispersing task demands (keeping resource-directing demands con-
stant) until these demands match those of the target task on which
the new task is based. This stage allows learners to automatize the L2
resources brought to bear in the first stage. The third stage then
involves increasing resource-directing task demands. These factors
bring the new task closer in line with the target task that learners need
to complete, destabilizing the linguistic strategies they have brought to
bear on the new task in the previous two stages and requiring learners
to analyse and restructure these L2 resources in line with the demands
of the full version of the target task, evidently complexifying the
language that they use in the process. If the task is particularly chal-
lenging, this cycle is repeated until learners can complete full versions
of the new task that are equal to the target real-world task it
represents.
An example of how Robinson’s SSARC model of L2 task sequen-
cing was put into practice in developing a task-based syllabus for a
EFL programme in Japan will be discussed later in this chapter. This
example will illustrate the high level of specification involved in an
illuminative task-based syllabus that aims to predict language learning

processes for all learners. Robinson’s approach to task-based syllabus

design was the ﬁrst with the potential to bridge the gap between the
approaches of Long, who focuses on training and assessing task ability
rather than language ability, and Prabhu, who focuses on language
ability but leaves the issue of how tasks relate to learning processes to
be negotiated by teachers in the classroom.

Ellis’ Task-Based Syllabus

In contrast to Long and Robinson, Ellis’ position on the task-based
syllabus might be seen as a return to the approach used by Prabhu.
Like Prabhu, his position is that a syllabus cannot anticipate all
sources of variation that will occur for different learners and across
different classrooms. Ellis argues that pedagogic tasks are conglomer-
ations of the features that interact with one another and with a full
range of individual and contextual variables to produce the learning
processes that take place in the classroom. A research agenda aimed at
predicting performance and learning based on tasks would have to
identify the effects of myriad variables in all of the possible combin-
ations in which they occur in tasks when used with different learners
and in different contexts. Robinson (2011), for example, identifies
thirty-six variables related to task–learner interactions. It is not diffi-
cult to imagine a research agenda of this complexity expanding expo-
nentially with each new variable and receding beyond the event
horizon of ISLA. In addition, however, Ellis’ position, again like
Prabhu’s, is that an illuminative syllabus would be undesirable, even
if it were possible, as it limits what teachers and learners bring to the
learning process in terms of the intuitive decisions and adjustments
that they make in optimizing learners’ mastery of syllabus content.
A third issue Ellis mentions is feasibility. He points out that it is not
clear what an illuminative syllabus connecting tasks to language learn-
ing might look like. For example, it is not clear how tasks might be
matched to learner profiles in practice. Even if researchers were able to
demonstrate a finite number of learner types, it would be unrealistic in
most language programmes to develop and implement separate sylla-
buses for each of them in order to control for the effects of individual
differences on performance and predict the specific learning processes
that learners will engage in while performing tasks.

Scope
Ellis (2018a; see also Chapter 10) thus returns to Prabhu’s position on
the role of the syllabus in the language curriculum, advocating a

syllabus which functions as an operational construct. Ellis makes a

clear distinction between task-as-workplan and task-as-process (see
Chapters 1 and 4). The former is the domain of syllabus design while
the latter is the domain of methodology. Ellis thus limits the role of the
syllabus to a workplan for teachers to help their learners achieve
competence in the language rather than training learners to perform
a specific set of tasks. Unlike Prabhu, however, who is interested in
grammatical competence, Ellis sees the purpose of task-based instruc-
tion (TBI) as developing interactive and pragmatic competence so that
learners can negotiate understanding effectively and appropriately
on the range of tasks that they may face in the future. For this goal
to be reached in the classroom, it is necessary for teachers to adjust the
task online as it is being performed to take account not only of
individual difference but of fluctuations in classroom dynamics in
order to ensure learning. Specifically, Ellis (2003) distinguishes
between ‘performance options’ and ‘process options’. The former
refers to options that can be planned in advance, whereas the latter
can only figure in situ as the task is performed. Examples of the latter
variables might be: (1) whether, when and how long specific learners
need to plan, (2) whether and how they can best benefit from repeating
tasks, (3) whether they can better benefit from imposing a time limit on
task performance or whether they should be allowed to do it at their
own pace, (4) whether and to what extent they can benefit from
performing receptive versions of pedagogic tasks and what the focus
of these tasks should be, (5) how receptive versions of tasks can best be
integrated with productive versions of tasks for a given group of
learners, (6) whether the social roles and conditions of tasks should
be adapted to learners’ affective dispositions, and (7) when and how to
provide effective feedback for learners of different aptitudes and
motivations within a given class. In short, rather than attempting to
ensure specific learning processes (automatization, restructuring, etc.)
in advance, Ellis’ position is that teachers can best ensure these pro-
cesses by matching tasks to learners’ individual differences as the tasks
are performed in the classroom.
Unlike Long and Robinson, Ellis (2009a) defines ‘task’ not in real-
world terms, but as a language learning tool. He argues that for tasks
to develop interactive and pragmatic L2 competence, they must: (1)
focus learners on meaning rather than form, (2) have a communicative
gap that necessitates communication, (3) require learners to draw
on their own resources in completing them, and (4) result in a com-
municative outcome beyond the use language for its own sake (see
Chapter 1). Ellis’ definition of task thus focuses on the conditions
necessary for authentic processing of language and the development

of a general capacity to use language that is not be limited to speciﬁc

tasks, but which he argues will transfer to any number of tasks that
learners may complete in the future. The integrity of tasks, according
to this four-part deﬁnition, ensures that teachers’ workplans involve
tasks. If any of these four criteria are missing, the resulting activity
may correspond to what Ellis refers to as a situational grammar
exercise rather than a task (Ellis, 2009a).

Selection
Ellis (2018a; see also Chapter 10) argues that different approaches to
selecting specific tasks will be required for different learners. In the
case of learners with specific occupational needs involving the target
language (e.g. nurses, flight attendants, hotel staff, receptionists, sales
staff, etc.), tasks should be selected based on a needs analysis as
suggested by Long (2015). However, in the case of children or
general-purpose language learners, tasks might be selected based on
criteria similar to Prabhu’s: (1) the intrinsic interest that tasks for
learners, and (2) learners’ previous experience or familiarity with
them. A typical example of general-purpose learners might be English
majors in Japanese universities. These learners frequently have very
little contact with English after graduating and entering the Japanese
workplace. However, they are very motivated to develop social and
interactive competence in English for personal development and are
keen to use it in any situation that affords them the opportunity. Ellis
argues that, provided tasks have the four criteria to ensure their
integrity as learning tools, any of the approaches to task selection
discussed in this chapter may be appropriate depending on the goals
and interests of the learners in the programme for which the syllabus is
being designed. In other words, Ellis recognizes the case for selecting
tasks that cater to the target needs of learners where this is possible,
but that otherwise different criteria to select tasks for a general-
purpose language course are needed.

Sequencing
Whether tasks are selected based on real-world tasks or whether
generic tasks are selected which engage learners’ interest and back-
grounds, once the list of pedagogic tasks to be included in the syllabus
has been determined, Ellis (2003) suggests grading and sequencing
tasks in terms of factors relating to four aspects of their structure:
(1) the input they provide, (2) the interactive conditions they entail, (3)
the reasoning they require, and (4) their resulting outcomes. Based on

Table 7.2 Ellis’ criteria for task sequencing

Easy Difﬁcult
Input factors
1 Non-verbal input Written input Aural input
2 High-frequency lexis Low-frequency lexis
3 Shorter, simple Longer, complex
sentences sentences
4 Static information Dynamic information Abstract information
5 Few elements/ Many elements/
relationships relationships
6 Structured Unstructured
7 Here and now There and then
8 Familiar Unfamiliar
Interactive factors
9 Two-way One-way
10 Single task Dual task
11 Dialogue Monologue
Reasoning factors
12 Information gap Reasoning gap Opinion gap
13 Few steps Many steps
Outcome factors
14 Pictures Written Oral
15 Closed Open
16 Descriptions Instructions/narratives Arguments

Source: Ellis (2003)

a review of the literature, Ellis suggests sixteen potential factors which

might be used as a basis for sequencing decisions in these four areas
(see Table 7.2).
It should be remembered, however, that Ellis does not suggest that
these criteria be prescribed in a ﬁxed format in developing programmes,
as Robinson does, but rather that they function as a set of resources and
guidelines. According to Ellis, teachers will develop a task-as-workplan
and then carefully monitor the resultant behaviours in the classroom
(i.e. the task-as-process) and make adjustments to ensure that task
sequences work to provide the opportunities necessary for learners to
reach task outcomes and develop competence. The resources that they
use to do this will vary in different contexts and with learners of
different aptitudes and affective dispositions. Ellis’ task-based syllabus,
like the syllabus proposed by Prabhu (1987), thus provides content for
teachers to promote L2 learning and leaves the critical issue of

supporting learners of differing motivations and aptitudes to be dealt

with by teachers, based on their experience with their learners and the
behaviours that occur as tasks are performed in the classroom.
Ellis (2018a; see also Chapter 10) also provides a rationale for a
modular syllabus which incorporates modules based on a task-based
syllabus along with modules based on a traditional structural syllabus
organized in terms of targeted grammatical and lexical forms. He argues
for an approach in which a task-based syllabus is used at the lower level
of proficiency until learners gain fluency in the language. The structural
syllabus then begins to play a progressively greater role at the intermedi-
ate and advanced levels. However, he also points out two other alterna-
tives for pairing structural and task-based syllabuses, which he favours
less. In the first, a structural syllabus is used at the early stages of
proficiency (as is typical of current practices in many programmes),
followed by a task-based syllabus at the later stages. In the second, both
types of syllabus are used in parallel at all levels of proficiency. If the third
approach is adopted, however, Ellis points out that it is essential that the
two syllabuses be kept separate (e.g. taught on different days). If the task-
based syllabus is used to practise what is taught in the structural syllabus,
the integrity of the task-based syllabus is compromised and the result is a
task-supported syllabus rather than a modular syllabus. In this way,
learners have the opportunity to access the full range of their own L2
resources to complete tasks. Ellis argues that his modular curriculum
represents a feasible way to begin using TBLT in many international
contexts. Thus, although Ellis’ approach to task-based syllabus design
does not attempt to predict performance or provide accountability for the
learning that will take place in the classroom, it has the advantage of
being more practical and feasible in many contexts internationally in
which teachers and programme designers might otherwise struggle with
a task-based approach or not have the freedom to use one at all.

Applications
Before concluding, it will be helpful to consider how some of the
approaches that have been outlined in this chapter have been put into
practice. Two actual task-based syllabuses that have been used with
EFL learners in Japanese universities will be considered. The ﬁrst
example (Lambert and Robinson, 2014) will illustrate a task-based
syllabus which is based on an analysis of the future needs of Japanese
English majors at public university in Japan. The syllabus is intended
to function as an illuminative construct within the L2 curriculum in
that it is highly structured and aims to account for the speciﬁc learning
processes (automaticity, restructuring etc.) that take place at different

stages for all learners. In contrast, the second example (Kelly and
Kelly, 1996) will illustrate a task-based syllabus in which tasks were
selected based on the materials designer’s experience and intuition of
tasks that are intrinsically interesting to Japanese EFL learners and
that are likely to develop their general interactive competence in using
English. Unlike Lambert and Robinson (2014), this syllabus functions
as an operational construct within the L2 curriculum in that it has a
relatively low degree of structure. In addition to the main task
sequence, it provides optional exercises for teachers to draw on in
supporting different learners in different contexts.

Example 1:
The SSARC model of L2 task sequencing

This example is a task-based syllabus used with English majors at a

Japanese university (Lambert and Robinson, 2014). A task-based needs
analysis revealed that summarizing English-language materials was one
of five critical task types for English majors at the university (Lambert,
2010). In the module, task sequences were developed to train learners in
summarizing English-language short stories. The picture strips used in
the first four lessons of the syllabus were taken from Trondheim (2007)
and adapted for the pedagogic tasks used in each sequence. This chil-
dren’s book consists of thirty highly comparable sixty-frame pictures
stories involving a protagonist (Mr I) and his series of frustrated
attempts to steal a piece of a pie cooling in a window.
The example consists of a module of six lessons. The first two were
based on simplified versions of the stories in which frames relating to
mental states were removed so that the narrative consisted only of a
series of physical actions and motion paths. The first lesson was
intended to activate learners’ current L2 resources in connection with
the task and draw their attention to resources that they had already
learned for conjoining clauses and expressing different types of motion
in English. It was thus connected with the simplify-and-stabilize phase
of Robinson’s (2010) SSARC model. The second lesson was intended
to automatize the resources that learners brought to bear in Lesson
1 and was connected with the automatize phase of the SSARC model.
The focus was on extensive, time-pressured use of relevant language
forms in different contexts.
The third and fourth lessons were based on longer picture strips
(sixty frames), which included frames representing both actions and
the mental states of the characters involved in the narrative. The third
lesson was intended to force learners to restructure and complexify

(the final phase of the SSARC model) the resources that learners
brought to bear in Lesson 1 and automatized in Lesson 2 in line with
the demands of adding mental states to their narratives and the verbal
subordination processes that facilitate such narratives. The fourth
lesson was then intended to provide more extensive practice of the
resources that learners had brought to bear in Lesson 3 and develop
flexibility in applying them in new and diverse contexts, returning to
the automatize phase of the SSARC model.
Finally, the fifth and sixth lessons were based on authentic English-
language short stories. In the fifth lesson, learners summarized Japan-
ese short stories in English translation, and in the sixth lesson, they
summarized original English short stories. The purpose of the fifth
lesson was again to provide them with exposure to language used with
each story and notice the gaps between their own production and the
production of their native-speaking peers. It was thus connected with
further restructuring and complexifying their language for summariz-
ing stories. Learners were expected to pick out the important events in
the story, say what they thought the characters were thinking and
feeling to behave the ways that they did, and try to provide some
conclusion of their own regarding what the stories mean. In the sixth
lesson, the purpose was again to automatize the resources brought to
bear in Lesson 5 and develop fluency and flexibility in applying what
each learner had acquired in different contexts.
The task sequences in each lesson illustrate how the syllabus relates
directly to the learning processes that are intended to occur:
Lesson 1: Activate and stabilize current L2 resources for narrating
physical actions
Task 1: Each learner in a group of four has a different twenty-four-
frame picture strip story limited to overt actions. They summar-
ize their stories as best as they can so that the others can
sequence scrambled sets of pictures. They are given time to
prepare and make notes next to each frame.
Task 2: Learners complete cloze activities based on transcriptions
of their native-speaking peers completing the narratives from
Task 1. The focus is on coordinating conjunctions used to
combine clauses. They first guess the missing conjunctions, then
listen to check their answers.
Task 3: Learners repeat Task 1 eight frames at a time. Next to each
frame, options for verbs of motion are provided (e.g. came back,
walked over, snuck over, etc.) for learners to choose from as they
tell the story. They perform the task three times in pairs,
changing partners each time.

Task 4: Learners listen to two different native speakers narrate a

new story. They use cloze transcripts to guess missing conjunc-
tions and verb particles for verbs of motion then listen and check
their answers.
Task 5: Learners repeat Task 1 after ten minutes of planning time.
Revision: Learners read the transcripts of native speaking peers
telling the four stories they told in class, identify useful language
and expressions, and write a narrative of a new story incorpor-
ating this language.
Lesson 2: Automatize current L2 resources for narrating physical
actions
Task 1: Learners retell the stories from Lesson 1 (Tasks 1 and 5)
three times each to different partners each time under time
pressure. Listeners make notes to tell each story back to the
original speaker, who conﬁrms or corrects key facts.
Task 2: Learners repeat the procedures for Task 1 with the new
stories that they completed for homework under progressively
increasing time pressures.
Task 3: Learners read cloze versions of transcripts of the stories in
Task 2 with bold type used to focus their attention on subordin-
ating conjunctions for combining clauses.
Task 4: The procedures in Tasks 1 and 2 are repeated in telling and
retelling new stories. Learners are given ten minutes to plan.
Revision: Learners read transcripts of the stories told in class,
identify useful language and expressions, and write two letters
to a friend summarizing two of these stories. They include what
they believed each character was thinking and why they behaved
in the way they did to activate their current L2 resources for the
subsequent lesson.
Lesson 3: Restructure and complexify L2 resources for narrating
physical actions and mental states
Task 1: Learners summarize the stories they had prepared for
homework, telling them three times each to different partners
each time.
Task 2: Learners listen to a narration of a full (sixty-frame) story
consisting of frames connected with actions and mental states.
They listen statement by statement, writing the number of the
statement connected to each frame. Multi-clause statements
sometimes correspond to multiple frames.
Task 3: Learners guess the missing verbs in twenty statements made
that attribute mental states to characters (Mr I noticed that . . .,
He was surprised that . . ., etc.), and listen to check their answers.

Task 4: Learners complete a dictation task of ﬁve statements made

and label the function of each statement (provide background,
establish goal, state outcome, draw a conclusion etc.).
Revision: Learners study transcripts of statements from Task 2 in
relation to the frames of the story they described. The examples
involve subordinated constructions relating to mental states and
expressing motivations for the story. They prepare a narration of
the full sixty-frame version of the stories that they had summar-
ized in Lesson 1 (Task 1) with mental state frames included.
Lesson 4: Automatize new L2 resources for narrating physical actions
and mental states
Task 1: Learners tell the stories they had prepared for homework in
pairs and make notes on ways to improve them based on lan-
guage their partner used in telling her story.
Task 2: Learners listen to narrations of the full sixty-frame versions
of the stories they prepared for homework and note how these
narratives differ from their own.
Task 3: Learners read twenty statements expressing cause-and-
effect relationships used with gapped connector words (i.e. in
order that, so, as a result), guess the missing words and listen
to check.
Task 4: Learners summarize new sixty-frame stories (actions, inten-
tions, motivations). They have ﬁve minutes to prepare.
Revision: Learners review transcripts and write 1,000-word sum-
maries of a new sixty-frame story. They also read one of four
short stories by Haruki Murakami and prepare to summarize it
in the next class.
Lesson 5: Restructure and complexify L2 resources for summarizing
authentic short stories
Task 1: Learners take turns summarizing their Murakami stories in
groups of four.
Task 2: Learners listen to a recording of a proﬁcient speaker
summarizing of one of the Murakami stories and identify which
of the forty-eight words were used to explain the mental states
(e.g. remembers, wonders, hopes, anticipates).
Task 3: Learners listen to a summary of a second story. They
identify events from the story which were mentioned or left out.
Task 4: Learners listen to a summary of a third story. They identify
statements that are true and false.
Task 5: Learners listen to a summary of a fourth story. They
identify the conclusion made.
Task 6: Learners repeat Task 1 after ten minutes of planning time.

Revision: Learners review transcripts of the Murakami summaries

from class with key elements highlighted. They also read one of
four short stories by Ernest Hemingway and prepare to summar-
ize it in the next class. They pick out the important events in the
story, say what they think the characters are thinking and feeling
to motivate their actions, and provide a conclusion on the story’s
meaning.
Lesson 6: Automatize L2 resources for summarizing authentic short
stories
Task 1: Learners form groups of four and work in pairs to sum-
marize their Hemingway stories three times each to different
partners each time without time pressure. They thus tell their
story three times and hear three other stories once each. Each
time they make notes and retell the story that they heard back to
their partner, who checks and corrects any problems.
Task 2: Learners change to a new group of four. They repeat Task 1
under increasing time pressure.

Example 2:
A modular approach to the task-based syllabus

The second example of a task-based syllabus is from a published

textbook for use in courses aimed at developing oral English skills in
Japanese universities (Kelly and Kelly, 1991). The book provides
an interesting example of how task-based and task-supported struc-
tural syllabuses might be used in parallel for learners of differing
motivations and aptitudes based on enjoyable and cognitively
engaging content. The task sequences in the book are organized
around solving mysteries through detective work (e.g. catch a killer,
disarm a bomb, identify counterfeits, perform rescues, catch a thief
etc.). The tasks were not selected based on an analysis of the situations
that learners might face in the future, but based on the materials
designer’s experience of what is intrinsically interesting to the targeted
student population and the opportunities that they provide for
learners to develop the interactive skills that they need in order to
complete the range of tasks that they will face in the future. As none of
the tasks could be argued to be more frequent or critical for learners,
the arrangement of task types within the syllabus seems to have been a
matter of convenience or intuition.
The example is taken from Unit 2 of the book and involves a
sequence of tasks which involves ﬁnding and disarming a bomb.
Learners work together in pairs to share information in order to

complete a three-task sequence: (1) ﬁnd the apartment in which the

bomb is hidden, (2) find the piece of furniture in which the bomb is
hidden, and (3) follow instructions for disarming the bomb. In the first
task, learners each see the outside of six apartments. One learner has
the view of these apartments in the morning, and the other has the
view in the afternoon. They are told that two changes were made
outside of the apartment containing the bomb between morning and
afternoon. They work together based on their own L2 resources to
describe each of the entrances and find the one with two differences. In
the second task, they move to the interior of the apartment to identify
the piece of furniture in which the bomb is hidden. One has a picture
representing a perspective on the interior of the apartment, and the
other has a floorplan from a rotated perspective. They must describe
and compare the ten objects in the room in relationship to one another
in order establish a shared perspective and identify what the shape on
the floorplan containing the bomb corresponds to in the room. Finally,
in the third task, they disarm the bomb. Both of the learners have
pictures of the same bomb consisting of several different colour wires,
a timer, a battery and some explosives. However, one learner has a
precise eight-step sequence for disarming it labelled on this diagram
(cutting, disconnecting, turning off, etc.), and the other does not. The
speaker must explain these steps to the listener, who must ask any
questions in order to execute each step accurately by marking the
diagram.
As learners complete each task sequence in the book, they have a
checklist to write their answers. When they are finished with the
sequence, the teacher evaluates this checklist and allocates points for
correct answers. The teacher often has some flexibility in the level of
precision required in learners’ answers so that higher-proficiency
learners can be pushed to be more precise than lower-proficiency
learners. Students then enter their points for each task sequence into
the back of the book to track their progress across the course. In the
end, they are awarded a level of achievement based on the quality of
their detective work rather than their English. The entire book is set up
as a detective school in which they must solve multiple mysteries, and
the focus throughout is thus on achieving these task outcomes rather
than on the language used.
In addition to the task sequences and the goal-tracking system, an
additional sequence of task-supported language exercises is provided
for each task sequence in a separate part of the book. Based on the
motivation, aptitude and ability of their learners, teachers can choose
to leave this material out entirely or draw on it in part or in whole
before, during or after the task sequences, based on the needs of each

group. The resources consist of vocabulary items which might ease

learners’ cognitive load while completing the tasks, input-based tasks
and slot-and-frame structures which draw attention to key functions
for negotiating meaning (asking for more information, confirming,
clarifying, asking to repeat, expressing uncertainty, etc.) when com-
pleting each task sequence. Finally, two comparatively simple versions
of the tasks in the main task sequence are provided for learners to
practice using these resources. The first is typically a simplified version
of one of the tasks in the main task sequence (identify T-shirts by
describing shapes in relationship to one another), and the second is
typically an open task allowing learners to discuss content of their own
(design a T-shirt and describe it so that you partner can draw it).
Clearly, the purpose of the course provided in Kelly and Kelly
(1991) is not to train learners to complete the specific tasks involved
in locating and disarming bombs in the future, but to provide engaging
pedagogic tasks which develop their interactive competence in using
English to negotiate understanding on a range of tasks in the future.
Likewise, the syllabus is not intended to determine the learning pro-
cesses that will take place for all learners, but to support the teacher in
ways in which learners at different levels of proficiency, motivation
and aptitude profiles might be aided in successfully completing the
task sequences in the syllabus, thereby improving their interactive
competence in English.

Conclusion
It should be clear from the discussions in this chapter, that each of the
approaches to task-based L2 syllabus design that have been covered
are with the essential principles of TBLT outlined in Chapter 1, and
that they all have advantages and disadvantages which will make them
more or less appropriate to the needs of teachers and learners in
different instructional contexts. In concluding, it will be helpful to
compare the strengths and limitations of each.
In the approach advocated by Long (1985, 2015), the focus is on
learners’ ability to perform speciﬁc real-world tasks that they are likely
to encounter outside of the classroom. As in other forms of workplace
training, syllabus content is selected and organized in terms of the
frequency and criticality of these target tasks that learners need to
complete, and learners are trained with initially simple, progressively
more demanding, versions of each task so that they can master what
performance of the full task requires. At the syllabus level, L2 devel-
opment is seen as a by-product of task learning. Learners acquire the
language necessary to complete tasks incidentally as a function of

learning to complete tasks effectively and efﬁciently. Focusing on task

learning rather than language in this way avoids the problems of
variability in SLA and provides criterion-based learning outcomes
for all learners that directly reflect what they need to accomplish
outside of the classroom.
However, L2 learners and policy makers are often interested in
developing more than the ability to accomplish specific tasks. Many
stakeholders in language education are interested in developing the
general competences that proficient speakers bring to bear on tasks.
Robinson’s (2010) approach to task-based syllabus design provides a
solution to accounting for broader L2 development as a product of
performing tasks. By attempting to specify the cognitive and affective
demands of tasks in addition to identifying the tasks themselves,
Robinson’s approach aims to predict the language learning processes
that will take place on tasks. While an admirable goal for research, this
approach might not be feasible for L2 practitioners in contexts with
mandated tests and curricular content.
In this regard, Ellis (2018a, Chapter 10) argues that a modular
approach to task-based language teaching will not only allow practi-
tioners to experiment with tasks within the context of current lan-
guage curricula, but that this combination of approaches might
complement one another and lead to learning outcomes that neither
approach can achieve easily on its own. Like Prabhu (1987), however,
Ellis advocates an operational syllabus over an illuminative syllabus,
even if an illuminative syllabus were possible in a given context. He
argues that seeing to the details of L2 development on tasks is best
done by teachers in situ with specific groups of learners. Ellis argues
that the syllabus should not function to dictate the procedures used in
the classroom, but that it should provide teachers with resources and
freedom to address the needs of learners differing in motivation and
aptitudes as well as fluctuations in classroom dynamics.

Introduction
The methodology of task-based teaching refers to the various options
for task implementation that may affect the affective and linguistic
aspects of task performance and the learning that results. This chapter
seeks to explore options and ideas that are empirically informed
and/or theoretically justified, to stimulate thoughts for further
research, and to expand pedagogical choice building on available
research and current practices. The chapter will foreground the
importance of making evidence-based decisions, reflecting ‘the
centrality for research’ in a task-based approach in comparison
with ‘communicative language teaching’ that is comprised of
some loose concepts devoid of theoretical and empirical support
(Skehan et al., 2012). Three types of evidence are drawn upon in
this chapter, including: (1) empirical evidence, which consists of the
methodological options reported in the research and the findings
regarding the effects these have on the process and product aspects
of task-based learning; (2) practical evidence, which takes the form of
pedagogical recommendations proposed by teacher guides; and (3)
theoretical claims on the benefits and limitations of task-related
procedures.
The three types of information roughly correspond with Long’s
(2015) evaluation criteria for the validity of the methodological prin-
ciples he proposed for the implementation of task-based language
teaching (TBLT). Long claimed that the principles are based on at
least one of the three major criteria: theoretical motivation, empirical
support and logical argumentation. The three criteria were illustrated
through the example of corrective feedback (CF), the utility of which is
supported by multiple theories (theoretical motivation), a large
amount of research (empirical support) and arguments (logical argu-
ments, not claims of independent, well-established theories) such as

208

that of Lydia White (1987), who worried about the impossibility of

unlearning errors related to first language influence merely through
positive evidence. Long states that due to various constraints, there has
been a lack of research in certain areas, but language teaching cannot
be ‘put off for a few years while the research is carried out’ (p. 301).
Therefore, alternative criteria such as theoretical motivation and
logical argumentation are in order.
In his discussion of the methodology of task-based teaching, Long
(2015) made a distinction between methodological principles and
pedagogical procedures. Methodological principles are language
teaching universals, informing teachers of what should be done; peda-
gogical procedures are specific steps teachers follow in implementing
the principles, specifying how it should be done. Long (2015) only
elaborated some methodological principles (‘elaborate input’, ‘provide
negative feedback’, etc.) without discussing pedagogical procedures,
on the grounds that there are an infinite range of options to instantiate
the principles and the job of how to implement a principle should be
left to the local teacher. In this chapter, we do not make a distinction
between principles and procedures (although our discussion will
centre more on the latter); rather, we take a bottom-up approach
providing a detailed description of the various options available for
teachers at each stage of a TBLT lesson. We will consider these options
in terms of the three stages of a task cycle: pre-task, main task and
post-task (Ellis, 2003).
Dividing the task cycle into the three stages provides a convenient
framework to discuss the procedural aspects of a task-based lesson
(Willis, 1996; Ellis, 2003; see also Chapter 1). The pre-task stage
serves to prepare learners for the main task and the post-task stage is
where learners engage in follow-up activities after the task is com-
pleted. Ellis (2003) pointed out that among the three stages, only the
main task stage is required. However, as we will see, pre-task and
post-task activities are critical to achieving the goals of task-based
teaching, and there has been abundant research on the influence of
the various options, especially pre-task options, on task performance
and learning gains. It is necessary to clarify that although the three
stages are discussed in the context of a single task-based lesson, they
can occur in different lessons. For example, in preparation for a debate
on gun control, the teacher may brainstorm the topic with students in
the previous lesson to arouse their interest. Similarly, after completing
the main task, students may be asked to reflect on their task perform-
ance in the next lesson. Furthermore, one component that applies to
all three stages of a task cycle is participatory structure, which refers to
the pattern of interaction a task involves. In the following sections, we

will elaborate the various options for the three stages, followed by a
brief discussion of participatory structure.

Pre-Task Options
Various activities can be conducted in the pre-task stage (see Table 8.1
for a summary), some of which have been examined through empirical
research. To know what activities should be included in the pre-task
stage, it is important to understand the goals of pre-task activities.
First, pre-task activities serve to motivate learners, arousing their
interest and building up their expectations. To motivate learners,
teachers may inform students of the relevance of the task to their
personal life, the real world and/or their study goals (see Dembovs-
kaya (2009), discussed in Chapter 5, for example pre-task motivating
strategies). Giving students a reason to perform the task is especially
important in foreign language settings or for learners who have no
prior experience with tasks and who may have negative perceptions
about a task-based approach. Willis and Willis (2007) noted that
for comprehension (reading and listening) tasks, having students
make predictions about the content of the aural or written text, such
as by using the title or subtitles and accompanying visuals, is an
especially effective motivating strategy. Although task motivation
has signiﬁcant theoretical and pedagogical value for a task-based
approach, there has been very little research on this aspect of TBLT
(see Chapter 5).
The second goal of pre-task activities is to prepare learners for the
upcoming main task. Learners’ preparedness for a task means that (1)
they are clear about the task procedure and expected outcome, and (2)
they are equipped with the adequate resources required for task com-
pletion including the relevant linguistic and schematic knowledge –
background information about the topic. It should be clariﬁed that here
linguistic knowledge refers to lexical or vocabulary knowledge, not
knowledge about grammar or morphosyntax. Whereas vocabulary is
essential for task completion, grammar is often not important although
it may facilitate the accuracy of communication. The controversy over
whether to teach grammar in the pre-task stage will be addressed in the
next section. Furthermore, it is important to clarify that the purpose of
providing learners with the necessary vocabulary is to scaffold rather
than stipulate the language needed for task performance.
The third goal of pre-task activities is to provide learning opportun-
ities which may have an effect on learning outcomes or task perform-
ance. For example, responding to a survey or quiz about the topic of
the subsequent task may encourage learners to think and talk about

Table 8.1 Summary of pre-task methodological options

Options Description Recommendations

Pre-task planning Learners are given time to • 1 – 3 minutes of planning
plan the content and seem ideal, but teachers
language of subsequent should customize the
task performance length of planning
according to learner and
task characteristics
• During planning, teachers
may consider allowing
students to make notes,
work with language
materials, and plan in their
L1, L2 or both.
• No planning is an option
Pre-task focus on Pre-task grammar • Make sparing use of pre-
form instruction: learners are task grammar instruction
taught a grammar rule because it may affect
before task performance learners’ ﬂuency and the
complexity of their speech
production.
Pre-task modelling: learners • Be careful about what to
are provided with a model because the model
model performance has a signiﬁcant impact on
learners’ task performance.
Other options Topic preparation: provide • Too much background
content knowledge about knowledge reduces
the topic negotiation
Teacher scaffolding: • Help learners with task
perform a similar task procedure
together with learners
Vocabulary preparation • Teach vocabulary to
prepare learners for task
performance but do not
require them to use
prescribed words

grammar and to look up new words. Allowing learners to plan may

free up their attention, leaving more cognitive space for focusing on
linguistic forms during task performance. Modelling how to negotiate
meaning and form in the pre-task stage may increase the incidence of

language-related episodes (LREs) in learner–learner interaction (see

Chapter 4), which may lead to learning gains.

Pre-Task Planning
Pre-task planning, sometimes called strategic planning, is a peda-
gogical option that allows learners time to plan the content and
language of their task performance before the main task. Pre-task
planning is probably the most researched pre-task option in task-
based research, and the intensity of the interest is both theoretically
and empirically motivated. Theoretically, task planning has been
extensively investigated to verify Skehan’s Limited Attention Capacity
Hypothesis (LACH) (see Chapter 3), which states that pre-task
planning may ease the pressure on learners’ limited cognitive resources
during task performance and mitigate the trade-off between the vari-
ous aspects of speech production such as complexity and accuracy.
The findings of planning research have often been cited to support
Skehan’s theory and counter Robinson’s Cognition Hypothesis (CH),
which states that there is no trade-off between complexity and accur-
acy because they draw on different resources, and that what affects
learners’ task performance is the processing demands of the task,
rather than learners’ limited attentional capacity. Pedagogically, some
teachers consider it important to allow students some time to plan to
make them more prepared and to make performing a task in a second
language (L2) a more pleasant experience. Others, however, hold that
pre-task planning may deprive students of the opportunity to practise
how to use L2 knowledge in spontaneous communication. Planning is
also of relevance to language testing. Wigglesworth and Elder (2010)
argued that planning contributes to the fairness of an oral test by
reducing learners’ stress and anxiety to enable them to achieve their
best possible performance. In the following, we summarize planning
research with a view to demonstrating how the findings can inform
pedagogical decisions.
The Effects of Pre-Task Planning. The effects of pre-task planning
can be investigated in two ways: in terms of its influence on learners’
task performance and on the learning gains that result from perform-
ing the task. Task performance concerns what happens while learners
are performing a task, namely the process features of task-based
instruction (TBI). Learning gains, which are measured through pre-
tests and post-tests, pertain to the product aspect of TBI. To date, the
majority of studies have investigated task performance, and only one

study (Romanova, 2010) has examined the impact of planning on

learning gains. The study will be discussed in the section on main
task options because it also examined two during-task options:
within-task planning and online feedback.
Ellis (2009b) synthesized the research on the effects of pre-task
planning on task performance in terms of the complexity, accuracy
and fluency (CAF) of learners’ speech production. He reports that the
research has shown consistent effects for planning on fluency and
complexity but its effects on accuracy have been variable and unstable.
Ortega (1999) also noted mixed findings regarding the influence of
pre-task planning on accuracy: ‘When the accuracy findings for all
planning studies are taken together, it is also difficult to see any
consistent patterns’ (p. 134). Crookes (1989) was one of the first
studies examining the effects of pre-task planning. In this study, forty
adult L1 Japanese English as a Second Language (ESL) learners were
divided into two groups and performed two monologic tasks. The
planning group was allowed ten minutes to plan and the other group
performed the tasks without planning. The study found significant
effects for complexity in terms of lexical variety and syntactic
complexity measured by words per utterance, subordination and S-
nodes per utterance, but no difference was found for general accuracy
measured through the number of error-free T-units.
Why does pre-task planning have stronger effects on fluency and
complexity than accuracy? The primary reason seems to be that
learners spend most of the planning time organizing the content,
rather than language, of their task performance, which eases the
burden on the conceptualizer and leads to more fluent and complex
language. Sangarun (2005) reported that regardless of whether
learners’ attention was directed towards content or language, their
planning primarily focused on meaning. During task performance,
learners must match the planned content with the relevant linguistic
items, leading to more complex language. The elevated complexity
is probably obtained at the expense of accuracy, confirming
Skehan’s LACH.
Duration of Planning. In most studies investigating pre-task
planning (e.g. Ortega, 1999; Fu and Li, 2017), learners are allowed
ten minutes to plan – a practice based on Mehnert’s (1998) study,
which is probably the first to examine the impact of length of planning
on learners’ performance. In this study, thirty-one L1 English and L2
German learners were divided into four groups and performed two
oral tasks. The four groups were allowed to plan for zero, one, five
and ten minutes respectively. When the data for both tasks were

combined, the results showed that, for fluency, all planning groups
outperformed the no-planning group, and the effects were incremen-
tal, that is, longer planning time led to higher levels of fluency. For
accuracy, planning had a significant effect compared with no plan-
ning, but the effects were non-linear in that the one-minute planning
group showed higher accuracy than the five-minute and ten-minute
groups. For complexity, the ten-minute planners outperformed all
other groups but were only significantly better than the five-minute
planners; the non-planners showed higher scores than the one-minute
and five-minute planners. Overall, the results showed that the length
of planning time had the largest effect on fluency, less on accuracy and
least on complexity, and that the effects were non-linear. However, the
sample size of the study is small, with seven to eight participants in
each group. Therefore, the generalizability of the findings is limited.
Li, Chen and Sun’s (2015) study was conducted in a testing condi-
tion – learners were informed that their test performance counted
towards their final grades for the course. The planning lengths they
investigated were zero, one and five minutes, and thirty seconds. For
fluency, the effects were incremental but plateaued after three minutes,
and thirty seconds did not make a difference. For accuracy, the
one-minute, three-minute and five-minute groups outperformed the
zero-minute and one-minute groups but after one minute, there
was not much difference. For syntactic complexity, the thirty-second,
two-minute and five-minute groups outperformed other groups.
Finally, for lexical complexity, the one-minute group did best. Overall
the results showed that planning below one minute did not seem to
work and that the effects were non-linear and tended to plateau after
one or three minutes.
In another study that was claimed by the researchers (Wigglesworth
and Elder, 2010) to be conducted under a testing condition, three
planning lengths were examined: zero, one and two minutes. The
researchers failed to find any benefits for planning in terms of either
objective CAF measures or subjective ratings. Their explanations for
the null effects for planning were: (1) learners couldn’t remember what
they planned, and even if they did, planning might have affected only
the first few utterances, not the whole task performance; (2) the task
performance was monologic and unpressured, which allowed learners
to plan online and so made pre-task planning superfluous. Although
this study was conducted using tasks similar to those in an Inter-
national English Language Testing System (IELTS) test, it is not a
‘pure’ testing study because the learners’ task performance did not
affect their course grades or career development.

Whereas the tasks in these studies were monologic and the participants
were adult learners, Philp, Oliver and Mackey (2006) examined inter-
active tasks where the learners were children. The study showed that no
planning and two-minute planning led to more CF than five minutes, no
planning resulted in longer utterances and five-minute planning
enhanced grammatical complexity. The researchers argued that planning
made learners familiar with the content and removed the need for nego-
tiation, hence less feedback and shorter utterances. They also exemplified
how too much planning caused trouble for classroom management with
children, who can become restless more easily than adults.
The research reviewed above suggests that allowing students to plan
for one to three minutes (maximum five minutes) seem ideal and
practical, and this recommendation is based on the following grounds.
First, one to three minutes of planning seem effective in facilitating
CAF performance, and after three minutes, task performance seems to
plateau on some measures. Second, although in many studies ten
minutes is the default time, longer planning time does not always lead
to better performance. In fact, the effects of planning duration are often
non-linear, with shorter planning times showing larger effects. Third,
in real classroom settings where teachers are often under time pressure,
it is not always feasible to allocate ten minutes for pre-task planning.
Fourth, for tasks involving interaction, planning increases learners’
familiarity with the content and reduces the need for negotiation – a
feature that is a main source of learning, according to the interaction-
ists (Pica, 1987; Long, 1996; Gass, 1997; see Chapter 2). Fifth, in child
language classes, too much planning causes management difficulty. We
would like to point out that that we do not wish to make it a hard rule
for teachers to follow. Teachers are task executers, and they are able to
make the best decisions regarding the length of planning based on the
nature, difficulty and goals of the task and other idiosyncratic con-
straints imposed by the local instructional setting. We also need to
acknowledge that the results of the planning studies were very mixed,
which makes it difficult to arrive at definite conclusions.
Finally, in the research on the influence of the amount of time for
planning as well as the research on planning in general, the duration of
planning is invariably imposed by researchers, and learners are never given
the freedom to decide how much time they need for planning. Also, in
existing research, duration of planning is operationalized as a categorical
variable. If learners are allowed to plan as long as they wish or need to, the
duration of planning can be analysed as a continuous variable, and a
correlation analysis can be conducted to ascertain whether longer planning
time is associated with better task performance.

Other Options. There are a number of other aspects of pre-task

planning of relevance to pedagogy and research. The first concerns
notetaking during planning. In most studies (e.g. Park, 2010; Li and
Fu, 2017), learners are allowed to make notes without writing complete
sentences to ensure that (1) they did plan, and (2) they will not provide a
scripted performance. However, doubts about the value of scripted
performance have not been empirically tested, and it would be interest-
ing to investigate the effects of different types of scripted planning, such
as discrete phrases, complete sentences and complete scripts, on task
performance. Second, in most planning studies or studies investigating
the impact of the procedural and conceptual aspects of tasks on CAF,
the tasks are purely output-based, and learners have no linguistic input
prior to task performance. They were either asked to watch a silent
video clip (Skehan et al., 2012) or view a picture with no accompanying
L2 input (e.g. Ortega, 1999). While this practice may increase the
internal study of the research, it deprives learners of the opportunity
to learn from input, which is ‘the sine qua non of acquisition’ (Gass and
Mackey, 2015, p. 181). Third, the language of planning is of particular
pedagogical interest. In Park’s (2010) study, learners were allowed to
plan either in their L1 or L2, whereas in Mehnert (1998), only the L2
was allowed. In Ortega (1999), learners listened to a taped version of
the narrative in their L1 before starting to plan to ‘avoid too much
individual variation in the story lines . . . and to reduce the cognitive
load of the narrative task’ (p. 122). Park’s and Ortega’s studies show
how learners’ L1 can be judiciously used in the planning stage, given
that such use has been criticized as a concern for implementing a task-
based approach (Carless, 2004). Fourth, the research on planning, as
well as the underlying theory (Skehan, 2014a; see also Chapter 3 of this
volume), has primarily focused on CAF, not the acquisition of new
linguistic knowledge. Therefore, the benefit of planning seems to be
restricted to the retrieval, proceduralization and automatization of
previous knowledge, which is critical to ultimate L2 attainment. How-
ever, the scope of the research should be expanded to include various
options of form-focused instruction (FFI) (Ellis, 2016). We will discuss
an example of such research in the section on within-task options later
in this chapter.
To conclude this section, a final comment is in order: no planning is
an option. While planning has been advocated and shown to be a
useful pedagogical technique that leads to improved performance, an
alternative position is warranted, that is, no planning also has a place
in task-based methodology. Ortega (1999) reported that in her study
twelve out of the thirty-two learners disliked the idea of planning
because: (1) the tasks were simple enough, which made planning

superﬂuous, (2) task performance is a matter of L2 proﬁciency and

providing extra time does not necessarily help, (3) their planning notes
were taken away before task performance (although this is a problem
with the research design, not with planning per se), and (4) they
thought that planning introduced more anxiety because then they
would be expected to be more linguistically correct. Furthermore, as
Robinson (2011) argued, it is important to increase task complexity
along resource-dispersing variables (relating to the procedural aspects
of tasks) such as by not allowing time to plan in order to encourage
spontaneous communication.

Pre-Task Focus on Form

Skehan and Foster (2001) argued that the meaning-oriented nature
of communicative tasks may distract learners from attending to the
linguistic code and that some form-focusing strategies are needed to
attract learners’ attention to linguistic forms during task performance.
In a similar vein, Kim (2013) pointed out that the purpose of including
form-focusing techniques in the pre-task stage is ‘to raise learners’
awareness of these forms during planning time as well as during task
performance’ (p. 10). It is important to distinguish two broad types of
form-focusing strategies: explicit and implicit. Explicit strategies
overtly draw learners’ attention to linguistic forms, such as through
grammar explanation or input enhancement (e.g. by highlighting
instances of a particular structure in the input). Implicit strategies
induce learners’ attention to forms indirectly by pre-task planning,
which increases the chances for attention to form, or by providing a
model performance demonstrating how to negotiate forms during
group interaction (Kim, 2013). It must be pointed out that the
explicit–implicit distinction is best seen as a continuum, not a dichot-
omy. In the following, we discuss two pre-task form-focusing options
that have been researched: pre-task grammar instruction and pre-task
modelling.
Pre-Task Grammar Instruction. Whether to provide explicit gram-
mar explanation in the pre-task stage is of theoretical and pedagogical
significance. In Long’s (2015, 2016) model of TBLT, pre-task gram-
mar instruction is not an option on the grounds that learners may not
be developmentally ready for the preselected structure. Long argued
that focus on grammar must happen reactively when learners experi-
ence difficulty with linguistic forms required for meaning-making.
Skill acquisition theory (DeKeyser, 2013) holds that learners must
have declarative knowledge, which can be learned through pre-task
grammar instruction, proceduralized in subsequent skill-specific task

performance and then automatized through repeated practice. Ellis

(2003, 2017a) calls this approach task-supported language instruc-
tion, which he distinguishes from the purely task-based approach
advocated by Long.
From a pedagogical perspective, those opposing pre-task grammar
instruction (e.g. Willis, 1996) worry that explicit grammar instruction
predisposes learners to focus on the target structure, with the result
that they may treat a task as language practice rather than a site for
information exchange. This practice, it is argued, will subvert the
meaning-primary principle of task-based teaching. Proponents of
pre-task grammar instruction contend that learners need grammar
knowledge to perform a communicative task and that teachers prefer
to teach grammar before asking students to perform a task (Little-
wood, 2007; Shehadeh, 2012). Teachers’ preference for pre-task
grammar instruction has been examined in research. For example,
van de Guchte et al. (2017) interviewed five experienced ‘task-based
language teachers’ (p. 2) in a Dutch secondary school about the
necessity of such instruction. They all indicated that they taught
grammar explicitly before asking learners to perform an oral or writ-
ten task. Their argument was that most other Dutch teachers taught
grammar in their classes and so that they wanted to ensure that their
students knew the same grammar rules as other students. However, to
date, there is a lack of research on students’ perceptions about TBLT.
What has research demonstrated about the influence of pre-task
grammar instruction on task performance and learning effects? To
date, Ellis, Li and Zhu (2018) is the only study that has investigated
this question. In their study, ninety 8th-grade Chinese English as a
Foreign Language (EFL) learners were divided into three groups. One
group received a ten-minute grammar lesson on how to use the English
passive voice before performing two similar dictogloss tasks. During
each task, the teacher presented a story on PowerPoint, then the
students worked in pairs to practise retelling the story, and finally
they took turns to report the story to the rest of the class. A second
group performed the two tasks without receiving pre-task grammar
explanation. The researchers measured students’ task performance
during the reporting stage through CAF. The results suggest that
pre-task grammar instruction had a harmful influence on learners’
task performance in terms of the CAF of their speech production.
Probably the grammar instruction caused them to focus their attention
on the target structure at the sacrifice of overall performance. Also,
although the grammar instruction led to more frequent use of the
target structure, it failed to make the learners more accurate in produ-
cing it. The results of this study suggest that it might be better to avoid

pre-task grammar instruction. However, this does not mean grammar

should not be addressed at some point in a task-based lesson. Both
Ellis (2009a) and Long (2016) acknowledge the importance of focus
on form in TBLT. The question is when and how the focus on form
should happen.
Pre-Task Modelling. An alternative to grammar instruction in the
pre-task stage is modelling. This involves techniques that serve to: (1)
clarify task procedures, (2) demonstrate expected group dynamics (e.g.
how to negotiate with other group members), (3) provide linguistic
and pragmatic support, and (4) afford learning opportunities by
focusing on a particular linguistic structure. Modelling has many
variants, and it is important to attend to the details of the modelled
performance because they can influence how learners perform
the task.
Kim (2013) reported a study where Korean middle school EFL
learners participated in a pull-out programme designed to enhance
their oral proficiency. The learners were divided into two groups. One
group viewed a task modelling video before performing the task and
the other did not. Both groups attended three 45-minute treatment
sessions, and during each session they performed a focused task
targeting English question formation. The modelling group watched
a two-minute video clip showing how to deal with linguistic problems
relating to question formation, such as by providing feedback to each
other. For example,
A: How many types of pets?
B: No, you should use an auxiliary verb first (in students’ L1 – Korean)
A: How many types of pets do they have?
The modelling performance was provided by the researcher and the
learners’ own English teacher. The scripts were written by the
researcher. The instructional effects were measured by examining
LREs, defined as utterances where learners ‘talk about the language
they are producing, question their language use, or correct themselves
or others’ (Swain and Lapkin, 1998, p. 326) during planning and
task performance, and also by means of pre- and post-tests. L2 devel-
opment was operationalized as the number of learners who produced
questions at a higher developmental stage in the immediate and
delayed post-tests. The results showed that students in the modelling
group produced significantly more LREs than the control group and
the number of students who advanced to higher stages of question
formation was significantly greater than the control group.
Kim’s (2013) study is insightful for a number of reasons. First, it
shows how to draw learners’ attention to linguistic forms through

modelling. Used properly, modelling can serve as an effective alterna-

tive to grammar instruction, which, as we noted, may affect learners’
task performance. However, the study did not examine CAF, so it is
not clear whether modelling, like grammar instruction, has a negative
impact on learners’ task performance. Second, in Kim’s study, in the
video two teachers modelled the task that was to be carried out by the
students but this in effect removed the ‘gap’ in the task. It might have
been better if the teachers had modelled a similar task rather than the
same task. Third, in the modelling video, the teachers used the stu-
dents’ L1 to comment on each other’s language use, which led to the
students making extensive use of their L1 in the LREs during the pair
work. This study shows that modelling has a powerful influence on
how students perform a task.
Another study that provides valuable information on modelling is
van de Guchte et al. (2017). This study examined the influence of
language-focused and meaning-focused modelling on the use of the
target structure as well as on overall complexity, as represented by the
learners’ use of coordination and subordination. Forty-eight 9th-grade
L2 German learners at a Dutch school were divided into two groups,
both watching two videos of two girls describing a school cafeteria to
persuade potential students to enrol in the school. The target structure
was locative prepositions. While watching, the language-focused
group was asked to write twelve sentences that the actresses used to
describe the locations of the objects in the cafeteria, while the
meaning-oriented group was asked to answer eight questions about
how the presenters in the video persuaded prospective students to
enrol in the advertised school. After watching the videos, the students
were allowed to plan for ten minutes, writing down ten keywords,
which were later taken away. Then the students were asked to describe
a school cafeteria shown in a picture, which served as the post-test.
A similar task was completed before the treatment as the pre-test and
three weeks later as the delayed post-test. The results revealed an
overall trade-off between the use of the target structure and overall
task performance. The language-focused group attempted more fre-
quent and accurate use of the target structure but the meaning-focused
group showed greater linguistic complexity. The study did not meas-
ure overall accuracy or fluency so it is not clear what influence the two
different types of modelling had on those two aspects of overall task
performance.
The way the modelling was provided in this study has implications
for pedagogy. First, in contrast to the modelling in Kim (2013), which
was conducted by two teachers, in this study the modelling was
provided by two actresses whose ages were similar to the students’.

The researchers stated that peer (instead of teacher) modelling videos

were used ‘to increase the chance of identification, confidence, and
motivation’ (p. 7). The authors also pointed out that students in the
current cohort could be used to model performances for students in
future semesters. Second, this study shows the importance of carefully
designing and conducting the model performance. Modelling may
influence learners’ use of language, whether they orientate to meaning
or form when they perform the task, and the affective aspects of task
performance such as their confidence and motivation.

Other Pre-Task Activities

Other activities that might be included in the pre-task stage include
developing learners’ schematic knowledge of the task topic, providing
scaffolding and teaching vocabulary (Ellis, 2003). The purpose of
developing schematic knowledge is to increase learners’ familiarity with
the content domain of the task. This aspect of task familiarity – namely,
topic familiarity – is distinguishable from familiarity with the procedure
of a task. Topic familiarity frees up learners’ attentional resources
allowing them to allocate more resources to the formal aspects of a
task. In addition to the potential benefits in enhancing task perform-
ance, task familiarity may have a positive impact on motivation. In
research on reading, topic familiarity has been found to facilitate
vocabulary learning (e.g. Pulido, 2007) but it has not received much
attention in research on oral production tasks. Skehan (2009a) noted
that a task with familiar content (as in personal narratives) led to greater
fluency and accuracy but less complexity than a task with unfamiliar
content (as in a narrative and an argumentative task), suggesting a
positive effect for tasks with less familiar content. Therefore, while
teachers need to ensure that learners have sufficient background infor-
mation, there is also a case for tasks with less familiar content.
Teacher scaffolding refers to the level of assistance teachers provide
to enable a task to be completed successfully. Scaffolding is a concept
of the sociocultural theory (Lantolf, Thorne and Poehner, 2015; also
see Chapter 4). The type and amount of assistance is not predeter-
mined but needs to be determined in flight as the task unfolds. One
such example can be found in the Bangalore Project (Prabhu, 1987)
where the teacher performed a pre-task with the students by means of
question/answer before asking students to perform the main task
independently. This is not the same as modelling as the teacher does
not provide a model performance. Rather the teacher aims to provide
‘other-regulation’ to enable the development of ‘self-regulation’,
namely the ability of the learner to perform a task without assistance.

According to Willis and Willis (2007), learners should receive assist-

ance with the vocabulary required for task completion in the pre-task
stage. However, as Ellis (2003) warned, pre-teaching vocabulary
my result in learners treating the task as a practice activity for the
pre-selected words. Willis and Willis ﬂoat the idea of ‘linguistic
mining’ as an alternative to direct teaching. For example, the teacher
could give students a questionnaire on the topic of the task prior to
task performance and leave it to the students to take the initiative to
understand the meanings of key lexical items by consulting their peers
or a dictionary. Students can also draw up a list of words that they
deem important for performing the task individually and then com-
pare lists, resulting in a list of items that are of central importance for
performing the task. But teachers should avoid requiring students to
use speciﬁed words when performing the task to prevent it becoming a
vocabulary exercise.

Options for the Main Task

In this section, we examine different types of within-task focus on
form (see Table 8.2), particularly CF, which has been the focus of a
large amount of research since the late 1990s (see Chapter 2). We
conclude this section by exploring other considerations that may affect
the way learners perform a task and the resulting learning outcomes.

Within-Task Focus on Form

Within-task focus on form refers to attention to linguistic problems
while the task is ongoing. As explained in Chapter 2, it can be pre-
emptive or reactive (Ellis, Basturkmen and Loewen, 2002). In pre-
emptive focus on form, the teacher draws the learner’s attention to
form in anticipation of a linguistic problem or the learner makes a
language-related enquiry to the teacher or a peer. In reactive focus on
form, attempts to address linguistic forms are made in response to
errors learners produce in their task performance. Reactive focus on
form has been referred to as CF and is viewed by Long (2015) as the
primary way of doing focus on form. Within reactive focus on form,
Ellis et al. (2002) made a further distinction between conversational
and didactic focus on form, with the former referring to situations
where linguistic errors are dealt with to resolve communication break-
downs and the latter to situations where errors do not cause communi-
cation breakdowns and where language is treated as an object.
A related distinction is that between planned and incidental focus on
form. In planned focus on form, the teacher makes an a priori decision

Table 8.2 Summary of within-task methodological options

Options Description Recommendations

Within-task CF: provide feedback on • Use mixed feedback
focus on form errors arising during task • Start with a prompt and
performance provide a recast in the
absence of self-correction
Integrated focus on form • Interrupt a task to address
linguistic forms
Other options Within-task planning: no • Use within-task planning
time constraint is imposed with beginners
on task performance
Access to support • Provide access to linguistic
input such as word cues
• Do not provide access to
non-linguistic input such as
pictures
A surprise element • Introduce a new element to
increase amount of
production and motivation
Interim goals • Add interim goals to
increase task structure and
task accountability

to target a linguistic structure consistently; in incidental focus on form,

the teacher addresses linguistic forms in a contingent manner and it is
not planned ahead of time.
Another distinction, which has been the focus of the research by
Spada et al. (2014), is between integrated and isolated focus on form.
According to them, integrated focus on form is ‘provided within
communicative/content-based activities’, whereas isolated focus on
form is ‘provided separately from communicative/content-based activ-
ities’ (p. 457) and thus corresponds to the provision of explicit
instruction in the pre-task stage of a lesson. Thus the difference
between integrated and isolated focus on form is that the former is
embedded in communication and the latter occurs in separate blocks
of grammar instruction. Integrated focus on form, they further
explained, includes both CF and rule explanations during brief time-
outs while learners are performing a task. Thus, it would seem that
such form-focusing techniques include both online feedback which
does not interrupt the communication ﬂow as well as rule explan-
ations that temporarily put the ongoing communicative event on hold.

Within-task focus on form constitutes integrated focus on form in

Spada et al.’s terms.
Among the various options of within-task focus on form, CF has
received the most attention in research, because of its importance in
second language acquisition (SLA) and language pedagogy (Ellis and
Shintani, 2014). CF refers to responses to learners’ production errors
or comprehension problems that occurred during a task. Lyster and
Ranta’s (1997) seminal study identified six types of CF: recasts,
repetition, elicitation, clarification request, metalinguistic clue and
explicit correction. For example, for the error in the utterance ‘There
are many big livers in Africa’, where ‘rivers’ is mispronounced as
‘livers’, the teacher can respond by:
1) Using a recast – reformulating the wrong sentence: ‘There are many
big rivers in Africa.’
2) Repeating the error: ‘Liver?’
3) Eliciting the correct form from the learner: ‘There are many big . . .’
4) Making a clarification request: ‘Sorry?’
5) Providing a metalinguistic comment: ‘There is something wrong
with your pronunciation of “liver”.’
6) Alerting the learner to the presence of an error and supplying the
correct form: ‘No, not “liver”, “river”.’
These six CF types can be further divided into explicit and implicit
depending on whether learners’ attention is drawn to errors. Recasts,
for example, have been examined as an implicit CF type, and metalinguis-
tic clues as explicit feedback (e.g. Ellis, Loewen and Erlam, 2006). How-
ever, what determines the level of explicitness of CF is not just the type
but the context in which the CF occurs. The CF types can also be classified
as output-prompting (clarification, elicitation, repetition and metalinguis-
tic clue) versus input-providing (recasts and explicit correction) depending
on whether they encourage learner repairs (Lyster, 2004). These different
types of CF are discussed in greater detail in Chapter 2.
In terms of the participatory structure of CF in a classroom setting,
CF falls into two broad categories: peer CF and teacher CF. Peer CF
can be incidental, in which case learners respond to each other’s errors
in an ad hoc way during group work (Adams, Nuevo and Egi, 2011),
or planned, in which case errors are responded to consistently in a
controlled manner. Sato and Lyster’s study (2012) exemplified how
planned CF was provided. In their study, learners worked in groups of
three: a storyteller who told a narrative incorporating a variety of
errors, an error detector who had to provide CF (recasts or prompts)
on the errors, and an observer who recorded whether the errors were
corrected and reported the observations at the end.

With regard to teacher CF, the literature indicates three ways the
teacher may respond to students’ errors: during teacher–student inter-
action, during teacher–class interaction or during student–student
interaction. Teacher–student interaction can be further divided into
one-way and two-way interactions. In one-way teacher–student inter-
action (Sheen, 2010; Yang and Lyster, 2010; Li, Ellis and Zhu, 2016),
the teacher provides CF during the reporting stage of a communicative
task after students have completed group work. Two-way interaction
typically happens in a question/answer format where the teacher
interacts with a single student and provides CF on his/her errors (Yang
and Lyster, 2010; Goo, 2012). Teacher–class interaction occurs in
situations where the teacher performs the task together with the whole
class. For example, in one of the treatment activities in Lee and Lyster
(2016), the teacher made a series of commands, all students reacted,
and the teacher provided CF when students made mistakes in under-
standing his commands. Finally, CF can also be provided while
students work in groups as in van de Guchte et al. (2015), where the
teacher circulated the class, interrupted students and provided CF
while they were performing communicative tasks.
Next, we would like to highlight two points that teachers should heed
based on CF research. First and foremost, teachers should not hesitate
to provide CF, given clear evidence of its facilitative effects on L2
development (Ellis, 2010; Li, 2010; Lyster and Saito, 2010; Nassaji,
2016). Task-based teachers should recognize that CF is an ideal alter-
native to pre-task grammar instruction, which, as we have seen, can
affect learners’ overall task performance and makes learners treat the
task as a grammar exercise. Also, CF addresses linguistic forms during
communication and thus may facilitate the development of communi-
cative competence (Spada et al., 2014). It motivates the learner to attend
to the input while he/she is struggling to ﬁnd the correct form to convey
meaning (Long, 2015). Thus, linguistic knowledge acquired via CF is
proceduralized through immediate application in subsequent perform-
ance during the ongoing task (Li et al., 2016). One caveat is that
teachers may worry about the harmful effects that CF has on students’
motivation. However, Li’s (2017) meta-analysis of studies of teachers’
and students’ attitudes towards CF indicates that students are over-
whelmingly positive about CF (with an 89% endorsement rate for the
importance of CF), although teachers are much more hesitant (with a
mere 39% agreement rate). Furthermore, Zhang and Rahimi (2015)
found that students’ preference for CF was not related to their anxiety
level. Therefore, the detrimental impact that CF can have on the affect-
ive aspects of learning is perhaps less acute than teacher guides like to
suggest (see Ellis and Shintani, 2014).

Second, regarding which type of CF should be provided, we recom-

mend using a variety of types instead of a single type. The research has
shown an overall advantage for explicit feedback over implicit
feedback although the difference may disappear over time (Li, 2010).
Prompts have been found to be more effective than recasts in class-
room settings (Lyster and Saito, 2010). The comparative effects of
prompts vs. recasts also led to a debate between Lyster and Ranta
(2013) and Mackey and Goo (2013). Prompts push the learner to
reﬂect on his/her own language use, involve deeper cognitive process-
ing, and help learners become better at consolidating and automatiz-
ing previously learned linguistic knowledge. Recasts are generally
non-intrusive, provide both positive and negative evidence, and may
be especially useful for learning new linguistic structures. However,
the relative effects of different types of CF are context-dependent.
Recasts, for example, can work very well in classroom contexts where
students are oriented to attend to form. Instead of choosing one type
over the others, perhaps the best way is to provide mixed CF, for
example starting with a prompt to elicit self-correction followed by a
recast if the learner fails to self-correct (Li et al., 2016). A second
option is to start with recasts to establish an initial knowledge base
and then follow up with prompts to consolidate the new knowledge.
Similarly, the teacher may start with a few instances of explicit
feedback to direct students’ attention to the linguistic target but then
switch to implicit feedback which is less disruptive (Yilmaz, 2013a).
Teachers need to experiment with CF to ﬁnd out what works best for
them in their own instructional context.

Other Within-Task Options

Another within-task option is whether to allow learners to carry out
planning while they are performing a task. In a within-task planning
condition, learners are not given a time limit and so can perform a task
without time pressure. However, this feature has been operationalized
differently in the research. In Ahmadian and Tavakoli (2011) and Li
and Fu (2017), learners watched a video and were then given an
unlimited amount of time to complete the narratives. They were
encouraged to plan the content and language while performing the
tasks. In Wang’s study (reported in Skehan et al., 2012), however,
learners watched a slowed-down version of a video and told the
narrative simultaneously. The methodological differences between
these studies may explain their different ﬁndings. While Ahmadian
and Tavakoli (2011) and Li and Fu (2017) found an advantage for
within-task planning in enhancing accuracy and complexity at the

expense of ﬂuency compared with both no planning and strategic

planning, Wang failed to find any difference between within-task
planning and no planning.
Both of these studies examined CAF, not learning. Romanova
(2010), however, conducted a study examining the effects of
within-task planning on learning gains in comparison with pre-task
planning and no planning. L2 Russian (L1 English) learners per-
formed a picture description task where they told a narrative based
on pictures displayed on a computer screen. The within-task plan-
ners performed the task at their own pace but had no time to plan
before starting the narrative. The pre-task planners were allowed
five minutes to see the pictures beforehand but they performed the
task under time pressure, while the non-planners were not allowed
to plan either before or during task performance. All groups
received recasts on their third-person singular errors while perform-
ing the task. Both online and pre-task planning were found to be
more effective than no planning when recasts were provided, but
online planning was more effective than pre-task planning in con-
tributing to the effects of recasts and resulted in more noticing and
more modified output. This was because in the pre-task planning
condition ‘the unavailability of online planning time during inter-
action may inhibit acquisitional processes due to the limitations of
WM [working memory], just like in the NP [no-planning] condition’
(p. 865). To sum up, it would seem that within-task planning may
assist with the formal aspects of task performance and also have a
positive effect on learning. This might be especially true for begin-
ner learners who may need more time to plan and monitor their
speech performance.
A second possibility concerns the degree of support while per-
forming the task. This can take the form of planned notes, input
materials (images, videos, texts, etc.) or reference tools such as
dictionaries. In research on the effects of pre-task planning, typically
learners are not allowed to refer to their planned notes or any
reference tools during task performance, which Ortega (1999) con-
sidered undesirable from a pedagogical perspective. A related idea is
allowing learners to borrow from the input materials of the task as
it is performed, which Ellis (2003) suggests should be encouraged.
In Robinson (2007c), learners were required to complete three oral
tasks that differed in complexity. Learners had access to a list of key
phrases while performing the tasks. Robinson found that when
performing the more complex tasks, the learners referred to the
word list more frequently. He considered access to notes as a useful
way of helping learners perform more complex tasks, echoing

Ellis’ the idea about borrowing from input information. However,

asking learners to refer to task prompts such as photos (Révész,
2009) or videos (Skehan et al., 2012) increases learners’ processing
load and can affect task performance. Therefore, it would seem
that while access to linguistic input may afford learning oppor-
tunities, access to non-linguistic input such as picture cues may
have a negative effect. To avoid increasing learners’ processing
load, perhaps learners should be given the freedom to decide
whether they want to refer to available input or supporting
materials.
Other within-task options mentioned in TBLT guides (Ellis, 2003;
Willis and Willis, 2007) include using a clear task format (tables,
diagrams, etc.), introducing a surprise element or new information
half way through the task, and setting clear interim goals (that is,
making the outcome of each stage of a task clear and speciﬁc).
Although these options have not been subjected to empirical research,
they are pedagogically valuable in that they help learners to engage
with a task, lead to more output and so on.

Post-Task Options
Post-task activities are follow-up activities that build on the main task
(Table 8.3). The purpose is to provide learning opportunities by:
(1) asking learners to repeat a task, (2) addressing linguistic forms
that had been shown to be problematic for the learners in the main
task , and (3) engaging learners in reflective activities. The first option
encompasses different forms of task repetition, which has received
much attention in the research. The second option concerns the vari-
ous techniques teachers may draw on to address linguistic forms
explicitly. The third category involves activities encouraging learners
to reflect retrospectively and introspectively on their task performance.
In the following, we elaborate on each of these three broad categories
of post-task activities.

Task Repetition
A task can be repeated in different ways and, depending on whether
the repetition involves the content of the task, the task procedure or
both, three types of task repetition can be identiﬁed: exact repetition,
procedural repetition and content repetition. In exact repetition both
the content and procedure are repeated. This type of repetition has
been referred to as ‘task repetition’ in the literature (e.g. Kim and
Tracy-Ventura, 2013) but in order to distinguish task repetition as a

Table 8.3 Summary of post-task methodological options

Options Description Recommendations

Task repetition Learners are asked to repeat • Procedural repetition is
a task. Three types of ideal
repetition are possible: • May mix procedural and
exact repetition, content repetition
procedural repetition and • Feedback may be provided
content repetition. between repeated tasks
Explicit focus Post-task feedback • May ask learners to
on forms perform another task after
providing feedback on their
task performance
Providing a model • Learners may listen to the
audio recording of a task
• Learners may read a script
of a model performance
Other options • Learners may engage in
consciousness-raising (CR)
activities where they
extrapolate rules based on
given materials
• Learners may perform input
processing activities that
force learners to attend to
linguistic forms to process
meaning
• Learners may be asked to
‘grammatize’ a gapped text
where some linguistic
features (e.g. plural) are
missing
Reflection Reflective accounts • Ask learners to reflect on
various aspects of their task
performance
Transcription • Ask learners to transcribe
their task performance
• Integrate transcription and
other post-task options
such as teacher feedback

concept and task repetition as a technique, we refer to the latter as

exact repetition. As an example of exact repetition, Ahmadian and
Tavakoli (2011) asked learners to perform a narrative after watching a
ﬁfteen-minute silent video and then repeated the task one week later.
Another example is Gass et al. (1999), where the exact repetition
group (called content repetition by the authors) watched the same
Mr Bean video episode three times and performed a narrative after
each viewing.
In procedural repetition, learners follow the same steps each
time they perform the task but the content is different for each
performance. For example, in Gass et al. (1999), the procedural
repetition group (which the authors called ‘different content group’)
watched four different Mr Bean episodes about different topics. In
Patanasorn (2010), in the procedural repetition condition, pairs
of learners completed three decision-making tasks where they
followed similar steps but the topics for each task were different. For
example, in one task, learners exchanged information about two
suspects for a crime and discussed who was the perpetrator; in
another, they decided who was the greatest footballer based on infor-
mation about two candidates. In Kim and Tracy-Ventura (2013),
Korean EFL learners performed three information exchange tasks
following similar procedures but each was on a different topic: hosting
an American friend, describing school events/activities and discussing
mayoral candidates.
In content repetition, the repetition involves the same content but
the procedure is different. In Patanasorn (2010), the content repetition
group performed three different tasks involving the same information.
In the ﬁrst task, called story completion, they worked together to
complete a narrative about a prime ministerial candidate. In the
second task, which is called information exchange, they worked with
a new partner to tell their narratives. Finally they performed a
decision-making task with another new partner, arguing why their
own candidates should be elected. In this study, content repetition
happened through recycling the same information in a different format
with a different partner/interlocutor. Another procedural variation
involves setting different time limits for each repetition, as in Thai
and Boers (2015), where learners performed the same narrative task
three times, within three, two and one minutes respectively, each time
with a different listener. Still another option is to repeat the task in a
different modality. For example, after performing an oral debate task,
learners could be asked to summarize the main points of the debate in
writing. Similarly, after a listening task, learners could be asked to
read a passage on the same content/topic.

Several other methodological aspects of task repetition are worth

mentioning. One is the number of repetitions. In some studies (e.g.
Ahmadian and Tavakoli, 2011; Skehan et al., 2012), a task is repeated
only once and the second performance is treated as the outcome of
task repetition. In Kim and Tracy-Ventura’s (2013) study, in the
identical repetition condition, the learners performed the same task
three times. A second methodological variable is the interval of
repetition. In addition to immediate repetition (e.g. Thai and Boers,
2016), other intervals reported in the literature include one day (Kim
and Tracy-Ventura, 2013), two to three days (Gass et al., 1999), one
week (Ahmadian and Tavakoli, 2011) and fortnightly (Bygate, 2001).
A third aspect concerns whether task repetition happens by learners
performing the task individually (Ahmadian and Tavakoli, 2011) or in
pairs/groups (Kim and Tracy-Ventura, 2013).
An essential question is whether task repetition has any benefits for
L2 learning. First, with regard to identical repetition, the studies
involving monologic tasks show that it enhanced CAF on the same
task (Gass et al., 1999; Bygate, 2001; Ahmadian and Tavakoli, 2011;
Skehan et al., 2012). However, the effects tend not to transfer to new
tasks (Gass et al., 1999; Bygate, 2001). The studies involving group
interaction produced mixed results. Patanasorn (2010) found that
identical repetition did not have any effect while Kim and Tracy-
Ventura (2013) reported that identical repetition led to improved
accuracy in learners’ use of the English past tense but that this group’s
speech was less complex and fluent on the post-tests compared with
the pre-tests. Second, procedural repetition led to an improvement
in ratings of a performance of a new task (Gass et al., 1999), improved
accuracy in the English past tense (Patanasorn, 2010; Kim and
Tracy-Ventura, 2013) and improved linguistic complexity (Kim and
Tracy-Ventura, 2013). Third, content repetition led to improved
global fluency but reduced accuracy in the past tense (Patanasorn,
2010).
Based on the available evidence and bearing in mind that more
research is needed on the comparative effectiveness of the three
types of repetition, we suggest that exact repetition, which requires
learners to repeat the same task multiple times, is not the ideal option
in task-based teaching. Exact repetition not only results in unstable
effects on task performance or learning outcomes but it may also be
perceived negatively by learners (Plough and Gass, 1993; Kim and
Tracy-Ventura, 2013). Kim and Tracy-Ventura (2013) also reported
that exact repetition led to fewer LREs than procedural repetition. The
researchers suspected that this might have been because learners were
already familiar with the content, which removed the need for them to

negotiate for meaning. In contrast, procedural repetition, where

learners follow similar steps but work with different information,
has proven to be more useful and motivating. It has the potential to
provide new learning opportunities in the case of input-based tasks
when learners are exposed to more varied linguistic structures and
vocabulary. One possibility is to mix content and procedural
repetition by retaining part of the content and the whole or part of
the procedure. Finally, one option that is largely missing from the
research literature on task repetition is providing some kind of inter-
vention between the ﬁrst and second performance of a task – for
example, providing feedback on the ﬁrst performance.

Explicit Focus on Forms

Earlier in the chapter we argued that pre-task grammar instruction can
affect learners’ overall task performance by causing them to focus on
the linguistic target, with the result that the taskness of the task is
subverted. One way to address the limitation of pre-task grammar
instruction is to delay it to the post-task stage, as Willis and Wills
(2007) and Skehan (1998) have consistently advocated. Long (2016)
also argued that an explicit focus on forms is a useful option in the
post-task stage of a lesson because it is reactive (i.e. it addresses
attested problems with form). Post-task focus on form caters to learn-
ing by providing opportunities for learners to practise producing
structures they found difﬁcult when they performed the task. Teachers
can use the input materials to exemplify the usage of the linguistic
target and design various form-focused activities based on the mater-
ials. Also, addressing linguistic forms after learners have struggled to
use the appropriate forms to convey meaning enables them to see the
need to learn and is therefore more motivating than pre-task grammar
instruction. However, to date there has been little research on the
effects of post-task focus on form. In the following, we discuss the
options and beneﬁts of post-task feedback and then address some
other types of form-focused techniques.
Post-Task Feedback. One way to provide CF is to hold a plenary
feedback session addressing the typical errors that the teacher
observed the students making while they were performing the task.
Li et al. (2016) conducted a study on the relative effects of immediate
and delayed feedback. In the immediate feedback condition, learners
received feedback on English passive voice errors while they were
telling a story to the class, whereas in the delayed feedback condition
they received feedback on their errors after the task had been com-
pleted. When providing post-task feedback, the teacher quoted an

error she had noted and asked the learner to self-correct: ‘Josh, you
said “three people killed”. Can you say it again?’ If the learner failed
to self-correct, the teacher then provided a recast: ‘Three people were
killed.’ The researchers found that delayed feedback was effective but
not as effective as immediate/within-task feedback in facilitating the
learning of the target structure. They suspected that this was because,
unlike the learners who received immediate correction during task
performance, those who received delayed feedback did not have an
opportunity to apply the learned knowledge during the performance
of the task. The implication is that the effects of delayed feedback
might be boosted if learners were asked to perform another task after
the feedback session. However, this option needs to be empirically
examined.
Providing a model. Lynch (2009) proposed extending the scope of
feedback to include sample performances of the same tasks by more
competent speakers or native speakers so that learners can self-
correct by noticing the gap between their own performance and
the correct models. This can happen by asking learners to listen to
the audio recording of the task and/or read the script of the per-
formance. Lynch surveyed sixty international postgraduate students
at a British university and asked them to comment on the post-task
options that they thought would help them notice their errors in
their role plays. The respondents preferred to receive sample per-
formances (recordings or transcripts) after, rather than before, the
role plays. Encouraging self-correction by providing a model after
performance has several beneﬁts. First, it encourages a comparison
of the correct model with the learners’ own utterances, provides
opportunities for self-repair and fosters learner autonomy. Second,
the focus is only on learners noticing their errors rather than
mastering structure productively, which they may not be develop-
mentally ready for. Third, it can prevent learners feeling embar-
rassed as a result of being corrected by the teacher in front of other
students. However, post-task modelling also has limitations. For
example, some errors may be ignored because they are too compli-
cated and/or because learners just focus on meaning. Also error
correction happens in a haphazard rather than systematic way and
is unfocused.
Aside from CF, the teacher can also provide positive feedback. For
example, the teacher may note down useful words or advanced struc-
tures learners used during the task (Willis and Willis, 2007) and build
on that during the post-task stage by eliciting more examples.
Other Form-Focused Strategies. We do not intend to provide a full
account of the strategies for explicit FFI (for a detailed account of

various aspects of explicit instruction, see Ellis and Shintani, 2014).

Rather, we will highlight several input-based techniques. One such
strategy is called consciousness raising (CR), which Ellis (1997)
defined as ‘a pedagogic activity where the learners are provided with
L2 data in some form and required to perform some operation on it,
the purpose of which is to arrive at an explicit understanding of some
linguistic properties of the target language’ (p. 160). Essentially CR is
an inductive approach that caters to the learning of explicit linguistic
knowledge. In a typical CR activity, learners are provided with input
materials, either aural or written, from which they extrapolate linguis-
tic regularities relating to grammatical features that they failed to use
or used incorrectly when they performed the task. Ellis (2003) argued
that CR activities meet the criteria for tasks because they provide
opportunities for authentic communication as learners work together
to work out the rule for a grammatical feature. In other words, CR
tasks make grammar a topic to be talked about. In this sense, CR tasks
do not need to be limited to the post-task stage of a lesson. They can
figure as main tasks but they have a clearer role in the post-task stage
of a lesson.
Another strategy is called processing instruction, which is based on
VanPatten’s (2015) input-processing theory. A primary principle of
the theory is that learners tend to prioritize meaning over form when
processing linguistic input and therefore teachers need to design
activities that require learners to conduct form-meaning mapping.
Asking learners to fill in past forms in a text, for example, does not
require form-meaning mapping because learners can simply add ‘-ed’
to verbs without having to understand the connection between the
past forms and the meaning they encode. An alternative is to ask
learners to listen to a text with a mix of tenses such as the present, the
future and the past and identify just those events that happened in the
past. Similarly, an activity that provides clues that enable learners to
decode the meaning without processing the form obviates the need
for form-meaning mapping. For example, imagine an activity aimed
at helping the learner understand the English causative constructions
(i.e. ‘have somebody do something’, ‘have something done’, etc.)
where a learner is given the sentence, ‘Jill had her husband washthe
dishes.’ The sentence is accompanied with a photo showing a man
cleaning dishes. In this scenario, the learner could use the photo as a
clue to answer the question without having to process the target
structure. The alternative is to remove the photo so that the learner
has no clues to rely on to understand the meaning of the linguistic
structure.

The third strategy, consistently advocated by Willis and Willis

(2007), is called ‘grammatizing’, where learners are required to restore
the linguistic items or features that have been left out in a text. This is
an ideal strategy for integrating focus on form into the input materials
of a task. For example, the teacher could show a sentence: ‘He some-
times asked me to wash his car or cut the grass, but I was never forced
to do it.’ Then some words are erased: ‘He__ asked me__ wash his car
or cut __ grass, but I was never __ to do__.’ The number of words
deleted can be varied, making the activity more demanding – for
example, ‘He__ __ me __ wash his __ or ___ __ grass, but I ___ never
___ to do ___.’

Reflection
Learners can be asked to reflect on various aspects of the completed
task including the task itself and/or their own or their peers’ perform-
ance. We divide reflection activities into two types: reflective accounts
and transcription. Learners’ reflective accounts refer to self-reports
about what they think they learned during the task, their evaluation
of their task performance, their perceptions of the design features of
the task including its objective, nature and difficulty, their attitudes
towards the task and their opinions about how to improve it. This
involves asking learners to complete a simple questionnaire.
Transcription involves learners providing a transcript of their own
or a peer’s performance. Transcription can be used in various ways
and in conjunction with other post-task strategies such as post-task
modelling, CF, etc. For example, the learner could transcribe his/her
own performance, edit or reformulate the transcript, and/or compare
the transcript with the script of a sample performance. These activities
can be undertaken independently or in collaboration with peers and
can be integrated with teacher feedback. However, to date there has
been little research on the effects of different types of transcription on
learning outcomes.

Participatory Structure
Discussions of task-based teaching are often based on the assumption
that a task must be performed in pairs or small groups, but this is a
mistaken notion – there are different types of participatory structure.
Participatory structure refers to the way interaction occurs in a task,
which can be divided into four types: individual, student–student,
teacher–class and student–class (see Table 8.4) (Ellis, 2003). In an

Table 8.4 Participatory structure of task-based interaction

Prototypical form of
Participatory structure interaction Pattern of interaction
Individual Intrapersonal Individual student
Social Interpersonal Student–student
Student–class
Teacher–class

individual participatory structure, students work on a task or part of a

task independently. Student–student interaction occurs when students
engage in group or pair work. Student–class interaction occurs when
one student or a group perform a task with the rest of class. Teacher–
class interaction occurs when the teacher interacts with the whole
class. The following jigsaw task can be performed in different ways
involving any of the four different structures.
In this task, students are given ﬁve jumbled pictures depicting a
story. The task can be performed individually by asking students to
sequence the pictures independently before telling the story in oral or
written form. The task can also be performed in group work, where
each member is given one picture and describes what happens in his/
her picture; they then work together to arrange the pictures in the right
order. The task can also be performed via the student–class structure,
where one student is asked to describe the pictures in the right order to
the rest of the class, who arrange the pictures based on that student’s
narrative. Similarly, the task can be performed between the teacher
and the whole class, in which case the teacher tells the narrative and
the whole class sequence the pictures accordingly.

Conclusion
In this chapter, we have discussed various methodological options
available to teachers in the three phases of a task-based lesson. For
the pre-task stage, we reviewed the literature on pre-task planning,
pointing out that while this is a necessary and useful step, longer
planning time does not necessarily result in better task performance.
Based on the ﬁndings of the research and taking into consideration
the constraints imposed by local instructional settings, we recom-
mended three minutes as an optimal duration, which can be
shortened or extended depending on teachers’ own judgement. We
also recommended exploring the possibility of allowing learners the

freedom to decide on how much time they need to prepare for the
main task. With regard to pre-task grammar instruction, drawing on
empirical evidence and theoretical and pedagogical perspectives, we
argued against such practice. Teachers may experiment with alterna-
tive form-focused strategies such as pre-task modelling. Another
valuable pre-task strategy is to work with students and scaffold their
performance of a task before asking them to perform it (or a similar
task) independently.
For main task options, we considered ways in which teachers and
learners can focus on form as a task is being performed. We argued
that CF has proven to be an effective instructional device. However,
instead of a single feedback type, we proposed teachers use a variety of
corrective strategies, for example a prompt followed by a recast or
alternating between explicit and implicit feedback.
For the post-task stage, we started by identifying three types of task
repetition: exact repetition, procedural repetition and content
repetition. We went on to show, based on research, that exact repetition
is not an ideal strategy and that procedural repetition has proven to be
more useful and effective. We then explored different post-task options
such as modelling and various types of explicit FFI. Finally, we
explained different ways in which learners can be encouraged to reflect
on their performance of a task, including the use of transcriptions of the
students’ performance of the task.
Regarding participatory structure, we clarified that a task can be
performed in different ways involving different types of interaction
and that it is a misconception to consider group work the only way.
An example task-based lesson is provided in an appendix that exem-
plifies participatory structures and other options of the three stages of
a task cycle.

Appendix: An Example Task Lesson

Task Description
This is a dictogloss task adapted from Li, Ellis and Zhu (2016, 2017).
It is a focused task aiming to facilitate middle school EFL learners’
comprehension of the English passive voice and elicits their production
of the structure. The task requires learners to listen to a narrative
presented by the teacher, work in pairs to practise retelling the story
and take turns to report the story to the rest of the class after the pair
work. The lesson contains two tasks that are based on different
content but follow similar steps. This is called procedural repetition.

Li et al. (2016, 2017) examined the effects of different form-focused

options in different phases of the task cycle, and we include the
alternative options in the lesson plan described as follows.

Task Materials
The task materials include two narrative texts embedded with thirty
cases of passive use. They were tailored to the level of the learners in
terms of length and difﬁculty.

Task 1 Narrative: A Car Accident

There was a bad car accident yesterday. Three people were killed. Also one child
was injured. Her leg and arm were broken. Her face was seriously cut. She was
driven to the local hospital. Her injuries were treated there. The relatives of the
girl were told about the accident.
A witness said, ‘The car was hit by a big truck. It was badly damaged.’ The truck
was travelling on the wrong side of the road. The driver of the truck tried to run
away. But he was stopped, and he was arrested. He was taken to the police
station for questioning. Some bottles of beer were found in his car. He was
charged with drunk driving. He was locked in a police cell.

Task 2 Narrative: An Earthquake

Kiki was raised in a small house in the countryside. One day he was playing when
suddenly there was a big earthquake. He was knocked down by the falling
bricks. Then the walls fell down. He was trapped in the house. It was very dark.
Kiki was badly hurt and could not move. Later Kiki’s mom came back home.
She saw the house was destroyed. She thought her boy was buried in the house.
She shouted out to him. He could not hear her because he was covered with
bricks.
Some dogs were brought to search for him. Kiki was found. The bricks were
removed. Kiki was pulled out of the wreckage of the house. He was carried to
the local hospital. He was put in an emergency room for treatment. He was
given special food to help him recover. He was allowed to leave the hospital
after one month.

Lesson Plan

Participatory
Stages Procedure structure
Pre-task The teacher asks a few brainstorming Teacher–class
questions about drunk driving to arouse
students’ interest and provide background
knowledge.
The teacher introduces the instructions for the Teacher–class
task.
The teacher teaches vocabulary to facilitate Teacher–class
task performance, but students are not
required to use the words in their
subsequent task performance.
Optional: pre-task grammar instruction. The Teacher–class
teacher explains the use and formation of
the English passive followed by controlled
practice where students judge the
grammaticality of the ten passive sentences.
The teacher presents the narrative three times. Teacher–class
She/he reads it aloud, presents it on
PowerPoint, and reads it aloud again.
Main task Students work in pairs to practise retelling the Student–student
story by referring to given clues. They are
required to add an ending to the story.
They are asked to take turns to report the Student–class
narrative to the rest of the class, with
one student telling half of the narrative
before passing the speaker’s role to
the other. Other students listen and
compare their endings to the speakers’.
They vote for the best ending at the end
of the task.
Optional: the teacher provides CF on the Teacher–class
learners’ errors in their use of the passive
voice. The feedback package consists of a
prompt to encourage self-correction,
followed by a recast in the absence of self-
correction.
Post-task Option 1: the teacher provides feedback on Teacher–class
the students’ errors in passive use.
(continued)

(cont.)

Participatory
Stages Procedure structure
Option 2: the teacher provides grammar Teacher–class
explanation about the passive voice if this is
not done in the pre-task stage.
Option 3: the students receive the script of the Individual
narrative and reﬂect on their own
performance.
Option 4: students are asked to perform The above
another task following similar steps based patterns are
on the second narrative text. repeated

This chapter starts with a general introduction to the basics in testing

since these underpin a specifically task-based approach to assessment.
It then gives a brief survey of significant general theorizing in the field,
leading to an outline model of language testing and language test
performance. This is followed by the main section of the chapter,
describing the components of this model, focusing on the role of
‘ability for use’, the influence of the task and the factors that are
relevant to performance rating, all linked to the relevance of task
research. Several challenges for task-based testing are then addressed,
while the concluding section provides a survey and critique of four
practical attempts to test in a task-oriented way.

An Introduction to Testing
Essentially, testing consists of the elicitation of data so that deci-
sions can be made which are reliable, valid, fair and useful. Each of
these words is central to language testing. Elicitation of data could
mean completing a multiple-choice test or writing an essay or
undertaking a speaking task. Decisions may be about how much
learning has taken place (as with achievement tests), or whether
someone has the appropriate language to do a certain job (a form of
proﬁciency test), or even whether someone comes in the top 10 per
cent of a group (and can be offered a place on another course).
Tests also need to be reliable, in that they should lead to infor-
mation (and decisions) which are consistent and not clouded by the
chance factors in the data that have been elicited, or the scoring or
rating that is done by different individuals. Tests also need to be
valid, in that they measure what they are supposed to measure (and
not what it is more convenient to measure). And they need to be
useful, and make worthwhile contributions. This is less tangible
than measuring for reliability (for which all sorts of heavy-duty

241

statistical techniques are available), but just as important. It might

mean that a test score is not obtained just for the sake of it, to be
put in a filing cabinet (or computer file) and forgotten. It might also
mean that the test itself needs to have a positive influence, or
washback, on the world around it, and influence teaching or prep-
aration beneficially. Ideally, doing well on a test should connect
fairly clearly with how effectively one uses language in the real
world – a potential source of strength for task-based approaches.
Language testers, over the years, have developed a range of proced-
ures to structure the way all these claims can be examined,
regarding data elicitation, reliability, validity, fairness, utility and
washback. Indeed, one could characterize language testing as an
inherently self-reflective area (even if, on occasions, one may con-
sider that particular examples of testing could be subjected to
more searching scrutiny). All these issues are just as relevant when
assessment is task-based as when conventional testing procedures
are used.
Some other basic distinctions are relevant. The first of these is
between summative and formative testing. Summative evaluation gen-
erally comes at the end of a course, and is likely to be in numeric form,
a judgement of how well someone has done on a test. The information
from summative evaluation may well go to outsiders, e.g. school
principals, education authorities. It is not likely to have a direct impact
on the detail of instruction. Formative evaluation is likely to be col-
lected during a course, and its central purpose is to feed into ongoing
instruction. The information that it gathers is likely to have diagnostic
features, but it may not be individual-focused, instead more concerned
with finding out how a course as a whole can be made better while it is
still running. A second distinction concerns score interpretation. One
approach, norm-referencing, relates a particular score to the scores of
other people, e.g. when a score is shown to be higher than 60 per cent
of other test takers. A contrasting approach, criterion-referencing,
relates the score directly to the real world, and whether a score is
consistent with someone being about to do something specific, e.g.
function as a hotel receptionist. Norm-referenced approaches can also
be used to make connections with the real world, but these connec-
tions are necessarily indirect, and additional research is required to
link real-world performance to a norm-referenced base. For example,
it could be established that someone two standard deviations above
the norm on a particular language test is probably able to handle the
language demands of a university course. Criterion-referenced
approaches, in contrast, reduce the amount of inferencing involved,
because criterion-referenced language tests are likely to contain

procedures which have more natural relationships with the real world.
As we will see, a task-based approach to testing leans more towards a
criterion-referenced approach for interpreting the meaning of scores.

Underlying Language Testing Theory

Any approach to task-based testing needs good theoretical founda-
tions. Essentially, this basic theory has to deal with a number of issues
and tensions:
• the role of any underlying competences;
• they way such underlying competences are mobilized into actual
performance;
• the tension between generality (likely to be based on underlying
competences and abilities) and specificity (a need to be able to
analyse and sample particular examples of language use).
The major approaches in the literature to theorizing communicative
abilities all wrestle with these issues (Carroll, 1961). Arguably the first
major attempt at this challenge was Canale and Swain (1980),
followed by Canale (1983). Canale and Swain (1980) proposed three
underlying competences: grammatical, sociolinguistic and strategic.
These were later supplemented (Canale, 1983) with the addition of
discourse competence. Clearly these competences are meant to be
underlying in nature and to represent knowledge bases which can be
drawn on in actual communication. In other words, in addition to the
more conventional grammatical competence, there is the assumption
that there is a knowledge base relevant to sociolinguistic and discourse
issues, and that any rounded approach to measuring underlying
competences will need to sample each of these. Most interesting of
all is strategic competence. Canale and Swain (1980) proposed this as
a capacity for improvisation when other competences are lacking, e.g.
a vocabulary item or a discourse marker to help conversational flow.
Strategic competence would then be drawn on to overcome a resulting
communication problem. Sadly, empirical investigations have not
been particularly supportive of this version of communicative compe-
tence. Harley et al. (1990), in quite a large-scale study, did not find
much confirmatory evidence for the four-component structure or for
the functioning of strategic competence.
The Canale–Swain approach was overtaken by the more ambitious
formulation of Bachman (1990; Bachman and Palmer, 1996, 2010).
This approach also has an underlying competence structure, but with
important differences to Canale and Swain (1980). Overall communi-
cative competence is divided into organizational and pragmatic

competences. The former is then further divided into linguistic compe-

tence and textual competence, while the latter is subdivided into
sociolinguistic competence and discourse competence. Interestingly,
Bachman (1990) does provide some evidence supporting this structure
and the arrangement and interrelationships of the different sub-
competences, although this empirical support cannot be regarded as
extensive. In general, though, we have a more convincing view of
the structure of communicative competence. More interesting still in
Bachman’s account is the role of strategic competence. According to
Canale and Swain (1980) this only had a compensatory role. For
Bachman it is central, and effectively the mediator between underlying
competences and actual communicative language use. It is not primar-
ily compensatory, but at the heart of all normal communication. It is
concerned with things such as goal-setting, assessing a situation,
planning, drawing upon background knowledge and so on. So here
we have a feature, labelled a competence, which permeates
actual language use. The other competences may have important
foundational roles and act as resources for the communication that
takes place.
Bachman (1990) also made a distinction between what he termed
interactive-ability and real-life approaches to testing. The implicit
emphasis in the discussion so far has been on the interactive-ability
approach. To return to one of the foundational issues in language
testing, there is a tension between the limited amount of data that is
obtained in testing situations on the one hand, and the desire to
generalize as widely as possible on the other. An interactive-ability
approach attempts generalization by effectively sampling from under-
lying competences in a way which permits maximum generalization
potential, but with testing formats which meet interactional authenti-
city criteria. In other words, although the emphasis here is on compe-
tences, the testing contexts which are used need to trigger natural use
of language on the part of the test takers.
This perspective, then, contrasts with a real-life approach which
emphasizes the uses for language that a candidate will encounter.
Testing, in this view, needs to develop methods to systematically
research patterns of future language use, since these will be the basis
for test items, or tasks, that will be used in testing. Bachman (1990)
and Bachman and Palmer (1996, 2010) attempt to provide such a
systematic framework to structure the ways needs analyses are con-
ducted for testing purposes. Indeed, many other needs analysis
approaches to testing have taken a similar approach, as in vocational
or occupational or English for academic purposes (EAP) testing. This
is easier when a restricted set of circumscribed language situations and

Task
Underlying Ability for Raters and
which generates Score
Competences Use Rating scales
Performance

Figure 9.1 An outline model for task-based testing

uses are concerned, but more difﬁcult when the range of situations and
language demands is broader, even if such cases are more circum-
scribed than ‘general language use’.
These different proposals for language testing models (Canale and
Swain, 1980; Bachman, 1990) have foregrounded the need for assess-
ment to be based on communicative uses of language. Skehan (1996,
2001) attempted to make connection between such approaches and
task research by outlining a related model which places task centrally
within testing. A revised version is shown as Figure 9.1.
The task is central in this model and is the basis for the performance
which is generated. To the left of the task are two general areas,
relating to the test taker, which are mobilized and provide the founda-
tion for this task performance. These are underlying competences and
an ability to mobilize and access these competences, under communi-
cation conditions. To the right of the task are the judgements that are
made about the performance. First, in this regard, we have the raters,
but also the rating scales that they use. The outcome of this stage is the
score which is assigned to the performance.
Little further will be said here regarding underlying competences.
A formulation such as Bachman’s is assumed, comprising the different
components mentioned. It is assumed that test takers have such under-
lying competences, to varying degrees, that these have an important
impact on performance, and that the score that is assigned will in turn
be inﬂuenced by them. It is assumed that task research has relatively
little to say about such competences, at least at present. But next we
come to ability for use. This implicates a language user’s capacity to
mobilize underlying competences in naturalistic communication, in
real time and in actual contexts. It is assumed, for example, that two
people with similar underlying competences might differ in actual
performance because they vary in how effective they are in accessing
and mobilizing these competences in an appropriate manner. When
testing aims at predicting to real-world language use, competences are

not enough: we also have to consider capacity to use these compe-

tences. Task research, it is argued, may have a great deal to say here.
Even more clearly, the task itself, the next part of the model, is likely to
have an impact on performance. Tasks are not neutral devices – they
may themselves have an impact on performance, and so the perform-
ance which is rated may partly depend on the task which was chosen
and the conditions under which it was done. Once again, task research
may have a lot to contribute to transforming the model in Figure 9.1
from a schematic account to a more empirically grounded proposal.
Next we turn to raters and rating, significant foci for research within
language testing, as researchers have explored whether different raters
might use different standards in rating performance, or if different
emphases with rating scales might introduce a lack of systematicity
into the assigned scores. The relevance of such research unquestion-
able. But task researchers have put considerable effort into how task
performance might be measured, what its dimensions are and how
different levels of performance might be distinguished. This research,
too, may have relevance to testing, since it could, potentially, provide
empirical grounding for the rating scales which are used and the
judgements which are made.
Testing typically ends with a score, and it is assumed that the score
will reflect what we want it to reflect. What Figure 9.1 brings out
clearly is that the score which is assigned, in a task-based approach to
testing, may reflect underlying competences, and/or ability for use,
and/or the actual task which was undertaken and the conditions and
context that were used, and/or the particular raters and rating scales.
This is a formidable and confusing set of influences, and shows that
the practice of testing is fraught with difficulty. The point in itemizing
these different influences is to reveal that the model shown in
Figure 9.1 is rudimentary, and that a major contribution of task
research to testing is to bring some clarity to the contributions of these
different influences, as well as the potential each of them has for
introducing a lack of standardization in testing. The sorts of things
that task researchers have explored, in other words, can begin to fill in
details regarding the different components in the model.

Task Research and Task-Based Testing

The central three components from Figure 9.1, ability for use, task,
and raters and rating, have been presented as potential clouding inﬂu-
ences on the score which may be assigned in a test. If there is such a
potential, it is useful to explore if task research can clarify how such
clouding might function, and so this section will explore what

contribution such research can make to understanding testing. It will

examine tasks, including task conditions, ability for use, raters and,
more precisely, rating.

TASKS AND TASK CONDITIONS

Broadly there are two approaches in this section: generalizations

which emerge from narrative enquiry and findings from meta-
analyses. Regarding the first, if we consider that task research started
to gain momentum from the late 1980s, then we now have a range of
task studies that could form a basis for any claims that are made
through traditional narrative accounts. Such research has frequently
used a complexity, accuracy, lexis and fluency (CALF) framework,
which eases connections between studies, and it has also tended to
explore task types and characteristics, on the one hand, and task
conditions, on the other, as we have seen in other chapters in this
volume. The research, as we have seen, has had theoretical linkages,
often connected with the respective accounts provided by the Limited
Attentional Capacity (LAC) approach and the Cognition Hypothesis
(CH) (see Chapter 3).
We have seen, for example, task characteristics and influences on
performance such as:
• structured tasks raise accuracy and fluency, and sometimes
complexity;
• tasks requiring information transformation or integration raise
complexity;
• tasks based on familiar information raise fluency and accuracy;
• time perspective influences performance;
• tasks with more elements raise accuracy slightly;
• tasks with support raise fluency.
We have also seen research which clarifies the influence of task
conditions, such as:
• strategic planning raises complexity and fluency, and sometimes
accuracy;
• online planning (i.e. less time-pressured performance) raises accuracy;
• post-task conditions raise accuracy, and sometimes complexity;
• repetition raises complexity, accuracy and fluency (CAF), but does
not have a strong effect on lexis.
One would like to say that these generalizations are robust and strong,
but it has to be admitted that what is proposed here represents no
more than tendencies (as is discussed in the third part of the Issues in

Table 9.1 Overview of meta-analytic results

Task feature Complexity Accuracy Lexis Fluency

Resource-directing 0.13 0.13 0.28 0.09
Resource-dispersing 0.77 0.73 0.27 0.34

Task-Based Testing section, regarding a difference in consistency

between tasks themselves and task conditions). But even so, the rele-
vance for testing is clear. If test tasks vary in any of the listed charac-
teristics or conditions and this variation is not handled in a principled
way, then the interpretation of particular tests may be compromised,
since the basis for comparing performance on one set of tests with
another set of tests, or with real-world language use, can be ques-
tioned. For example, if two tests, intended for use at the intermediate
level (say), contrast one test using an unstructured task, with unfamil-
iar information, without support and without planning with another
test using a structured task based on familiar information, visual
support and ten minutes of planning time, comparison between the
two tests is going to be difficult, if not impossible.
This argument is supported if one considers the results of meta-
analyses. Table 9.1 is a reduced version of Table 3.3, based on Mal-
icka and Sasayama (2017), and shows results from the most extensive
meta-analysis that we have available in the field.
The meta-analytic results mostly confirm the generalizations which
have emerged from the narrative analyses. This study was more driven
by the CH, hence the prominent distinction between resource-
directing and resource-dispersing variables. These more macro cat-
egories suggest that resource-dispersing variables have much larger
and more consistent effects on CALF measures than do the resource-
directing variables. At a more detailed level (and see Table 3.3 for
details), planning, repetition, task structure and task support show
consistency with the narrative enquiry generalizations, with only
familiarity revealing a different pattern.
This pattern of results is also consistent with a claim made in Skehan
(2016), as already noted in Chapter 3. This is that task conditions tend
to produce more consistent and larger influences on CALF measures.
Task characteristics, he argues, do not generate comparable levels or
such dependable results. He suggests that this is understandable
through the easier connections between task conditions and the
stages within the Levelt model, as indicated through quantitative
and qualitative research with, for example, planning. Most of the

resource-dispersing variables in Table 9.1 are, essentially, task condi-

tion variables, whereas most of the resource-directing variables
are task characteristics. So essentially the meta-analytic results are
consistent with Skehan’s claim.
We come back to the point made earlier in this section – if task
characteristics and task conditions are built in to test tasks in an
uncontrolled and unsystematic way, there is a danger that the score
someone receives will not be any sort of transparent measure of
underlying competences or ability for use, but rather that the task
and task conditions which have been used have intruded into the
measurement. The potential for unfairness is clear.
There is, though, an important caveat. This analysis tacitly assumes
that task research findings are directly generalizable to testing con-
texts, in other words, that these results – for example, on structured
tasks or on planning – operate in the same way when the context is
testing. There are reasons, unfortunately, why reservations are needed.
Iwashita, McNamara and Elder (2001) used insights from task
research, with variables such as planning, perspective, immediacy
and adequacy, but in a testing context. They report that the sorts of
effects found in task research were not evident. Khabbazbashi (2017)
compared the Test of English as a Foreign Language Internet-Based
Test (TOEFL iBT) and real-life (academic-context) performances and
reports that the iBT performance was slightly more complex but less
accurate than more naturalistic performances, although informal lan-
guage was more evident in the real-life performances. Khabbazbashi
(2017) also reports small but inconsequential topic effects. Brooks and
Swain (2014), also researching the TOEFL iBT format, confirm the
greater complexity finding for the test-condition iBT, with slightly
lower accuracy and less informal language. So these studies suggest
that task effects do not translate easily, consistently or strongly to a
testing context.
Even so, one has to say that there are not many studies applying task
characteristic variables in the testing domain. Planning, though, has
been more extensively researched in relation to a testing context. We
saw that Iwashita et al. (2001) reported little planning effect with
TOEFL-origin tests. This was confirmed by Wigglesworth and Elder
(2010), who used International English Language Testing System
(IELTS)-like speaking tests. Brief planning periods (zero, one, two
minutes) produced no effects on ratings or on CALF measures, with
the possible exception of a small effect on complexity. Interestingly,
planning time was positively regarded by participants. Nitta and
Nakatsuhara (2014) also investigated planning, with First Certificate
tasks from the Cambridge system. Ratings showed a small but

significant effect for fluency measures and also for ratings. Nitta and
Nakatsuhara (2014) report that planning was associated with parallel
or asymmetric interaction styles (Galaczi, 2008), and non-planning
with a more effective collaborative style. The planning period was
three minutes, but all participants looked at visual items pre-task, thus
providing a form of planning across all conditions. The two planning
studies to show positive effects in a testing context are Tavakoli and
Skehan (2005) and Xi (2005), who both showed significantly elevated
performance under planning conditions.
The results of these planning studies are clearly mixed, and it is
certainly possible that variations in research design might be a factor
underlying inconsistencies in results, such as different lengths of time
for planning, since the testing-oriented studies tend to use shorter pre-
task planning periods. O’Grady (in press) goes some way to overcom-
ing these problems. He contrasted an exposition task with a narrative,
picture-based task, under planned and unplanned conditions, with
several planning intervals, ranging from thirty seconds to ten minutes.
He reports a planning effect for the five- and ten-minute conditions,
with this being more effective for the narrative (the possibly more
complex task) than the exposition task. But this effect, though detect-
able, is not large, and is reported as 0.36 of a logit, in the context of a
five logit levels being distinguishable in the dataset.
Summarizing this range of studies, it seems reasonable to conclude
that task and task condition effects are relevant for testing, but it
cannot be concluded that they suggest consistent or large influences.
The database is not extensive, and it is clear that we can learn
more about the particular operationalizations of task characteristics
and task conditions in the future, and that this may lead to more
impact on performance, whether this is in terms of ratings or CALF
measures. But in the main, the influence is not huge, and it seems the
sorts of effect sizes from planning research are not matched in testing
contexts.
One can interpret this in different ways. The most obvious is to say
that a testing context is a great leveller. Test takers, who are abun-
dantly aware that their performance is being scrutinized, may mobilize
attention to maximum advantage and perhaps try to emphasize more
challenging language, and especially, try to avoid error. Any research-
based effect for something like planning might then be attenuated. In
this case, testers might reasonably conclude that they have ‘license’ not
to be overly concerned about differences between tasks or differences
between conditions, and accordingly claim that test scores are fair
reflections of more underlying abilities. But an alternative interpret-
ation is to question the interactional authenticity of many testing

formats. It could be argued that if effects which are well established in

the task literature do not reproduce in a testing context, then this
questions the naturalness of the data which is elicited and the potential
this data has for generalization to real-world contexts. Of course, it
might be argued that task performance is not totally natural, but if it is
the case that such performance is more natural than what happens in
many testing contexts, then the test-elicited data may be a less depend-
able base for use outside a testing context.
We turn next to a relatively recent development in task research –
the use of tasks for the assessment of pragmatic ability (Roever, 2011).
Roever and Kasper (2018) argue that existing testing procedures have
a strong psycholinguistic emphasis. They suggest that pragmatic
abilities, i.e. the use of language in context to achieve pragmatic goals,
often in longer encounters and interactions, is not sufﬁciently reﬂected
in current procedures, and that the use of tasks, as an assessment
vehicle, may enable this imbalance to be redressed. Roever and Kasper
(2018) point to this aspect of communication ability as one urgently
needing research and development.
Pragmatic abilities, though, are wide-ranging, and incorporating
them within task-based testing certainly contains its challenges. We
range from a relatively circumscribed focus on speech acts, to wider
aspects of an interactional competence, such as the capacity to handle
sequential organization in more extensive encounters or texts, turn-
taking, implicature, a capacity to access appropriate routine formulae,
and the ability to ensure there is cohesion and coherence. Beyond this,
there is the capacity to handle breakdown and repair, and even to use
non-verbal means of communication. And even beyond that, scope to
engage in co-construction of discourse and to be effective in using
language contingent on the contributions of an interlocutor. The scope
for task research is considerable here, and the corresponding challenge
for testing equally large.
As with all testing, there is a ‘scope for generalization’ issue here.
The different aspects of pragmatic ability concern scale (how large the
language use task is, and hence, how much the language user has to
sustain performance); organization (and how important it is to clarify
linkages between one’s own contributions as well as those of an
interlocutor); context (with the wide range of different contexts where
language might be used, with the myriad differences they may
embody); and freedom (in degree of scope for an interaction (say) to
be co-constructed and take unforeseen paths). Any attempt to measure
pragmatic ability, and to be able to generalize to real-world perform-
ances, will need to sample amongst these possibilities in a systematic
and principled way.

The last aspect mentioned here, freedom, itself introduces additional

complexity. Tasks vary in how predictable the language they elicit
actually is. But if one wants to assess pragmatic ability it is likely that a
task will not simply provide opportunities for speakers to demonstrate
turn-taking ability or capacity to anticipate problems and repair, but
in addition they may co-construct the discourse that emerges so that a
task, quite naturally, evolves in unpredictable unforeseen ways. The
task will then have worked as a task, but it introduces the difficulty
that different candidates, doing the same task, push the task in differ-
ent directions. In testing, this may be problematic, since comparing
test-task performances may mean not comparing like with like. The
standardization that is a desirable quality in testing may not be attain-
able. As Timpe-Laughlin (2018) indicates, this creates difficulty not
simply for rating (do raters try to make allowances for this, and if so,
how?), but also for tasks used in parallel forms of tests, or for choice of
interlocutor and for interlocutor influences on a candidate’s perform-
ance. There may also be issues of fairness since some candidates may be
more equipped through previous experiences to have familiarity with a
variety of contexts (and context is central to pragmatic ability), with
this possibly being connected with issues of wealth and travelling. So
while the assessment of pragmatic abilities is very important, it is also
difficult, even compared to the challenges faced by testers elsewhere.
Despite this gloom, interesting work has been done. Concerning
speech acts, Ekiert et al. (2018) devised a series of (computer-
delivered) tasks, comparing making a complaint, giving advice and
delivering a refusal. They found that the complaint > advice > refusal
sequence indicated greater difficulty for their participants, but this
sequence only applied at a lower-proficiency level. Higher-proficiency
participants performed on the tasks at equivalent levels. The sugges-
tion is that refusal may be slightly more difficult to handle than the
other speech acts. Youn (2018) also researched speech acts (recom-
mendation, requests, refusals), all in an academic context. Youn
(2018) also used conversation analysis techniques to confirm that the
interactions that the tasks provoked had natural characteristics, an
interesting way of validating the functioning of the tasks concerned.
Role plays were used, structured to produce different speech acts such
as refusing, requesting, agreeing. It was found that the role play, which
required participants to write a ‘request letter’, was the most difficult.
Youn (2018) used item response analysis to establish this, while
FACETS results suggested that, broadly, tasks and scales were neutral
as regards the results obtained.
These two studies (Ekiert et al., 2018; Youn, 2018) are quite struc-
tured and targeted. Two other studies are much wider ranging, and

bring out the challenges that using tasks to assess pragmatic ability
entail. Norton (2013) was interested in the potential of tasks to elicit
co-construction, and particularly with large-scale testing formats
where there is a need for standardization. Using the Cambridge Main
Suite format (see the section ‘The Cambridge Main Suite of Tests’)
with a scripted interviewer performance, he was able to show that
there is some, though not extensive, deviation from the script, so some
degree of co-construction is possible in this context. He suggests
though, that less educated testees may be less effective in exploiting
these co-construction possibilities. As with Youn’s (2018) study, Lam
(2018) was also interested in wider interactional competence, and
focused on the potential for, and detection of, contingency in perform-
ance. In a Hong Kong context, with English-medium secondary school
students, Lam (2018) argues that two group interaction tasks pro-
voked many opportunities for interactional abilities to be used (and
provides a discussion of the different components of such abilities as in
formulating (paraphrasing), accounting (repackaging) or extending a
previous speaker’s utterance). He also suggests, through conversation
analysis (CA), that contingency moves were evident in the perform-
ances. He argues that the existence of such moves demonstrates the
richness and naturalness of the interactions, but simultaneously, the
difﬁculty of using such tasks for testing, since the importance of such
contingency shows how interactions can be co-constructed, and so
difﬁcult to compare in any standardized way.

Conclusions
We see from this section, then, that tasks and task conditions influence
performance. The bulleted points indicate what narrative accounts
have suggested, and, coupled with the largely complementary findings
from meta-analyses, give some empirical foundation for this claim.
Test constructors clearly need to take such results seriously, and we
will see examples of this in the section ‘Practical Approaches’. But we
have also seen the challenge in applying task research findings to
testing contexts. We have seen that this transfer often does not occur
in a straightforward manner, or effects seem to be smaller when testing
is involved compared to the broader area of task research. There are
important research design issues here, and possible non-application in
testing studies of the exact operationalizations used in task research.
But the challenge is real, and there is scope for additional research that
is obviously important. As a final point, we can also now see more
clearly that attempts to use tasks to assess pragmatic abilities have
considerable promise. This line of research is relatively new, and the

range of ﬁndings is not extensive and does not, as yet, present too
coherent a picture. There is, though, considerable potential, and
encouraging progress has been made. For all that, there is the point
that this may be the area where the limits of testing are most exposed.
The fundamental tension between standardization and comparability
on the one hand, and freedom to co-construct discourse and allow
natural and unpredictable development on the other, may be difﬁcult
to reconcile.

Ability for Use

Tasks themselves, though, are not the whole story. While actual
performance on a task depends on task characteristics and task
conditions and also draws on underlying competences, all of this is
mediated through the concept of ability for use. Bluntly put, this
means that underlying competences may be considerable, but if they
are not mobilized into actual performance, communication is
impaired. In contrast, if limited underlying competences are mar-
shalled effectively, the performance that results may go beyond what
might be expected on competences alone. It is useful, therefore, to
explore what ability for use might consist of. Two preliminary points
are helpful. First, we are dealing with something pretty close to Bach-
man’s strategic competence, with its stages of assessment, goal-setting
and planning. This too was focused on how actual communication is
achieved against the background of underlying competences and how
these competences are mobilized for performance. But Bachman’s
account is not specifically task oriented, nor does it embrace the
factors that can now be located within ability for use. Allied to this,
a second point is that the Levelt framework, covered in Chapter 3, is
highly relevant for the functioning of ability for use and clarifies how
the functioning of each of the Leveltian stages, as well as the relation-
ship between them, is very important.
Two influences within ability for use concern the Conceptualizer
stage. The first concerns the way the Conceptualizer handles issues
such as assessing a situation, selecting what to say, taking a stance
and then organizing propositions into the pre-verbal message (cf.
Bachman’s account, as covered in the section ‘Underlying Theory’).
Decisions about all these things influence the likely propositional
content of what is going to said and also the pragmatic considerations
which will influence these propositions. Context will be an important
driver here for how the underlying ideas are selected and organized.
Linked to this is the general knowledge base that can be drawn on to
underpin these processes, knowledge of the world and of interactions

within it, and also of sociolinguistic expectations and pragmatic con-

ventions. None of this directly implicates the underlying language-
oriented competences, but provides the foundation for the ideas that
will be expressed and, indeed, is what leads to the pre-verbal message.
It is assumed that people vary in how effectively they are able to
marshal Conceptualizer processes of this sort.
Then we have two inﬂuences more linked to the Formulator stage
and to the psycholinguistic processes that underpin actual language
production. First, the second language mental lexicon (SLML), dis-
cussed more extensively in Chapter 3, is, in itself, an underlying
competence, but ability for use is relevant when we consider aspects
of the lexicon which make it more effective in communication. Here
we have issues like the speed of operation of the SLML as well as
access skills, the richness of what is held in lemmas, the connections of
lemmas to one another and the completeness of what is retrieved. All
of these are going to vary between individuals, and are distinct from
more conventional views of what a competence consists of. In add-
ition, within the SLML, there is the range of formulaic language which
can be drawn on. As we saw in Chapter 3, formulaic language eases
processing demands and if the language user can draw upon ready-
made language chunks, the need to allocate attentional resources to
computation is reduced. The second Formulator inﬂuence here is
working memory, which concerns the resources and attention that
are available as a message is built. Those with larger working memor-
ies, other things being equal, will have more ability to organize ideas
and then access linguistic resources, with this being particularly
important under any communicative pressure. As indicated in
Chapter 3, it is interesting that working memory is an important
individual difference factor in online planning studies, but not in
strategic planning. It has to be concerned, that is, with the effectiveness
with which underlying competences are assessed and brought to bear
in actual communication.
The two remaining components of ability for use implicate both
Conceptualization and Formulation stages. A compensation ability
connects with the Canale and Swain (1980) framework, and also with
Hulstijn’s (2015) account of second language (L2) performance. It is
the ability to handle problems when they occur during communica-
tion, either because underlying competences are lacking or because
there has been some sort of misstep which needs to be retrieved. To
include this ability here suggests that this is a consistent capacity to
deal with the unforeseen. Linked with this resourcefulness is the
capacity to monitor more effectively, and to act upon that monitoring.
If things are going well, it is possible that monitoring, as a process,

may be close to inactive, but if someone is operating at close to or

outside their comfort zone, effective monitoring may pre-empt what
might become more serious problems later. Finally we have
metacognition, which is slightly more Conceptualizer oriented but
nonetheless relevant for Formulator operations also. It concerns the
L2 user’s insights into their underlying competences and other aspects
of ability for use, and the way these need to be used flexibly in a given
communication situation. Metacognition is pervasive in its use, as on
occasions the L2 user makes adaptations to be realistic about what
they can achieve, even changing the focus of Conceptualization
because of insights regarding limitations in ability and in the SLML
(Skehan, 2018). Pang and Skehan (2014) bring this out clearly from
their qualitative study of planning where it is clear that some learners
modified what they did during the planning period to make it more
likely that they would produce language commensurate with their
abilities and not allow over-ambition to lead them into difficulty. L2
speakers clearly vary in their skill with metacognition (as they vary
with other aspects of ability for use) and this then has an important
influence on communicative effectiveness.
If this analysis of ability for use as a mediator for underlying
competences in actual performance is accepted, the impact on effective
assessment is considerable. Of course, there are techniques to measure
underlying competences, possibly reliably and validly, and such
measurements have considerable use and meaning. But if ability for
use has this important mediating role, such competence-oriented
measurements have to be balanced by approaches to testing which
allow it to come into play. Not to do so is to miss testing a very
important part of communicative ability, and so generalizations as to
how people can use language in real situations are compromised.
Tasks then become an important vehicle to allow ability for use to
manifest itself, since they can potentially require the language user to
cope with time pressure, with a need to relate language to ongoing
discourse and context, and to use language appropriately.

Raters and Rating

There is an additional aspect of Figure 9.1 that requires more com-
ment, and which brings out another potential contribution of task
research. This is measurement, and the way task-related performance
is rated. The convention in assessing spoken and written performance
is to use global and analytic rating scales. Global scales tend to be an
amalgamation of the analytic, and so won’t be discussed separately
here. Analytic scales typically cover grammar, vocabulary, accuracy,

pronunciation, range, ﬂuency, organization, coherence, task fulﬁlment

and so on. The history of language testing makes frequent reference to
the construction of rating scales and the basis for providing descrip-
tions of performance at different levels which are valid and reliable,
as well as comprehensible to raters and consumers of language
test results. A considerable amount of expertise has been developed
in such scale construction, reflected in systems like the Common
European Framework of Reference for Languages (CEFR) (Council
of Europe, 2001, 2008) or the Foreign Service Institute system and so
on. But there have also been periodic calls to base scale construction
on a more secure empirical footing, both with respect to the dimen-
sions of performance and also the details in the scale descriptions
(Fulcher, 1996).
In parallel, task research has given considerable importance to the
measurement of CALF. There is a considerable degree of correspond-
ence here (although this is far from absolute), and so complexity (with
tasks) has a resemblance to grammar and range (in language testing),
accuracy is simply accuracy, lexis relates to vocabulary and fluency
maps onto fluency (and what is often termed ‘delivery’). This raises the
possibility that there are a number of areas where findings from task
research can make useful contributions to the content of rating scales
that are used in assessment and to their empirical base.
Skehan (2018) explores what task research has revealed in this
regard. For example, he shows that the two measures of complexity
most used in task research, subordination and words-per-clause do
not correlate particularly highly and appear to measure different
things. This confirms Inoue’s (2013) finding of the same effect, actu-
ally obtained in a testing context. The former reports a median correl-
ation across a number of studies (and different conditions within each)
of −0.03, and the latter, across a smaller number of tasks, of also close
to zero. This has the interesting implication that rating scale measures
of complexity in a testing context really need to address these two
areas separately, and include descriptors relevant to each. In contrast,
the task research findings concerning measures of accuracy suggests
that they broadly do the same thing, whether one looks at errors-per-
clause, or measures based on error gravity, or one examines measures
of length of clause that can be used accurately. So with accuracy the
results suggest that developing analytic rating scales is relatively
straightforward and likely to be consistent with measurement
approaches in task research.
With fluency things are less clear. The task research suggests that
speed, breakdown and repair form separate sub-dimensions of
fluency. This division might then be the basis for different aspects of

fluency rating scales, since it appears that the three areas have some
independence from one another. In addition, there seems something of
a tension, within fluency, between discourse and clause-linked influ-
ences (Skehan, 2018). The former are concerned with dysfluencies
which manifest themselves at analysis of speech (AS) or clause bound-
aries, whereas the latter connect more with within-clause problems,
indexed by such things as mid-clause silent and filled pauses and
repair. Tavakoli (2018) also presents interesting proposals suggesting
that different fluency measures are effective in distinguishing between
CEFR levels, with speed distinguishing between levels from A2 to
B2/C1; silent and filled pausing distinguishing between A2 and the
higher levels; and mid-clause pauses between A2/B1 and B2/C1. Once
again, the implications for the detail of rating scales which are used to
measure performance are intriguing. Finally, with lexis, task research-
ers broadly use measures of lexical diversity (i.e. corrected type-token
ratios) and lexical sophistication (typically the ‘penetration’ of a per-
formance by less frequent words). The former measures a speaker’s (or
writer’s) capacity to avoid recycling the same words, while the latter
seems to reflect a capacity to draw upon a wider underlying lexicon.
Interestingly, measures of each area do not particularly correlate
(Skehan, 2009b), and so these two aspects of lexical performance
(and so vocabulary ratings of test-task performance) need to be kept
distinct. Once again, task research has the potential to illuminate what
needs to be included in analytic rating scales.
Findings such as these have considerable potential to influence what
is covered in the rating scales that are used in language testing. They
may provide more valid and empirically based input to the wording of
scales and also, thereby, possibly bring rating scales and discourse
analytic measures into a greater degree of congruence. One example of
this could be O’Grady’s (in press) work in developing what he terms
EBB (empirically derived, binary choice-oriented (to simplify decision-
making) and boundary-oriented) scales, regarding the different levels
in rating scales. Detailed performance analysis in task research could
be an important impact in such an approach. In any case, returning to
Figure 9.1, it is clear that rating, and information for raters, are other
areas where task research has an important potential contribution to
make to language testing.

Conclusions
The model outlined in Figure 9.1 was schematic and preliminary. The
discussion since then has covered a great number of studies, and
shown that the model is now grounded in a good deal of research.

There is, of course, much still to do, but already we have a framework
which enables us to relate task research to testing, and not simply to
testing speaking. We know much more about task effects and task
condition effects. Measurement procedures from the task domain
have been shown to be clearly relevant to the process of rating test
performance. Ability for use, perhaps less grounded in research studies
than the other components of the model, nonetheless has clear rele-
vance for our understanding of effective testing. Above all, ability for
use provides a way of thinking about the relationship between
interactive-ability and real-life approaches. By focusing on the details
of processing, in addition to simply looking at task characteristics, we
may have a basis for generalizing more effectively from particular, and
necessarily limited, test tasks, provided that processing demands can
be linked to wider patterns of language use. It may be that, following a
task-based approach, the interactive-ability and real-life approaches to
testing are not as different as they may at ﬁrst sight seem.

Issues in Task-Based Language Testing

The previous section, ‘Task Research and Task-Based Testing’, has
tried to argue for a task-based approach to testing, to link such testing
with task research, and to sketch out a model which organizes the way
tasks can be related to the assessment situation. It has considerable
promise. But there are a number of issues which need discussion –
issues which indicate there is also progress yet to be made. We will
discuss four of these: the contrast between real-life and interactive-
ability approaches, given the role tasks can play in assessment; the
conundrum of task difﬁculty; the central issue of whether task per-
formance is predictable; and ﬁnally the enduring challenge of achieve-
ment testing.

Real-Life vs. Interactive-Ability Approaches

A thread running through this entire chapter so far is the contrast
between real-life and interactive-ability approaches to language testing
(Bachman, 1990). The former takes as a starting point real-life tasks,
and so the purpose of assessment (not to mention instruction) is to
identify relevant real-life tasks and then to model assessment tasks
upon them, as advocated by Long (2016) and Norris (2016), and
Bachman and Palmer’s (1996) Target Language Use (TLU) approach.
The emphasis here is on the real-life relationship between pedagogic
and assessment tasks, and some approaches (Brindley, 2013; Wiggles-
worth and Frost, 2017) see this as what task-based testing is.

This approach, assuming the accuracy of the real-life uses which are
identified, has the enormous advantage of a very good basis for
prediction to such situations, as well as the potential to devise scoring
procedures which reflect what is important in real-life tasks. Sampling,
in this view, is focused and is more likely to achieve comprehensive
coverage of a limited domain. Later in the chapter we will explore one
application of this approach.
Obviously the central strength of this approach – that there is a
clear relationship between test-tasks and real-world language uses –
presupposes that the range of real-world language uses is suffi-
ciently identifiable and restricted. In one sense, this strength is also a
weakness – the prediction to such real-life tasks may be good, but the
consequence is that prediction to other real-life tasks may be com-
promised precisely by the precision of the match that is the basis for
such effective prediction. The range of prediction may therefore neces-
sarily be narrow. Interestingly, another critique that is made of the
approach (Bachman, 2002) is that real-life contexts which are targeted
are rarely as uniform as the approach would require. Bachman (2002),
for instance, gives the example, from Norris et al. (1998) of two tasks
which may be both superficially concerned with the same theme, but
which differ markedly in other respects. This follows Skehan (1984)
who critiqued English for Specific Purposes (ESP) testing on a similar
basis – different domains contain vastly contrasting sub-specialities,
for example, psychiatry and dermatology are both within medicine but
very different in language use.
The alternative, the interactive-ability approach, also has its
strengths and weaknesses. Sampling, traditionally, has been on the
basis of underlying competences, perhaps linked to some target-
language use system, although this, necessarily, will be nothing like
as precise as the alternative real-life approach. This sampling will
attempt representative coverage, but of a much larger domain, and
in many cases the domain involved, as with general proficiency tests,
can only be very considerable indeed. But if the sampling is maximally
effective, then the generalization to real-world performances can be
wide-ranging, even if less exact about particular real-life events. Even
so, the problem is that the relationship between the sampling and the
ultimate real-language use situations is not clear.
In his formulation of this distinction, Bachman (1990) makes it clear
that the tasks which are used in testing need to generate interactional
authenticity, in other words, a use of language which resembles real-
life language use and which engages abilities in a similar way to what
happens in such contexts. In addition, he proposes that strategic
competence, as he formulates it, is important in making a bridge

between underlying competences and actual communication. What we

have seen in this chapter is that tasks and the reconceptualized ability
for use could be an important basis for sampling and the generaliza-
tion that is then enabled. By trying to draw upon a variety of tasks and
task conditions, the data base which will emerge will be a more robust
basis for connecting with a range of language use situations. What will
be discovered is underlying processing capacities, and these are likely
to provide a better basis for predicting how well people will do in
future unspecified situations. This needs to be coupled with sampling
of ability for use, for probing how effectively test takers can handle
issues such as effectiveness of Conceptualization, speed and effective-
ness of SLML access, capacity to use working memory, metacognitive
processes, compensation and so on. The specific tasks that are used
may not be general in nature, but the underlying competences that
they trigger could be a better basis for generalizing than simply
targeting underlying competences. In this way, a task-based approach
to testing, it could be argued, is a better justification for using
an interactive-ability approach, and also, possibly for avoiding the
limitations of a real-life approach.
Of course, there is a lot of work that needs to be done here. The
findings on task characteristics are suggestive more than definitive as a
base for predicting actual performance, despite the generalizations
mentioned. Conditions, perhaps, are more promising, but even here
there is the issue of replicating task condition research findings in a
testing context. Further, while ability for use as a construct is interest-
ing and has great potential, tying down its different components and
establishing the detail of their influences is a challenge for the future.
But these factors do have considerable potential relevance for testing,
and it is to be hoped that in the future they will motivate a great deal of
testing research.

The Difﬁculty of Test Tasks

One aspect of Figure 9.1 requires further discussion. The task box in
the figure is highly likely to be influenced by both Conceptualizer and
Formulator issues. These are important for the general influences that
they have on performance. They are also relevant to the concept of
difficulty in language testing. Conventional testing, armed as it is with
some complex statistical procedures, is able to arrange test items on a
cline of empirically established difficulty, where that difficulty is based
on information regarding the proportion of test takers who pass an
item, and the extent to which individual items conform to a ‘model’ of
difficulty (McNamara, 1996). With task-based approaches to testing,

we are well beyond individual items which can straightforwardly be

passed or failed. The performances that are elicited are multidimen-
sional and complex, and do not lend themselves to simple passing and
failing decisions. Pollitt (1991) notes that in assessment there is some-
thing of a division between approaches which count things (as with the
items in conventional tests) and approaches which rate performance. It
is the latter approach which is relevant to testing, as the multidimen-
sional nature of performance is likely to be rated for things such as
range (close to complexity in task-based performance), accuracy,
vocabulary, pronunciation, and outcome or task fulfilment, as we
have just seen. This richness of performance, coupled with the fact
that with tasks one cannot have large numbers of ‘items’, means that
measuring performance is much more complex, and in turn, that
making decisions about difficulty is also much more complex, since
tasks may push different aspects of performance in varying directions,
and so the multidimensional nature of what is said or written makes it
difficult to develop a one-dimensional scale of difficulty.
Skehan (2018) proposes in this regard that it is useful to separate
notions of task difficulty in testing into those which are
Conceptualizer-linked and those which are Formulator-linked. In
the former case what would be manipulated are the ideas that
make some tasks more conceptually difficult than others, such as
information which is more extensive or more abstract, or operations
which are more complex. Then with Formulator influences one would
be looking to vary performance features which have an impact on
access to the SLML, and the speed and ease with which language
can then be built and syntactic frames assembled. Skehan (2018)
proposes that the influences on making tasks more difficult or complex
in these two cases are different, and that if one is developing a series of
tasks as the sampling basis for assessment, it is helpful to keep them
apart, both conceptually and with actual test-task development. In this
way one can perhaps obtain a more rounded view of performance
and a better basis for generalization. Table 9.2 exemplifies how this
might work.
The essential point here is that sampling combinations of easy and
difficult tasks and conditions for each of the stages of speech produc-
tion can be more effective in giving a comprehensive account of a
candidate’s communicative competence. The details in the different
cells in this figure are sketchy and incomplete, essentially only exem-
plifying, but as task research accumulates, we will have a better idea as
to how to build in different degrees of difficulty in a testing situation
and how to probe different aspects of ability. The figure offers a
framework that could be very useful in testing.

Table 9.2 Level of difﬁculty linked to Conceptualizer and Formulator

inﬂuences

Conceptualizer easy Conceptualizer difﬁcult

Formulator easy • Unpressured • Unpressured, small-scale
communication communication
• Familiar, structured • Extending, planning,
information, only reasoning,
requiring retrieval transformation
Formulator difﬁcult • Pressured • Pressured
communication, heavy communication; heavy
input, monologic, non- input; monologic; non-
negotiable negotiable
• Familiar, structured • Extending, planning,
information; emphasis reasoning and
on retrieval transformation

The Predictability of Tasks and Task Conditions

In Chapter 3, and also earlier in this chapter, the predictability of tasks
and task conditions have been discussed. Indeed, two approaches
which try to account for regularities in task and task condition effects
have emerged. The LAC approach proposes that these are numerous,
and that actual performance is the conjoint result of the combinations
of inﬂuence which may operate in any one task, such as when a
structured task (and such tasks lead typically to greater accuracy) with
planning opportunities (generally associated with higher complexity
and ﬂuency) produce raised performance in the three CAF areas. The
CH, through its resource-directing, resource-dispersing distinction,
makes predictions about the effects of the various examples of each
of these categories. Each of the approaches has generated a consider-
able amount of research and the database we have available is exten-
sive, as meta-analyses, such as that of Malicka and Sasayama (2017),
make clear.
Task research, then, through theory and through empirical work,
has provided clear indications that there are important generalizations
one can draw on. But Skehan (2016) expresses words of caution.
Reviewing the available research, he suggests that the consistency in
effects of task characteristics (in general, and not just associated with
the two more theoretical approaches just mentioned) is disappointing.
There are many interesting avenues to explore (task structure, task
complexity, information type and organization and so on) but the

evidence is not strong that they produce effects, let alone substantial
effects. He suggests that task conditions are associated with more
consistent findings, with this applying to pre-task planning, online
planning (i.e. lack of time pressure during actual performance) and
repetition.
It is clearly the case that language testing needs to take account of
this theorizing and research in L2 task performance. But we have also
seen that when language testing studies have tried to incorporate
variables from task research, there has been far from a simple replica-
tion of what happens. We have seen that comparing TOEFL iBT
performance to real-life performance sees the latter associated with
slightly higher complexity but lower accuracy (Brooks and Swain,
2014; Khabbazbashi, 2017). Studies of planning, in general, do
show an effect on performance, in a testing context, but the effect
tends to be small, and likely more important with CALF measures
than with performance ratings (Wigglesworth and Elder, 2010;
Nitta and Nakatsuhara, 2014; O’Grady, in press). It is also reported
that the effect is not so great as to lead to different rating levels being
assigned (from FACETS analyses). There are important research
design issues with these results, in that we have not seen a systematic
examination of the relevance of variables identified in task research
for a testing context, e.g. task complexity or time pressure. Nor
have variables which have been used, principally planning, been
operationalized in the same way in testing as in task research (with
perhaps the exception of O’Grady). So the jury is still out. But one
does have to say that so far task research can only be suggestive as
regards a testing context. There is considerable promise but it has yet
to be realized.
This discussion relates to two important challenges to task-based
testing raised by Bachman (2002). First, he suggested that progress in
task-based testing requires that tasks which function in the same way
generally be identified, so that they are dependable influences
upon performance. As we have seen, the findings on this issue since
Bachman’s critique was published have not been replete with sugges-
tions which meet his criterion. Some are, possibly, such as the usual
suspects of planning, time pressure during performance and some task
characteristics such as structure. But generally, his point is still rele-
vant. His second critique, and this was perhaps in response to the
Hawai’ian EAP project, was to suggest that the prospects for a task-
based approach to testing, or at least real-life testing, are limited
because a key issue is that tasks (and presumably task conditions)
interact with individuals such that one cannot assume that a given
task, condition or task condition bundle will work in the same way

with different people. There will be differences in background know-

ledge, training, experience, preferences, which will mean that a par-
ticular task, even if it were of a particular ‘difficulty’, would not be
responded to in the same way by different individuals. Short of identi-
fying task characteristics which operate in a ‘candidate-independent’
way, the hope of regarding a task as a dependable unit of testing is
limited.
This is an interesting criticism. There are two reasons, though, why
it may not have so much force given the view of task-based testing that
is promoted here. First, there is the issue that a lot of task research has
been done since the critique was originally made. We are now in a
position, as noted, to propose some reasonably powerful generaliza-
tions, even if task researchers would like to be able to report more!
Second, and returning to Figure 9.1, the clearer prominence and
understanding we have regarding ability for use is another force that
clarifies some of the variance in test scores which might otherwise be
regarded as error. Performance does not depend solely on tasks and
task conditions. If we can understand more about how the different
aspects of ability for use function, we can design more effective assess-
ment procedures and have a clearer basis for generalizing to real-life
performance. After all, to take one example, if working memory
interacts with processing conditions and time pressure, we have first
of all identified what might otherwise be regarded as an unforeseen
interaction, and second of all, we are focusing on something which
would be relevant to performance in the real world. If working
memory, for example, confers an advantage in some real-life language
use, then we have to be attuned to that effect.
But we are left with what might be the greatest challenge of all to the
predictability of tasks (and test tasks) – the importance of co-
construction. As we saw in the discussion of pragmatic abilities,
effective communication requires much more than knowledge of an
underlying system. There is also a need to understand how L2
speakers handle pragmatic demands and how they function effectively
in terms of interactional abilities. This, naturally, requires test formats
to incorporate interaction, and this leads to a considerable problem in
that once a candidate in a testing situation is confronted by the need to
interact, there is the possibility that the interlocutor will exert an
influence on the candidate’s performance. But even more challenging,
if the interaction is allowed to develop naturally, in whatever direction
seems appropriate to the interactants, then a lack of standardization is
introduced. In other words, co-construction does present very serious
challenges (Timpe-Laughlin, 2018) for any capacity to compare per-
formances of different candidates (since they may not have been doing

the same thing even though the starting point was the same) or parallel
forms of tests (since this follows from the previous difﬁculty – if a
given starting point can lead to different development, two (or more)
different starting points can only make matters worse). In the second
part of this chapter, in the section ‘Practical Approaches to Task-Based
Testing’, we will see how attempts to overcome this difﬁculty have
been made. For now, it has to be recognized as a serious concern – if
tasks are to meet criteria for naturalness and unpredictability, their use
in testing is not at all straightforward.

Achievement Testing
Long (2015) suggests that the most urgent area for progress in task-
based testing is that of achievement testing, and indeed there are a
number of ways of justifying this. The first is that this is the area which
has shown least development since the late 1980s or so. Proficiency
testing, whether interactive ability or real life, has seen some important
developments, but testing linked to teaching has languished in com-
parison. There is great scope for improvement, in other words. But
even more important is the consequence of lack of development in this
area. Washback, as we have seen, is an important test quality. If task-
based teaching is not matched by task-based assessment, and even
worse, if assessment focuses on more formal aspects of language,
the achievements of a task-based approach will be compromised,
as learners (and parents and educational systems) will inevitably pri-
oritize what is tested. So the need for educational systems to match
task-based instruction (TBI) with task-based assessment is vital.
There are, of course, many reasons for this lack of progress. Test
construction and validation are expensive and time-consuming and so
it is no coincidence that the greatest activity is associated with profi-
ciency tests, especially where these are well funded, as with major
international testing organizations. Teachers are far less resourced in
this regard, and also typically lack the same level of technical and
computing expertise that is routinely used in effective test validation
processes. Qualities such as reliability, validity, fairness, usefulness are
just as important with achievement tests as with proficiency tests, but
are more difficult to attain given the resourcing typically available.
There is also, perhaps, the issue of a lack of theory for the development
of achievement tests, something which reflects a disconnect between
L2 acquisition research, on the one hand, and classroom and
textbook-based instruction, on the other. Achievement testing needs
to be based on what has been taught, but many approaches to teach-
ing, and this may well include some versions of TBI, may not be based

on a clear view of development and acquisitional progress, which does

not help in designing tests which are meant to reflect progress.
In one respect, though, achievement testing ought to be simpler. The
‘target’ of such testing is what has been taught, and so the challenge in
designing tests should be to devise effective ways of sampling so that
tests cover the different aspects of teaching in a systematic way. The
emphasis in such cases is to establish content validity. More traditional
approaches to teaching lend themselves to this fairly well, in that the
items of a syllabus can be sampled, and performances on test items can
then be totalled, even with the prospect of giving diagnostic infor-
mation about aspects of the syllabus which have been covered. Things
are more difficult with task-based approaches. Test tasks are likely to
require performances to be generated, which are time-consuming
(enabling fewer ‘items’) and require more complex rating procedures.
In addition, the units in task-based teaching are less likely to lend
themselves to generating representative items. So the tester has a
number of additional problems to face, not least as (as Chapter 7
demonstrated) notions of syllabus in the task-based area are less
definitive or precise. In fact, compared to the textbook dependence
of much conventional teaching, task-based achievement tests may
require classroom observation to discover what has actually been
taught, or even individual-tailored tests, given the freedom and non-
prescriptive approach in task-based achievement tests.
The difficulties are clear, but how might they be overcome? Obvi-
ously a first and very important approach is for teachers to develop
achievement tests on the basis of the task-based classes they have
taught. This is feasible, and even minimal investigation of the func-
tioning of such tests will be valuable to establish some degree of
reliability and effective measurement. But this will be demanding on
people with a very limited amount of time. More realistic, given
the likelihood of coursebooks being used, is for publishers to make
greater efforts to develop task-based tests for coursebooks which make
claims about the use of tasks as part of their methodology. Large
coursebook series now routinely do contain testing material, but this
is often fairly conventional in nature. Publishers need to take more
responsibility for matching the approach to testing with a claimed
approach to instruction. They have considerable resources, and such
coursebook series have very large budgets. Producing a range of
testing materials which teachers can draw upon, as needed, should
become the norm.
An alternative approach, if teaching is carried out under the aegis of
a state educational authority, is for collaboration with institutions of
higher education or with testing bodies. In this way professional

expertise can come from the institution concerned but there can be
collaboration with teachers involved in the process of test develop-
ment itself. Consequently, hard-pressed teachers can work within a
wider framework, but be supported in their efforts. This promotes the
possibility that tests will be locally relevant and reflect the reality of
TBI, but that the tests themselves can meet professional standards.
Indeed, more widely collaborative action can lead to the production of
resource banks for testing which, cumulatively, can grow and provide
a basis for selection of tasks of known qualities.
One final initiative might be of note. Harrison (1982) describes a
system where sets of different coursebooks were analysed for the tasks
that they used and then these were taxonomized through content and
functions/notions. Test tasks were then developed targeting different
content areas and, within this area, a range of mini-tasks were devised.
These were assembled into a ‘challenge sheet’, with claimed differences
in difficulty. This was done in collaboration with actual teachers. But
what was innovative in this work is that the challenge sheets were then
given to the students themselves, who were given responsibility for
saying when they were ready to take on the detailed items within each
challenge sheet. The interlocutor was another student, who was
responsible for the ‘validation’ that the mini-task had been done
adequately. Each challenge sheet contained ten or so mini-tasks, and
the student in question could only approach the teacher when all ten
tasks had been validated by other students. In this way, considerable
assessment was carried out, but the teacher’s workload was consider-
ably eased. The system required significant centralized development
work, but it then led to considerable washback as students were given
much more responsibility than is usual for language process within the
task-based system.
The final section of the chapter will include a couple of examples of
achievement-testing projects.

Practical Approaches to Task-Based Testing

This section will present four case studies in task-based testing. These
bring out the relevance of the preceding discussion, and at the same
time indicate how practical problems can be solved to make a task-
based approach a reality. They comprise a project to develop a testing
procedure in a university EAP system (a real-life approach); a set of
publicly available commercial tests (an interactive-ability approach);
some tests to accompany a coursebook series; and a series of testing
ventures within an educational system (with these last two examples
both functioning as achievement tests).

The Hawai’i Performance Assessment Project

Many of the principles and difficulties of a task-based approach to
testing are illustrated by a testing project which originated at the
University of Hawai’i, and which is documented in two book publica-
tions. Norris et al. (1998) describe the development of the bank of test
items and provide a rationale for the approach. Brown et al. (2002)
provide an account of considerable practical work with this item bank
and their work contains rich statistical data. The purpose of the
project was to develop a wide range of prototypical tasks which would
‘represent real-world tasks that might face university students studying
a second or foreign language’ (Norris et al., 1998, p. 71). In other
words, this is clearly a real-life approach. On the basis of an extensive
needs analysis, a very large number of such tasks was generated. The
tasks fell into seven general categories:
• health and recreation/entertainment (eighteen tasks);
• travel (sixteen tasks);
• food and dining (thirteen tasks);
• at work (seventeen tasks);
• at the university (nineteen tasks);
• domesticity (fifteen tasks);
• environment/politics (six tasks).
The test tasks which were developed, more than one hundred in total,
covered this range of themes. They each of them met criteria for ‘taski-
ness’ (e.g. giving medical advice; enquiring about financial support (in a
university context); advertising for a housemate; planning a presenta-
tion), and were mostly associated with relevant support materials. The
authors developed a difficulty rating scheme for each task, based on three
areas, with two subcategories of each (loosely based on Skehan, 1996):
■ Code complexity:
○ range;
○ number of input sources.
■ Cognitive complexity:
○ organization of input;
○ availability of input.
■ Communicative demands:
○ mode;
○ response level.
Each subcategory had two values (− and +), yielding six ratings in
total, generating a score range of 1 (the easiest level of task, based on
six minuses) to 6 (the most difficult, with six plusses). Each general

and speciﬁc category were provided with copious descriptive material,

and high reliability was achieved in the ratings, which were calculated
by category difficulty and overall test task difficulty.
At first sight the project is more obviously consistent with Bach-
man’s real-life approach, in that the different tasks outlined in the
Hawai’i project relate to studying at a university, and so have clear
real-world relevance within that sphere. In that respect, effective sam-
pling of behaviour can be claimed quite straightforwardly through
surveys of language use in such contexts, systematically sampled, as
they were. But Norris et al. (1998) were introducing a scheme to assess
difficulty which has resemblances to what I have proposed here. It also
has connections with Bachman’s interactive-ability approach to
testing, since these categories concern more general processing. The
approach, that is, melds a real-life task choice approach with an
interactive-ability notion of difficulty.
The research team also developed two types of rating scale. The first
was task specific and, using five steps ranging from clearly inadequate
to adept, provided clear descriptions of the different levels of skill
needed to complete a task. Impressively, a different task-specific set
of descriptors, at the five levels, was developed for each separate task.
In addition, a task independent scale was developed, also with five
steps. Generic descriptors were provided for code command, cognitive
operations and communicative adaptation. Self-ratings were also col-
lected as part of the research project.
The project was largely successful in what it set out to achieve.
Extensive analyses of the dataset showed that performance ratings,
both task specific and task independent, were highly reliable, so that
judgements of performance on the different tasks were consistent and
dependable. The project demonstrated, in other words, that although
the tasks and the performances might be considered complex, the
judgements made about those performances amply met conventional
statistical standards. In addition, there was important validity evi-
dence. Different reference groups took the different tests, and these
reference groups had fairly clear differentiation in general ability
levels. The data patterned strongly in accordance with differences
between the various groups.
The one area where the research could be regarded as disappointing
(and this is an important disappointment) concerns the relationship
between estimated task difficulty and actual performances on tasks of
different proposed difficulty. The authors indicate that they had hoped
that a clear relationship in this area would have been the basis for
claims of generality. The capacity to relate tasks to difficulty in a
principled and empirical way would enable more powerful claims to

be made about generality and how test tasks could be chosen to make
more precise statements, and then perhaps go beyond the particular test
tasks chosen. So the weak-to-moderate relationship that is reported is a
disappointment. The correlation between predicted difficulty and
actual performance that is reported is −0.43 (negative being appropri-
ate here), and so this suggests there is room to improve the predictability
of difficulty with tasks. This is consistent with discussions elsewhere in
this book suggesting that linking task characteristics to predictable
language use, including task difficulty, is often extremely problematic.
It also connects with the discussions relating to Table 9.2, where a
somewhat different account of task difficulty is proposed.
But this point should not diminish the achievements of the Hawai’ian
research group. They have produced a wide range of tasks which
address a particular assessment problem in a systematic way. They have
demonstrated very impressive validity and reliability evidence. They
have also contributed in a major way to the development of rating
scales, and added to our understanding of the relationship between
adapted and general rating scales. The research was extensive and
practical, and has demonstrated how much effort is required to meet
conventional testing standards within a task-based framework. It was
essentially a collaborative research project, and this is the only way,
frankly, outside a testing organization, where such thoroughness could
be possible. What has been achieved is a resource which can be drawn
on by others. The project is a milestone in our field. It also, interestingly,
fits in very well with the task definition from Chapters 1 and 13:
meaning primacy; a goal which needs to be addressed; learners relying
on their own language resources; and a clearly defined outcome.

The Cambridge Main Suite of Tests

The Cambridge Main Suite is an important set of public examinations,
used all over the world. In fact it can be regarded as a set of separate
tests, each appropriate for a particular proficiency level; however, in
more recent years the relationship between the different tests has been
articulated more clearly and, in addition, the connection of these tests
to the CEFR has been clarified. The tests (and each is associated with a
CEFR level) comprise the Key English Test (KET) (CEFR A2, Way-
stage), the Preliminary English Test (PET) (B1, Threshold), the First
Certificate in English (FCE) (B2, Vantage) and the Certificate in
Advanced English (CAE) (Effective Operational Proficiency: C1).
(There is also the Certificate of Proficiency in English, CPE (Mastery:
C2), which will not be considered here, since it effectively functions at
close to native-speaker level.) The stakes are high with these tests.

They are available worldwide, and have to attain very high standards
of test construction, scoring and reporting. In addition, they need to be
available in multiple forms (and regular new forms) to ensure security.
They are, broadly, interactive-ability tests that try to assess more
general ability in English. Since they function from CEFR A2 Way-
stage up to C1, Operational Proficiency (and beyond, with C2, Mas-
tery), they have to deal with a very considerable range in ability.
Relatively recently they have undergone a major process of construct
validation and are related to Weir’s (2005) socio-cognitive model,
particularly in relation to cognitive and contextual validity factors.
The former largely derives from the application of the Levelt (1989,
1999) model, and the latter is concerned with setting factors (task and
administration) and demands (linguistic and interlocutor). Extensive
analyses are provided of how research relates to the different aspects
of the socio-cognitive model (e.g. Galaczi and ffrench, 2010), and how
these different factors are exemplified in the different tests within the
main suite (Weir, Vidakovic and Galaczi, 2014).
What is relevant here is whether the tests at the different levels, and
the subsections which comprise them, represent a task-based approach
to testing. To explore this question, we will focus here on the speaking
test component of the different tests (Galaczi and ffrench, 2010), and
the illustrations will be in terms of the publicly available material,
drawn from the Cambridge English website (Weir et al., 2014).
Broadly, it cannot be said that the details of what is done in the
different tests comes directly from the task literature, but at the same
time, a great deal of task research is drawn on when the sub-tests are
discussed, and familiar variables from the task literature are in evi-
dence. Galaczi and ffrench (2010, p. 170) suggest that complexifica-
tion within the main suite set of tests involves the following features:
• a move from controlled to semi-controlled to open-ended response
formats;
• a move from greater to lesser support (visual and otherwise);
• a move from familiar topics, with stronger examiner influence, to
more open-ended topics and more general topics;
• an increasing amount of time for each task type (note: not a differ-
ence in time pressure here, but more a need for more extended
language use);
• a move from factual to evaluative discourse;
• a move from persuasion and description to exposition and
argumentation;
• a move from personal and concrete information to non-personal
and abstract information.

Each of these bullet points is based on variables which, one way or

another, have featured in task research and been the basis for perform-
ance effects. So there is some degree of congruence between the
preoccupations and theorizing of task researchers and what is done
to develop greater complexity within the main suite tests. There are, of
course, other factors in task research not so prominently represented.
The distinction made earlier in this chapter between Conceptualizer
and Formulator influences on difficulty and complexity is not so clear.
Other variables, such as planning, are not represented (although
they are mentioned in relation to cognitive validity in the Cambridge
tests). Broadly, though, there are clear connections between the design
features evident with the tests and the approach taken within task
research.
A brief description of testing material for the speaking tests, taken
from the Cambridge English website, will make some of these issues
clearer. The KET, identified at A2 level, has two parts and is con-
ducted with an interlocutor and a pair of candidates. In the first part,
the interlocutor asks candidates fairly easy questions such as ‘Do you
think English will be useful for you in the future? What did you do
yesterday evening/last weekend?’ In the second part each candidate is
given information, such as about a museum or a bookshop, and the
other candidate in each case is required to ask questions about that
information. Brief prompt material is provided. The PET (CEFR B1)
has three sections. The first section is broadly similar to the KET. In
the second part there is more complex input material, entirely visual,
and this is the basis for the two candidates to talk to each other,
drawing on the visual suggestions, to formulate advice, for example,
on somebody’s visit to a city for the weekend. So the task is supported,
contextualized and has a clear outcome. But it also requires conversa-
tion management. The third part also has a visual prompt, for
example, requiring people to respond to pictures about things that
people are doing at home. The intention is to push candidates to
taking a longer turn. Each candidate completes the picture description
task, and then they have to talk to one another about what they
themselves do at home.
The FCE is regarded as B2 level, and has four parts. The first is
similar to the previous levels, in that the interlocutor asks questions,
although the focus for these questions is more wide-ranging, even
though the personal nature remains. The second also has similarities
to the KET, but in this case a degree of comparison is involved, as well
as the need to respond to more general and judgement-provoking
questions from the interlocutor. The third part requires interaction

based on input, which requires some processing (such as how to

attract tourists to a town), and then some reasoning and justification
in connection with this. Part Four in this example, also interactive,
develops the tourism theme but moves to holidays more generally and
issues connected with them, requiring the capacity to express opinions
and justify them. Finally, the CAE is regarded as C1 level, and like the
FCE has four parts. Part One, similar to FCE, is interlocutor-driven
and could be considered ‘extended personal’. Part Two is similar to
FCE Part Two, but perhaps with a little bit more justification in the
responses made to pictures. There is also greater prominence for
talking about the feelings and motives of those in the pictures. Part
Three uses a visual prompt (in the form of a splash diagram) and
requires conversation between the candidates about decisions, e.g.
choosing a university. The focus is not on the decision itself but on
the process of decision-making, the factors which are important and
how decisions can be justified. Part Four develops this theme, going
beyond the concrete and specific decisions covered in the third part,
and is concerned with decision-making in life generally. It is more
abstract, requiring the student to link this to their own personal
experiences and to offer more difficult justifications.
It is clear in all the subsections of the tests that language is being
elicited so that it can be evaluated. The examiner is clearly in charge
and orchestrates events either closely (Part One) or with some degree
of freedom for the candidates (Parts Two to Four). The candidates are
given every opportunity to display a range of language, likely to the
limit of what they can do, with this being partly in response to direct
questions from the examiner or in interaction with another candidate.
As the bulleted points indicate, there is also organized development in
what candidates are given the opportunity to do across the levels of the
main suite. It is evident that we are dealing with an interactive-ability
approach to testing, and that the formats provide many occasions
where ability for use needs to be brought into play. But this analysis
does not, in itself, establish ‘taskiness’. At the very least, there is a
strong communicative dimension to all the encounters in the speaking
test. Perhaps the most prominent and obvious connection with tasks
comes at the lower levels (KET and PET) where there are clearer
outcomes for the tasks which are given (cf. the museum and bookshop
tasks, the visit to a city). At higher levels there are structured encoun-
ters, interaction, longer turns and the opportunity to use a range of
language abilities, but not the same focus on task achievement as at
lower levels. This is obviously good in provoking wider sampling and
opportunity to stretch language, and so may be inevitable in testing at
higher levels. Perhaps the one other thing of note is that when the

examiner is involved, there is restriction in the spontaneity of what is

said. Interlocutor contributions are scripted, and so there can be little
doubt that unequal discourse is involved and lack of scope for the
candidate to shape what is said. But this may well be inevitable if
challenges to fairness and equal opportunities are to be met. All in all,
this range of tests does indicate that, where the testing of speaking is
concerned, there is considerable prominence for the use of tasks in a
very widely used public examination. To return to a point made
earlier, the washback that is likely from these tests is vital: one
could not prepare for these tests by learning lists of grammar or
vocabulary items.

Developing Task-Based Tests to Accompany a Coursebook

Major coursebook series have been one of the most significant devel-
opments in the teaching of English as a Foreign Language (EFL), and
since the late 1970s such coursebook series have changed in interesting
ways. They are now far more comprehensive, and typically contain a
considerable quantity of integrated supplementary material, much of it
delivered through multimedia. In addition, the ‘input’ to the classroom
teacher for each lesson or sequence of lessons is far more extensive
now, with the result that the teacher has, arguably, less of a develop-
mental, or mediating or adapting role. Such coursebook series tend to
claim communicative approaches almost routinely, but importantly
often contain a lot of more traditional material.
It is intriguing that, despite the massive financial commitments made
by publishers towards such series, one area has remained surprisingly
unchanged – that of assessment. Where coursebook series provide
testing materials, even when major claims are made towards commu-
nicativeness, the testing materials are remarkably traditional. In view
of that, we will go back some years in the case study in this section,
and examine the testing materials from the COBUILD course (Willis
and Willis, 1988), a series consisting of three volumes (at three profi-
ciency levels), which claims to be task-based (with important lexical
connections also). The series is distinguished by having a separate
testing book (Fried-Booth, 1989), and so this provides an interesting
opportunity to explore to what extent the testing materials respond to
the nature of the materials themselves.
The testing book for the three-volume series contains ten tests,
three for the first volume, three for the second and four for the third.
The tests have a standardized format, and contain sections on vocabu-
lary and pronunciation, grammar, reading, writing and listening.
In addition, at the third level the reading tests sometimes contain

material on dictionary skills. No statistical information is provided on

any of the tests concerned.
The vocabulary, pronunciation and grammar sections of each test
are fairly traditional in nature, with multiple choice, fill in the blanks
and matching formats. The only additional thing to say here is that the
COBUILD course has vocabulary features (linked to a Birmingham
University corpus of spoken and written English), with the course
materials written to incorporate designated (and important) sets of
words that arose from this corpus. The tests then focus on this mater-
ial, and so there can be claims of content validity. The reading and
listening components of each test can be described as communicative,
rather than task-based. With reading, interesting and challenging
material is presented. This often has a magazine-article quality, and
then meaning-oriented questions are asked about such material. The
focus is very much on meaning, although on occasions there are
sections which ask for synonyms to be found from within the reading
passages. So while the material is real-world based, and the questions
asked are genuinely meaningful, these are not real-world tasks. The
same can be said about the listening tests. These are based on record-
ings, where the test item may range from choosing amongst multiple-
choice illustrations, to filling in blanks in a form, to making choices
from several statements. The listening material is given at a reasonable
but not excessive pace, and the decisions that have to made are totally
meaning-focused, but once again, it is difficult to claim a real-world
relationship. The most task-based section of the various tests is con-
sistently the one on writing. Here the test contains similar sorts of
tasks to those which figure in the Hawai’ian project, with context,
input materials and a specific need to communicate something which
has a real-world connection. Even so, some of the time there are
transformation exercises and more communicative activities which
are more generic in nature. But compared to the other sections, criteria
of ‘taskiness’ are more clearly met.
On balance, then, the COBUILD tests are more communicative than
completely task-based. Even so, it should be said that, compared to
other major coursebook series, they are more concerned with tasks
and communication than most. Most such series claim to be communi-
cative, and even task-based, but the mismatch between professed aims
and any associated testing material is much greater. If, then, we are
concerned with washback, the COBUILD tests just reviewed do com-
plement, to a considerable degree, the aims of the broader series and
the units which compose it. Returning to the general task definition
(meaning primacy, gap to be addressed, learners use their own
resources, clear outcome), the tests in question go some way to

achieving this, and they demonstrate that task-oriented coursebook

series can achieve useful testing. But the match here is not as great as it
was with the Hawai’ian tests. The Cobuild tests are less convincing
with all of the individual criteria at some point or another. So we have
an interesting contrast between two sets of ‘task-based’ tests. On the
other hand, they do sample from the coursebook materials reasonably
effectively, and this may well be more important.

Task-Based Testing within an Educational System

The case studies we have examined so far concern what might be
considered to be elective contexts – in each case learners choose to
study a language. The fourth case is slightly different in that it con-
cerns a mandated education system, in Belgium. Colpin and Gysen
(2006) describe a whole series of task-based tests developed within this
context, for learners of various ages and for different contexts. They
also discuss many of the associated issues in developing task-
based tests.
The range of tests that were developed included (for more examples,
see Colpin and Gysen, 2006):
1. Six-year-old immigrant children who have finished kindergarten
and are entering primary education. The test is intended to indicate
to their new teachers their new classes’ Dutch language proficiency.
2. A test for twelve- to sixteen-year-old children assessing their Dutch
language proficiency at the end of their first year in reception in
Belgium. The test was designed to establish functional language
proficiency in Dutch.
3. An adult education setting, aiming at discovering whether a basic
level of Dutch L2 proficiency has been attained. It is used to decide
if non-native Dutch speakers have reached a level to enable them to
enter vocational training.
Many interesting and vital points emerge from the development of
these (and related) tests. The first of these is the importance of collab-
orative effort. The driving force for the development of these tests was
the Centre for Language and Education (CLE) in Brussels. This is a
government-funded centre, which then receives commissions for the
development of tests to achieve particular goals. The CLE has consid-
erable expertise in TBI and in task-based assessment. It was also
resourced sufficiently to produce such tests, which then became avail-
able within the Belgian education system for widespread use. So the
tests had impact, but they required an organizational structure and
adequate financing for their production. It was not teachers in

particular educational contexts who were able to produce the tests.

A second point which flows from this first one is that the test develop-
ment teams could bring considerable sophistication to bear in what
they did. The CLE team were very aware, for example, of the threats
to reliability in performance testing, and made strong efforts to coun-
ter sources of error in the test results that they obtained (such as taking
great care that instructions were clear and were understood, and
incorporating multiple tasks, to the extent that time pressures allowed,
to gain a more rounded estimate of performance). They were also
aware of validity-linked developments in the field, and so drew upon
the task-based literature in the way they tried to deal with task
difficulty, for example.
The test development team was also concerned with curriculum
issues. One consequence of this was a strong awareness of washback.
A related feature of their work was that research was conducted to
gather feedback from teachers on their reactions to a move towards
more task-based assessment. This research revealed clear reservations
on the part of teachers regarding what could be achieved if assessment
involved tasks. What this meant was that if washback was to be
positive, it was important to work with teachers so that they were
convinced that not using traditional testing techniques did not mean
effective assessment was not taking place. In other words, merely
introducing a more task-based approach to assessment is not enough.
For washback to be positive, so that approaches to teaching itself
become more task-based, working with teachers becomes essential.

An Evaluation of the Case Studies

The four case studies cover a range of language testing contexts. They
are united in the way that they demonstrate the feasibility of task-
based testing, with prospective university students, with general
learners looking for certification, with students following a major
coursebook series and with learners within a state educational system.
In each case test developers solved a range of problems, and produced
sets of tests which are task-based and which address the different
educational contexts. This alone is important, because it shows what
can be done.
But there are differences between the four case studies which go
beyond the obvious, the things such as differences in age, in first
language (L1)/L2 combination and in the uses of language that are
important in each case. One concerns the number of people involved
in test development. The Hawai’ian, Cambridge and Belgian tests were
the products of considerable collaborative work. In the first of these

cases, a set of researchers worked together; in the second, massive

numbers of specialists (system developers, task developers, production
departments, international administration systems, statistical experts)
all worked together to develop and stockpile many test tasks; in the
third, a centre was at the heart of test development, and it is clear that
teams of staff worked on the tests in question. The COBUILD tests
were a smaller scale affair, with one lead developer, plus the two
authors of the actual coursebook series. In addition, the Hawai’ian,
Cambridge and Belgian tests are associated with statistical investiga-
tion while the COBUILD tests do not seem to be. Even so, given the
content orientation of these actual tests, the argument for validity
is less statistical and more concerned with content itself. But the
different test initiatives do give some indication of the resources
required (principally time) to develop and validate task-based tests.
Connected with this, the Hawai’ian, Cambridge and Belgian teams
were more ambitious in what they attempted. In the first case a testing
system was developed which interestingly combines interactive-ability
and real-life elements. In the Cambridge tests, separate tests were
combined to function within a wider proficiency range, and were
essentially interactive-ability in nature. In the third, a series of essen-
tially distinct tests were produced, with a strong achievement connec-
tion but also with summative qualities. It is not surprising, therefore,
that larger numbers of test developers were involved.
All the approaches to testing, though, share some qualities. It is clear
that while all of them have a summative emphasis, each lends itself to
the generation of formative information which can feed back into
instruction. The tests generate numbers and these numbers can enable
important summative decisions to be made, but the information con-
tained through the test formats can provide clear indications of areas
of weakness, both for individuals and for courses of instruction.
Related to this, even though all of them generate numeric scores, they
also have a criterion-referenced element, in that completing the tests is
likely to provide direct information as to how language can be used in
the real world and with tasks which connect with real-world events.
These qualities, formative evaluation and criterion-referencing, are
significant strengths for the tests concerned.

Conclusion
This chapter has covered a considerable amount of work in task-based
assessment. We have seen that task-based research is highly relevant to
the area of testing, providing a framework and principled basis for
what would otherwise be a loosely focused communicative approach

to assessment. Of course, a great deal more research is needed, but it is

already clear that task research is relevant to the design of assessment
procedures and to the interpretation of results. There is also the
important claim that a task-oriented view of assessment may go some
way to reconciling the interactive-ability and real-life approaches by
showing how tasks can be based on researched characteristics and
conditions, and by analysing the processing demands required for task
completion. This may provide a better basis for making test-linked
generalizations.
We have also seen some impressive practical developments in task-
based assessment, and these have shown what can be done. But these
successful practical initiatives also bring out the need for adequate
resourcing if task-based tests are to be constructed and validated.
Testing is not cheap, and so major developments are most likely to
be associated with situations where resources are available. It is to be
hoped that consortia, such as that described in task-based testing in
Belgium, can go some way to mitigating the importance of ﬁnance
and resourcing. This would start to address the point made earlier
in this chapter – that the single most important area for development
is progress in the construction of task-based achievement tests
(Long, 2015).

By and large, the research we have addressed so far has been theoret-
ically motivated, carried out by researchers and concerned with the
performance of individual teaching or assessment tasks. Clearly,
though, there is a need to investigate complete task-based programmes
from the perspective of teaching, the stakeholders involved in them
and the learning that results. This section seeks to do this by address-
ing the following key questions:
1. How effective is task-based language teaching (TBLT) in compari-
son to more traditional approaches?
2. How can TBLT programmes be evaluated?
3. What do evaluations of TBLT programmes tell us about their
viability and effectiveness?
4. What practical problems arise in introducing a TBLT programme
and how can these be addressed?
Chapter 10 seeks an answer to the first of these questions by
reviewing comparative method studies. These are studies that have
investigated the relative effectiveness of TBLT vis-à-vis some other,
more traditional approach such as presentation–practice–production
(PPP). Such studies are notoriously difficult to design and often suffer
from design flaws. Overall, to date the studies do point to the superior-
ity of TBLT in a variety of instructional contexts – e.g. state schools in
India and English-for-specific-purpose courses in the United States –
but there is a clear need for more studies. The chapter concludes with
some guidelines for the design of future studies.
Chapter 11 seeks answers to Questions 2, 3 and 4 by evaluating
(rather than ‘researching’) TBLT programmes. The studies that have
undertaken this are not comparative in nature and do not aim to
contribute to the theories that underpin TBLT. Rather they are
practice-oriented, addressing whether TBLT ‘works’ and what might
be done to make it work more effectively. The chapter begins with a

281

discussion of TBLT as an innovation and illustrates how the general

factors that are known to affect the success of educational innovations
can be used to predict and explain the success or failure of particular
TBLT programmes. The chapter then reviews a number of actual
evaluations – both macro-evaluations where complete TBLT pro-
grammes were studied and micro- evaluations where teachers (not
researchers) carried out evaluations of speciﬁc tasks in their own
classrooms. Drawing on these evaluations, the chapter ends with a
discussion of the major problems that teachers face in implementing
TBLT and also suggests some solutions.

Introduction
Long (2015) distinguished two kinds of programme evaluations.
Descriptive studies aim to investigate whether a particular programme
is achieving its goals and whether it should be abandoned or, more
likely, modiﬁed in the light of the evaluation. Such studies consider
programs in situ and focus on the nature of the instructional processes
that result from the implementation of the programme and on students’,
teachers’ and stakeholders’ evaluations of it. In contrast, experimental
studies are comparative in nature and examine the learning outcomes of
the different approaches by means of tests. Ideally, they have a pre-test/
post-test design and randomly formed groups. However, as Long
acknowledged, true experimental designs are often not possible in insti-
tutional settings as it is usually only possible to use intact classes. In this
book we have elected to examine these two types of programme evalu-
ations in separate chapters. In this chapter we focus on studies that
have compared task-based language teaching (TBLT) with some other
teaching approaches and have brieﬂy included a post-test. In Chapter 11
we will consider descriptive programme evaluations.

Comparative Method Studies

Comparative method studies seek to determine which of two or more
language teaching approaches is most effective. Such approaches are
notoriously difﬁcult to implement and many suffer from a number of
design problems. Ellis and Shintani (2014) listed the design features of
a sound method comparison study (see also Long (2015)):
1. Pre-tests are needed to establish that there are no differences in the
groups taught by the different methods at the outset.
2. The study should include a control group that just completes the
testing regime in order to establish that learning takes place in the

283

experimental groups. Without a control group is not possible to

determine whether any gains from pre- to post-testing are simply
the result of a test practice effect or of general exposure to the target
language.
3. The study should establish through observation that the process
features which are evident when the approaches under study are
implemented correspond to the external descriptions of these
approaches. If there is no difference in the instructional processes,
then the study lacks internal validity.
4. The instruction should be carried out by the same teacher to guard
against the possibility that any learning differences which are found
are the result of variations in the teachers’ background and general
teaching abilities. It is also necessary to ensure that the learners in
the different groups are equivalent in background.
5. The study needs to ensure that the tests used to measure learning
outcomes are not biased in favour of one of the methods being
compared. It is, however, very difﬁcult to ensure content-fair
testing (Beretta, 1990).
6. Ideally, too, the study should investigate the effect of the different
methods on individual learners as well as on groups in order to see
whether individual learner factors mediate the effectiveness of each
method.
Few of the studies we will consider satisfy all these criteria. In the
concluding section we will evaluate the studies in terms of them.
We also acknowledge that the construct of ‘method’ or ‘approach’ is
problematic in several ways. Kumaravadivelu (2001), for example,
has critiqued ‘method’ on the grounds that it requires teachers to
simply implement a set of externally described prescriptions about
how to teach and thus disempowers them. He argued that teachers
need to be free to write their own ‘script’ rather than perform a script
given to them. Larsen-Freeman (2000) also argued that methods
cannot be understood as prescriptions for classroom behaviours and
imposed on teachers as a strict set of procedures to follow. The
method/approach construct is problematic in another way. No
method is monolithic and so narrowly prescribed that there are no
options for how it can be implemented. This is especially true of the
two approaches we will be examining in this chapter – presentation–
practice–production (PPP) and TBLT. Each stage of a PPP lesson can
be implemented in a variety of ways. For example, the presentation of
the target grammatical structure can be inductive or deductive. The
availability of options is equally true for TBLT, as was made clear in
Chapter 1. For example, there are different views about whether

learners’ attention should or should not be directed to form while they

are performing a task (see for example, Willis and Willis (2007) and
Long (2015)) or whether the tasks selected should mirror the target
tasks that learners need to perform in real life (see Chapters 1 and 7).
These problems have led some commentators to propose that
method comparisons should be discontinued. However, advocates of
both PPP and TBLT have argued the need for such studies. Advocates
of PPP such as R. Sheen (2006) insist that it is dangerous to propose
new approaches such as TBLT purely on theoretical grounds without
‘long-term trialling in normal classrooms’ (p. 273) and without evi-
dence that they result in superior learning to well-established
approaches such as PPP. While Sheen is wrong to claim there has been
no long-term trialling of TBLT (see, for example, Van den Branden,
2006), Long (2015, 2016), one of the main advocates of TBLT, has
also accepted the need for comparative studies.
Long (2015) argued that experimental studies should first be carried
out under laboratory conditions where it is possible to control poten-
tially confounding variables and then extended to classrooms to check
if the findings of the laboratory studies have ecological validity. How-
ever, we considered laboratory-based studies in Part II of the book, so
we have elected to focus only on classroom-based studies in this
chapter. Ultimately we believe that only studies carried out in class-
rooms will enable a true comparison of approaches that were designed
for the classroom. We acknowledge, however, that the findings of such
studies, even when well designed, are limited because of the difficulty
of controlling the cluster of variables that are inevitably present in the
holistic setting of a classroom.

Method Comparison Studies

Studies comparing TBLT and another approach are of two basic types.
In some studies, which we will call ‘general programme comparisons’,
TBLT is compared to some pre-existing programme. In these studies,
the TBLT programme constitutes an innovation – that is, TBLT has
been introduced into a teaching context where a more traditional
approach has been the norm. A good example of such a study is
Beretta and Davies’ (1985) comparison of Prabhu’s (1987) Communi-
cational Language Teaching Project (CLTP) and the well-established
Structural-Oral Situational Method. In such studies, it is common for
multiple classes to be allocated to each approach, all taught by differ-
ent teachers and with no random assignation of students to classes or
even of classes to type of instruction. Often there was no pre-testing of
the students; comparisons were made solely on the basis of post-tests.

In other studies, however, the comparison between TBLT and another

approach was built into the design of the study. That is, separate
groups were formed and the types of instruction to be compared were
devised based on explicitly formulated theoretical and methodological
principles (e.g. ‘focus on forms’ versus ‘focus on form’) and imple-
mented in accordance with them. These studies investigated the effects
of the instruction on the acquisition of speciﬁc linguistic features
(vocabulary or grammar) by means of both pre- and post-tests. We
refer to such studies as ‘focused comparative studies’. A good example
is Shintani’s (2016) comparative study of TBLT and PPP, which we
will discuss later. We will ﬁrst consider programme comparisons and
then focused comparison studies.

General Programme Comparisons

A characteristic of the four programme comparisons we will consider
here is that they all involved instructional programmes that were not
specifically designed for purposes of research – that is, the courses that
were compared were all institutionally scheduled. The contexts and
learners of these courses were very different – beginner-level learners
in high schools in India (Beretta and Davies, 1985), relatively
advanced learners of Spanish in a university programme for business
and economics in Belgium (de Ridder, Vangeheucten and Gomez,
2007), limited proficiency students in a Spanish course for border
control agents in the United States (González-Lloret and Nielson,
2015) and a computer-delivered programme for intermediate-
level high school students learning Turkish as a foreign language
(Arslanyilmaz, 2013). The obvious differences in these programme
comparisons preclude reaching clear conclusions but they can be seen
as affording insights into the relative effectiveness of TBLT in very
different instructional settings.
Beretta and Davies (1985) compared Prabhu’s (1987) CLTP (i.e.
TBLT) with the Structural-Oral Situational Method, a version of PPP
(a teacher-centred, form-oriented instructional method). The study
took place in India with beginner-level secondary school students.
The TBLT programme involved a series of communicative tasks with
no direct teaching of specific linguistic items. The PPP courses con-
sisted of the presentation and practice of sentence patterns through
drills and situational grammar activities with the underlying aim of
developing automaticity in their use. Beretta and Davies measured
learning through a battery of tests that included a test favouring the
TBLT experimental group (i.e. a task-based test), one favouring the
PPP comparison group (i.e. a structural test) and three neutral tests

(i.e. contextualized grammar, dictation and listening/reading compre-

hension tests). These tests were only administered at the end of the
school year (i.e. there was no pre-test). The results showed that
the task-based group performed better than the comparison group
on the task-based test whereas the comparison group performed better
on the structural test. However, on the three neutral tests, the task-
based group showed a clear advantage over the structural group.
Beretta and Davies argued that the task-based learners demonstrated
superior acquisition as even though these learners had not been expli-
citly taught specific grammatical structures they were able to deploy
the grammar they had learned more readily.
Ridder et al. (2007) compared two Spanish programmes for third-
and fourth-year university business and economics students, who were
assigned randomly to two groups. Both groups participated in course
components consisting of form-focused instruction (FFI) involving
presentation followed by practice with communicative activities
(apparently a version of PPP). The comparison group then worked
individually on compiling an individual dossier about twelve Spanish
companies followed by an oral test where they gave a brief presenta-
tion about these companies. The experimental group worked in
pairs in a series of task-based workshops directed at preparing an
advertisement for a brand-new project. Ridder et al. argued that
whereas the comparison’s assignment required them to use their
acquired knowledge in a similar context to the rest of the programme,
the experimental group needed to make creative use of their acquired
knowledge in a new context. Measurements of the learners’ oral
performance were taken at the end of the programme using ratings
of pronunciation, fluency, intonation, sociolinguistic competence,
lexical competence and grammatical competence. The experimental
group outperformed the comparison group on grammar, vocabulary
and social adequacy but there was no difference on fluency and the
comparison group performed better on pronunciation and intonation.
A problem with this programme comparison – acknowledged by the
authors – is that the final oral activity used to measure learning
outcomes was not equivalent; the learners in the comparison group
carried out the oral presentation individually with an examiner
whereas the data for learners in the experimental group were derived
from conversations they had with their partners.
González-Lloret and Nielson (2015) evaluated a newly introduced
task-based Spanish for Specific Purposes programme designed for
students preparing to become border patrol agents in the United
States. The project involved three empirical studies. The first included
twenty students from the TBLT programme, who were compared with

nineteen students who had participated in the previously taught

grammar-based programme. The participants in both courses com-
pleted an oral picture-guided narration task after they had finished the
course, which lasted for approximately 150 hours over seven weeks.
Their narratives were audio recorded and analysed in terms of fluency,
lexical complexity, syntactic complexity and grammatical accuracy.
The results showed that those students who had received the task-
based instruction (TBI) displayed significantly greater fluency and
syntactic complexity. However, there were no significant differences
between the two groups in lexical complexity and grammatical
accuracy. González-Lloret and Nielson argued that the TBLT course
was superior in developing those abilities that were the main aims of
the TBLT approach (i.e. oral fluency and complexity) and just as
effective in developing grammatical accuracy, which was the principal
aim of the traditional, grammar-based approach. The second study
was not comparative but we will consider it briefly here. A whole
cohort (n = 256) of TBLT programme participants took a computer-
scored oral proficiency test as pre-test and post-test. The test afforded
scores for overall proficiency, sentence mastery, vocabulary, fluency
and pronunciation. The results showed that participants improved
significantly in all categories and that the benefits of the instruction
were evident at all levels of proficiency. These two studies show that
the TBI resulted in all-round development of second language (L2)
proficiency and was more effective than the conventional structure-
based approach.
Arslanyilmaz (2013) investigated the comparative effects of
computer-delivered TBI and FFI by analysing the learners’ production
during the computer-mediated lessons. The researcher assigned thirty-
eight high school students in two intact classes to the TBI or the FFI
group and held seven 40-minute computer-mediated sessions over
seven days. The students in the TBI group engaged in two-way infor-
mation-gap tasks where pairs of learners interacted in a chat room to
complete the task. These learners were encouraged to refer to a video-
recorded model of a conversation between two native speakers for
guidance. The FFI group listened to the same audio samples of native
speakers as the TBI group but then completed controlled activities (e.g.
fill-in-the-gap activities) with the teacher. Arslanyilmaz (2013) audio
recorded and transcribed the lessons. The FFI class was ‘dominated by
the teacher’s talk’ (p. 310) with limited opportunity for students to
speak. An analysis of the learners’ production in terms of accuracy,
fluency and lexical complexity showed that the learners in the TBI
group were more fluent. There were no statistically significant differ-
ences between the two groups in lexical complexity or accuracy but

Arslanyilmaz argued that the TBI group scored higher in both aspects
and that the differences did not reach statistical significance only
because of the small sample size.
These studies have a number of methodological limitations. As
Beretta and Davies’ (1985) study was a post hoc evaluation, the
participants were from different age groups, and the length of instruc-
tion varied depending on the four schools involved in the study. Also,
as there was no pre-test, we do not know whether differences existed
between the groups before they started their instructional pro-
grammes. De Ridder et al.’s (2007) study seems a very ad hoc evalu-
ation; not only was there no pre-test but the learning outcomes were
measured using non-equivalent methods. González-Lloret and Nielson
(2015) carefully planned their studies, including a pre-post assessment
of the learners’ general oral proficiency as well as group comparisons
between the TBLT and grammar-based courses. However, as Gonzá-
lez-Lloret and Nielson acknowledged, the study was limited in a
number of ways, particularly the lack of a pre-test in the first study
and of a comparison group in the second. Arslanyilmaz’s (2013) study
only examined the learners’ speech production within the lessons and,
as such, did not address whether the TBLT instruction resulted in
better L2 acquisition. All in all, then, although they do provide evi-
dence of the effectiveness of TBLT, no clear conclusions about the
relative effectiveness of TBLT and a more traditional approach are
possible. In this respect, the focused comparative studies we consider
next are more insightful.

Focused Comparative Studies

Focused comparative studies are more theoretically oriented. These
studies compared two instructional approaches (TBLT and PPP) in
order to investigate two clearly deﬁned constructs – focus-on-form
(FonF), where the primary focus is on meaning but with attention to
form when communicative or linguistic problems arise, and FonFs,
where the instruction explicitly addresses predetermined linguistic
forms. We will examine four such studies: one is longitudinal (Sheen,
2005) and three are relatively short term (de la Fuente 2006; Shintani,
2013, 2015). As discussed here, Sheen’s study was unsatisfactory in a
number of ways, making clear conclusions impossible. However, the
three short-term studies afford clearer evidence in support of the FonF
approach.
Sheen (2005) compared the effects of a one semester-long (seven-
month) ‘strong communicative language teaching (SCLT)’ course with
a FonFs course involving PPP on the learning of two grammatical

structures (i.e. WH interrogatives and adverb placement). The learners

were grade 6 elementary French-speaking students in Canada aged
eleven and twelve. The PPP instruction consisted of explicit descrip-
tions of the target structures, controlled practice exercises and oppor-
tunities for free production. Different teachers taught the TBLT and
PPP classes. Sheen administered three tests – an aural written compre-
hension test, an oral interview (scored for correct use of the target
structures) and a grammaticality judgement test (GJT). However, he
presented results for only the latter two. In the oral production test,
which was conducted as a pre-test, post-test and delayed post-test, the
PPP group demonstrated some improvement in the two target struc-
tures whereas the TBLT group failed to show any improvement. In the
GJT, which the researcher administered at the end of the study, only
the PPP group improved their scores. Sheen concluded that the PPP
instruction was more effective than the TBLT instruction. The major
problem with Sheen’s (2005) study was that while he provided
detailed descriptions of the PPP instruction, he offered no account of
the procedures for the TBLT instruction except to say that the students
participated in ‘enjoyable tasks and game activities’ (p. 288) with
‘FonF opportunities’ (p. 289). However, there is no information about
how often the students in the TBLT class experienced focus on form
directed at the target structures and, in fact, Sheen noted that the
lessons offered few opportunities for corrective feedback on the use
of either target structure. In other words, it would seem that in fact
there was no or very little focus on form in the TBLT lessons. There
are also a number of design issues that are problematic. Sheen’s testing
clearly favoured the PPP group and none of the tests examined the
learners’ ability to produce the target structures in spontaneous com-
municative speech – arguably the best way of measuring the effect of
instruction. Also the two groups received instruction from different
teachers, with Sheen, an advocate of PPP (see Sheen, 1994), teaching
the PPP class.
We will now examine three short-term studies that investigated
TBLT: two of these examined vocabulary acquisition (de la Fuente,
2006; Shintani, 2013) and one examined incidental grammar acquisi-
tion (Shintani, 2015). These studies controlled instructional factors
more strictly and also provided more detailed descriptions of the
research procedures and the instruction. For these reasons they afford
more convincing results than Sheen’s study.
De la Fuente (2006) included process descriptions of the interactions
that took place in the FonF and FonFs instruction as well as measuring
learning outcomes. Thirty university students in an elementary Span-
ish class were divided into three groups: the PPP group (FonFs), the

TB-NEF group (task-based without explicit instruction) and the TB-EF

group (task-based with explicit instruction). The explicit instruction
was provided after the learners had completed the task, and this
focused on the morphological, phonological and spelling problems
the learners had manifested when performing the task. The TB-NEF
group simply repeated the same task. The lessons targeted fifteen
Spanish words, with which all the participants were unfamiliar.
The PPP learners received fifty minutes of instruction consisting of
the three phases of PPP – explanation of the new words (presentation),
controlled oral and written production exercises (practice) and a
role play performed in pairs (free production). The students in the
two TBLT conditions worked on a restaurant task in pairs, where
students needed to negotiate the meaning of the target words in order
to complete the task. The TB-NEF group repeated the same task
twice in the lesson, whereas the TB-EF group received the same
controlled exercises as the PPP group instead of repeating the same
task. The researcher measured acquisition using a discrete-item oral
production test.
The different types of instruction resulted in no statistically signifi-
cant differences in the immediate post-test. However, in the delayed
post-test, the two TBLT groups outperformed the PPP group. Exam-
ining the interactions that occurred in each group, de la Fuente (2006)
found that the TBLT instruction provided more opportunities for
learners to negotiate for meaning, produce the target words and
retrieve the target words online than did the PPP lessons. By examining
the process features of the instruction, de la Fuente was able to offer an
explanation of why the TBLT groups learned the target words better
than the PPP group. However, it could be argued that it was
the opportunity to produce the target words that can account for the
TBLT groups outperforming the PPP group. Without examining the
learners’ actual exposure to the target words or the opportunities to
produce them during the instruction, it is not clear that the inter-
actional processes de la Fuente identified were responsible for the
advantage found for the TBLT groups. The study also had some
methodological limitations. As de la Fuente admitted, the learning
outcomes were only measured by means of a discrete-item oral pro-
duction test and there was no control group. However, this study is
illustrative of the kind of process-product approach that is needed to
investigate the comparative effects of instruction.
Shintani’s (2013) study also included a process analysis. She ana-
lysed both the nature of the interactions that took place during
instruction, the extent to which the learners were exposed to the target
items in the input and whether they produced them in their output.

The learners were forty-five Japanese children, all aged six, with no
prior experience of any L2 learning. Shintani divided the learners into
three groups: FonFs (i.e. PPP), FonF (i.e. TBLT) and control. The two
experimental groups received nine 30-minute lessons over five weeks.
The PPP group completed five activities in one lesson, which followed
the three phases of PPP. The learners first saw the thirty-six target
words on flashcards with Japanese translations. The students then
engaged in ‘fun activities’ requiring the students to produce the words
repeatedly. The TBLT learners completed three input-based tasks that
required them to listen to the teacher’s commands and demonstrate
that they had understood them by selecting the appropriate flashcards.
However, the teacher used flashcards only for nouns; adjectives only
appeared in the teacher’s spontaneous utterances when she sought to
help the students to understand her commands and select the appro-
priate cards – for example, by referring to the size or colour of an
object. The thirty-six target words included twenty-four nouns
and twelve adjectives. The researcher measured both receptive and
productive knowledge of the target items in a pre-test, a post-test
conducted one week later and a delayed post-test conducted four
weeks later. The production tests included a discrete-point test and a
task-based test.
Both the PPP and TBLT groups outperformed the control group in
the acquisition of nouns. The TBLT group outperformed both the PPP
and the control groups in the acquisition of adjectives. The process
analysis demonstrated that although the PPP group produced the
target words during the lessons more frequently than the TBLT group,
the quality of their production differed dramatically. In the PPP group
initiation–response–feedback (IRF) exchanges predominated, so that
the learners’ production of the target items was almost always teacher-
initiated. In contrast, the TBLT group engaged frequently in student-
initiated negotiation where they were exposed to and in some cases
voluntarily produced both the nouns and the adjectives.
Shintani (2015) drew on Laufer and Hulstijn’s (2001) claims about
the importance of cognitive load for vocabulary learning to explain
her results. She used the process data to evaluate the ‘need’, ‘search’
and ‘evaluation’ of the use of the target words in the two groups, and
she concluded that the interactions that comprised the instruction
imposed a greater cognitive load for the adjectives in the TBLT lessons
than in the PPP lessons. Both the PPP and the TBLT instruction
involved ‘need’ (i.e. an externally imposed or self-imposed task
requirement to attend to a word). But whereas the TBLT group
engaged in ‘search’ (i.e. the attempt to find the meaning of an
unknown L2 word or to find the L2 word form expressing a concept),

the PPP group did not do so because the teacher explained the mean-
ing of the target words at the beginning of each lesson. ‘Evaluation’
(i.e. the comparison of a target word with other words and assessing
whether a word ﬁts its context) was required in the PPP but only in a
limited way when the learners responded to the teacher’s feedback.
The TBLT learners needed to engage more deeply in ‘evaluation’
because they were required to infer the meaning of an adjective when
the teacher elaborated her commands after the learners had failed to
comprehend her initial command, and also because the students used
adjectives to seek clariﬁcation when negotiating understanding a com-
mand. Shintani (2015) argued that this constituted ‘strong evaluation’
as it required generative rather than just selective choice. Table 10.1
summarizes the differences in the TBLT and PPP in terms of ‘need’,
‘search’ and ‘evaluation’. Shintani concluded that it was the difference
in the cognitive load imposed by the two types of instruction that led
to the TBLT group outscoring the PPP group for adjectives.
As part of the same study, Shintani (2015) also investigated the
incidental acquisition of two grammatical features – plural-s and
copula be –neither of which was explicitly taught in either the TBLT
or the PPP lessons. As with the vocabulary study there were nine
repeated lessons in the FonF (i.e. TBLT) group and in the FonFs (i.e.
PPP) group. The learners in both groups were exposed to multiple
exemplars of the two structures in the teacher’s utterances but were
not required to produce the forms. Acquisition of plural-s was meas-
ured by means of a task-based production test and discrete-point
comprehension and production tests. The acquisition of copula be
was measured by discrete-point and task-based production tests. The
results demonstrated that the TBLT group acquired plural-s but not
copula be. The learners in the PPP group acquired neither structure.
Analysing the classroom interactions, Shintani (2015) showed that
there was a functional need for the learners to attend to plural-s in
the TBLT classroom but not in the PPP classroom. Copula be is a
redundant feature and thus there was no functional need to attend to it
in either classroom. Shintani argued that the focus on form that
occurred quite naturally in the TBLT instruction enabled the learners
to distinguish the meanings of plural and singular nouns. In the PPP
instruction, however, the feedback only focused on whether the
learners had produced the correct noun form, which did not enable
the learners to make a form-meaning mapping for plural -s.
The focused comparative studies we have considered are a mixed
bunch and certainly do not provide conclusive evidence of the super-
iority of TBLT. Sheen’s (2015) study reported results that suggested
PPP was more effective than TBLT in enabling learners to acquire

Table 10.1 ‘Need’, ‘search’ and ‘evaluation’ in PPP and TBLT

PPP TBLT
Nouns and adjectives Nouns Adjectives
Need + Production was necessary to + Comprehension of the nouns + Comprehension of the adjectives
complete the activities. was necessary to complete the was useful to complete the
tasks. tasks.
+ The use of the adjectives was
motivated by the tasks.
Search – The pictorial image of the words + Engaging in or observing + Inferring the meaning of the
(word meaning) was provided. negotiation of meaning. adjectives.
Evaluation + Positive and negative feedback on + Choosing the correct noun card. + Feedback on the noun choices
production was provided. + Feedback on the noun choices was provided.
was provided. ++ Finding the appropriate
adjectives for negotiation.

Note: – : an absence of an involvement factor; + : moderate presence of an involvement factor; ++: strong presence of an involvement
factor.
Source: Shintani (2015, p. 58).
Comparative Method Studies 295

speciﬁc grammatical features but, as we have pointed out, the design

problems with this were so great as to preclude any conclusion. The
other studies by de la Fuente (2006) and Shintani (2015) were better
designed and point to a clear advantage of TBLT, especially where
longer-term learning is concerned. These studies also involved com-
parisons of the process features of the types of instruction being
compared and were able to show that active engagement by learners
in the TBLT lessons promotes acquisition. Clearly, though, more
studies that investigate learners with different levels of proﬁciency
and different target features, and that, crucially, are longitudinal, are
needed.

Conclusion
Tables 10.2 and 10.3 summarize the comparative studies we have
examined and also indicate to what extent they took account of the
key design factors that were outlined earlier in this chapter. The
programme comparisons are clearly very limited – as indeed the
researchers who conducted them admitted. They satisfy few of the
good design criteria for such studies. Only González-Lloret and Niel-
son (2015) investigated the instructional processes in the second of
their studies but did not investigate learning outcomes. Thus we
cannot be sure that there were in fact clear differences in the types of
instruction being compared or that these differences were responsible
for the learning outcomes. In these general programme comparisons
there were no pre-tests or control groups and teacher and learners
factors were not controlled for. Only group results were reported.
Only Beretta and Davies (1985) attempted to avoid testing bias.
However, these studies do at least point to the potential advantage
of TBLT over more traditional forms of instruction in authentic class-
room contexts.
The focused comparison studies varied in the extent to which they
satisﬁed the criteria for well-designed comparative studies. Sheen’s
(2006) study included a pre-test and also controlled for learner factors.
But it did not examine process features (except in anecdotal com-
ments), it did not control for the teacher factor, it only examined
group differences between the TBLT and PPP learning outcomes, there
was no control group and the tests were arguably biased in favour of
the PPP group. De la Fuente’s (2006) study included discussion of the
instructional processes but there was no pre-test, it did not control for
learner factors, only group comparisons were reported and only one
type of test was administered – a discrete-point test – which was biased
in favour of the PPP group, although, in fact, this group performed less

Table 10.2 General programme comparisons between TBLT and PPP

Study Beretta and Davies Ridder et al. González-Lloret González-Lloret and Arslanyilmaz
(1985) (2007) and Nielson Nielson (2015): (2013)
(2015): Study 2
Study 1
No. of participants 390 (4 schools) 68 256 39 28
Treatment duration 1–3 years 1 full academic 8 weeks 8 weeks 7 days (7
year sessions)
Target features General proficiency General General Oral CAF Oral CAF
proficiency proficiency
Measurements Task-based test: Ratings of oral Oral proficiency ■ fluency, lexical ■ CAF in oral
■ structural test production: test: complexity, syntactic production
■ contextualized ■ pronunciation ■ overall complexity and during the
grammar ■ fluency proficiency accuracy on an oral tasks
■ dictation ■ intonation ■ sentence production
■ listening/reading ■ sociolinguistic mastery
comprehension adequacy ■ vocabulary
■ vocabulary ■ fluency
■ grammar ■ pronunciation
Results Mixed but overall Mixed but TBLT > PPP TBLT > PPP TBLT > PPP
TBLT > PPP overall TBLT
> PPP
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.016
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:31, subject to the

Instructional No No No Yes No
processes examined
Pre-test No No No No No
Teacher factor No Not clear No No Not clear
controlled
Learner factors No Yes – randomly No No No
controlled formed groups
Testing bias avoided Yes Yes Yes (measured accuracy) No (only free
production)
Inclusion of a control No No No No No
group
Effect on individual No No No No No
learners as well as
on groups
examined
Teacher factor No Not clear Yes No
controlled
298 Investigating Task-Based Programmes

Table 10.3 Focused comparison studies

Study features Sheen (2005) De la Fuente Shintani (2013)

(2006)
No. of 48 30 45
participants
Treatment 6 months (4 hours 1 day 5 weeks (9 lessons)
duration per week) (50 minutes)
Target features Interrogative 15 words 36 words
form, adverb
placement
Measurements ■ oral interview ■ discrete-item ■ discrete-point
GJT oral production oral production
test test
classroom ■ task-based
interaction oral production
test
■ classroom
interaction
amount of
input and
output
during the
lessons
Results PPP > TBLT TBLT > PPP TBLT > PPP
Instructional No Yes Yes
processes
examined
Pre-test Yes Yes Yes
Teacher factor No Not clear Yes
controlled
Learner factors Yes Yes Yes
controlled
Testing bias No (only discrete No (only discrete Yes
point) point)
Inclusion of a No No Yes
control group
Effect on No No Yes
individual
learners as
well as
groups
examined

well than the TBLT groups on it. Both of Shintani’s (2013, 2015)
studies satisfied all the criteria.
De la Fuentes’ (2006) and Shintani’s (2013, 2015) studies point to
the superiority of TBLT for both vocabulary and incidental grammar
learning. However, caveats need to be acknowledged. First, these
studies operationalized TBLT in very different ways. De la Fuente,
for example, investigated oral tasks that required learners to produce
language spontaneously and negotiate for meaning to complete them.
Shintani’s studies, in contrast, employed input-based tasks where stu-
dent production was optional whereas her comparison (PPP) group
completed output-based tasks. Thus two aspects were confounded –
TBLT versus PPP and input-based versus production-based instruc-
tion. However, arguably this is inevitable given that TBLT can only be
implemented for complete beginners by means of input-based tasks.
As was pointed out in the Authors’ Preface and in Chapter 1, TBLT
does not constitute a monolithic approach. Different versions exist, as
is clearly evident in how TBLT was operationalized in the studies we
have examined. Thus care must be taken in not overgeneralizing from
the results of studies that investigated a particular version of TBLT.
PPP is also not monolithic and can be implemented in different ways.
Thus, all that these studies show is that the particular version of TBLT
they investigated led to superior outcomes than the particular version
of PPP with which it was compared. This reflects the inherent problem
with comparative method studies that we mentioned in the ‘Introduc-
tion’ to this chapter. It is difficult to generalize the findings of such
studies to make claims about the superiority of one type of instruction
given that ‘method’ is an external construct which will become mani-
fest in variable ways when it is implemented.
There is a final issue that needs to be considered. The well-designed
studies showed that TBLT was advantageous where the acquisition of
specific linguistic forms was concerned. That is, they constituted
examples of what Ellis (2012) called ‘local’ comparative studies. With
the exception of González-Lloret and Nielson’s (2015) study, they did
not investigate other aspects of L2 development – complexity,
accuracy and fluency (CAF), for example. Clearly, though, if it is to
be argued that TBLT is superior to traditional forms of teaching such
as PPP, it is necessary to provide evidence that it is not just superior in
enabling learners to acquire specific linguistic features but also super-
ior in helping the development of their overall proficiency. This means
that researchers need to investigate the incidental acquisition that
inevitably occurs in the exposure to the ‘input’, ‘output’ and ‘inter-
action’ that any type of instruction provides. One of the essential
claims of TBLT is that such incidental acquisition is enhanced when

learners perform tasks. Shintani’s (2015) study did compare the inci-
dental acquisition that occurred in the TBLT and PPP lessons but it
focused narrowly on two grammatical features. What is needed are
longitudinal studies that can inform about the relative effectiveness of
different approaches on general proficiency – in effect, carefully
designed programme evaluations. But as we have seen these are likely
to be plagued by design problems, in particular the difficulty of con-
trolling for all the variables that impact on classroom instruction.
Thus it may be that while comparative studies (especially longitudinal
ones) are needed to address the doubts of traditionalists such as Swan
(2005a) about TBLT, even then clear answers about the relative
effectiveness of the different approaches will be difficult to achieve.
The assumption of a comparative method study is that it can pro-
vide a generalized picture of which of two methods is the more
effective. We have seen that the studies that have compared TBLT
with PPP manifest a number of methodological problems. Perhaps the
biggest challenge lies in the very nature of comparative method studies
as they tend to (and perhaps must) prioritize ecological validity over
predictive validity. Implementing different approaches inevitably
involves a variety of factors that can potentially affect learning. Con-
trolling all these factors is not realistic, considering that the ultimate
goal of comparative method studies is to provide practical advice to
teachers. We can ask, therefore, whether there is a case for conducting
comparative method studies. We believe there is. In order to accept a
new teaching approach, teachers need to know whether it is likely to
be more effective than their current approach.
How, then, can we ensure that comparative method studies are well
designed and also informative for teachers? Ellis (2012) suggested that
researchers should design ‘local’ comparative method studies involv-
ing relatively short periods of instruction and focused on specific
linguistic features rather than ‘global’ studies involving long periods
of instruction and assessing general language proficiency or achieve-
ment. We have seen that those researchers who conduct longitudinal
studies have struggled to control for the various factors that can
influence outcomes. Although many of the short-term studies had the
same methodological problems as the longitudinal ones, a few have
provided some clear findings. These studies were premised on the
assumption that externally defined constructs (for example, FonF
and FonFs) result in different processes when they are implemented
by teachers in a particular pedagogical context. When a comparative
method study includes an examination of how the methods were
implemented and how the learners responded to them, the results are
arguably more convincing and more informative. The analysis of the

process features in de la Fuente (2006) and Shintani (2013, 2015)

demonstrated that when learners have the opportunities for FonF
while engaging with tasks it has advantages over PPP.
Arguably then, it is the comparison of process features that is most
revealing. As Kumaravadivelu (2001) pointed out, for ‘language peda-
gogy to be relevant, it must be sensitive to a particular group of
teachers teaching a particular group of learners pursuing a particular
set of goals within a particular institutional context embedded in a
particular sociocultural milieu’ (p. 538). Seedhouse (2005a) empha-
sized the difference between the task as ‘workplan’ and the task as
‘process’ (i.e. what is actually happens in a particular teaching con-
text). Larsen-Freeman (2000) suggested that methods should not be
understood as prescriptions for classroom behaviours and imposed on
teachers as a strict set of procedures to follow. We need to develop an
understanding of how different manifestations of TBLT result in dif-
ferent classroom processes and how these processes affect the learning
that takes place.
We will conclude this chapter by reviewing the guidelines for con-
ducting comparative studies and the extent to which the studies sum-
marized in Tables 10.2 and 10.3 conformed to them.
• Studies should examine the process features of instruction (what
happens in the classroom) along with the product of the instruction
(what the students learned). Shintani (2015) and de la Fuente (2006)
attempted to do this.
• Studies need to control for factors other than instructional vari-
ables – such as the teacher and learners. Controlling for teacher
factors can be achieved by ensuring the same teachers instruct both
groups of learners, as in Shintani’s (2005) study. The learner factors
can be controlled for by conducting pre-treatment tests as in de la
Fuente (2006) and Shintani (2005), to establish the equivalence of
the two groups being compared.
• Studies should guard against test bias. Four of the studies only used
discrete-point tests, which are likely to favour explicit types of
instruction. Arslanyilmaz’s (2013) assessed learners’ production in
free production activities, which showed that TBLT instruction was
superior. Three studies used multiple tests to guard against bias.
According to Beretta and Davies (1985) , TBLT instruction pos-
sessed advantages not only in the test that was biased in its favour
but also in tests designed to be ‘neutral’. González-Lloret and Niel-
son (2015) included ﬂuency and complexity measurements that they
consider favour TBLT and accuracy measurements in the grammar-
based programme. Shintani (2005) demonstrated that the two

groups showed similar gains in both types of tests for nouns but that
the TBLT outperformed the PPP in all tests for adjectives.
• Studies should include a control group. Most of the studies looked
at in this chapter did not include a control group; as such, it is only
possible to interpret the results in terms of comparative effects, not
in terms of either method’s effectiveness. The only exceptions were
Shintani’s studies.
• Studies should examine incidental acquisition (i.e. they should not
just measure the learning of the linguistic items targeted by the
instruction but also the acquisition of non-targeted features that
arise naturally as the different approaches are implemented). Only
Shintani (2005) did this.
Overall, the comparative method studies do indicate that TBLT is
more effective than traditional approaches. However, given the prob-
lems with comparative method studies, greater insight as to the
strengths and limitations of TBLT and PPP might be achieved through
descriptive programme evaluations. Such evaluations focus on com-
plete courses. They examine the instructional processes that arise
when a particular approach is implemented more narrowly and they
often elicit the subjective evaluations of the participants. Descriptive
programme evaluations may also help our understanding of how
different versions of TBLT result in different classroom processes. In
Chapter 11 we review descriptive evaluations of TBLT.

Introduction
For many teachers task-based language teaching (TBLT) constitutes an
innovation. This is especially the case in instructional contexts where
the established approach to teaching involves a structural syllabus,
explicit instruction and controlled practice exercises (i.e. what Long
(1991a) called ‘focus-on-forms’). Introducing task-based teaching into
such contexts is likely to pose a number of problems for both teachers
and students. It is for this reason that there is a need to move beyond
the kinds of controlled experimental comparative studies we con-
sidered in Chapter 10 to examine how teachers and students respond
to TBLT in specific instructional contexts. This calls for evaluation
studies rather than research.
Evaluation has fundamentally different goals from formal research,
which is directed at either testing or building theoretical positions. In
the case of tasks, the research has typically been experimental, often
laboratory-based, and studies have covered a relatively short period
of time (not even a complete lesson in most cases), although, as we saw
in Chapter 10, there have been some longer comparative studies (e.g.
Shintani, 2016). Skehan (2003) commented that ‘applications of research
findings do not really make sufficient connection with most classroom
decision-making’ (p. 9). Arguably, evaluation studies are better equipped
to speak to the issues that teachers and learners face as they constitute
‘a more encompassing and more contextually relevant approach’ (Norris,
2015, p. 28) where there is ‘systematic attention to the constellation of
factors that make up a learning environment’ (p. 35). Evaluations are
not concerned with theoretical issues (although they may still inform
about them). Rather they are directed at accountability (i.e. showing how
TBLT works under what circumstances and with what effects) and also
at development (i.e. identifying how aspects of the approach can be
improved for subsequent use in the same instructional context).

303

Norris (2009b) provides a brief historical sketch of educational

evaluation. Early evaluations of language programmes in the 1960s
and 1970s were summative in nature, aimed at demonstrating the
effectiveness of a method (often in comparison with another method)
by administering language proficiency tests at the end of the period of
instruction. By the 1980s, driven in part by funding agencies such as the
British Council, the focus switched to investigating the accountability
of specific training programmes and began to look beyond outcomes
and products by gathering evidence about the processes involved in
implementing a programme using a variety of data collection instru-
ments. In the 1990s a more pragmatic approach to evaluation emerged
that took account of the uses to be made of an evaluation study (see, for
example, Alderson and Beretta, 1992). Norris notes that this pragmatic
orientation continues to the present day. The evaluations of TBLT we
will consider in this chapter are essentially pragmatic in nature; that is,
they addressed how particular tasks or programmes were carried out
in particular instructional contexts, with what achievements and with
what problems.
First, though, we will consider what insights can be gained by viewing
TBLT as an innovation. Then we will describe different types of evalu-
ation. This will provide a basis for reporting the findings of empirical
evaluations of both complete TBLT programmes and of individual
tasks. The chapter concludes with a discussion of the problems of
implementing TBLT that these evaluations have identified, along with
suggestions for how they might be addressed.

TBLT as Innovation
One approach to evaluating TBLT is predicting whether it is likely to
prove a success when it is introduced into a particular instructional
context. This involves asking what factors are likely to impact on the
successful introduction of TBLT. Table 11.1 lists the general factors
that evaluations of innovations have found can inﬂuence their uptake.
Such a list is helpful in two ways. First, it can be used to explain why
an innovation was successfully implemented. Second, it helps innov-
ators to develop a good understanding about an innovation they are
planning, which, as Van den Branden (2009) noted, ‘constitutes the
crucial ﬁrst step in a process that may ultimately lead to the imple-
mentation of an innovation’ (p. 662).
Let us consider two different teaching contexts and apply the char-
acteristics shown in Table 11.1 to try to explain why one innovation
did in fact prove successful while a second projected innovation is less
likely to succeed.

Table 11.1 Factors determining the success of innovations

Attribute Deﬁnition
Initial The level of dissatisfaction that teachers experience with some
dissatisfaction aspect of their existing teaching.
Feasibility The extent to which the innovation is seen as implementable
given the conditions in which teachers work.
Acceptability The extent to which the innovation is seen as compatible with
teachers’ existing teaching style and ideology.
Relevance The extent to which the innovation is viewed as matching the
students’ needs.
Complexity The extent to which the innovation is easy to grasp.
Explicitness The extent to which the rationale for the innovation is clear
and convincing.
Trialability The extent to which the innovation can be easily tried out in
stages.
Observability The extent to which the results of the innovation are visible to
others.
Originality The extent to which the teachers are expected to demonstrate
a high level of originality in order to implement the
innovation (e.g. by preparing special materials).
Ownership The extent to which teachers come to feel they ‘possess’ the
innovation.

Source: From Ellis (1997, p. 29).

TBLT in Flanders
In this case we examine an actual project. The Educational Priority
Policy issued by the Flemish government in Belgium aimed at enhan-
cing the quality of Dutch language education at primary, secondary
and adult levels, with a view, in particular, to enable pupils at risk and
adult immigrants to beneﬁt from the educational and occupational
opportunities open to them. The predominant approach at the begin-
ning of the project was teacher-centred and audiolingual. Responsi-
bility for introducing TBLT was assigned to the Centre for Language
and Education at the Katholieke Universiteit of Leuven, which under-
took the design of programmes and the training of teachers. Members
of the centre were available to work with large schools teams, educa-
tional counsellors, policy makers and educationalists to support the
step-by-step introduction of TBLT (Van den Branden, 2006).
There are a number of factors listed in Table 11.1 that can explain
the success of this innovation. There was a general dissatisfaction with
the audiolingual approach, teachers were helped to see that TBLT

matched their students’ needs, the rationale for the innovation was
made clear to teachers, it was introduced step by step and teaching
materials were provided, reducing the level of originality required by
the teachers. Publications relating to the project have shown that the
introduction of TBLT in this context was largely successful (see Van
den Branden, Van Gorp and Verheist, 2007).

Elementary Schools in Japan

In this case we describe a context where TBLT has not yet been
introduced but might well be. English was introduced in elementary
schools in Japan in 2002 and became compulsory for 5th and 6th
grades in 2006, but only as a ‘foreign language activity’. In 2020 it will
become a ‘school subject’ with designated textbooks and formal
assessment (Aoki, 2016). A communicative approach has been man-
dated. It is expected that elementary schoolteachers will take responsi-
bility for teaching English although it is likely that many will lack oral
proficiency. In all probability, too, the teachers will have had no
experience of learning English through TBLT.
An inspection of Table 11.1 suggests that these teachers will struggle
to implement TBLT. While it may be the case they were dissatisfied
with the traditional approach they experienced in their own language
education (a positive factor), there are many probable negative factors.
For example, implementing TBLT in the conditions that the teachers
work in will be difficult, TBLT may not be seen as relevant if the methods
of assessment are the traditional ones common in Japan and teachers will
need to demonstrate a high degree of originality unless teaching materials
are made available to them. Thus, although TBLT is the ideal approach
for introducing English at the elementary level in Japan (Ellis, 2017b), it
is questionable whether it will be successful unless substantial support is
provided. Carless’ (2004) evaluation of the introduction of TBLT into
Hong Kong elementary schools (discussed later) found many problems
even though, in many respects, Hong Kong constitutes a more favourable
context for its introduction than does Japan.
These two contexts have some similarities but also some conspicu-
ous differences. In both cases, dissatisfaction with existing approaches
provides the platform for innovation and in both innovation was
officially mandated. The main difference lies in the level of support
teachers receive. In Japan it remains to be seen what kinds of support
will be provided – but we are not optimistic – whereas in Flanders
external support and a communication network were established from
the start, enabling the teachers to develop a sense of ownership with
regard to the innovation.

Viewing TBLT as an innovation and applying the criteria for

successful uptake of an innovation is instructive because it allows for
the effective planning of a TBLT programme. However, as Van den
Branden (2009) pointed out, ‘the power of innovations should be
measured by whether learners are making more progress than before
an innovation was implemented’ (p. 669). This calls for the evaluation
of programmes. In the rest of this chapter we will consider reports of
actual evaluations of TBLT. We begin with a consideration of what
evaluation involves and the different types of evaluation.

Types of Evaluation
Some of the studies we considered in Chapter 10 can be thought of
as evaluation studies. For example, Long (2015) and Norris (2015)
consider Beretta and Davies (1985) an evaluation study. However, this
evaluation did not account for the holistic, pragmatic and program-
matic nature of TBLT that is the goal of evaluation studies. In this book
we have elected to remain more faithful to the aims of evaluation studies
by distinguishing those studies that are purely product-focused – such as
Beretta and Davies’ study – which we considered in Chapter 10 and
evaluation studies that address how TBLT is implemented in this
chapter.
Evaluators have in general been concerned with the evaluation of
whole programmes – see, for example, Norris (2009b) and Alderson
and Scott (1992). R. Ellis (2011), however, pointed out that such
evaluations are not necessarily in accord with teachers’ ideas of what
evaluation involves. He suggested that teachers are less likely to
focus on whole courses or programmes and more likely to be interested
in whether speciﬁc activities ‘work’ in the context of a particular
lesson. A distinction can be made, therefore, between macro- and micro-
evaluations. Figure 11.1 shows the different types of evaluation based
on this distinction. The broken line indicates that macro- and micro-
evaluations are potentially connected.
A macro-evaluation can be carried out: (1) to establish to what
extent a programme/project is effective and efﬁcient in meeting its
goals, or (2) to identify in what ways it might be improved. This is
what Weir and Roberts (1994) refer to respectively as an ‘accountability
evaluation’ and a ‘development evaluation’. Applied to TBLT, a macro-
evaluation looks at a complete TBLT course. Most evaluations of
language teaching have involved macro-evaluation. In early evaluations
(e.g. Alderson and Beretta, 1992; Weir and Roberts, 1994; Kiele and
Rea-Dickens, 2005, and the special issue of Language Teaching
(vol. 13.1 on ‘Understanding and improving language education through

Accountability-oriented

Macro -evaluation (summative)

Development-oriented
Evaluation

Action research

Micro -evaluation (formative) Exploratory practice

Task-focused

Figure 11.1 Types of evaluation

program evaluation’) reports of macro-evaluations dominated but few

concerned TBLT. In Alderson and Beretta (1992), for example, there
was only Beretta’s own study of Prabhu’s TBLT project. Subsequently,
reflecting the increased interest in TBLT since the 1990s, there have been
a number of macro-evaluations of TBLT programmes (see, for example,
Carless, 2004; Van den Branden et al., 2007; East, 2012; Hu, 2013;
Nielson, 2014).
A micro-evaluation has a narrow focus on some specific aspect
of the curriculum. Probably the most well-known type is action research
(Burns, 2010) but we will also consider exploratory practice (EP)
(Allwright, 2003, 2005) and an approach that takes as its starting point
a specific task (Ellis, 2015a). Micro- evaluations have investigated such
issues as the motivational value of using tasks (Loumpourdi, 2005),
how to incorporate tasks into lessons based on traditional textbook
materials (Muller, 2005), the extent to which it is possible to predict
the language that will be needed to perform specific tasks (Cox, 2005),
the effect of students’ reporting the outcome of a task publicly on the
quality of their language output (Johnston, 2005) and the effect of
training students to use meaning-negotiation skills on task performance
(Lee, 2005).1 While such evaluations often draw on the theories that
inform task-based research, they reflect the specific questions that
teachers are interested in and are action-oriented.
The distinction between macro- and micro-evaluations correlates in
part with another distinction – that between summative and formative
evaluation. In a summative evaluation, a programme is assessed on its

completion to decide whether it should be discontinued or continued

and in the case of the latter what modifications are needed. In a
formative evaluation, a programme is assessed during its development,
with a view to providing information that can be used to modify it as it
continues. Winke (2014) identified the key aspects of formative evalu-
ation: students’ progress is monitored regularly based on work
they complete in class, it aims to provide continuous feedback to both
the teacher and students, and it encourages reflection. A summative
evaluation is typically macro in type while a formative evaluation is
micro. A summative evaluation is generally conducted by an outsider
while a formative evaluation is more likely to be undertaken by a single
teacher or perhaps collaboratively with other teachers. A formative
evaluation will necessarily involve the evaluation of individual lessons
and, perhaps also, individual tasks. A summative evaluation can involve
the collection of evidence not just at the end of a programme but
throughout its implementation. Thus, a series of micro-evaluations
carried out for formative purposes can contribute to the final summative
evaluation.
Irrespective of the type of evaluation, effective evaluations need to
collect a variety of evidence – documentary, qualitative and quantita-
tive. R. Ellis (2011) suggests that an evaluation can be student-based,
response-based or learning-based (or any combination of these).
A student-based evaluation collects information about how the stu-
dents responded to the teaching (i.e. how engaged they were) using
self-report methods such as rating slips, questionnaires, interviews,
focus groups and written commentaries. Similarly, a teacher-based
evaluation obtains information about teachers’ perceptions about
TBLT. A response-based evaluation addresses whether the students
responded to the task-based activities as intended by the designer of
the materials or the teacher(s) using them and whether the students
were successful in achieving the intended outcomes of the tasks. It
involves observing or recording task-based lessons and also collecting
documentary records of task outcomes. A learning-based evaluation
looks for evidence that some learning has taken place in terms of new
linguistic knowledge and greater control over existing linguistic
resources or discourse skills. It may require the administration of tests
but evidence of learning can also be obtained through the detailed
analysis of lesson transcripts.

Macro-Evaluations of TBLT
Table 11.2 provides a summary of a selection of macro-evaluations
of TBLT. These evaluations cover TBLT programmes in a variety of

Table 11.2 A selection of evaluations of TBLT programmes

Study Context Purpose Evaluation type Findings

Beretta (1989, Secondary schools in To establish to what Response-based: Very mixed – most of the
1990) southern India extent TBLT was being recordings and teachers not directly
Language – English (Prabhu’s implemented in observations of involved in the project
Communicational accordance with the selected lessons. failed to implement
Language Teaching CLTP’s principles. TBLT effectively due
Project (CLTP). to lack of oral
proﬁciency in English.
Carless (2004) Primary school To see how teachers were Response-based: data Overall, the tasks
Language – English classrooms in Hong attempting to collected by means of represented language
Kong. implement TBLT in classroom observation practice activities
their classrooms and of lessons and semi- rather than genuine
what their attitudes structured interviews communication and
were towards TBLT. with teachers. the teachers
manifested a poor
understanding of what
a task was. Teachers
expressed concern
about use of ﬁrst
language (L1) and
discipline (noise)
issues.
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.017
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:30, subject to the

Watson-Todd University students in To evaluate the Response-based by Various changes made to

(2006) Thailand – TBLT continuation of the inspecting the course over time;
Language – English course developed by TBLT course, documentation from fewer tasks, greater
teachers. including changes the course; teacher- emphasis placed on
made to it as it evolved based through explicit instruction and
from one year to the interviews and records increased emphasis
next. of end-of-year given to examinations
meetings to discuss the rather than continuous
course. assessment. These
reﬂected ‘a move away
from a “pure” version
of task-based learning
towards a more mixed
methodology’ (p. 9).
McDonough and Innovative task-based To investigate what Response-based and Learner independence
Chaikitmongkol course for students at teachers’ and students’ teacher/student-based: increased; teachers
(2007) Chiang Mai attitudes to the new a variety of data became more positive
Language – English University, Thailand, TBLT course were and collection instruments about lack of grammar
developed by teachers what concern teachers including task and over time; students
involved in the had about its course evaluations, recognized the real-
programme. introduction. observations by world relevance of the
teacher participants course. Changes
and researcher, and subsequently made to
interviews with course to address the
teachers. need to help students’
adjust, to provide
support through
supplementary
materials and to
reduce the number of
activities in each
lesson.
(continued)
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.017
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:30, subject to the

Table 11.2 (cont.)

Study Context Purpose Evaluation type Findings

Hu (2013) Introduction of TBLT in To investigate teachers’ Response- and teacher- Three levels of
Language – English Chinese elementary beliefs and practices in based: implementation of
and high schools. using TBLT in their classroom observations; TBLT identified: (1)
classrooms. interviews with denial (teachers
teachers; documentary continued to function
evidence (e.g. as examiners), (2)
instructional passive acceptance
materials). (teaching based on
textbook), (3) active
application (teachers
expressed positive
views about TBLT but
by and large
implemented only a
weak form of TBLT).
Nielson (2014) Online task-based course To investigate whether Student-based and Overall participation
Language – for high school the performance-based learning-based: declined except for
Chinese students in USA based assessments (PBAs) results of the PBA tasks; conversation partner
on needs analysis. built into the course Likert scale survey sessions; the PBAs
worked, whether the completed five times were effective in
students’ Chinese during the course. assessing students’
proficiency improved level of ability in
and how the tasks performing particular
could be improved. tasks; task completion
not an adequate
measure of task
performance.
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.017
Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:30, subject to the

González-Lloret Eight-week programme To investigate whether Learning-based and Overall Spanish oral
and Nielson for US Border Patrol the students’ Spanish student-based: proficiency improved
(2015) Agency based on needs proficiency improved pre- and post-tests of in both the more
Language – analysis. as a result of the task- students’ oral advanced and novice
Spanish based programme and proficiency; Likert students; students held
to gauge students’ scale surveys positive views about
opinions about the administered to both the course and believed
course. students in programme it had helped them to
and those who had left in their jobs.
and were in
employment as border
patrol guards.
314 Investigating Task-Based Programmes

languages (English, Chinese and Spanish) and instructional contexts

(elementary and secondary public schools, universities, specific-
purpose contexts, online and face-to-face in classrooms). There is a
focus on TBLT as an innovation, as in all the studies TBLT is a recent
introduction. For this reason, perhaps, most of the studies are con-
cerned with the extent to which TBLT was effectively implemented
in the different contexts and teachers’ attitudes to its introduction. As a
result, most of the evaluations were response and/or student/teacher-based
in their methodology. However, two of the most recent evaluations
(Nielson, 2014; González-Lloret and Nielson, 2015) were also learning-
based. The comparative aspect of González-Lloret and Nielson’s evalu-
ation was considered in Chapter 10; here we also consider information
about the student-based component of this evaluation. Data collection
methods in the evaluation studies listed in Table 11.2 were diverse,
including classroom observation, questionnaires and interviews, and tests.
We will first examine three of these evaluations in detail before attempting
a summary of their findings.

English in Indian Secondary Schools

To the best of our knowledge, the ﬁrst macro-evaluation of a task-
based programme was Beretta’s (1989; 1990) evaluation of Prabhu’s
(1987) CLTP. Beretta’s evaluation was a follow-up to Beretta and
Davies’ (1985) comparative study of the learning outcomes of the
CLTP and the traditional Structural-Oral Situational Approach which
we considered in Chapter 10. Beretta (1989) focused on teachers’ error
correction practices. He reported that ‘the treatment of linguistic error
was largely consonant with the project’s statements about the kinds of
attention that are appropriate to a focus on meaning and that this
could be distinguished from the ways of treating linguistic error that
are attributable to a focus on form’ (p. 283). However, he went on to
suggest that this might have been because the input tasks used in this
project meant that there was limited learner production, reducing the
likelihood of learners making errors and thus the need for teacher
correction. Beretta (1990) examined the historical narratives from
ﬁfteen teachers and rated them according to three levels of implemen-
tation of the task-based approach: (1) orientation (i.e. the teacher
demonstrated a lack of understanding of task-based instruction and
failed to implement it), (2) routine (i.e. the teacher understood the
rationale of the project and was able to implement it effectively), and
(3) renewal (i.e. the teacher had adopted a critical perspective and
could demonstrate awareness of its strengths and weaknesses). Beretta
found that 40 per cent of the teachers were at Level 1, 47 per cent at

Level 2 and 13 per cent at Level 3. However, when he distinguished

between regular and non-regular teachers involved in the project, he
found that three out of four regular teachers (yet none of the non-
regular teachers) were at Level 1. Beretta concluded that task-based
instruction of the kind practised in the project is not easily assimilated
by regular classroom teachers in southern India. He pointed to these
teachers’ lack of English proﬁciency as one reason for their failure
to adopt task-based teaching. Beretta’s qualitative evaluations of the
classroom processes evident in the CLTP suggested that it did not
consistently result in the kinds of teaching described in Prabhu (1987).2

English in Primary Schools in Hong Kong

Carless’ (2004) evaluation took the form of case studies of three
primary schoolteachers, who were described as ‘young, capable and
open-minded’. The context was Hong Kong, where a new ‘target-
oriented curriculum’ had been recently introduced into primary-level
schools, requiring a shift away from traditional, teacher-fronted
instruction to task-based instruction involving group work. Data were
collected through classroom observations (seventeen per teacher) and
interviews. Carless presented his evaluation in terms of three class-
room episodes (one for each teacher), which he used to illustrate his
general ﬁndings. In Episode A selected students came out to the front
of the class, were blindfolded and then tried to guess what fruit they
were holding. In Episode B students worked in groups to answer
questions about some photographs they were given. In Episode C the
class was divided into groups of six and then tried to draw a picture
based on the group leader’s description of what he could see when he
looked out of the classroom window. Carless reported that while two
of the teachers had a fairly clear idea of what a task was, one teacher
was vague, offering this deﬁnition: ‘mainly has objectives and it can
link pupil ability and understanding, conceptualizing, that kind of
communication’ (p. 648). The teachers varied in their views of the
mother tongue, one believing that the lesson should be conducted
entirely in English but the other two seeing some value in allowing
students to use the L1 (e.g. to work out how to do a task). The teachers
experienced tension between their desire to carry out the tasks and
their wish to maintain a quiet and orderly classroom, with the result
that ‘concerns over noise and discipline inhibited implementing task-
based language teaching’ (p. 656). Finally, Carless noted that too
much time was spent on non-linguistic activities such as drawing.
Carless concluded by suggesting that a task-supported approach might
be more effective in Hong Kong primary schools on the grounds that

‘if English language structures are not pretaught, then beginning

learners will probably not have sufﬁcient English to use during tasks’
(p. 658).

Spanish Special-Purpose Programme

González-Lloret and Nielson (2015) reported a comprehensive
evaluation of an eight-week Spanish specific-purpose programme
designed for the US Border Patrol Academy. TBLT was introduced
as an alternative to the existing grammar-based course. Along the lines
advocated by Long (1985, 2015), the course was based on a needs
analysis of the target tasks that border agents are required to perform
as part of their job. González-Lloret and Nielson (2015) conducted
three separate evaluative studies. The first, which involved a compari-
son of the grammar-based and TBLT courses, was considered in
Chapter 10. The second was a learning-based evaluation focusing on
whether the TBLT course was successful in developing the students’
oral proficiency in Spanish. The students completed a short oral
proficiency test prior to the start of the programme and on its comple-
tion. Overall proficiency scores improved by 7.47 points. There was
also a statistically significant improvement in vocabulary, fluency and
pronunciation. The results of the test further indicated that both the
most advanced students and novice students improved their level of
proficiency. The third study was student-based. Both in-service and
pre-service agents completed an electronic survey designed to investi-
gate their perceptions of the strengths and weaknesses of the pro-
gramme. The students reported they had enjoyed the course, they
found it relevant to their jobs and they expressed a desire to keep on
learning Spanish. However, many still felt they were not yet ready to
talk to native speakers. Their comments on specific aspects of the
course indicated their approval of specific task types such as role-
playing but also their disapproval of a video game designed to provide
an out-of-class learning experience. Commenting on this evaluation,
Norris (2015) noted that it illustrates ‘the diverse purposes to which
an evaluation can and should be put’ (p. 50). He emphasized that such
evaluations of TBLT programmes are important in ‘enabling a better
understanding of how they work’.

Practitioner Research: A Place for Micro-Evaluations

of TBLT
Macro-evaluations of whole programmes require careful planning and
are time-consuming to carry out. For these reasons they are typically

not carried out by teachers but by external evaluators. In contrast, the

micro-evaluations of tasks that we will consider in this section were
small scale and carried out by teachers in their own classrooms. They
are examples of practitioner research. We consider three types of
practitioner research that have been used to investigate TBLT – action
research, exploratory practice (EP) and the micro-evaluations of tasks.

Action Research
Action research (Burns, 2010) is probably the best-known type of
practitioner research. It involves a research cycle where teachers work
on investigating small-scale aspects of their own practice in their own
classrooms. The cycle involves a planning stage, an action phase where
the plan is implemented, an observation phase where data is collected
to investigate whether the plan has worked, and a reflection phase
which can lead to further critically informed action (Carr and Kemmis,
1986). Action research is often problem oriented, with the teacher
systematically researching solutions to a teaching problem that he/
she has identified.
A good example of an action-research study investigating TBLT is
Calvert and Sheen (2015). This study reports a teacher’s attempt to
design and use a task about housekeeping with low-literacy adult
refugees. The effectiveness of the task was evaluated by inspecting
the task outcome (a handout where the students had to enter infor-
mation about housekeeping duties). This indicated that most of the
students failed to complete the task successfully. Also, the students’
responses to a post-task questionnaire revealed they had a negative
attitude to the task. In the reflection phase of the study, the teacher
identified a number of factors responsible for the failure – in particular,
the learners’ lack of familiarity with the format of the task and
the linguistic demands the task placed on them. This led to several
modifications being made to the task, reducing the amount of infor-
mation the learners had to handle and simplifying the language. In
addition, a new pre-task activity was introduced where the teacher first
performed the task with the whole class, scaffolding the actions required
of the learners and the language involved. When the students performed
the modified task they were able to achieve the outcome successfully
and this time evaluated it more positively. This study is informative
because it reveals the kinds of problems that novice teachers have
in designing and teaching tasks and also the value of action research
in helping a teacher to solve initial problems. Another advantage of
conducting the action evaluation is that it led the teacher (Calvert) to
revise her initially unfavourable view of TBLT.

Weaver (2012) reported on an action-research study he carried out

in a Japanese university. The focus here was on a formative assessment
cycle, which Weaver incorporated into the implementation of his
task. This cycle involved a needs analysis, task description (using the
framework in Ellis, 2003), the development of assessment criteria
based on both performance processes and the task product, the actual
assessment of the students’ performance of the task, and the provision
of feedback directed at both the students and the teacher himself.
Weaver (2012) claimed that the formative assessment of the kind he
conducted ‘can help teachers establish a framework for systematically
implementing TBLT in their classrooms’ (p. 307).3
However, doubts have been expressed about whether action
research is the best way of conducting practitioner research. Formu-
lating clear research questions is not something that teachers always
ﬁnd easy (Nunan, 1990). Also, the requirement that it is cyclical places
a burden on teachers. It is difﬁcult to see many teachers carrying out
the study reported by Calvert and Sheen (2015), for example. Like
formal research, action research is also technicist (i.e. it requires
technical skills required to design the research and to collect and
analyse the data). Rainey (2000) noted that, despite teacher educators’
advocacy of action research as a way of creating a bridge between
theory and practice, it has not been widely adopted.

Exploratory Practice
An alternative approach is EP, which Allwright (2003, 2005) pro-
moted on the grounds that it is less likely to result in teacher/researcher
burn-out. EP is not directed at solving problems but at developing an
understanding of some aspect of ‘the quality of life’ in a speciﬁc class-
room by integrating enquiry into actual classroom practice. In EP,
students as well as teachers are practitioner-researchers. EP focuses on
investigating ‘puzzles’ rather than ‘problems’. These include puzzles
that learners acknowledge about themselves (e.g. ‘Why I don’t speak
English after nine years of study’) as well as about teachers (e.g. ‘Why
do teachers have no time to answer students’ questions’) . Allwright
(2005) listed the general principles that inform EP – e.g. ‘involve every-
body’ and ‘work for mutual development’ – and offered some practical
suggestions – e.g. ‘integrate the work for understanding into the
working life of the classroom’ (Allwright, 2003, p. 130). A detailed
account of how to conduct EP can be found in Allwright and Hanks
(2009) along with narrative accounts of EP in practice.
Slimani-Rolls’ (2005) study may not qualify as fully-ﬂedged EP but
it illustrates a number of the principles that inform it – in particular,

incorporating data collection as part of the instructional activities of

the lesson and emphasizing the contribution that learners can make to
both their own and their teachers’ understanding. Slimani-Rolls was
interested in researching how task design factors affect the negotiation
of meaning and the conversational adjustments that arise during it,
which, as we saw in Chapter 2, has been the focus of a number of
research studies. She investigated three tasks – a one-way information-
gap task, a two-way information-gap task and a decision-making task.
A quantitative analysis confirmed the results of other studies, namely
that the two-way task resulted in the most conversational adjustments.
But Slimani-Rolls also noted that the individual students differed
widely both within the same task and across the three task types. She
included the students in her investigation by asking them to explain
their motives and attitudes to engaging in the negotiation of meaning
and found that their behaviour was highly idiosyncratic and unpredict-
able, reflecting a host of affective, social and cognitive variables. The
students very clearly had their own ideas about how to handle the tasks.
Slimani-Rolls concluded by emphasizing that task-based research can
benefit from treating a task performance as a puzzle which can be best
understood through teacher and students engaging collaboratively in
practitioner enquiry.
Allwright (2005) conceived of EP as involving long-term action
on the part of the teacher and students. One of his key principles is
‘Make the work a continuous enterprise.’ (Allwright, 2003, p. 130). It
is therefore perhaps not best suited to the in-depth micro-evaluation of
tasks. I turn now to an alternative form of micro-evaluation – one that
takes as its starting point a specific task and one that has potential to
shed light on how tasks open up spaces for teacher development.

Micro-Evaluations of Tasks
A micro-evaluation of a task addresses a simple question ‘Does a task
work?’ Ellis (2015a) outlined three general ways of addressing this
question corresponding to the three general types of evaluation men-
tioned earlier in this chapter – student-based, response-based and
learning-based. A student-based evaluation is the easiest to carry out
as it does not involve any interruption to the normal conduct of the
task. A response-based evaluation can also be relatively undemanding if
it only involves collecting documentary evidence of the task outcome
(e.g. the map showing the route students have drawn when following
instructions). However, it becomes more demanding if the aim is to
examine how students actually performed the task. This requires spe-
cially designed checklists or alternatively recording and transcribing the

interactions that result from performing the task. The most demanding
type of micro-evaluation is learning-based as it is difficult to find of way
of measuring whether a single task results in discernible learning.
R. Ellis (2011) outlined a procedure for conducting a micro-
evaluation of a task. The starting point is a clear description of the
task. This helps the evaluator to see what needs to be evaluated and
thus guides the planning of the evaluation. The objectives of the
evaluation need to be stated. These will address whether the task is
successful in achieving the goals a teacher has set for the task and also,
perhaps, whether unexpected benefits occur (Eckerth, 2008). Data
collection can take place before the lesson starts (e.g. by administering
a short test) but is more likely to occur during the lesson as the task is
performed (e.g. by observing or recording the learners’ performance of
the task) or after (e.g. by means of a short questionnaire to tap into
students’ perceptions of the task). Documentary evidence of the task
completion also needs to be collected. Data analysis can involve both
qualitative and quantitative methods but teachers may find it easier
and more informative to analyse the data qualitatively (e.g. by inspect-
ing a recording of the task performance to identify episodes that
capture important aspects). However, simple descriptive statistics help
to give a general picture. The analyses of the data guide the conclu-
sions, where the teacher comments on whether the task worked and
what changes are needed to improve it.
Hoogwerf’s evaluation of a consciousness-raising (CR) task
(reported in Ellis, 2008) serves as an example of a micro-evaluation.
Hoogwerf’s students were second-year Japanese college students
enrolled in an eight-month study-abroad college programme. Hoog-
werf designed a grammar CR task focusing on subject-verb agreement.
The task consisted of: (1) a statement of the subject-verb agreement
rule, (2) sentences serving as examples of the rule, and (3) sentences
taken from the students’ own writings to be completed by the students
using choices provided. The students were asked to read the explan-
ation of the rule and the examples provided and underline the subjects
of the sentences and then supply the correct verb form from the
choices provided. She identified two main objectives for the task:
(1) to raise the students’ awareness of subject-verb agreement, and
(2) to enhance the students’ motivation to attend to what she con-
sidered a fossilized error. Hoogwerf undertook both an objectives
model evaluation and a developmental evaluation; that is, she wanted
to see whether her task had ‘worked’ and how she might improve it.
She undertook a student-based evaluation by examining the comments
that the students made about the task in their journals, a response-
based evaluation by observing the students while they performed the

task and by collecting in their written answers, and a learning-based

evaluation by examining the students’ free writing before and after the
task to see if there was any evidence of increased accuracy in subject-
verb agreement.
The students’ journals indicated they were positive about the task.
The students commented on the fact they liked composing their own
sentences and many of them asked for more CR tasks. The response-
based part of the evaluation revealed that the students worked quickly
and eagerly on the task in a manner notably different from their
normal classroom behaviour. Also, the students’ answers were nearly
all correct. However, with the exception of the top three writers in the
class, there was no evidence of improved accuracy in the students’
subsequent writing. Hoogwerf concluded that the task had worked in
so far as it raised students’ awareness of subject-verb agreement and
increased their motivation to attend to fossilized errors in their writing
but she acknowledged that it did not lead to improved accuracy in the
students’ free writing. Overall, Hoogwerf felt that her experience of
evaluating the task was valuable and stated that she planned to use
more CR tasks in the future.
Ellis (2015a) reported a number of other micro-evaluations of tasks.
These were undertaken by teachers enrolled in a TBLT course as part
of an MA programme. The teachers first designed their own tasks,
planned how to evaluate them and then collected data for the evalu-
ations while implementing the tasks in their classrooms. Ellis reports
that the teachers chose the tasks to evaluate for very different reasons –
their perceptions of their students’ needs, to fulfil the aims of the
course they were teaching, to test their understanding of research
findings or simply through a desire to engage their learners more fully.
They also approached how they developed their tasks differently.
Johnson (2000) found that less experienced teachers used ‘task frame’
(e.g. the participatory organization or the skills to be practised) as a
starting point for designing a task, whereas more experienced teachers
opted mainly for ‘task genre’ (e.g. information gap) or ‘task function’
(e.g. describing a person). He also noted that the teachers varied in
terms of whether they opted for a ‘language-oriented’ or ‘task-oriented’
approach. The teachers in Ellis (2015a) adopted a similar mixed
approach to developing their tasks. However, whereas Johnson found
that the teachers he investigated did not draw on research involving
tasks, Ellis reported the some of his teachers did make use of the
research-based knowledge they had acquired on their MA course.
Ellis noted that his teachers were very aware that every task has a
context and that this can influence both how the task is performed and
what is learned as a result. For example, the teachers recognized the

importance of sequencing activities in order to ensure that a task

works effectively. Thus, even though they were asked to evaluate only
the main task, their evaluation reports revealed the importance they
attached to pre-task and post-task activities. In other words, they saw
their tasks as part of a complete lesson. In this respect their approach
differed from that of many researchers who have focused narrowly on
the performance of a single, specific task.
Another general finding of the micro-evaluations was that a task
does not result in the same ‘activity’ when it is performed by different
students – as claimed by sociocultural theory (see Chapter 4). One
teacher noted that there were marked differences in the students’ negoti-
ation of meaning. One pair engaged much more extensively in the
negotiation of meaning than another pair, worked harder to resolve
the communication problems that arose and was more successful in
doing so. There was also a difference in how these pairs negotiated.
The pair that negotiated extensively did so by means of clarification
requests whereas the other pair employed confirmation checks. Clearly,
the same ‘task’ resulted in very different interactional processes for these
two pairs of students. The students also varied in their perceptions of the
purpose of the tasks they were asked to perform. In sociocultural terms,
they had very different ‘motives’ for the ‘actions’ they carried out. For
example, two of the students performing a dictogloss task thought it was
a communicative activity whereas a third student in the same group saw
it as a grammar practice activity. Micro-evaluations of tasks are valuable
because they help teachers to understand the factors that are responsible
for the differences that arise in how the same task is performed.
The response-based part of a micro-evaluation involves examining
both the ‘product’ (i.e. whether the learners are successful in achieving
the outcome of the task) and the ‘process’ (i.e. what transpires when
the task is being performed). One of the conclusions that Ellis
(2015a) came to after examining a number of micro-evaluations was
that closed tasks are advantageous because they allow teachers a quick
and easy method of determining whether a task has ‘worked’. All the
teachers’ micro-evaluations investigated task processes using a variety
of techniques drawn from the task-based literature (e.g. negotiation of
meaning; language-related episodes). In particular, they were inter-
ested in the level of learners’ ‘engagement’ with the task. One teacher,
for example, developed an observation chart to record behaviours that
were indicative of motivation, enjoyment and task fatigue The same
teacher also used a ‘reflection questionnaire’ to record his own
responses about the students’ participation in the task once the lesson
was over. This concern with ‘engagement’ lends support to current
work on task engagement which we discussed in Chapter 6.

All the micro-evaluations in Ellis (2015a) included a student-based

component – mainly by asking the students to complete a short
questionnaire about their perceptions of the task. In this respect the
evaluations differed again from research studies, where collecting such
data is rare. Arguably ﬁnding out how learners orientate to a task is
crucial because it will affect how they perform it. It is also worth
pointing out that asking students to complete a short questionnaire
can be seen as part and parcel of the pedagogy of task-based teaching
(i.e. it constitutes a legitimate post-task activity).
Micro-evaluation of tasks is a convenient way to conduct practi-
tioner research. It is arguably easier for teachers to take as their
starting point a particular teaching activity they are interested in
than a ‘problem’ or a ‘puzzle’. The teachers who conducted the
micro-evaluations in Ellis (2015a) reported they found them of
value in developing their understanding of TBLT. However, they also
noted that such evaluations were time-consuming and that they
were unlikely to undertake them in their day-to-day teaching. Micro-
evaluations of tasks clearly have value in teacher education pro-
grammes, however. One way of making them more practical for
teachers might be to develop a set of guiding principles along the lines
that Allwright (2003) has done for exploratory practice. Ellis (2015a)
concluded with a list of such principles (e.g. Data should be collected
in a way that is practical and that enhances the pedagogic value of the
task and the evaluation should be seen as a mutual enterprise.

Final Comment
Samuda (2015) sees tasks as devices for creating pedagogical spaces.
She commented ‘a task has the capacity to open up a space for learning
and teaching’ (p. 282). The value of the practitioner research con-
sidered in this section is that it focuses on how tasks create pedagogical
spaces – that is, how tasks are interpreted and reshaped as they are
performed. Practitioner research inevitably addresses how context
determines the choice of task and how the enactment of the task
dynamically constructs the context as it is performed. We have sug-
gested that micro-evaluations of tasks carried out by teachers in their
own classrooms have special value in this respect. As Van den Branden
et al. (2007) observed ‘tasks on paper and tasks in real classrooms
may differ from each other in an astonishing number of ways’ (p. 3).
Micro-evaluations of tasks are ideally suited to investigating this. They
take ‘task’ – the fundamental element in TBLT – as the starting point
and then investigate how a speciﬁc task mediates the teaching and the
learning that take place.

Implementing TBLT: Problems and Some Solutions

The evaluation studies we have considered point to a number of
problems that can arise in implementing TBLT – especially when it
constitutes an innovation and teachers are struggling to make it ﬁt into
their instructional context. It is, however, unwise to overemphasize
these problems. Long (2016), for example, dismisses the claim of
Bruton (2002a) and Swan (2005a) that TBLT is not suited to foreign
language contexts because of the limited number of hours of instruction
each week in typical programmes. He commented ‘inadequate instruc-
tional time and lack of L2 [second language] exposure outside the
classroom are real problems for all kinds of LT (language teaching),
not just TBLT, and not just in foreign language settings’ (Long, 2016,
p. 26). Long’s point is well made. In identifying the problems with
TBLT it is important not to assume that these only apply to TBLT.
Nevertheless, the evaluation literature does point to a number of prob-
lems. These can be grouped according to whether they concern the
teacher, the students or structural issues within the education system.

Problems Involving Teachers

One problem that figures in a number of evaluation studies is that
teachers do not always have a clear grasp of what a ‘task’ is. We noted
this in Carless’ (2004) evaluation of TBLT in Hong Kong elementary
schools. Hu (2013) also noted that the Chinese public schoolteachers
of English he investigated had very different ideas of what a task was –
some simply equated it with exercises in their textbook. Lin and Wu
(2012) likewise found that none of the Taiwanese teachers they inter-
viewed could give a clear definition of a task. This problem can be put
down to inadequate training but even when this is available teachers
still seem to struggle to fully grasp what a task is. Erlam (2016) reports
on an in-service teacher education programme for teachers of foreign
languages in New Zealand. As part of this programme she asked the
teachers to design their own tasks, which she evaluated in terms of
whether they satisfied the four criteria for a task proposed by Ellis
(2003) – see Chapter 1. She found that the teachers’ tasks more often
than not failed to satisfy criterion (3) – namely, ‘learners rely mainly
on their own linguistic and non-linguistic resources’. In other words,
the tasks the teachers developed often involved the explicit presenta-
tion of target language.
The difficulty teachers experience in defining what a task is
reflects another more general problem. Teachers implementing TBLT
frequently express concern about their students’ grammatical

development (Watson-Todd, 2006; McDonough and Chaikitmong-

kol, 2007; East, 2014; McDonough, 2015). But this problem is argu-
ably more imagined than real as there is clear evidence that
grammatical development does take place in TBLT (Beretta and
Davies, 1985; Shintani, 2016). The real problem here is the difficulty
that teachers have in recognizing that grammar can be acquired inci-
dentally and that incidental acquisition can be more effective for the
development of communicative abilities than the intentional lan-
guage learning to which traditional language teaching caters. Unless
teachers are ready to accept the utility of incidental acquisition they
are likely to opt for a weak version of TBLT (i.e. task-supported
language teaching), which incorporates explicit instruction in the
pre-task phase of a lesson. In other words, the fundamental issue is
the theory of language learning that teachers have internalized.
Unless teachers buy into the theory of language learning that informs
TBLT (i.e. a theory of incidental acquisition) they are unlikely to
implement it effectively. Helping teachers to understand how lan-
guages are acquired ‘naturally’ should be a major goal of teacher
education for TBLT.
The tendency teachers have to resort to a weak version of TBLT and
to fall back on traditional modes of teaching is exacerbated when
teachers lack proficiency in the target language or lack confidence in
their proficiency (Butler, 2004; Jeon and Hahn, 2006). Again, though,
lack of proficiency is a problem no matter what the teaching approach.
The solution is again effective teacher training programmes. It is also
worth noting, however, that one way teachers can improve their own
ability in the L2 is by using it in task-based teaching!
Several commentators (Littlewood, 2007, 2014; Butler, 2011) have
pointed to the conflicts that exist between TBLT and traditional
culturally embedded teaching approaches. This is especially the case
in Asian contexts where it is claimed that Confucian notions that
emphasize knowledge as residing in books and the teacher as the
primary source of knowledge are incompatible with an approach that
emphasizes experiential learning. Samimy and Kobayashi (2004), for
example, claimed there is a cultural mismatch between communicative
language teaching and the Japanese culture of learning. However, as
Butler (2011) pointed out, this problem has been overstated. She noted
that the cultural backdrop of Asian language classrooms varies con-
siderably; for example, the culture of primary classrooms in Japan is in
fact well suited to the introduction of TBLT. Also, it should be noted
that TBLT does not require the total abandonment of traditional roles.
The post-task phase of a lesson, for example, affords opportunities for
a more traditional approach to teaching. What is needed is training

that helps teachers to modify the role they adopt in accordance with
the varying purpose of instructional activities in a TBLT lesson.
Teachers have sometimes been found to experience problems
in managing task-based lessons, especially when the students work
in small groups. Carless (2004) found that primary schoolteachers in
Hong Kong were concerned that TBLT resulted in noise and discipline
problems. However, this can be a problem with group work in gen-
eral, not just with TBLT. Also, TBLT does not necessitate group work.
Where students are unused to performing tasks, it might be advisable
to begin with input-based tasks with the whole class, where the teacher
has greater control over what transpires.
A ﬁnal problem concerns teachers’ perception that planning TBLT
lessons imposes too great a workload on them. East (2014), for
example, found that this was one of the main ‘negative characteristics’
of TBLT that teachers in an in-service teacher education programme in
New Zealand mentioned. The teachers pointed to the time required to
design tasks and to monitor students’ performance. This problem
is very real given the lack of published task-based teaching materials.
It can be addressed if teachers collaboratively develop a bank of
task-based materials that individual teachers can draw on as needed –
a proposal that Candlin (1987) made in his seminal article many years
ago. This would, however, require some central organization of the
kind found in the Belgium TBLT project (Van den Branden, 2006)
or in the TBLT course that McDonough and Chaikmongkol (2007)
evaluated.

Problems Involving Students

The evaluation studies indicate that students sometimes have problems
adapting to TBLT (McDonough and Chaikmongkol, 2007; Butler,
2011). This can arise when students are used to teacher-centred
instruction, where language is treated an object rather than as a tool,
and as a consequence fail to see the point of performing tasks that
cater to incidental rather than intentional language learning. This
problem is not universal, however. Students in some contexts welcome
TBLT. González-Lloret and Nielson (2015), for example, reported
that the border patrol agents in the programme they evaluated wel-
comed a task-based approach as it equipped them with the Spanish
they needed for their work. Bao and Kirkebaek (2013) also reported
that Danish university students responded positively to a task-based
Chinese course because they saw it as increasing their engagement,
creating an interactive and interesting learning environment, and
leading to effective learning. Students’ perceptions of TBLT derive

from their prior cultural experiences of classroom learning and are

inﬂuenced by their motives for learning a language. In cases where
negative views prevail, some learner training may be needed to
develop students’ awareness about the nature of language learning.
In addition it may be helpful if teachers help students to see that they
have learned some new language as the result of performing a task
(e.g. by asking them to note down any new language at the end of a
task or by means of an occasional oral or written quiz).
Another commonly voiced belief is that beginner-level learners
cannot perform tasks because they have not developed the English
proﬁciency needed to communicate. This problem, however, derives
from the misconception that TBLT must necessarily involve ‘speak-
ing’. Clearly students cannot speak in English if they don’t know any
English. The solution is to introduce TBLT to beginner-level learners
through input-based tasks where the teacher works to make the input
comprehensible to the students. Shintani (2016) provides detailed
advice about how input-based tasks can be successfully used with
complete beginners.
A third perceived problem is that students resort to the use of the L1
when faced with communication problems that they cannot solve in
English, resulting in the overuse of the L1 in the TBLT classroom
(Carless, 2004). We saw in Chapter 4 that there are legitimate uses
of the L1 in TBLT but, as Butler (2011) noted, it is not easy for
teachers inexperienced in implementing TBLT to make judgements
about when and how much L1 use to allow. This also needs to be a
focus of teacher-education. Teachers can be introduced to some useful
strategies for dealing with L1 overuse by students: (1) using input-based
tasks where the students’ responses are non-verbal, (2) performing
speaking tasks in a whole class context rather than in small group
work so the teacher can exert control over the use of the L1, (3)
allowing pre-task planning time, which, as we saw in Chapter 3, can
help students to formulate the language they need to perform a task,
and (4) ensuring that the task poses a reasonable level of challenge to
the students, arguably the most important strategy for guarding against
L1 overuse.

Structural Problems
Structural problems are problems that arise as a result of the external
requirements imposed on teachers that they are relatively powerless
to change. There are two kinds – classroom-level constraints and
societal-institutional-level constraints (Butler, 2011). A commonly
mentioned classroom-level constraint is the large size of classes that

make small group work difficult and allow little opportunity for
individual students to use English (Li, 1998; Samimy and Kobayashi,
2004). This, however, is a problem with any teaching approach.
Arguably, if TBLT is skilfully implemented it is more likely to ameli-
orate this problem than accentuate it. Again, input-based tasks have
an important role here as they maximize students’ exposure to the
target language. Societal-institutional problems include the require-
ment placed on teachers to follow a structural syllabus and the preva-
lence of traditional examinations that emphasize grammatical
accuracy over communicative proficiency and that assess by means
of discrete-item tests. In some settings a structural syllabus is officially
mandated even when TBLT is the recommended approach. Trad-
itional tests seem to have a life of their own, stubbornly continuing
to exist even when communicative language teaching has received
official approval. TBLT can only be successfully introduced if:
(1) the syllabus is task-based, and (2) students are assessed by means
of performance-based tests involving the use of the same kinds of task
as for teaching.
If teachers are constrained by a structural syllabus, TBLT is not
possible. But it is still possible for them to employ tasks if they adopt a
task-supported approach. We have seen in Chapter 1 that task-
supported language teaching has its advocates and is supported by
skill-learning theory. Samuda (2015) suggested that there is a need to
investigate how tasks work out in both task-based and task-supported
language teaching – a view we endorse. Key issues are whether pref-
acing the performance of focused tasks by explicit instruction directed
at specific target structures actually results in the use of these structures
when learners perform the tasks and also whether it affects the overall
quality of the language used. Ellis, Li and Zhu (2018) reported an
experimental study that showed explicit instruction did not result in
greater accuracy in the use of the target structure and had a negative
overall effect on the language produced. Micro-evaluations of tasks
can shed further light on these issues.
Traditional methods of assessment involving indirect, system-
referenced tests are arguably the greatest barrier to TBLT. The prob-
lem here is that teachers will teach to the tests. However, as Long
(2016) pointed out there is evidence to show that TBLT enables
learners to perform well in such tests. Also, even if teachers are faced
with preparing their students for a high-stakes traditional test, they
still have the option of incorporating performance-based assessment
tasks into their courses. Two evaluation studies have shown the via-
bility of this approach to assessment. Nielson (2014) included such
tasks in her online task-based Chinese course and showed how they

could support the task-based content of the course. Winke (2014)

described a course where task-based assessments involving students
self-assessing their own performance were incorporated into a Chinese
language course. She was able to show a high level of correspondence
between the students’ and the instructor’s rating of their performance
on the tasks, pointing to the validity of learner self-assessment.
Self-assessment also provides an answer to a question that teachers’
often pose – namely, how to assess learners’ speaking proﬁciency. It is
often not practical for teachers to assess individual students’ speech
but students can be helped to self-assess in ways that will enhance
reliability. Nevertheless, as Butler (2011) concluded, if TBLT is to
thrive ‘not only are changes in the exam system required, but also
drastic changes toward learning and assessment in general in society
are needed’ (p. 46).

Conclusion
Throughout this chapter we have pointed to differences between
researching and evaluating TBLT. Both are useful, of course, but
arguably evaluation studies are especially valuable as they provide
information about how TBLT works in actual classrooms. Evaluation
studies shed light on whole courses (in the case of macro-evaluations)
or on specific lessons or tasks (in the case of micro-evaluations). They
show how teachers and students respond to TBLT, whether the task-
based activities result in the kinds of behaviour intended and whether
(and to what extent) learning takes place. The accumulated evidence
points clearly to the effectiveness of TBLT in a wide range of instruc-
tional contexts. They show that teachers are capable of real change
provided that they receive support. The factors influencing the success
of innovations shown in Table 11.1 provides a useful checklist of
the support needed to ensure that when TBLT is introduced it is
successful.
The evaluations have also pointed out the problems that can arise.
We identified a number of issues involving teachers and students along
with some serious structural problems. While recognizing that many
of these problems are not specific to TBLT, it is clear that serious
consideration needs to be given to how they can be addressed. One is
to acknowledge the contextual constraints and accept that a communi-
cative approach is not appropriate or desirable in some settings (Bax,
2003). Another is to opt for a weak form of communicative language
teaching. This is the solution favoured by Littlewood (2014) and
Butler (2011). They argued that task-supported language teaching is
more compatible with existing practices and more closely aligned with

current methods of assessment in Asian contexts. Several of the

evaluation studies summarized in Table 11.2 (e.g. Watson-Todd,
2006) reported that task-based courses mutated over time into
task-supported courses. Teachers clearly do need to take account of
contextual factors and make adjustments when needed but they also
need to consider the evidence from comparative method studies (see
Chapter 10), which indicated that presentation–practice–production
(PPP) may be less effective than TBLT if the aim is to develop commu-
nicative abilities. Also, the macro-evaluation studies considered in this
chapter (e.g. Van den Branden, 2006; McDonough and Chaikmong-
kol, 2007; González-Lloret and Nielson, 2015) provide convincing
evidence that TBLT can be implemented successfully. Thus, while we
acknowledge there can be a place for task-supported language teach-
ing (TSLT), we believe that ideally it should not supplant TBLT.
The answer is to address the problems of implementing TBLT, not
to abandon it.
Evaluation studies do not just tell us what works and what doesn’t.
They also serve a teacher development function. This is especially true
of micro-evaluations. They document actual TBLT lessons and how
speciﬁc tasks work out in real teaching contexts. Unlike programme
evaluations that are typically carried out by external evaluators,
micro-evaluations are conducted by teachers in their own classrooms.
They serve as one of the most effective ways of developing teachers’
understanding of TBLT and how to implement it.

In this ﬁnal part of the book we address the following questions:

1. What criticisms of TBLT have been made and how can these be
addressed?
2. What lines of research should researchers pursue to further under-
stand the relationship between task and learning?
3. In what ways can theory and the praxis of TBLT be mutually
informing?
4. In what ways can task-based research and task-based teaching
profitably interface?
In line with the main aim of the book, we continue to explore the
interface between research-based and pedagogical-oriented perspec-
tives on TBLT.
Task-based teaching has been subjected to considerable criticism
from teachers and teacher-educators who maintain allegiance to
more traditional approaches to teaching. Chapter 12 considers these
outsider-criticisms (i.e. criticisms that originate from opponents of
TBLT) and the misunderstandings that underlie them, while also
acknowledging a number of real issues they raise that need to be
addressed if TBLT is to progress in the future. The chapter also looks
at insider-critiques (i.e. issues raised by advocates of TBLT). One such
issue concerns how to define and investigate complexity so as to arrive
at a principled basis for sequencing tasks in a syllabus. We will see that
there are different positions regarding such issues.
In Chapter 13 we take stock and suggest the future directions that
we think TBLT should take. Theory and research about TBLT show
no sign of losing vitality. We can expect researchers to both draw on
and contribute to theories of L2 acquisition. We foresee that influences
from neighbouring disciplines will grow. We also predict that TBLT
will be increasingly mandated by educational authorities so there will
be continued pressure for teachers at all levels (primary, secondary and

331

tertiary) to adopt TBLT. Crucial to the successful uptake of TBLT

will be the development of a testing regime that is compatible with it;
so we envisage that task-based assessment will receive increased atten-
tion. The increasing importance of technology-mediated language
instruction will also lead to the development of suitable task-based
materials for online use and of research programmes directed at
investigating these.

Introduction
As might be expected in the case of a new approach that differs radically
from mainstream approaches, TBLT has aroused considerable criticism.
In part, this criticism has its origin in the suspicion in which research in
general is held by some members of the language teaching profession.
Swan (2005a), for example, talked of ‘legislation by hypothesis’, argu-
ing that SLA researchers have foisted TBLT onto the profession, ignor-
ing the realities of most classrooms. Hadley (2013) saw TBLT as the
‘new orthodoxy’ and as ‘a new religion’ and talked of ‘the disconnect
between scholarly proponents and classroom practitioners’ (p. 194).
This view that TBLT is just the product of armchair SLA researchers
is mistaken, however. For a start, many of these researcher-advocates of
TBLT (including the authors of this book) were once teachers them-
selves, and their advocacy of TBLT derives from their experience of the
limitations of mainstream approaches juts as much as from research
and theory. Also, TBLT has spokespersons from among teachers and
teacher-educators (e.g. Estaire and Zanon, 1994; Willis, 1996; Cutrone
and Beh, 2014), who do not see themselves as members of the SLA
research community. It is unwise to dismiss research and to make a case
against TBLT just because it is advocated by SLA researchers.1
TBLT critics sometimes fail to recognize that TBLT is an ‘approach’
rather than a well-deﬁned ‘method’ (Richards and Rodgers, 1986).
That is, it involves a quite varied set of theoretical principles, curricu-
lar designs, lesson plans and methodological strategies as we have
tried to make clear in this book. There is a difference between task-
based and task-supported language teaching (see Chapter 1) but even
this distinction is not watertight, given that both can involve explicit
instructional strategies. Critics, however, have ignored this variation
and straitjacketed TBLT in order to set up their critiques. In so doing,
they end up misrepresenting TBLT. By way of example, Table 12.1

333

Table 12.1 The characteristics of TBLT according to Swan (2005a)

Swan’s list Commentary

‘Natural’ or ‘naturalistic’ language use This is a key principle. All versions of
involving a primary focus on TBLT (and also some types of TSLT)
meaning rather than language. see tasks as affording opportunities
for ‘natural’ language use where the
primary focus is on meaning.
Learner-centredness rather than This is a misrepresentation. Not all
teacher control. task-based lessons need involve
group work. Tasks can be performed
by the teacher working with the
whole class. Group work can also be
followed by teacher-led activities.
Naturalistic learning does not This is partially correct. Learners can
guarantee target-like accuracy, so acquire much language
some intervention is needed. ‘naturalistically’ (i.e. while focused
on meaning) but some forms are
‘blocked’ (Ellis, 2006) by the L1 or
by their lack of saliency and for this
reason intervention is needed.
Intervention can be best achieved by ‘Focus-on-form’ is an essential feature
focusing on form while the of TBLT. However, intervention
overriding focus is on meaning. involving focus-on-form need not be
restricted to the online performance
of tasks. Ellis (2016) outlined
various ways of achieving a
focus-on-form (see also Chapter 2).
Communicative tasks are the best Tasks are central to both TBLT and
vehicle for this approach. TSLT. All ways of focusing on form
involve tasks. However, the pre- and
post-task stages of a lesson need not
be task-based (see Chapter 8).
Pre- and post-task activities can prime This misrepresents because it is
or boost noticing during incomplete. Focus-on-form is
communication. directed at inducing ‘noticing’ while
tasks are being performed. Pre- and
post-take activities serve a number of
different purposes including
planning the language needed to
perform a task in the pre-task stage
and explicit learning in the post-task
stage. ‘Noticing’ is just one of the
psycholinguistic processes that TBLT
aims to activate.
(continued)

Table 12.1 (cont.)

Swan’s list Commentary

Traditional approaches are ineffective Some advocates of TBLT (e.g. Willis,
and undesirable. Long and Skehan) do reject
traditional approaches. Others (e.g.
Ellis) propose a modular curriculum
consisting of both a synthetic
component (as in traditional
approaches) and an analytic
component (as in TBLT) – see
Chapter 7.

summarizes the characteristics of TBLT that Swan (2005a) proposed

as the basis for his critique along with a commentary on each charac-
teristic he mentions. Swan correctly identiﬁed the two key tenets found
in all versions of TBLT, namely the centrality of ‘natural’ language use
where the primary focus is on meaning and the need for intervention
involving focus-on-form. But he misrepresents TBLT in other respects
(e.g. he fails to recognize the importance of teacher control) and under-
represents it in others (e.g. by failing to recognize that focus-on-form is
not just directed at ‘noticing’). He seizes on the view espoused by some
advocates of TBLT (e.g. Willis and Willis, 2007; Long, 2015) that
traditional forms of instruction need to be abandoned in favour of
TBLT and he ignores the proposals of other advocates (e.g. Ellis, 2003;
Klapper, 2003) for a modular approach involving both TBLT and
traditional approaches. Swan mistakenly assumes that there is a single
set of principles that characterize TBLT and as a result many of his
criticisms are unwarranted.
We ﬁnd the same problem in Bruton’s (2002a, 2002b) critiques of
TBLT. He assumes that TBLT necessarily involves learners interacting
in pairs or small groups and that the sole purpose of using tasks is to
generate oral interaction. For Bruton, tasks ‘exclude listening approaches’
(2002a, p. 285) and there is no recognition that they can include reading
and writing too. From this totally mistaken standpoint, Bruton went on to
claim that TBLT is only tenable if the learners ‘have some existing target
language ability’ (p. 285). Also, like Swan, he sees TBLT as marginalizing
the intervention of the teacher. He fails to recognize that tasks can be
performed in a whole-class participatory structure where the teacher plays
the dual roles of a co-performer of the task and an instructor.
There are two groups of critics of TBLT. Outsider critics approach
TBLT from an adversarial position. Included in this group are Sheen

(1994, 2003), Bruton (2002a, 2005), Widdowson (2003), Swan

(2005a, 2005b) and Littlewood (2007, 2014). There are also insider
critics – advocates of TBLT who are their own critics. They dispute
key issues, such as the relative merits of real-world and pedagogic
tasks and how focus-on-form can best be executed. They also acknow-
ledge the problems that can impede the implementation of TBLT. We
will ﬁrst take a look at what the outsider critics have had to say and
then consider the issues that concern the insider critics.

The Outsider Critics

In his defence of tasks and TBLT, Long (2016) distinguished ‘non-
issues’ and ‘issues’. The non-issues were those raised by the outsider
critics, while the real issues were those that required serious consider-
ation. Our view is a little different. Like Long, we see many of outsider
critics’ problems as non-issues, but we also see some of the problems
they identiﬁed as real issues. We organize this section on the outsider
critics accordingly.

Non-Issues
Problem 1: Tasks cannot serve as the units of a syllabus. Arguably the
most serious problem identified by the outsider critics rests in the claim
that task behaviours are not predictable and therefore cannot serve as
syllabus specifiers. Bruton commented ‘the lack of predictability of the
language generated in many tasks make (sic) it likely that any form of
planning for assimilation will be rather arbitrary’ (Bruton, 2002a,
p. 285). Seedhouse (2005b) argued in a similar vein by noting that
the interaction that transpires when learners perform a task frequently
does not match that intended by designers of the task. He claimed that
for this reason it is impossible to plan a language course based on
tasks-as-workplans.2 There is in fact evidence to support such a pos-
ition. In Chapter 4 we considered research based on sociocultural
theory (e.g. Coughlan and Duff, 1994) that showed that the ‘activity’
that results from a ‘task’ varies according to the specific motives that
different learners have for performing the task.
This problem is overstated, however. While the relationship
between task-as-workplan and the activity it gives rise to is not a
perfect one, a task can have predictive value, as shown in research
that investigated how the design features of tasks can influence
the complexity, accuracy and fluency of learners’ production (see
Chapter 3). In addition, focused tasks can be successful in eliciting
the linguistic features they target (see Ellis, 2003). Thirdly, it is easy to

design input-based tasks in a way that obligates learners to process

pre-determined linguistic features contained in them. Fourthly, Bygate
(2018) proposed that tasks can display ‘pragmatic predictability’ (i.e.
the discourse phases that characterize the trajectory learners follow
when they perform a particular type of task). Bygate investigated
how learners performed a picture story task with each learner in the
group holding one of the six pictures that made up the story. There
was considerable consistency in how these learners moved through
discourse phases involving description, comparison, interpreting,
sequencing and ﬁnally narrating the story.
But predictability, especially of linguistic features, is a non-issue.
Irrespective of whether it is or is not possible to predict the language
use that occurs when learners perform a task, TBLT is premised on the
assumption that it is not possible to predict what individual learners
will learn from a task. In other words, the design of a task-based
syllabus does not need to be informed by what Bruton called ‘planning
for assimilation’. Indeed, the fundamental rationale for TBLT is
that such planning is not possible given that learners acquire language
in a non-linear way, working on a number of different features con-
currently, and moving only gradually towards mastery of them.
The strength of the task-based syllabus is that it acknowledges the
processes by which language is acquired naturally, removing the need
to pre-determine what and when learners will acquire speciﬁc linguis-
tic features. To dispute the validity of tasks as a basis for a syllabus
it would be necessary to dispute the theory of acquisition upon
which TBLT is based. This is exactly what Swan (2005a) tried to do
by attacking the hypotheses which he claimed the theory is based on.
In so doing, however, Swan ignored the large body of research that
supports these hypotheses – as Long (2016) so admirably demon-
strated in his defence of TBLT.
There is, however, a real issue behind the problem stated by Bruton
and Seedhouse. If it is not possible (or desirable) to base the design of a
task-based syllabus on linguistic grounds, the question arises as to
what principles should inform the selection, grading and sequencing
of tasks in a syllabus. This is an issue which will be considered in the
later section dealing with insider critic issues.
Problem 2: TBLT does not help learners learn ‘new language’.
A common complaint of outsider critiques is that because TBLT
emphasizes incidental language learning while performing tasks, it is
inferior to traditional approaches that cater to intentional language
learning through the presentation and practice of what Swan (2005a)
terms ‘new language’. Outsider critics, however, are not entirely dis-
missive of tasks as they see them as supplementary to linguistic

approaches in helping to develop learners’ oral ﬂuency. As Swan

(2005b) put it ‘TBLT is not primarily a language teaching approach’
as it is ‘mainly concerned with improving learners’ command of what
they already know’ (p. 254) – a view that reflects how tasks figured in
communicative language teaching in the 1980s (see Chapter 1).
Whether learners can learn ‘new language’ incidentally through
performing tasks is an important issue. If it is not possible – or if it only
takes place inefficiently – then a case can be made for direct, teacher-led
instruction of the kind favoured by the outsider critics. In fact, though,
there is plenty of evidence that learners do acquire language incidentally
as a result of performing tasks. Studies based on input-based tasks (e.g.
Ellis, 2001; Shintani, 2016) show that when learners perform tasks,
they acquire both new vocabulary and new grammatical structures (see
Chapter 2). Other studies involving output-based tasks (e.g. Mackey,
1999; Spada and Lightbown, 1999) show that performing interactive
tasks enables learners to progress along an acquisition sequence.
Corrective feedback studies (e.g. Ellis, Loewen and Erlam, 2006; Yil-
maz, 2013b) show that learners’ grammatical accuracy improves when
they perform tasks and receive feedback on their output. In fact, the
SLA literature is replete with studies that demonstrate that incidental
acquisition does take place when learners perform tasks and that this is
true irrespective of whether acquisition is defined as involving new
language, progression from one developmental stage to another or
greater accuracy. Meta-analyses (e.g. Norris and Ortega, 2000; Goo
et al., 2015) also provide clear evidence of the adequacy of incidental
acquisition. As Long (2015) concluded ‘incidental focus on form works’
(p. 13).
Outsider critics might still argue, however, that incidental acquisi-
tion is less efficient than intentional learning and that, therefore, an
approach involving the direct teaching of ‘new’ language is preferable
to TBLT. The meta-analyses referred to in this section, however, show
that focus-on-form is not statistically less effective than traditional
focus-on-forms instruction. In Chapter 10, we examined a number
of studies comparing TBLT with traditional approaches, such as PPP.
We noted that these studies suffer from a number of methodological
problems but that on balance they point to the superiority of TBLT
over PPP, especially when acquisition is measured in ways that tap into
learners’ ability to use language in free oral communication.
Problem 3: There is no grammar in TBLT. This putative problem is
really an extension of the previous one. Sheen (2003) claimed that in
task-based language teaching there is ‘no grammar syllabus’ and went
on argue that proponents of TBLT ‘generally offer little more than a
brief list of suggestions regarding the selection and presentation of new

language’ (p. 391). He was also critical of the fact that in TBLT any
treatment of grammar only takes the form of quick, corrective feed-
back, allowing for minimal interruption of the task activity. Swan
(2005a) insisted that TBLT ‘outlaws’ the grammar syllabus. Perhaps
this view that that there is no grammar in TBLT originates in Willis’
version of TBLT. Willis rejects focusing on form during the main task
phase, arguing that to do so would interfere with fluency. However,
other versions of TBLT see attention to linguistic form as a necessary
feature of task performance.
In fact, in one way or another there is plenty of grammar in TBLT.
In Ellis’s (2003) version of TBLT, there is the possibility of designing
focused tasks to address learners’ specific grammatical problems. Swain
and her co-researchers (e.g. Swain and Lapkin, 1995) have shown that
when learners perform a focused task in pairs or small groups ‘language-
related episodes’ occur frequently and that these are often successfully
resolved and contribute to learning (see Chapter 4). Opportunities for
focusing on grammatical form also occur in the pre-task and post-task
stages of a lesson. Guided pre-task planning allows learners the oppor-
tunity to consider what grammar they will need before they start to
perform the task. In the post-task stage, teachers can deal explicitly with
observed grammatical difficulties (see Chapter 8).
Problem 4: Performing tasks encourages indexical and minimal use
of the L2. Seedhouse (1999) criticized task-based language teaching on
the grounds that the performance of tasks results in indexicalized and
pidginized language because learners are over-reliant on context and
thus do not need to stretch their linguistic resources. Bruton (2002a)
also thought that asking learners to perform tasks leads to the devel-
opment of a classroom pidgin. Widdowson (2003) noted that learners
may be successful in achieving the communicative outcome of a task
without any need to attend to their actual use of the L2.
Clearly a task can result in language consisting of single words and
formulaic chunks – especially if learners are beginners. However, there
is also plenty of evidence to show that tasks can give rise to much more
complex and accurate use of the L2. Much depends on the nature of
the task and the way it is implemented. Opinion-gap tasks can elicit
more complex language use than information-gap tasks (Rulon and
McCreary, 1986). Giving learners opportunities to plan before they
perform the task also has a notable effect on the complexity – and in
some cases accuracy – of the language used. Thus, the claim that tasks
will inevitably result in impoverished learner output is unjustified.
Much of the research on tasks has been directed at identifying the
design features and implementation options that will attract learners’
attention to form and motivate language that is more complex and/or

more accurate (see Chapter 3). Furthermore, as Long (2015) pointed

out, the kind of interaction typical in traditional language teaching –
the ubiquitous initiate, response, feedback (IRF) exchange – is in
fact much more likely to involve the ‘narrow and restricted variety
of communication’ that Seedhouse (1999, p. 155) believed tasks
resulted in.
Problem 5: TBLT is not suited to low-proficiency learners. Little-
wood (2007) commented that speaking tasks are difficult for learners
of low proficiency. However, this problem seems to derive from the
commonly held belief that TBLT only involves speaking tasks. It is
obvious that beginner-level learners cannot engage in the free speaking
that oral tasks require. However, they can perform listening tasks.
Ellis (2001) reviewed a number of studies involving what he called
Listen-and-Do Tasks such as the Kitchen Task, where the learners
listened to instructions about where to locate kitchen objects (repre-
sented in numbered pictures) in a kitchen (represented by a diagram
of the kitchen). They demonstrated their understanding of the instruc-
tions by writing the numbers of the correct pictures in the correct
position in the diagram of the kitchen. Ellis was able to show that
performing this task resulted in learners learning the words for kitchen
objects. Shintani (2016) showed that Listen-and-Do tasks can be used
successfully with complete beginners and can contribute to the acqui-
sition of grammatical features.
Learners do not begin the process of acquiring their L1 by speaking
it. Rather they match what they hear in the input to the objects and
actions around them and thereby acquire new words and grammatical
forms. Input-based tasks cater to this process. The L2 resources that
are built up through performing such tasks can then later be used in
speaking tasks. It is important to note, however, that input-based tasks
do not prohibit learners from speaking; they simply do not require it.
Therefore, learners who wish to try speaking from the outset – and
there are some – are free to do so.
TBLT is not just well suited to teaching learners of low proficiency
but in many respects is better suited than more traditional methods
that require speaking from the outset. The only way that traditional
methods can elicit speaking from learners is by carefully controlling
their output using models and slot-and-fill and substitution exercises.
Input-based tasks provide learners when an opportunity to experience
natural language use from the start and to build up procedural L2
knowledge which will then be available for speaking.
Problem 6: TBLT assigns the teacher a very limited role. Swan
(2005a) complained that ‘the thrust of TBLT is to cast the teacher in
the role of manager and facilitator of communicative activity rather

than as an important source of new language’ (p. 391). He based this

claim on the fact that TBLT is implemented through small group
work, rather than in teacher-centred instruction. However, as noted
above, the claim that TBLT requires the performance of tasks in
groups is mistaken. It has probably arisen from the fact that much of
the research has investigated speaking tasks performed in pairs or
groups. In fact, as noted previously, group work is not an essential
characteristic of TBLT. Prabhu (1987), for example, argued that
effective task-based teaching needs to expose learners to good models
of the L2 and that this requires the teacher to take charge of the task
in a whole-class context. Input-based tasks require a teacher–class
participatory structure. Even speaking tasks can be carried out with
the teacher interacting with the whole class. In an information-gap
task, such as Spot the Difference, the information can be split between
the teacher and the students.
Nor is the teacher just a ‘manager and facilitator of communicative
activity’. The teacher serves as a major source of input in TBLT. In the
case of input-based tasks, teachers need to function as ‘navigators’
(Shintani, 2016) by helping the students to understand the input through
repetition and non-verbal clues. In the case of output-based tasks, the
teacher has a major role to play in providing corrective feedback while
students are performing a task. At times, the teacher may need to step
outside the task to provide some brief explicit instruction in order to
guide learners to make the link between a grammatical form and its
meaning. In fact, in TBLT, the teacher needs to perform multiple roles.
Long (2015), addressing the same non-issue, noted that ‘the teacher’s
role in TBLT requires greater expertise, and is more important, more
demanding and certainly more communicative than in PPP’ (p. 24).

Some Real Issues

We turn now to some of the other problems mentioned by the outsider
critics. These issues we consider ‘real’ in the sense that they raise
important questions which need to be addressed.
Problem 1: ‘Task’ is an ill-defined construct. Widdowson (2003)
claimed that ‘the criteria that are proposed as defining features of tasks
are . . . so loosely formulated . . . that they do not distinguish tasks
from other more traditional classroom activities’ (p. 126). Widdowson
seized on the definition provided by Skehan (1998), who identified
four key criteria for a task:
• meaning is primary
• there is a goal that needs to be worked towards

• the activity is outcome evaluated

• there is a real-world relationship
According to Widdowson the key terms such as ‘meaning’, ‘goal’,
‘outcome’ and ‘real-world relationship’ are ill-defined. Ellis (2009a)
attempted to clarify; he pointed out that a task involves both semantic
and pragmatic meaning (whereas an exercise involves only semantic
meaning), that the ‘goal’ means a communicative goal, that the ‘out-
come’ that is evaluated is again a communicative (not linguistic) one,
and the ‘real-world relationship’ refers to language use that is inter-
actionally rather than situationally authentic. In Chapter 1, we offered
an extended definition of a task that we believe is tighter and better
able to distinguish tasks from exercises. We included an additional
criterion, namely ‘learners rely mainly on their own linguistic and non-
linguistic resources’ to capture the fact that a task needs to involve
learners in text creation not just text manipulation.
Nevertheless, problems still exist with the definition of a task. Van
den Branden (2006) distinguished seventeen definitions of a task,
which he divided into those that define a task in terms of language
learning goals and those that define it as an educational activity. The
proliferation of definitions is itself an indication of a definitional
problem. In the applied linguistics literature, the term ‘task’ is often
used to refer to just about any type of pedagogic or assessment activity.
Questions can also be raised as to whether the tasks used in various
research studies always qualify as tasks. For example, do text reproduc-
tion activities such as dictogloss require learners to use their own
linguistic resources? Questions might also be asked about picture-
composition tasks, which figure in many studies. Is telling a story when
there is no communicative purpose for doing so a ‘task’? Text-
enhancement studies also raise questions. Are activities where all that
the learners need to do is read a text with some words highlighted and
then answer some comprehension questions really ‘tasks’? Where is the
gap in such activities? Where is the communicative outcome? Perhaps
the distinction between a task and an exercise is less clear cut than has
sometimes been claimed or assumed. Some activities – for example, cue-
card dialogues (Revell, 1979) – seem to be task-like rather than fully
fledged tasks. It is perhaps not so surprising that teachers are often very
uncertain about what constitutes a ‘task’ (see Chapter 11).
It may not be possible to arrive at a watertight definition of a task
and perhaps such a definition is not needed. Perhaps it is sufficient
to simply call any activity a ‘task’ that requires the receptive or
productive use of the L2 in a meaning-focused activity that involves
text creation at some level. Nevertheless, synthesizing the results of

task-based studies is going to be problematic given that the studies

have been conducted with very different notions of what a task is.
Problem 2: Learners will resort to their L1 when performing tasks.
A common criticism is that if learners are focused on achieving the
task outcome, they will simply resort to the use of their shared L1 as
the most efficient way of achieving this. Carless (2003), for example,
reported that mother tongue use was one of the key issues raised by
both teachers and teacher-educators when task-based language teaching
was introduced into Hong Kong schools. Carless also noted that most
of the teachers he interviewed took a pragmatic view on mother tongue
use on the grounds that it was needed to maintain students’ attention
and involvement. He speculated that when teachers became concerned
about the overuse of the mother tongue in group work tasks, they
would simply resort to traditional, teacher-fronted teaching.
One response to this problem is to argue that it is not a problem at all
and that the use of the L1 is entirely legitimate in TBLT. Sociocultural
theorists have argued that the L1 can serve as a mediating tool for
performing tasks in the L2 (see Chapter 4). Good practice in TBLT
reflects current views about the value of the L1 for performing such
functions as task management, task clarification, discussing vocabulary
and meaning and even presenting grammar points (Cook, 2001). It
acknowledges the naturalness of code-switching when there is a shared
L1 (Macaro, 2001).
Nevertheless, overuse of the L1 in TBLT constitutes a real problem –
especially in monolingual classrooms such as those in Hong Kong.
However, it does not warrant the abandonment of TBLT. For a start,
L1 overuse can arise in any approach to teaching. Also, there are
various ways of mitigating it. The teachers and teacher-educators that
Carless interviewed mentioned several – appointing language moni-
tors to remind their classmates to use English, using a reward system
for using the L2, performing tasks as a whole class or in pairs rather in
than groups. The task-based research points to other ways. One is task
repetition. Shintani (2011) found that when her beginner-level learners
performed a task for the first time, they naturally drew on their L1, but
that over time, as the task was repeated and the children grew familiar
with both the procedures and the English needed to perform the task,
they increasingly used English. In pre-task planning. learners can draw
on their L1 to conceptualize what they want to say, which can ease the
processing burden and give them time to formulate when they perform
the task in the L2.
Problem 3: TBLT is not suited to ‘acquisition-poor’ environments.
Swan (2005a) claimed that TBLT is not suited to ‘acquisition-poor’
environments, by which he meant foreign-language contexts where

learners are dependent on the classroom for learning a language. He

argued that in such contexts, learners need a more structured approach
to ensure they learn the grammatical resources required for communi-
cating. However, while grammar is certainly helpful for communicating,
it is not essential. Research on early L2 acquisition both outside the
classroom (e.g. Klein and Perdue, 1997) and inside (e.g. Ellis, 1984)
has shown that L2 learners manage to communicate with no or very
minimal grammar. They draw on a limited range of formulaic expres-
sions and vocabulary in conjunction with context to express what they
want to say. Grammar, in fact, comes later, driven by the need to express
more complex ideas that require more advanced language. It would
follow then, that if the aim is to foster communicative ability in learners
from the start, an approach that focuses on lexis would be most effective –
for example, the Lexical Approach (Lewis, 1993). Such an approach is
entirely compatible with TBLT as tasks can serve as a vehicle for provid-
ing lexical input (see Ellis, 2001).
Nevertheless, as we have already noted, there is a concern – voiced
by both the critics of TBLT and by many teachers – that there is no
grammar in TBLT and reservations regarding TBLT seem to centre on
this point. If grammar learning has a place in courses for beginner-
level learners, however, it is not in the classroom but in self-study
materials that can be made available for use outside the classroom (e.g.
for homework). The learning that results from such materials is
unlikely to feed directly into the learners’ ability to communicate.
Rather it results in explicit knowledge which according to some theor-
ies (e.g. Ellis, 1994) serves as a resource for monitoring output and for
priming attention to grammatical forms that learners are exposed to in
task-based input and thus indirectly – and only in due time – facilitating
the implicit knowledge needed for communication. However, for
beginner-level learners in acquisition-poor environments, their only
access to communicative input is the classroom; for that, tasks – in
particular input-based ones – are needed.

Insider Critics
As we noted earlier in this chapter, there is considerable diversity in
TBLT. This diversity manifests itself in debate among TBLT advocates
on a number of issues.

Issue 1: Task-Based Research Is Limited in

a Number of Respects
One of Swan’s major objections to TBLT was that the theoretical
principles that underpin TBLT lacked supporting evidence from

research. However, as Long (2015) pointed out, Swan’s claim has little
merit as there is in fact ample evidence to support each of the hypoth-
eses (online, noticing and teachability) that Swan rejected. Neverthe-
less, there are some notable gaps in the research agenda that informs
TBLT. Révész (2017) pointed to several. She noted that researchers
have focused on the effect that task design and implementation
variables have on task performance, but have generally neglected to
investigate the relationship between task performance and develop-
mental outcomes. She also pointed to the predominant focus on
output-based tasks and the paucity of research that has investigated
input-based tasks. A side effect of this is that misunderstandings arise
about what TBLT entails, leading to false claims about its unsuitability
for beginner-level learners. Révész also noted that young learners have
received little attention and that ‘relatively few task-based studies have
been conducted in actual classroom settings, even though the aim of
TBLT is to inform L2 pedagogy’ (p. 9).
TBLT clearly needs to justify itself as a researched pedagogy and
while some progress has been made in this direction (see Chapters 10
and 11), there are obvious gaps in need of attention. If TBLT is to
claim it is relevant to all learners in all teaching contexts, then it needs
to be able to demonstrate its practicality and effectiveness in a range of
instructional settings. Starts have been made in this direction (see, for
example, the articles in the special issue ‘Complementary theoretical
perspectives on task-based classroom realities’, TESOL Quarterly, 51,
3) and in Shintani (2016). However, much more research is needed.
Advocates of TBLT agree on this but differences exist in the directions
they think this research should take.

Issue 2: What Types of Tasks Should Figure in

a Task-Based Course?
One issue where there is a clear divergence of opinion among TBLT
advocates concerns the types of tasks that should figure in a task-based
course. Differences exist regarding whether they should be real-world
tasks or pedagogic tasks (a distinction originally made by Nunan,
1989) and also whether focused (as opposed to unfocused tasks) have
a role to play in a TBLT programme.
In Long’s version of TBLT (Long, 1985), a needs analysis is first
conducted to establish the target tasks that a particular group of
learners need to master. These then serve as a basis for developing
the tasks in a syllabus. Long (2015) has continued to maintain this
position, which is clearly the ideal way to proceed for those groups of
learners who have well-defined target needs, such the US Border Patrol
Agency students in González-Lloret and Nielson’s (2015) study (see

Chapters 10 and 11). But for many learners – for example, those in
state schools learning a foreign language – there are no immediate
target needs and a needs analysis would have to be based on putative
future requirements. Lambert (2012) attempted this with a group of
Japanese university students only to find that the tasks he identified
and designed did not motivate the students (see Chapter 7). For many
students, a better approach is surely to build a course around peda-
gogic tasks that engage them.
The use of focused tasks is another point of controversy among
TBLT advocates. Both Long (2016) and Skehan (1998) see no need for
them in TBLT. Long dismisses focused tasks on the grounds that they
belong to a focus-on-forms approach. Skehan sees the aim in TBLT as
ensuring a balanced development in terms of complexity, accuracy
and fluency, which can be achieved through the adroit selection of
tasks to induce a variable focus on these aspects of language produc-
tion. Skehan argued that only unfocused tasks are needed to achieve
this. Ellis (2003), however, proposed that focused tasks do have a
place in both a language programme and in research designed to
inform pedagogy. Learners are likely to experience problems with
certain linguistic features (e.g. subject–verb agreement and complex
structures such as hypothetical conditionals) even at an advanced
stage of development. The selective use of focused tasks directed at
these structures can help learners overcome their learning problems.
Focused tasks are useful in research because they make it possible to
pre- and post-test learners in order to investigate whether performing
tasks results in the learning of the targeted features. Investigating the
learning that might result from unfocused tasks is much more prob-
lematic because pre-testing is not possible.3
A particular type of focused task is a consciousness-raising (CR)
task (Fotos and Ellis, 1991; Ellis, 1993). This makes a linguistic
feature (typically grammatical or pragmatic) the topic of the task
and aims to help learners achieve a metalinguistic understanding of a
rule or regularity by guiding them through an analysis of data that
illustrates the forms and uses of the target feature. Long (2016)
dismisses CR tasks on the grounds that they are ‘components in the
delivery of a traditional linguistic syllabus’ (p. 6). Ellis (2018b)
responded to Long’s critique of CR tasks by pointing out that his idea
was not that they should comprise the entire syllabus but rather that
they could provide the means for developing explicit knowledge to
help learners overcome specific and persistent learning problems. Thus
CR tasks would only figure at a stage in learners’ development when
they were able to communicate reasonably effectively in the L2 but
with restricted accuracy. Ellis also pointed to another advantage of CR

tasks. He suggested that they can provide opportunities for communi-

cating if learners work collaboratively to solve the linguistic problems
they pose. It is in for this reason that they can still be called ‘tasks’. In a
CR task, language becomes the topic of talk and for serious-minded
learners language is surely a relevant topic. An interesting example of the
use of CR tasks can be found in Williams (2008), who drew on Concept-
based Language Teaching (Lantolf and Thorne, 2006) to enable learners
to develop a scientiﬁc concept for choice of auxiliaries in L2 French
(a persistent problem for learners), which they then applied in a task that
required them to verbalize their grammatical choices.

Issue 3: What Makes a Task Complex and How Can Tasks

Be Sequenced Effectively?
Ahmadian and Mayo (2018), in their introduction to a book examin-
ing recent trends in task-based language teaching, observed that little
attention has been paid to how tasks can be used as the basic units of
language teaching syllabi even though Van den Branden (2006) had
pointed to this as a major problem many years previously. Central to
this issue is the question of how to determine the complexity of tasks in
order to ensure a progression from ‘simple’ to ‘complex’ tasks. Long
(2016) saw this as a ‘real issue’. He noted that while ‘much good work
has been published on task complexity . . . the overall yield has been
disappointing’ (p. 27). He saw the solution as more research to ‘help
make ﬁndings cumulative, encourage replication studies, increase
productivity, and generally speed up progress on this issue’.
In the case of input-based tasks, however, the problem is less ‘real’
as there are established ways of determining the complexity of input –
for example, standard measures of readability such as the Dale and
Chall formula (1948). These measures were developed to measure the
readability of written texts but they may also be relevant for oral texts
as research has shown that the cognitive processes involved in reading
and listening are quite similar (Jobard, Crivello and Tzourio-Mazoyer,
2006). Word-frequency lists derived from corpora also provide a basis
for grading the oral and written texts used in input-based tasks. The
extensive work on simplifying input to make it comprehensible to
learners can also help (see Ellis, 2008, chapter 6). Long and Ross
(1993), for example, pointed to the advantage of what they called
‘elaborative simpliﬁcation’, which involves the restructuring of the
propositional content rather than the language of a text.
Determining the complexity of output tasks is, however, much
more challenging, as was noted in Chapter 1. This is what Long
(2016) was referring to when he considered the overall results

disappointing. The most developed framework for sequencing tasks

from simple to complex – and the one that has attracted the most
research – is Robinson’s (2001, 2011) Cognition Hypothesis. This
was discussed in Chapters 3 and 7 where we noted some obvious
problems. It cannot be assumed that a task designated as ‘complex’
in terms of Robinson’s framework actually entails a greater cognitive
load when it is performed. However, this can be addressed
by obtaining independent measures of the cognitive load posed by
a task – for example, by utilizing a dual-task methodology or by
collecting ratings from learners who perform the task, as in
Sasayama (2016). However, this may not solve the fundamental
problem of determining the complexity of production tasks. As noted
in Chapter 3, tasks are holistic and involve conglomerates of factors.
Research that investigates just one or two design variables hypothe-
sized to affect task complexity (e.g. +/ reasoning; +/ familiar
topic) may succeed in showing how these variables affect complexity
in these tasks but there is no guarantee that they will work in the
same way in a set of tasks that involve different clusters of features.
There is currently no theory of how the myriad design variables
interact to affect task complexity. The problem is exacerbated by
the fact that the complexity of any single task depends not just on its
design but also on how the task is implemented. Whether learners
have the opportunity to plan before or during the performance of a
task, for example, will affect the cognitive load that the task places
on the learner. Thus, the impact that design variables have on task
complexity cannot be determined independently of implementation
conditions. Skehan (2016), while not rejecting the importance of
design variables, felt that the research that has focused on them has
failed to ‘generate consistency or robust generalizations’ (p. 37) and
went on to suggest that studies that have investigated the effect of
task conditions – such as pre-task planning – have proved more
insightful. In short, we believe that we should not be optimistic that
a research agenda directed at identifying the design factors that
determine task complexity will bear much fruit.
The way forward is for course designers to make use of what
research has shown about task complexity but also to draw on their
own experience and intuition about what constitutes the right type
and level of task for a particular group of learners. The fact that this
may result in a less than precise sequencing of tasks in terms of their
intrinsic difﬁculty may matter less if due regard is given to how a task
is implemented with a particular group of learners. In other words,
methodology may help to overcome the problems of grading and
sequencing tasks in a syllabus.

Issue 4: What Is the Role of Explicit Instruction?

TBLT advocates agree that explicit instruction has a role to play in the
post-task stage of a lesson. There is some disagreement about whether
it has a role in the main task phase. Willis and Willis (2007) argue
that it will subvert the purpose of the task by causing learners to focus-
on-form rather than meaning. In contrast, Long (1991b) and Ellis
(2003) point to the need for drawing learners’ attention to form during
the performance of a task. This often involves corrective feedback but
can also include explicit instruction directed at linguistic problems if
these arise. A good example of how this can be carried out can be
found in the lesson based on the Things in Pocket Task reported in
Samuda (2001). The main point of disagreement, however, lies in
whether explicit instruction has a place in the pre-task stage. In task-
supported language instruction (TSLT) it clearly does. TSLT draws on
an inventory of linguistic forms which are first taught explicitly and
then practised in ‘real operating conditions’ using tasks in accordance
with skill-learning theory (DeKeyser, 1998). However, in the kind of
‘pure’ TBLT that Long (2015) promotes, it does not. Long argues that
there is no room for explicit instruction preceding the performance of
a task as this would constitute a return to ‘focus-on-forms’, which he
sees as antithetical to the principle that ‘learners, not teachers, have
most control over their development’ (p. 24).
In pitting ‘focus-on-form’ against ‘focus-on-forms’ some obvious
questions need to be asked. One such question is whether focus-on-
forms cannot result in true implicit knowledge, as Long claims. This
question is notoriously difficult to address given the difficulty in devis-
ing tests that afford separate measures of explicit and implicit know-
ledge It is now clear, however, that TSLT can result in the ability to
deploy the target of instruction in free production (see Norris and
Ortega’s, 2000, meta-analysis of form-focused studies). However,
TSLT studies that report a positive effect for functional grammar
teaching (e.g. Harley, 1989; Day and Shapson, 1991), typically
involve massive amounts of practice directed at the target feature
and, as Long points out, this is not feasible if teachers have to cover
all the linguistic features listed in the syllabus. But this does not justify
the wholesale rejection of TSLT, as it may still be of value in helping
learners acquire those linguistic features that are not learned ‘natur-
ally’, such as the target structures that figured in the studies of Day and
Shapson and Harley. In other words, while TSLT cannot replace
TBLT, it might usefully complement it (Ellis, 2018b).
There is, however, a possible drawback to providing explicit
instruction prior to the performance of a task. As Willis and Willis

(2007) suggested, it may interfere with how learners perform a task by

leading them to treat the task as a practice exercise rather than as
opportunity to communicate naturally. There has been surprisingly
little research that has addressed this question. Mochizuki and Ortega
(2008) reported that a group that received a handout providing expli-
cit information about English relative clauses produced more than
twice as many relative clauses when they performed the focused task
than a group that did not receive the handout. They also found that
there was no difference in the global complexity and fluency of the
production of the two groups. In the second study, Li et al. (2017),
also found that learners who had received explicit instruction in the
target structure (past passive) were more likely to try to use it when
they performed the tasks than learners who did not receive this
instruction. In this study, however, the explicit instruction resulted in
production that was less complex and fluent and tended to be less
accurate overall. Whether and to what extent explicit instruction
‘interferes’ in the way a task is performed is an important issue for
TSLT. If it does interfere, then clearly tasks are not functioning in
the same way as in TBLT. Even if they result in better learning of the
target structure, they will not provide the same opportunities for
natural language use and thus, in the long term, may be less successful
in developing all-round proficiency in the L2. Clearly, though, this is
an issue in need of further investigation.

Issue 5: Teachers and Students’ Negative Perceptions

about TBLT
The ﬁnal issue concerns how teachers and students perceive TBLT.
Where perceptions are negative it is unlikely that TBLT will work
effectively. The outsider critics’ critiques of TBLT derive from their
own negative perceptions about TBLT. But there is also plenty of
research evidence to point to the concerns that teachers and students
have about TBLT. East (2018) reported on this via his interviews with
teachers and teacher-educators in New Zealand. He found that
they were uncertain about the extent to which they needed to ‘let go’
in TBLT, noting that a tension arose when ‘teachers struggled to
reconcile the primary emphasis on communication with the support-
ing requirement to develop language knowledge’ (p. 224). The
teachers felt a need for explicit instruction as a complement to TBLT.
East concluded that the ‘teachers’ practices are most aligned with a
task-supported form of TBLT’ (p. 229). East (2012) also found that
the teachers were concerned about the amount of planning that TBLT
required of them, as there were few existing task-based resources.

Students can also have negative perceptions of TBLT if they are

wedded to an approach that treats language as a set of linguistic
objects to be mastered and so they therefore resist performing tasks
or just treat them as fun activities rather than as serious activities that
help them learn the language (Foster, 1998).
The literature on TBLT is in fact replete with the problems that
teachers have experienced in attempting to introduce TBLT into their
classrooms (see Chapter 11). These include structural problems that
arise from classroom-level and societal–institutional level constraints
(Butler, 2011) – for example large class sizes that make small group
difficult to manage (Li, 1998; Samimy and Kobayashi, 2004). Other
constraints include the structural syllabus that teachers may still be
required to teach to and discrete item tests emphasizing linguistic
accuracy that their students need to pass. Several commentators
(Littlewood, 2007, 2014; Butler, 2011) pointed to the conflicts that
exist in Asia between TBLT and culturally embedded traditional
teaching approaches although Lai (2015) warned against ‘essentialist
statements about cultural inappropriateness of TBLT in Asia’ (p. 14).
The problems that teachers face shape their perceptions of the
desirability and viability of TBLT. When students lack an understand-
ing of how tasks can help them develop proficiency in an L2 they will
resist performing them. Long (2016) rightly points out that many of
the problems associated with TBLT are also evident in structurally
based language teaching. But some are new, and, in any case, it is
much easier for teachers to tackle problems from the standpoint of a
teaching approach they are familiar with than from one that is new to
them.

Conclusion
In this chapter we have discussed a number of critiques of TBLT –
some derived from outsider critics who view TBLT as an inadequate
replacement for traditional mainstream approaches where language is
taught directly and explicitly, and others arising from what we have
called insider critics who advocate for TBLT but also see issues that
need to be addressed. Table 12.2 summarizes the various issues. We
have presented arguments to show that many of the outsider critics’
issues (points 1, 2, 3, 4, 5 and 6) are in fact non-issues arising from a
misunderstanding of the nature of TBLT (in particular its diversity)
and/or ignorance or rejection of the relevant research that gives sup-
port to TBLT. However, we have also acknowledged that some of the
issues they raise (points 7, 8 and 9) do need to be addressed. Of the
issues raised by insider critics, some derive from differences in their

Table 12.2 Critiques of TBLT

Outsider critics Insider critics

1. Tasks cannot serve as the units of a 1. Task-based research is limited in a
syllabus. number of respects.
2. TBLT does not help learners learn 2. What types of tasks should figure in a
‘new’ language. task-based course?
3. There is no grammar in TBLT. 3. What makes a task complex and how
can tasks be sequenced effectively?
4. Performing tasks encourages 4. What is the role of explicit instruction
indexical and minimal use of the in TBLT?
L2.
5. TBLT is not suited to low- 5. Teachers’ and students’ negative
proficiency learners. perceptions about TBLT.
6. TBLT assigns the teacher a very
limited role.
7. ‘Task’ is an ill-defined construct.
8. Learners will resort to using their
L1 when performing tasks.
9. TBLT is not suited to ‘acquisition-
poor’ environments.

views about what TBLT is (e.g. points 2, 3 and 4), while others are
generally acknowledged (e.g. points 1 and 5).
There is the obvious danger that focusing on problems and difﬁcul-
ties will lead readers to the conclusion that TBLT is not worth the
effort. However, we believe that it is essential to look critically at
TBLT and to distinguish the non-issues from the real issues. In doing
this, we can identify the problems that need to be looked at. In
particular, we see the need to consider what additional research is still
needed, what role task-based and task-supported teaching play in a
language curriculum and how the constraints that affect the introduc-
tion of TBLT in many instructional contexts can be addressed.

The final chapter is in three sections. The first returns to the questions
which were posed at the end of Chapter 1, and attempts answers to
each of them. The second section identifies a set of challenges facing
task-based learning and teaching, and outlines responses to these
challenges. The conclusion then offers brief suggestions for the future.

Section 1: Questions
The ten questions which ended Chapter 1 are wide-ranging in nature,
and set out the major areas which this book seeks to address. This
section tries to explore where the discussions in the various chapters of
the book have taken us. In some cases, the answers which can be
provided are fairly comprehensive, while in others it is clear there is
still scope for greater progress to be made (and hence the Challenges
section which follows).

Question 1: How should the central unit of task-based teaching – the

task – be defined?
Although Chapter 1 included this question in the set of ten questions,
in fact, that very same chapter had already offered a response: that
tasks are activities which make meaning primary, which include some
kind of gap which needs to be addressed and hopefully resolved,
which require learners to rely on their own language resources and
which have a clearly defined outcome. We feel this definition is still
operative after the coverage from the book. Its purpose was to distin-
guish tasks from exercises, and we feel that it does that. It is note-
worthy that the definition does not include explicit reference to the real
world (although other approaches might well make this a very import-
ant component of the definition). We feel that the existing components
of the definition ensure that there will be a real-world relationship in
the way a task is carried out, and it will be sufficiently authentic in the
way language is used.

353

Question 2: What kinds of tasks are appropriate for different groups

of learners? Is a needs-based approach for identifying target tasks
appropriate for all learners? Question 4: How can task-based teaching
be made to work for beginner-level learners who have no or very little
knowledge of the L2?
Here we are bundling two questions from Chapter 1 because, essen-
tially, we can consider Question 4 to be a special case within Question
2. There are a number of parts to the answer to this question. First of
all, while there is still room for progress, we have learned a great deal
about the factors which influence task difficulty, and so we have a
basis for finding tasks which are appropriate for different proficiency
levels, and different ages. Similarly, the importance of task conditions
has emerged very clearly and this also provides a way to increase the
pressure within a task, or more likely, to reduce it, for example,
through planning time provision, visual support and so on. Then it
is very important to stress that not all tasks are speaking tasks.
Shintani (2016) shows that input-based tasks with an emphasis on
listening work very effectively with young, beginner Japanese learners.
There are many options with tasks and how they are used, and this
openness has to be exploited to ensure the relevance and practicality
of tasks in different situations. Question 2 also raised the issue of
whether needs-based tasks are always appropriate. Here we have seen
different perspectives. Long (2015) advocates the use of needs analyses
followed by the development of associated pedagogic tasks. Through-
out this book, though, there has been a slightly different emphasis in
task definition. There does need to be a real-world relationship in a
task, to some degree, and learners have to use their own linguistic
resources. But the starting point for task design is not seen as inevit-
ably based on needs analyses. Pedagogic issues, previous experience
and learner interests are all likely to be independent motivators for
task design. The central issue is how the task fits into a task-based
approach to pedagogy and whether it will realize appropriate peda-
gogic goals. Needs analysis-based tasks are fine, and often totally
appropriate, but they are not the only way to identify desirable tasks.

Question 3: How can the problems of determining the complexity of

tasks be resolved to ensure that learners of different levels of proﬁ-
ciency are faced with tasks that pose a reasonable challenge?
The response to this question has to be that progress has been made,
but that the future looks even more promising. As discussed at various
points (Chapters 2, 3 and 9) a great deal of research has been con-
ducted into task complexity, with mixed results. There is some degree
of agreement, and a range of task features have been identiﬁed which
do have some consistency in the way they are used, in practical

settings, to organize task sequences (see Chapter 7). But there is also a
degree of unpredictability with tasks, as discussed in Chapter 3. This is
perhaps less evident with task conditions, although such conditions do
not relate to task complexity in the same direct way. There is, though,
still no convincing and uniﬁed theoretical account of task complexity
that is empirically grounded, only a range of (very interesting) general-
izations about particular task characteristics. The greatest hope for the
future has to come from two related areas. First, relating task complexity
to cognitive load, and in turn, to an examination of this construct
through the use of secondary measures (time estimation, secondary
tasks, etc.) may well enable an independent account of complexity to
emerge. Second, there may be a shift in the future from predominantly
quantitative approaches to research to the use of qualitative methods.
Each of these is taken up in the Challenges section. A point repeated at
several points in the book, however, is that it is the conditions under
which a task is performed, rather than the complexity of the task-as-
workplan, that has the greatest impact on the resulting activity.

Question 5: Is there a role for focused as well as unfocused tasks, and

if so, how should focused tasks be incorporated into a task-based
syllabus?
In general, it would seem a reasonable claim that tasks will be pre-
dominantly unfocused in nature. Such tasks are more typical of
research, certainly, and of much pedagogy. But we have seen that
there are powerful arguments for the use, at least some of the time,
of focused tasks. Chapters 2, 4 and 7 all discussed focused tasks in one
way or another. Samuda (2001, 2015) examines how tasks can create
pedagogic spaces and shows, through clever task design, how this can
be achieved. Ellis (2003) has argued that consciousness-raising tasks,
where the task itself concerns language, can be effective as tasks by
motivating learners to reﬂect upon aspects of language itself. Swain
(2006), through her concept of languaging, has also suggested that
tasks can elicit talk about language leading to greater understanding of
language itself. Then, in Chapter 7, we have seen how such tasks can
be incorporated into a syllabus. In addition, one would have to say,
that tasks which incorporate an integral focus on language also have
their convenience since they may enable more conventional syllabus
planning to be accomplished, resulting in greater systematicity in what
is covered. The key point is that what is done does not compromise the
‘taskiness’ of the activity.

Question 6: Is there merit in a modular curriculum that includes both

a task-based component and a traditional structural component? How
should such a curriculum be organized?

Any answer to this question has to have strong connections to other

questions, such as Questions 2 and 4 (covered together) and Question 5.
The issue was discussed in Chapters 7 and 12. The major point is that
one needs to relate a task-based syllabus to context, and recognize the
challenge that might arise in certain circumstances, with contexts such as
beginners, low-proficiency learners, acquisition-poor environments and
culturally challenging contexts. A modular approach then enables the
unfamiliarity and potentially threatening nature of a complete task-
based approach to be mitigated flexibly. As was described in Chapter 7,
this approach has many advantages, and might enable a task-based
approach to be acceptable where otherwise it might never be used. The
proposal is that there would need to be clear separation in time between
the two co-existing approaches (e.g. different days) and that a decision
would need to be made whether to start off with a more conventional
approach at earlier levels and move towards a task-based approach as
proficiency increases or vice-versa, where the aim would be to establish
basic proficiency with a task-based approach first and introduce a
structural component with focused tasks later on (see Ellis, 2018b).

Question 7: What alternatives are there for the organization of a task-

based lesson? Is the lesson format proposed by Willis (1996), which
has proved very influential, the only way?
The Willis approach contains pre-task, during-task and post-task
phases. Of these, only the during-task phase is required. As a starting
point here, it is clear that three phases of this sort have to be con-
sidered the norm for task-based and many other sorts of learning.
Using three phases strongly suggests that tasks need to be prepared for
and contextualized, and that simply completing a task is not enough
without some sort of reflection, extension or evaluation. The Willis
model, though, is specific in what is likely to happen at each of these
stages. For one thing, particular suggestions are proposed for the pre-
task and post-task phases, and it is inevitable that individual teachers
will make their own decisions at each of these stages, may not follow
Willis-type suggestions or may prefer to replace them. But the essential
point is that good task-based teaching is likely to prepare the ground
(with some form of input, model, topic preparation, vocabulary), and
the post-task phase is likely to take language made salient by the task
itself, and then do something with that language. Chapter 8 discusses
these issues at some length, and provides an example teaching
sequence. So broadly (if not in total detail) there is agreement between
the Willis approach and general practice.
The major area of potential difference concerns what happens
during the task phase itself. Willis advocates doing the task, then

building on what has been learned through planning and then com-
pletion of a linked, non-identical task. In addition, she rejects a focus-
on-form in the main task phase, arguing that learners should be entirely
focused on meaning. In contrast, most task research has emphasized
other factors that might be operative at the task phase, including a
focus-on-form (see Chapter 2) and explored other options like time
pressure (and opportunity for online planning), effect of support material
available, provision of feedback and so on (see Chapter 8). This contrast
brings out the central response to this question – teaching means decision
making – and embraces inﬁnite variety. It is likely that teachers will tread
their own path, and approaches such as that of Willis, or any other set of
proposals in task-based teaching, will have inﬂuence but will be adapted
in any particular teaching encounter.

Question 8: How can a focus-on-form be best incorporated into a

task-based lesson?
The basic point, covered at several places within the book, but espe-
cially in Chapters 2, 4 and 8, is that tasks (in which meaning has to be
primary) provide an arena within which it is essential that form does
not lose focus (while a task is being transacted) but that engineering
such a focus-on-form (FonF) is not always easy. Much research and
pedagogic innovation has been directed towards achieving such a
FonF. Clearest in this regard, and a signiﬁcant success within task-
based research, is our greater understanding of the range and variety
of feedback that can be provided and of the comparative effectiveness
of these different types of feedback (see Chapters 2 and 8). It is clear
that while a task is running, there are opportunities to provide FonF
without disturbing the interaction or compromising the primacy of
meaning. Focused tasks, as well as approaches such as languaging,
also show how form can come into focus. In addition, pre- and post-
work can also be relevant. In the ﬁrst case, pre-task work can make
aspects of form salient (Mochizuki and Ortega, 2008). In the latter,
post-task activities such as public performance or transcription of
one’s performance (Foster and Skehan, 2013; Li, 2014) can change
how attention was allocated when the task was done and make form
more salient. Post-task activities can also exploit the form which has
emerged during task completion and consolidate and develop it.

Question 9: How can teachers carry out formative assessments of

task-based lessons to gather evidence of whether learning is taking
place and what changes may be needed to the task?
We would like to see a wider range of possibilities in answer to this
question, and it connects with the need to develop more achievement

tests (see the next section). But the activity in this area has not been so
extensive. Chapter 11 discusses issues connected with programme
evaluation, and formative assessment certainly ﬁts in with this. But
there are some other initiatives which are worth mentioning. Two,
interestingly, involve self-assessment. Harrison (1982), covered in
Chapter 9, used a system of task-cards, each containing ten mini-
tasks. Each learner then worked through the mini-tasks by asking a
classmate to ‘sign off’ when a task was completed satisfactorily, and
the student only approached the teacher for the tenth and ﬁnal task. In
this way, important formative feedback was obtained about progress.
Winke (2014), with a CALL-based course, also had provision for
learners to self-assess, and, interestingly, these self-assessments showed
high agreement with teacher-based assessments. In this way learners
were engaging in assessment that provided them with fairly direct
information about their own progress. One other example that has
an element of formative assessment comes from the discussion in
Chapter 4 of dynamic assessment – this, by its nature, with scaffolding
being provided as an integral part of the assessment, is necessarily
formative in nature.

Question 10: What problems do teachers face in implementing task-

based teaching and how can these be addressed?
The problems here are numerous. They range across uncertainly as
to what ‘task’ actually means, whether grammatical development
might be neglected, how different learner types and contexts might
be inappropriate for a task-based approach, a general lack of
resources, the challenge of the different teacher role a task-based
approach introduces, whether external pressures (institutional or
assessment-linked) will render such an approach impossible or at
least very difﬁcult, excessive workload and even whether this
approach may lead to excessive noise! We have looked at some of
these already, and will revisit them in the Challenges section. There
have been successes, for example, with regard to different types
of learners (Chapters 7 and 8); with the options available to
ensure that grammatical development is not neglected (covered in
Chapters 2 and 8); and with assessment, to a limited degree (in
Chapter 9).
These responses are very important. But at a more general level,
three sorts of response hold some promise. First, very clearly, is the
dissemination of good practice. All the problems mentioned in this
section have been solved somewhere. The challenge is to convey to
other prospective task-based teachers what has been achieved in
sufﬁcient detail that the success has a chance of being replicated.
Second, and connected to this, there is the development of

international organizations, such as the International Association of

Task-Based Language Teaching (http://www.tblt.org/about/), which
can act as a clearinghouses for such good practice. Third, there is the
need to develop teacher training so that it prepares teachers for the
challenges that are likely to occur when task-based teaching is used.
Taken together, these three responses have a great deal to offer. We
will return to them in the next section particularly when we discuss
teacher training.

Section Two: Challenges for Task-Based Approaches

The previous section tried to address the questions from Chapter 1.
But arising from them, and also from the various chapters in the book
more generally, a number of challenges can be identified for task-based
learning and teaching. This section explores these challenges. In some
cases, the outline of a response is clear. In others it is not, and that is
part of the excitement of this approach – we have to wait and see how
the field develops.
Broadly speaking, three types of challenge are identified: theoretical,
research and practical. The categories are not mutually exclusive but
represent tendencies, and follow from different sections of the book.

Theoretical Challenges
1. COMPLEXITY AND COGNITIVE LOAD

Task complexity has had a considerable impact on task research, not

least because it is so important to two of the accounts (the Cognition
Hypothesis and the Limited Attention Capacity approach) that have
been proposed in this area. It is also important in more general
research terms. Any proposals for choosing tasks need to be grounded
on a clear view of complexity and the demands that will be placed on
the learner, and so it is relevant to any research study which uses tasks,
to language testing and to pedagogic decision making. As we have
touched upon, research findings in this area have had mixed success.
There have been achievements, but there are many examples of
research results not turning out as expected. Clearly a first reaction
here would be to hope that future research will overcome these prob-
lems and deliver more consistent and even predicted results. We have
learned through the problems which have already been identified, such
as participants in studies finding certain factors confusing and making
unforeseen interpretations (Inoue, 2013; Sasayama, 2015). The key
question is whether tasks are inherently unpredictable, or if it is simply
the case that research designs need to be improved. There are a couple

of developments which give grounds for optimism, and may pave the
way for generalizations to emerge more clearly. First, there are the
insights emerging from the use of qualitative research methods. These
provide a window to understand interpretations of task demands
more deeply (Pang and Skehan, 2014; Sasayama, 2015). Second, there
is the general area of using independent measures to establish cognitive
load, referred to earlier. In the past, claims have been made about task
complexity simply on the basis of researcher analysis. Such analyses
may be limited in effectiveness – excellent starting points but needing
to be supplemented by more objective measures. The range of possi-
bilities that we are now seeing in the ﬁeld, increasingly on a routine
basis, may be a crucial development. This knowledge will enable more
focused research designs which can (a) explore the relationship
between complexity and dimensions of performance and (b) provide
a sounder basis for theorizing, which includes tasks themselves and
also task conditions.

2. TASKS AND ACQUISITION

The emphasis in much task research is on performance itself. There

are suggestions within the Cognition Hypothesis and the Limited
Attention Capacity approach which do speak to acquisition (more
in the former than the latter) but they are limited. Of course, nurtur-
ing performance is important, and so all this research is relevant to
the ways changes in underlying language systems can be transformed
into effective real-time language use. But accounting for acquisition
itself has not been a major component of most task-based work. The
Interaction Hypothesis and the Cognition Hypothesis are stronger on
the process of change (through feedback) rather than on the course of
development or any account of underlying system change. Coupled
with this, task-research is often quite short-term in nature, and one
can make the oft-repeated (yet totally justified) critique that more
longitudinal research is needed. So one would hope to see, within
task research, longer-term research designs, and evidence that sus-
tained task work leads to development and increased proficiency.
The chapters on evaluation are encouraging in this regard. It would
be nice, though, to see connections between task research and under-
lying processes, to clarify whether we are dealing with explicit or
implicit processes (or, of course, most likely both); whether the focus
is automaticity and access, or alternatively underlying system
change. A model to account for these issues, specifically in
the context of task-based approaches, would be of enormous value.
Chapter 2 raised these issues. There is still a need for progress.

Research Challenges
The challenges in this section do not have strong theoretical underpin-
ning but are included simply because research gaps are evident. Existing
research is often suggestive as to where additional data would be worth
collecting to ﬂesh out a broader picture. Theory is not unimportant, but
the focus is more on research gaps which have been identiﬁed.

1. MEASUREMENT: CALF AND BEYOND

Complexity, accuracy and ﬂuency have been used in many of the

psycholinguistic studies covered in Chapter 3. More recently lexical
performance (diversity and sophistication) has been added to this basic
set of measures. In addition, Interaction Hypothesis studies have
tended to use the sorts of feedback types covered in Chapters 2 and
8. Finally, sociocultural theory approaches have tended to use a dif-
ferent approach, emphasizing patterns of interaction and particularly
language-related episodes (LREs), as well as evidence of dynamic
assessment. These various measures have served us well, and have
had a major impact on our understanding of performance. But in a
sense, they capture what has been learned from what might be termed
the ﬁrst phase of task research. We now have additional possibilities
we need to consider. Chapter 3, for example, discussed the greater
understanding we now have of sub-dimensions of CALF, and how this
knowledge needs to be reﬂected in more sophisticated performance
measures. In addition, there is clearly an urgent need, following
Révész, Ekiert and Torgersen (2016), to incorporate indices of
outcome and of acquisition. Other areas, relating to interaction, can
also explore new measures that might be used, of contingency, for
example. Such developments will be important. Task effects and con-
dition effects have already been detected with the measures we have
available. It is to be hoped that in the future newer, subtler, more valid
measures may do an even better job at detecting the impacts that are
hypothesized in task research designs.

2. TASKS AND INDIVIDUAL DIFFERENCES

Chapter 5 used Robinson’s Cognition Hypothesis and its triarchic

structure to locate the relevance of individual differences for task
research, particularly regarding claims about task difficulty. The fun-
damental point to make is that the model effectively provides a list,
only, of possible influences on task performance and task difficulty.
But the research base relevant to this is not very extensive, and so

currently, proposals are often speculative. One can wonder whether

different aspects of aptitude have greater or lesser influence on per-
formance with different sorts of task, as well as different (e.g. explicit
vs. implicit) types of learning. We need to know more about working
memory and its connection with different types of planning.
Motivation (and see also the discussion of involvement in Chapter 6)
may link importantly with the use of different types of task, and
possibly in different learning environments. There is also a case to
include additional ID variables, such as personality. Tasks involve
interaction, and it could be that different personality profiles respond
differently to various sorts of tasks. Even beyond this, there may be
potential for aptitude-treatment interaction studies which probe the
ways particular types of task, or task-driven syllabuses, for example,
interact with different sorts of ID. Much fundamental research in this
area needs to be done if we are to better understand how tasks
function with different learners and how we can gain greater under-
standing of task difficulty.

3. LONGITUDINAL RESEARCH

Pedagogy takes place over an extended time period. Research, very

frequently, takes place over a shorter time span, probably lasting
between as little as thirty minutes and rarely longer than two to four
weeks. The TBLT field has many findings and generalizations, and
these have been covered extensively in this book. But as mentioned
elsewhere, a major challenge is to forge greater connections between
theory-motivated research, on the one hand, and the practice of TBLT,
on the other. There is considerable scope to explore research designs
which last longer than two to four weeks, and come to resemble more
the lengths of contact that teachers typically have with their students.
What would be the consequences, for example, of researching
planning, or post-task interventions or repetition over something like
a fifteen-week period? Would planning lose or gain effectiveness, for
example? Could learners be trained to exploit planning more effect-
ively? Could planning be selectively directed at performance, or alter-
natively and more challenging, to promote acquisition? Similarly, one
can ask additional questions about a task feature such as structure,
given the positive findings which have been accumulated. Would
learners, with regular exposure to this variable, come to realize that
imposing their own structure is beneficial for performance? Again,
there is considerable scope here. A solid database of such longer-term
findings would enable more straightforward connections to be made
with pedagogy.

Pedagogic Challenges
1. THE TASK CYCLE AND HOW BEST TO SET UP (PRE-TASK WORK) AND
THEN CONSOLIDATE (POST-TASK WORK)

This issue was raised as Question 7, at the end of Chapter 1. Broadly,

as we saw, it is uncontroversial to suggest a three-part structure for
teaching (pre-, during-, post-), and this links nicely with much task
research. Researchers try to manipulate and control, in an attempt to
identify the effects of particular variables. But the real challenge is to
identify optimal environments for teaching, and this means judicious
combinations of pre-, during and post- activities which go well beyond
the conﬁnes of particular, focused research studies. The alternatives
for pre-task work were set out in Chapter 8, and we need to learn
more about their practical effectiveness in classrooms. Similarly, the
during-task options are extensive, and so we need to learn more about
useful advice for teachers regarding, for example, feedback options,
timing options and so on. Finally, the post-task stage may be the most
complex of all. Tasks make language salient, and so the post-task stage
may be ideal to capitalize upon this. How this is best done becomes a
major challenge in task work. But all along there is a tension between
maximizing the chances that there will be a focus on language, and
most ambitiously, conditions where new language becomes salient,
and then developing more automatized skills with that new language.
With pre-, during- and post-task phases, the tension is essentially the
same: Should the general aim be to push for change and development
or should it be to achieve greater levels of control? The challenge is to
ﬁnd optimal ways of achieving this. We have learned a great deal
about the different options available – now we need to learn how they
can be combined.

2. TASKS AND EXTENDED TEACHING: ARE THERE OPTIMAL SYLLABUSES?

Educational contexts vary. Elsewhere in this book we have discussed

ways that tasks can be shown to be relevant for a variety of learners
(young, in acquisition-poor contexts and so on). We have also seen
that there are circumstances where a needs-based approach to syllabus
design is appropriate. But the general educational context is a chal-
lenge, and in many parts of the world, even though using tasks may be
seen as desirable, there are also countervailing forces which argue for
more traditional methods. In such circumstances to move to a fully
ﬂedged TBLT approach may be impractical, and so a modular
approach may be more realistic as an option. It is important to say
that the goal of a TBLT approach is not lost, but by taking a more

nuanced approach, greater practicality is achieved to work towards

the same general aim. We also need to move on and consider whether
different versions of a task-based approach are better than others,
particularly when we relate different approaches to different contexts.
This, too, is a considerable and important challenge for the future.

3. THE NEED FOR TEACHERS TO FIND AND/OR TO GENERATE TASKS

Tasks need to be appropriate to particular groups of learners and this

means that teachers are less likely to be able to simply rely on course
book tasks (and see answer to Question 10 in the previous section).
Their choices are either to find suitable tasks, or to generate the tasks
themselves. That is not to say that course book series do not have
useful tasks. It is more that since task-based courses are generally
learner-centred, it is likely that there will be a need for tasks which
are suitable for particular groups, and this may go beyond textbooks
(and be necessary in situations where there are no textbooks). If the
choices are finding or generating, then with the first of these, there is
considerable merit in ‘task banks’ and it is likely that, through the
aegis of bodies like IATBLT, these will grow in number and in organ-
ization so that it may be possible to search databases of tasks to find
ones that are appropriate for particular groups of learners, of particu-
lar ages and proficiency levels, in particular locations and so on.
Regarding the generation of tasks, in Chapter 1 the approach taken
by Jane Willis was described illustrating a number of pedagogic task
types: listing, ordering and sorting, comparing and classifying, prob-
lem solving, sharing personal experiences, creative tasks. These
options could be applied to endless topics. Imagine, for example,
trying to produce six tasks, following each of the possible options,
for dating, and for pets. This is a practical method of generating tasks
and undoubtedly there are others. Below we discuss the issue of
teacher training, and it is clear that in the context of task-based work,
a teacher training component needs to focus on methods such as
Willis’s in order to help teachers to meet this challenge during their
professional careers.

4. HOW CAN TASKS WORK IN DIFFERENT CONTEXTS AND WITH

DIFFERENT LEARNERS

We have explored this issue at different points during the book, and
indeed this was the focus for Questions 2 and 4 at the end of Chapter 1.
We have seen several ways in which problems can be addressed, with
respect to young learners, beginners and also in difﬁcult contexts such

as acquisition-poor settings and culturally challenging contexts. In

general, these initiatives have shown that these educational settings
do not preclude the use of tasks, but at the same time, it is not as if
there is a profusion of examples of how the various problems have
been overcome. Perhaps young learners and beginners have been the
most successful in this regard. Creative teachers (see the previous
section on generating tasks) and creative textbook writers have pro-
duced relevant materials in each case. In addition, by interpreting tasks
more broadly, and not exclusively focusing on speaking, it is clear that
judicious use of contextual support can provide interesting and chal-
lenging educational environments. In addition, the use of a modular
approach to syllabus design can go some way to reducing the scale of
these problems.
Acquisition-poor environments are indeed an issue but, as was
pointed out earlier, this problem is not exclusive to task-based
approaches – it is general, and only disguised in ‘organized’ and
teacher-dominated classrooms. Two approaches to addressing this
problem are, in fact, more available with a task-based context. First,
there is a natural connection between task-based approaches and
project work. In other words, it is natural, within a task-based
context, to transcend the limits of the classroom and to plan how
additional language opportunities that do exist can be exploited
(Fried-Booth, 1986). This links with a second response to this same
problem – the potential of technology to provide language use oppor-
tunities even in acquisition-poor environments. The internet (and see
the section on CALL) provides a wealth of resources, with audio as
well as print materials. If projects are designed to use such materials,
there will be many opportunities for listening and reading tasks, and
even writing tasks also. Task-based approaches, with the autonomy
they foster, can enable learners of many ages and proficiencies, pro-
vided they are stimulated with appropriate tasks, to go well beyond
the limitations of their own particular learning context.
A related worry is that a task-based approach is inappropriate
because of a mismatch between the processes and procedures of using
tasks and cultural expectations regarding the role of the teacher and
the nature of learning. We saw discussion of this earlier in relation, for
example, to some Asian cultures where it has been argued that the
authoritative role of the teacher and a view of learning as transmissible
knowledge is seen as provoking difficulty (Littlewood, 2007). Indeed
there are accounts of difficulties when task-based approaches have
been tried in such contexts (e.g. Carless, 2004). But first of all, we
need to consider whether other issues were at play, not simply cultural
mismatch. Lack of appropriate teacher training (and see the section

‘Teacher Roles and Training’) might be one such factor, as would a

mismatch between task-based teaching and the nature of the examin-
ation system, or indeed, an unrealistic expectation about the speed
with which an innovation can be introduced.
We would argue that these are early days regarding the impact of
any cultural mismatch, and that there are grounds for optimism,
provided that realism and some willingness for persistence are
involved. In this book we have seen examples of task-based
approaches being successfully implemented in Japan, for example. In
China, too, there are important initiatives. Yafu Gong (2014) has
published a major task-based coursebook series, following extensive
research into the sorts of tasks that are appropriate for mainland
China school-age learners. He has also provided a book-length ration-
ale for this work (Gong, 2015) and this shows that task-based insights
are relevant even in what might be considered to be the most challen-
ging of circumstances.
Even so, these are serious challenges for a task-based approach. To
be able to point to impressive examples of how these challenges have
been responded to by committed teachers and educational groups is
one thing; for the approach to gain general currency is quite another.
The real challenge here is the dissemination of good practice so that
those who might not otherwise consider using a task-based approach
become more enthusiastic about doing so.

5. CALL AND THE USE OF TECHNOLOGY

The major issue here is that technology is pervasive, and so not to

exploit the possibilities it enables would probably be regarded as very
odd by most students! The possibilities themselves are endless. There is
scope to access text, audio and video sources on the web or through
social media; there are opportunities to contact native L2 speakers
directly, and in real-time; it will also undoubtedly become common-
place to access artiﬁcial intelligence-driven resources, for information
and for interaction. All these are changing language learning possibil-
ities rapidly (not simply in relation to tasks), and recasting what an
acquisition-poor environment means. Teachers and learners focusing
on listening skills can access huge quantities of ever-changing mater-
ials, speaking can be facilitated by international interactions, reading
has the resources of the entire web available and writing can be
enhanced by massive access to relevant input material as well as
support software for the process of writing itself. Tasks, clearly, have
to be alert to all these possibilities, and the evidence is that they are, for
research and, more relevantly here, for pedagogy.

The use of technology as a source of input and support is exempli-

ﬁed by Oskoz and Elola (2014), who were concerned with students
learning to write expository and argumentative texts. They show how
technology was used to scaffold learners through the stages of writing
(planning, drafting, getting feedback and revising). They also show
how the use of computer chats and wikis had a major role in driving
the content of what was written. As a contrast to this, several research-
ers report on initiatives to develop virtual worlds which enable
learners to interact with the locations and the characters in these
worlds to develop their language abilities. Learners are given tasks
that structure the way they can develop language, and it is interesting
to see how task theorizing and research can impact on the design of
these encounters. For example, in a task-based context, Sykes (2014)
explored the untiring nature of the virtual world to enable learners to
restart encounters intended to develop ability with the speech acts of
apologizing and requesting. Restarting was seen as highly useful in
automatizing the speech act skills concerned. Sykes also reported that
if one takes such a technological path, the real-world norms of gaming
have to be respected – encounters need to be player – rather than
learning-driven, for example, even if the real intention is that learning
should occur. Canto, De Graaf and Jauregi (2014) also uses a virtual
world, but theirs is intended to enhance capacity to handle intercul-
tural meanings, and show that tasks and technology can work
together in areas beyond simply language code. Gutierrez (2014)
shows the potential of the virtual world approach by basing its design
in her research on sociocultural theory – yet another twist on linking
tasks, technology and applied linguistic theory.
The web also contains incredible quantities of unstructured
resources. The relevance for linguistic input is very great, but one of
the major challenges will be how to organize the ways language
learners access this wealth of material. It is clear that the structure
provided by task-driven learning is likely to be the major way that this
happens, but that the value of doing so will depend on how effectively
this potential is curated – another major challenge for the future.

6. TEACHER ROLES AND TEACHER TRAINING

One of the criticisms made of tasks (covered in Chapter 12) is that the
approach leaves the teacher with nothing to do. As indicated earlier,
we ﬁnd this a curious claim. In some ways we would ﬁnd it easier to
understand the opposite critique, namely that a task-based approach is
very demanding of teachers! The central issue, of course, is that a task-
based approach necessarily encourages a greater degree of learner

autonomy, and then the question concerns the ways a teacher con-
tinues to exert control, to monitor what is happening and to design
future work (and keep in mind institutional pressures such as examin-
ation systems) while learners function relatively independently within
tasks. Added to this, teachers may need to design tasks on a regular
basis, or at least adapt existing ones. Then they need to deal
with alternatives at the post-task stage, where choices of extending,
analysing, consolidating and automatizing all come into play. Taking
a slightly longer timescale, there may be a need to design effective
formative assessments. The set of challenges is forbidding!
A major issue we then have to consider is the way teachers are
prepared for such challenges, and in that respect the focus switches
towards the procedures within pre-service or in-service teacher educa-
tion. Much current teacher education attempts to develop sets of skills
which will serve teachers well during their careers. But the emphasis,
generally, is towards teachers as deliverers of materials within conven-
tional educational systems. The wider range of challenges that a task-
based approach requires does not have a corresponding emphasis. As
a consequence, a major issue for the future will be the development of
changes within teacher training systems to equip teachers for such a
contrasting approach. Until that is done, there will be a considerable
impediment to the wider use of tasks. Task-based teachers have to be
knowledgeable, resourceful and ﬂexible, and teacher training needs to
provide them with the beginnings of the skills they will need to develop
further.

7. ACHIEVEMENT TESTING

This is certainly the area where a magic wand is most needed. Long’s
(2015) claim, as we saw, is that progress with task-based achievement
testing would be the single most effective advance for the widespread
use of task-based approaches. It was also clear in Chapter 9 that while
there have been interesting initiatives in this area, progress is limited.
To develop a point made there, the way forward would seem to be the
establishment of collaboration and networks of teachers. Testing is
time-consuming and requires unavoidable standards, e.g. for
reliability and validity. In addition, unlike testing in other areas where
a set of test questions can be devised and reused, it is inherent with
task-based approaches that developing achievement tests based on
tasks-as-workplans is not sufficient. What happens in classrooms
may have a relationship to tasks-as-workplans, but it is going to go
well beyond it, and achievement testing will necessarily need to reflect
this. Therefore, this is going to be a serious challenge for the field, so

there is a need to ﬁnd frameworks to promote and to catalyse collabor-

ation. We brieﬂy saw two examples. Colpin and Gysen (2006) report
on the development of a series of achievement tests for several groups
of quite different learners. Harrison (1982) was associated with the
development of achievement tests for French with secondary (high)
school-age children in the UK. Interestingly Colpin and Gysen worked
in a university education department and Harrison worked for an
examination board, with all these authors working with teachers in
the development of the achievement tests. This may be an important
approach for achievement testing – initiatives bringing together people
with general testing expertise and actual teachers able to shape tests
that are appropriate to particular local circumstances. Obviously there
needs to be compromise here. The tests which are developed in these
cases have to have some generality, but at the same time they should
address the detail and variation in the different teaching contexts.

Section Three: Theory and Praxis

It is said that responses to innovations go through three stages, as a
marker of acceptability:
1 This innovation will never work.
2 This innovation will work but it is unimportant, marginal, minor!
3 That’s not an innovation. I’ve been doing it for a long, long time!
We think it reasonable to place task-based approaches between
numbers 2) and 3) on this ‘scale’, and realistically, closer to 2)! This,
of course, means that such approaches are taken seriously and have
some importance, but are far from mainstream. One can wonder,
therefore, what will happen next.
It is our contention that the vitality represented in this book, the
research, the practical applications and the variety of approaches will
all contribute to the continuing importance of a task-based approach.
There is the fundamental advantage that this approach aims at
developing the capacity to use language effectively, and that this is
what will characterize language education in the twenty-ﬁrst century.
The capacity to be effective in languages is likely to be even more
important in a future characterized by increasing internationalization.
Task-based approaches are appropriate to the challenge this
represents.
The book has shown that task-based approaches have certainly
come of age. Signiﬁcant progress has been made in the area of theory,
with alternative accounts each offering interesting interpretations of
what happens when tasks are used, and how these relate to underlying

disciplines like linguistics, psycholinguistics, sociolinguistics, educa-

tion and so on. A considerable amount of relevant research has been
completed, and even more important, research methods have been
refined and have evolved as greater understanding of issues like task
complexity and performance measurement has been reached. There
have also been impressive accounts of syllabus and methodology, as
well as examples of task-based approaches in operation in a variety of
circumstances. Comparative evaluations have been published (with
encouraging results) –see Chapter 10 – and increasingly task practi-
tioners have developed methods of critiquing and examining what
they have done in class. In addition, the formation of the IATBLT
group – which organizes a biennial international conference, has a
growing website and provides an important source of contact for task-
based researchers and teachers alike – is a major development. Its
growth is a testament to the vitality of the field. Local TBLT organiza-
tions, such as those in Japan, have also flourished.
A slight worry one might mention is the relationship between theory
and practice. Both have been covered in the book, but it has to be said
that in the field more generally, there has been something of a separ-
ation between the two. In an ideal world a closer and clearer connec-
tion between theory and practice would have been better. But it is
possible to state the opposite argument. Both areas have developed
strongly, and rather independently, but have been aware of the other’s
existence, and frequently bridges have been built between them. It may
be, in fact, that it has been better to have a degree of separation as each
of them has developed through important stages. Now, with each in a
position of some strength, the way they can come together may lead to
greater synergies and progress both theoretical and practical.

1 The Pedagogic Background to Task-Based

Language Teaching
1. Prabhu’s Communicational Teaching Project actually started in 1979.
Another early programme that could lay claim to be being task-based was
in Malaysia. The Malaysian Language Syllabus (see Richards, 1984) specified
objectives in task-like terms and suggested ‘procedures’ for realizing them.
2. As a result of experience with the Communicational Teaching Project, Prabhu
did offer some quite concrete suggestions for grading and sequencing tasks.
Beginning tasks needed to be input- rather than output-based. Information-
gap tasks were easier than reasoning-gap tasks, which in turn were easier than
opinion-gap tasks.
3. Loschky and Bley-Vroman (1993) acknowledged that it is very difficult to
design a structure-based task that makes the production of a predetermined
linguistic feature essential given that the choice of what linguistic resources to
use when performing the task is still left to the learner. Achieving task
essentialness is much easier with input-based tasks that learners can only
perform successfully if they are able to process the target structure.
4. The conviction that in TBLT the traditional distinction between syllabus and
methodology is no longer relevant dies hard, however. Van den Branden (2006)
continued to claim that the ‘what’ and the ‘how’ are blurred in TBLT (see p. 6).
5. Long (2005), however, claimed that ‘every language course should be con-
sidered a course for specific purposes, varying only (and considerably, to be
sure) in the precision with which learner needs can be specified’ (p. 1).
6. Richards and Rodgers (2014) also concluded that TBLT was unlikely to
provide a basis for national teaching programmes but suggested it could
constitute a ‘partial approach’ alongside a traditional language-based syllabus.

2 Cognitive-Interactionist Perspectives
1. Lyster and Ranta’s list is not exhaustive. Nassaji and Fu (2016) identiﬁed
twelve types of CF in ten hours of classroom interactions in an adult Chinese-
as-a-foreign-language classroom – recasts, delayed recasts, clariﬁcation

371

requests, translation, metalinguistic feedback, elicitation, explicit correction,

asking a direct question, repetition, directing a question to other students,
re-asks and using L1 English.
2. Long (2015) repeatedly links incidental and implicit learning.
3. Long considers it still an open question whether explicit learning is necessary
for L2 adults. He accepts, however, that ‘intentional and explicit learning are
likely to speed up acquisition and so becomes a legitimate component of a
theory of ISLA (instructed language acquisition)’ (2015, p. 65).
4. In fact, though, input-based tasks can still allow for two-way interaction as
learners may elect to speak even though the task does not require them to do so.
5. There is, of course, also non-interactive input – for example, when someone
reads a text, listens to a tape or the radio or watches TV. Tasks can provide
non-interactive input but this is not the focus of this chapter.
6. Ellis et al. (2001) proposed a similar construct to LREs – FFEs. By and large
studies of small group interactions have preferred LREs while studies of
whole classroom interaction have investigated form-focused episodes.
7. Robinson (2007b) also argued that task complexity affects the CAF of L2
production. This is considered in Chapter 3.
8. Uptake-with-repair certainly indicates that learners have noticed but may not
constitute evidence of deep processing if the learners have just echoed the
corrected utterance.
9. Learners need multiple exposures to lexical items to learn them but extensive
exposure is even more necessary for grammatical features. The learners in
Shintani (2016) demonstrated better productive knowledge of the target
feature than did the learners in Erlam and Ellis (2018), reﬂecting the greater
exposure in her study (nine as opposed to two lessons).
10. In fact, one reading of the CF experimental research is that what it shows is
not that CF is effective but that explicit instruction + CF is effective.

4 Sociocultural Perspectives
1. ‘Tasks’ here refers to any stimulus for talk given to a learner. It includes tasks
in the technical sense of the term in this book but it also includes what would
qualify as ‘exercises’ (see Chapter 1 for this distinction).
2. Wertsch, Minick and Arns (1984) pointed out that the motives of the teachers
and rural mothers were culturally determined. Rural mothers, for example,
were socialized to engage in activities (e.g. producing and selling artefacts
such as clothes and pottery) that necessitated the avoidance of error.
3. The use of the L1 is not restricted to establishing shared goals for a task.
Storch and Wigglesworth (2003), reported that six pairs of students (each
with a shared L1) in a task used their L1 for four functions: (1) task manage-
ment, (2) task clariﬁcation, (3) discussing vocabulary and meaning, and
(4) deliberating about grammar points.
4. Dynamic assessment studies have typically investigated grammar. However,
van Compernolle et al. (2016) and Qin and van Compernolle (in press) also
applied dynamic assessment to socio-pragmatic aspects of language.

5. In fact, for Lantolf, SCT does not provide a rationale for TBLT because TBLT
does not aim to develop the learner’s scientiﬁc concepts of language.

11 Evaluating Task-Based Language Teaching

1. This list of the kinds of issues that can be addressed in micro-evaluations
comes from the collection of practitioner research studies reported in
Edwards and Willis (2005).
2. Prabhu (1990) pointed out a number of problems with Beretta’s interpret-
ation of the project.
3. Other examples of de facto action research studies of TBLT can be found in
Edwards and Willis (2005), where teachers focused on some speciﬁc aspect of
TBT that interested them and then investigated this in the context of their
own classrooms.

12 Responding to the Critics of Task-Based Language

Teaching
1. TBLT is not just informed by SLA research. It is also supported by the
principles of sound education – see Chapter 6.
2. Despite his doubts about TBLT, Seedhouse (2017) has been closely involved
in one of the most innovatory task-based projects in recent years where
learners participate in a digital kitchen to carry out instructions about how
to prepare different food dishes in a foreign language.
3. Skehan (1998) did suggest that tasks could be designed that would involve
clusters of speciﬁc linguistic features. Such tasks could be considered semi-
focused. It might be possible to pre-test in this case.

Abbs, B. & Freebairn, I. (1982). Opening Strategies. Harlow: Longman.

Ableeva, R. & Lantolf, J. (2011). Mediated dialogue and microgenesis of second
language listening comprehension. Assessment in Education: Principles, Policy
and Practice 18, 133–49.
Adams, R., Nuevo, A. & Egi, T. (2011). Explicit and implicit feedback, modified
output, and SLA: Does explicit and implicit feedback promote learning and
learner-learner interaction? Modern Language Journal 95, 42–63.
Ahmadian, M. J. (2012). The relationship between working memory capacity and
L2 oral performance under task-based careful online planning condition.
TESOL Quarterly 46, 165–75.
Ahmadian, M. J. & Mayo, M. (2018). Introduction recent trends in task-based
language teaching and learning. In Ahmadian, M. & Mayo, M. (eds.), Recent
Perspectives on Task-Based Language Teaching (pp. 1–8). Boston/Berlin:
Mouton de Gruyter.
Ahmadian, M. J. & Tavakoli, M. (2011). The effects of simultaneous use of careful
online planning and task repetition on accuracy, complexity, and fluency in EFL
learners’ oral production. Language Teaching Research 15, 35–59.
Aida, Y. (1994). Examination of Horwitz, Horwitz, and Cope’s construct of
foreign language anxiety: The case of students of Japanese. The Modern
Language Journal 78(2), 155–68.
Al Khalil, M. (2011). Second language motivation: Its relationship to noticing,
affect, and production in task-based interaction. PhD dissertation, Georgetown
University.
Alderson, J. & Beretta, A. (eds.) (1992). Evaluating Second Language Education.
Cambridge: Cambridge University Press.
Alderson, J., Clapham, C. & Steel, D. (1997). Metalinguistic knowledge, language
aptitude and language proficiency. Language Teaching Research 1, 93–121.
Alderson, J. & Scott, M. (1992). Insiders, outsiders and participatory evaluation.
In Alderson, J. & Beretta, A. (eds.), Evaluating Second Language Education
(pp. 25–58). Cambridge: Cambridge University Press.
Aljaafreh, A. & Lantolf, J. (1994). Negative feedback as regulation and second
language learning in the Zone of Proximal Development. The Modern Lan-
guage Journal 78, 465–83.

374

Allwright, D. (2003). Exploratory practice: Rethinking practitioner research in

language teaching. Language Teaching Research 7, 113–41.
Allwright, D. (2005). Developing principles for practitioner research: The case for
exploratory practice. Modern Language Journal 89, 353–66.
Allwright, D. & Hanks, J. (2009). The Developing Learner: An Introduction to
Exploratory Practice. Basingstoke: Palgrave Macmillan.
Ammar, A. & Spada, N. (2006). One size fits all? Recasts, prompts, and L2
learning. Studies in Second Language Acquisition 28, 543–74.
Aoki, M. (2016). English heads for elementary school in 2020 but hurdles
abound. Japan Times, 5 September.
Appel, C. & Gilabert, R. (2002). Motivation and task performance in a task-
based web-based tandem project. ReCALL 14, 16–31.
Arslanyilmaz, A. (2013). Computer-assisted foreign language instruction: Task
based vs. form focused. Journal of Computer Assisted Learning 29, 303–18.
Aubrey, S. (2017a). Inter-cultural contact and flow in a task-based Japanese EFL
classroom. Language Teaching Research 21, 717–34.
Aubrey, S. (2017b). Measuring flow in the EFL classroom: Learners’ perceptions
of inter-cultural and intra-cultural task-based interactions. TESOL Quarterly
51, 661–92.
Avermaet, P. & Gysen, S. (2006). From needs to tasks: Language learning needs in
a task-based approach. In Van den Branbden, K. (ed.), Task-Based Language
Education: From Theory to Practice (pp. 17–46). Cambridge: Cambridge
University Press.
Bachman, L. F. (1990). Fundamental Considerations in Language Testing.
Oxford: Oxford University Press.
Bachman, L. F. (2002). Some reflections on task-based language performance
assessment. Language Testing 19, 453–76.
Bachman, L. F. & Palmer, A. S. (1996). Language Testing in Practice. Oxford:
Oxford University Press.
Bachman, L. F. & Palmer, A. S. (2010). Language Testing in Practice, second
edition. Oxford: Oxford University Press.
Baddeley, A. (2007). Working Memory, Thought, and Action. Oxford: Oxford
University Press.
Bao, R. & Kirkebaek, M. (2013). Danish students’ perceptions of task-based
teaching in Chinese. In Kirkebæk, M., Du X.-Y. & Jensen A. (eds.), Teaching
and Learning Culture (pp. 61–78). Rotterdam: Sense Publishers.
Baralt, M. (2014). Task complexity and task sequencing in traditional versus
online language classes. In Baralt, M., Gilabert, R. & Robinson, P. (eds.), Task
Sequencing and Instructed Second Language Learning (pp. 59–122). New
York: Bloomsbury Academic.
Baralt, M. & Gurzynski-Weiss, L. (2011). Comparing learners’ state anxiety
during task-based interaction in computer-mediated and face-to-face commu-
nication. Language Teaching Research 15(2), 201–29.
Baralt, M., Gurzynski-Weiss, L. & Kim, Y. (2016). Engagement with language:
How examining learners’ affective and social engagement explains successful
learner-generated attention to form. In Sato, M. & Ballinger, S. (eds.), Peer

Interaction and Second Language Learning: Pedagogical Potential and

Research Agenda (pp. 209–40). Amsterdam: John Benjamins.
Bax, S. (2003). The end of CLT: A context approach to language teaching. ELT
Journal 57, 278–87.
Benevides, M. & Valvona, C. (2008). Widgets: A Task-Based Course in Practical
English. Hong Kong: Pearson Longman Asia ELT.
Beretta, A. (1989). Attention to form or meaning? Error treatment in the Banga-
lore project. TESOL Quarterly 23, 283–303.
Beretta, A. (1990). Implementation of the Bangalore project. Applied Linguistics
11, 321–37.
Beretta, A. & Davies, A. (1985). Evaluation of the Bangalore project. ELT Journal
39, 121–7.
Berwick, R. (1990). Task Variation and Repair in English as a Foreign Language.
Kobe: Kobe University of Commerce/Institute of Economic Research.
Bialystok, E. (1994). Analysis and control in the development of second language
proﬁciency. Studies in Second Language Acquisition, 16, 157–68.
Blake, R. (2000). Computer-mediated communication: A window on L2 Spanish
interlanguage. Language Learning & Technology 4, 120–36.
Block, D. (2003). The Social Turn in Second Language Acquisition. Edinburgh:
Edinburgh University Press.
Breen, M. (1984). Processes in syllabus design. In Brumﬁt, C. (ed.), General
English Syllabus Design (pp. 47–60). Oxford: Pergamon.
Breen, M. (1989). The evaluation cycle for language learning tasks. In Johnson,
R. K. (ed.), The Second Language Curriculum (pp. 187–206). Cambridge:
Cambridge University Press.
Brindley, G. (2013). Task-based assessment. In Chapelle, C. A. (ed.), The Encyclo-
pedia of Applied Linguistics (online). Oxford: Blackwell.
Brooks, F. & Donato, R. (1994). Vygotskyan approaches to understanding foreign
language learner discourse during communicative tasks. Hispania 77, 262–74.
Brooks, L. & Swain, M. (2014). Contextualizing performances: Comparing per-
formances during TOEFL iBTTM and real-life academic speaking activities.
Language Assessment Quarterly, 11, 353–73.
Brown, G. & Yule, G. (1983). Teaching the Spoken Language: An Approach
Based on the Analysis of Conversational English. Cambridge: Cambridge
University Press.
Brown, G., Andersen, Shilcock & Yule, G. (1984). Teaching Talk: Strategies for
Production and Assessment. Cambridge: Cambridge University Press.
Brown, J. D., Hudson, T., Norris, J. & Bonk, W. J. (2002). An Investigation
of Second Language Task-Based Performance Assessments. Honolulu, HI:
University of Hawai’i Press.
Bruner, J. S. (1960). The Process of Education. Cambridge, MA: Harvard
University Press.
Bruton, A. (2002a). From tasking purposes to purposing tasks. ELT Journal 56,
280–8.
Bruton, A. (2002b). When and how language development in TBI? ELT Journal
56, 296–7.

Burns, A. (2010). Doing Action Research for English Language Teachers:

A Guide for Practitioners. New York: Routledge.
Butler, Y. (2004). What level of English proficiency do elementary school teachers
need to attain to teach EFL? Case studies from Korea, Taiwan, and Japan.
TESOL Quarterly 38, 245–78.
Butler, Y. (2011). The implementation of communicative and task-based teaching
in the Asia-Pacific region. Annual Review of Applied Linguistics 31, 36–57.
Butler, Y. (2017a). Young learners’ rationales for self-assessing their performance
with and without tasks. Unpublished paper in 7th International Conference on
Task-Based Language Teaching, Barcelona.
Butler, Y. (2017b). Motivational elements of digital instructional games:
A study of young L2 learners game designs. Language Teaching Research
21, 735–50.
Bygate, M. (2001). Effects of task repetition on the structure and control of
oral language. In Bygate, M., Skehan, P. & Swain, M. (eds.), Researching
Pedagogic Tasks: Second Language Learning, Teaching and Testing
(pp. 23–48). Harlow: Longman.
Bygate, M. (2018). Dynamic systems: Theory and the issue of predictability
in task-based language: Some implications for research and practice in TBLT.
In Ahmadian, M. & Mayo, M. (eds.), Recent Perspectives on Task-Based
Language Teaching (pp. 146–66). Boston/Berlin: Mouton de Gruyter.
Calvert, M. & Sheen, Y. (2015). Task-based language learning and teaching:
An action-research study. Language Teaching Research 19, 226–44.
Cambridge English. (n.d.). Available at: www.cambridgeenglish.org.
Cameron, L. (2001). Teaching Languages to Young Children. Cambridge:
Cambridge University Press.
Canale, M. (1983). On some dimensions of language proficiency. In Oller, J. (ed.),
Issues in Language Testing Research (pp. 333–42). Rowley, MA: Newbury
House.
Canale, M. & Swain, M. (1980). Theoretical bases of communicative approaches
to second language teaching and testing. Applied Linguistics 1, 1–47.
Cancino, H., Rosansky, E. & Schumann, J. (1978). The acquisition of English
negatives and interrogatives by native Spanish speakers. In Hatch, E. (ed.),
Second Language Acquisition. Rowley, MA: Newbury House.
Candlin, C. (1987). Towards task-based language learning. In Candlin, C. &
Murphy, D. (eds.), Language Learning Tasks. Englewood Cliffs, NJ: Prentice
Hall International.
Canto, S., De Graaf, R. & Jauregi, K. (2014). Collaborative tasks for negotiation
of intercultural meaning in virtual worlds and video-web communication.
In González-Lloret, M. & Ortega, L. (eds.), Technology-Mediated
TBLT: Researching Technology and Tasks (pp. 183–212). Amsterdam: John
Benjamins.
Carless, D. (2003). Student use of the mother tongue in the task-based classroom.
ELT Journal 62, 331–8.
Carless, D. (2004). Issues in teachers’ reinterpretation of a task-based innovation
in primary schools. TESOL Quarterly 38, 639–62.

Carr, W. & Kemmis, S. (1986). Becoming Critical: Education, Knowledge and

Action Research. London: The Falmer Press.
Carroll, J. (1961). Fundamental considerations in testing for English language
proficiency of foreign students. In Center for Applied Linguistics (ed.), Testing
the English Proficiency of Foreign Students (pp. 30–40). Washington, DC:
Center for Applied Linguistics.
Carroll, J. (1981). Twenty-five years of research on foreign language aptitude. In
Diller, K. (ed.), Individual Differences and Universals in Language Learning
Aptitude. Rowley, MA: Newbury House.
Carroll, J. & Sapon, S. (1959). Modern Language Aptitude Test. New York:
The Psychological Corporation/Harcourt Brace Jovanovich.
Catani, M., Dell’Acqua, F. & De Schotten, M. (2013). A revised limbic system
model for memory, emotion and behaviour. Neuroscience and Biobehavioral
Reviews 37, 1724–37.
Cheng, Y., Horwitz, E. K. & Schallert, D. L. (1999). Language anxiety: Differen-
tiating writing and speaking components. Language Learning 49(3), 417–46.
Colpin, M. & Gysen, S. (2006). Developing and introducing task-based language
tests. In Van den Branden, K. (ed.), Task-Based Language Education: From
Theory to Practice (pp. 151–74). Cambridge: Cambridge University Press.
Conway, A., Kane, M., Bunting, M., Hambrick, D. & Engle, R. (2005). Working
memory span tasks: A methodological review and user’s guide. Psychonomic
Bulletin & Review 12, 769–86.
Cook, V. (2001). Using the first language in the classroom. Canadian Modern
Language Review 57, 402–23.
Corder, S. P. (1967). The significance of learners’ errors. International Review of
Applied Linguistics 5, 161–9.
Côté, S. & Gaffney, C. (2018). The effect of synchronous computer-mediated
communication on beginner L2 learners’ foreign language anxiety and partici-
pation. The Language Learning Journal First view: 1–12.
Coughlan, P. & Duff, P. (1994). Same task, different activities: Analysis of a SLA
task from an activity theory perspective. In Lantolf, J. & Appel, G. (eds.),
Vygotskyan Approaches to Second Language Research (pp. 173–93).
Norwood, NJ: Ablex.
Council of Europe. (2001). Common European Framework of Reference
for Language Learning, Teaching, and Assessment. Cambridge: Cambridge
University Press and Council of Europe.
Council of Europe. (2008). Relating Language Examinations to the Common Euro-
pean Framework of Reference for Language Learning, Teaching, Assessment.
Available online at: www.coe.int/t/dg4/portfolio/documents/exampleswriting.pdg.
Cowan, N. (2015). Second-language use, theories of working memory, and the
Vennian mind. In Wen, Z., Mota, M. B. & McNeill, A. (eds.), Working
Memory in Second Language Acquisition and Processing: Theories, Research,
and Commentaries (pp. 29–40). Clevedon: Multilingual Matters.
Cox, D. (2005). Can we predict language items for open tasks? In Edwards, C. &
Willis, J. (eds.), Teachers Exploring Tasks in English Language Teaching (pp.
171–86). Basingstoke: Palgrave Macmillan.

Crespo, M. (2011). The effects of task complexity on L2 oral production as

mediated by differences in working memory capacity. MA thesis, University
of Barcelona.
Crookes, G. (1989). Planning and interlanguage variation. Studies in Second
Language Acquisition 55, 367–83.
Crookes, G. & Schmidt, R. (1991). Motivation: Reopening the research agenda.
Language Learning 41, 469–512.
Csikszentmihalyi, M. (1975). Beyond Boredom and Anxiety. San Francisco:
Jossey-Bass.
Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience.
New York: Harper Perennial.
Cutrone, P. & Beh, S. (2014). Welcome to Kyushu: A Task-Based Approach to
EFL Learning Using Authentic Dialogues. Tokyo: Shohakusha.
Dale, E. & Chall, J. (1948). A formula for predicting readability: Instructions.
Educational Research Bulletin 27, 37–54.
Day, E. & Shapson, S. (1991). Integrating formal and functional approaches to
language teaching in French immersion: An experimental study. Language
Learning 41, 25–58.
De Bot, K. (1992). A bilingual production model: Levelt’s ‘Speaking’ model
adapted. Applied Linguistics 13, 1–24.
De Bot, K. & Larsen-Freeman, D. (2011). Researching second language develop-
ment from a dynamic systems theory perspective. In Verspoor, M., de Bot, K.
& Lowie, W. (eds.), A Dynamic Approach to Second Language Development.
Amsterdam: John Benjamins.
De Graaf. (1997). Effects of Explicit Instruction on Second Language Acquisition.
Netherlands: Holland Institute of Generative Linguistics.
De Jong, N. H., Groenhout, R., Schoonen, R. & Hulstijin, J. (2013). L2
fluency: Speaking style or proficiency? Correcting measures of L2 fluency for
L1 behaviour. Applied Psycholinguistics 36, 1–23.
De la Fuente, M. J. (2006). Classroom L2 vocabulary acquisition: Investigating
the role of pedagogical tasks and form-focused instruction. Language Teaching
Research 10, 263–95.
DeKeyser, R. (1998). Beyond focus on form: Cognitive perspectives on learning
and practicing second language grammar. In Doughty, C. & Williams, J. (eds.),
Focus on Form in Classroom Second Language Acquisition (pp. 42–63).
Cambridge: Cambridge University Press.
DeKeyser, R. (2007). Skill acquisition theory. In VanPatten, B. & Williams, J.
(eds.), Theories of Second Language Acquisition: An Introduction
(pp. 94–112). New York: Routledge.
Dembovskaya, S. (2009). Task-based instruction: The effect of motivational and
cognitive pre-tasks on second language oral production. PhD thesis, University
of Iowa.
De Ridder, I., Vangehuchten, L. & Gomez, M. (2007). Enhancing automaticity
through task-based language learning. Applied Linguistics 28, 309–15.
Dewey, J. (1913). Interest and Effort in Education. Boston: Houghton Mifflin.
Dewey, J. (1938). Experience and Education. New York: Simon and Schuster.

Donato, R. (1994). Collective scaffolding in second language earning. In Lantolf,

J. & Appel, G. (eds.), Vygotskian Approaches to Second Language Research
(pp. 33–56). Norwood, NJ: Ablex.
Donato, R. (2000). Sociocultural contributions to understanding the foreign
and second language classroom. In Lantolf, J. (ed.), Sociocultural Theory
and Second Language Learning (pp. 27–50). Oxford: Oxford University
Press.
Dörnyei, Z. (2002). The motivational basis of language learning tasks. In Robin-
son, P. (ed.), Individual Differences and Instructed Language Learning.
Amsterdam: John Benjamins.
Dörnyei, Z. (2005). The Psychology of the Language Learner: Individual Differ-
ences in Second Language Acquisition. Mahwah, NJ: Lawrence Erlbaum
Associates.
Dörnyei, Z. & Kormos, J. (2000). The role of individual and social variables in
oral task performance. Language Teaching Research 4, 275–300.
Dörnyei, Z. & Ottó, I. (1998). Motivation in action: A process model of L2
motivation. Working Papers in Applied Linguistics (Thames Valley University,
London) 4, 43–69.
Dörnyei, Z. & Ushioda, E. (eds.). (2009). Motivation, Language Identity and the
L2 Self. Clevedon: Multilingual Matters.
Doughty, C. (2001). Cognitive underpinnings of focus on form. In Robinson, P.
(ed.), Cognition and Second Language Instruction. Cambridge: Cambridge
University Press.
Doughty, C. (2003). Instructed SLA: Constraints, compensation, and enhance-
ment. In Doughty, C. & Long, M. (eds.), Handbook of Second Language
Acquisition (pp. 256–310). New York: Blackwell.
Douglas, D. (2000). Assessing Languages for Speciﬁc Purposes. Cambridge:
Cambridge University Press.
Dulay, H. & Burt, M. (1973). Should we teach children syntax? Language
Learning 23, 245–58.
Dunn, W. & Lantolf, J. (1998). Vygotsky’s zone of proximal development
and Krashen’s i +1: Incommensurable constructs; incommensurable theories.
Language Learning 48, 411–42.
Duran, G. & Ramaut, G. (2006). Tasks for absolute beginners and beyond:
Developing and sequencing tasks at basic proﬁciency levels. In Van den Bran-
den, K. (ed.), Task-Based Language Education: From Theory to Practice
(pp. 47–75). Cambridge: Cambridge University Press.
East, M. (2012). Task-Based Language Teaching from the Teachers’ Perspective:
Insights from New Zealand. Amsterdam: John Benjamins.
East, M. (2014). Encouraging innovation in a modern foreign language initial
teacher education programme: What do beginning teachers make of task-based
language teaching? The Language Learning Journal 42, 261–74.
East, M. (2018). ‘If it is all about tasks, will they learn anything?’ Perspectives
on grammar instruction in the task-oriented classroom. In Ahmadian, M. &
Mayo, M. (eds.), Recent Perspectives on Task-Based Language Teaching
(pp. 217–31). Boston/Berlin: Mouton de Gruyter.

Eckerth, J. (2008). Task-based language learning and teaching – Old wine in new
bottlers? In Eckerth, J. & Siekmann, S. (eds.), Task-Based Language Learning
and Teaching: Theoretical, Methodological, and Pedagogical Perspectives
(pp. 13–46). Frankfurt am Main: Peter Lang.
Edwards, C. & Willis, J. (eds.). (2005). Teachers Exploring Tasks in English
Language Teaching. Basingstoke: Palgrave Macmillan.
Egbert, J. (2003). A study of flow theory in the foreign language classroom.
Modern Language Journal 87, 499–518.
Egi, T. (2007). Interpreting recasts as linguistic evidence: The roles of linguistic
target, length, and degree of change. Studies in Second Language Acquisition
29, 511–37.
Ehrman, M. E. & Oxford, R. L. (1995). Cognition plus: Correlates of language
learning success. The Modern Language Journal 79, 67–89.
Eisenstein, M. (1980). Childhood bilingualism and adult language learning
aptitude. Applied Psychology 29, 159–72.
Ekiert, M., Lampropoulou, S., Révész, A. & Torgerson, E. (2018). The
effects of task-type and L2 proficiency on discourse appropriacy in oral task
performance. In Taguchi, N. & Kim, Y. (eds.), Task-Based Approaches
to Teaching and Assessing Pragmatics (pp. 247–64). Amsterdam: John
Benjamins.
Elkhafaifi, H. (2005). Listening comprehension and anxiety in the Arabic
language classroom. The Modern Language Journal 89(2), 206–20.
Ellis, D. (2011). The role of task complexity in the linguistic complexity of
native-speaker output. Qualifying paper, PhD in Second Language Acquisition
Program, College, Park, MD: University of Maryland.
Ellis, N. (2005). At the interface: Dynamic interactions of explicit and
implicit language knowledge. Studies in Second Language Acquisition 27,
305–52.
Ellis, N. & Robinson, P. (2008). An introduction to cognitive linguistics and
second language acquisition. In Robinson, P. & Ellis, N. (eds.), Handbook of
Cognitive Linguistics and Second Language Acquisition (pp. 3–24). London:
Routledge.
Ellis, R. (1984). Classroom Second Language Development. Oxford: Pergamon.
Ellis, R. (1987). Interlanguage variability in narrative discourse: Style shifting in
the past tense. Studies in Second Language Acquisition 9, 1–19.
Ellis, R. (1993). Second language acquisition and the structural syllabus. TESOL
Quarterly 27, 91–113.
Ellis, R. (1994). A theory of instructed second language acquisition. In Ellis, N.
(ed.), Implicit and Explicit Learning of Languages (pp. 79–114). San Diego,
CA: Academic Press
Ellis, R. (1997). SLA Research and Language Teaching. Oxford: Oxford Univer-
sity Press.
Ellis, R. (1999). Learning a Second Language through Interaction. Amsterdam:
John Benjamins.
Ellis, R. (2000). Task-based research and language pedagogy. Language Teaching
Research 4, 193–220.

Ellis, R. (2001). Non-reciprocal tasks, comprehension and second language

acquisition. In Bygate, M., Skehan, P. & Swain, M. (eds.), Researching
Pedagogic Tasks: Second Language Learning, Teaching, and Testing
(pp. 49–74). Harlow: Longman.
Ellis, R. (2003). Task-Based Language Learning and Teaching. Oxford: Oxford
University Press.
Ellis, R. (2005). Planning and task-based research: Theory and research. In Ellis,
R. (ed.), Planning and Task Performance in a Second Language. Philadelphia:
John Benjamins Publishing Company.
Ellis, R. (2007). The differential effects of corrective feedback on two grammatical
structures. In Mackey, A. (ed.), Conversational Interaction and Second
Language Acquisition: A Series of Empirical Studies. Oxford: Oxford Univer-
sity Press.
Ellis, R. (2008). Evaluating and researching grammar consciousness-
raising tasks. In Rea-Dickens, P. & Germaine, K. (eds.), Managing Evaluation
and Innovation in Language Teaching (pp. 220–52). London: Longman.
Ellis, R. (2009a). Task-based language teaching: Sorting out the misunderstand-
ings. International Journal of Applied Linguistics 16, 221–46.
Ellis, R. (2009b). The differential effects of three types of task planning on the
ﬂuency, complexity, and accuracy in L2 oral production. Applied Linguistics
30, 474–509.
Ellis, R. (2010). Cognitive, social, and psychological dimensions of
corrective feedback. In Batstone, R. (ed.), Sociocognitive Perspectives
on Language Use and Language Learning (pp. 151–65). Oxford: Oxford
University Press.
Ellis, R. (2011). Macro- and micro-evaluations of task-based teaching. In Tom-
linson, B. (ed.), Materials Development in Language Teaching, second edition
(pp. 212–35). Cambridge: Cambridge University Press.
Ellis, R. (2012). Language Teaching Research and Language Pedagogy. Malden,
MA: Wiley-Blackwell.
Ellis, R. (2015a). Teachers evaluating tasks. In Bygate, M. (ed.), Domains and
Directions in the Development of TBLT (pp. 247–70). Amsterdam: John
Benjamins.
Ellis, R. (2015b). Understanding Second Language Acquisition, second edition.
Oxford: Oxford University Press.
Ellis, R. (2016). Focus on form: A critical review. Language Teaching Research
20, 405–28.
Ellis, R. (2017a). Moving task-based language teaching forward. Language
Teaching 50, 441–82.
Ellis, R. (2017b). The case for introducing task-based language teaching in Asian
primary schools. Invited talk at Kansai University, Japan (4 July).
Ellis, R. (2018a). Reﬂections on Task-Based Language Teaching. Clevedon:
Multilingual Matters.
Ellis, R. (2018b). Towards a modular curriculum for using tasks. Language
Teaching Research. First view: https://doi.org/10.1177/1362168818
765315.

Ellis, R. (2019). Explicit versus implicit oral corrective feedback. In Hossein, N. &
Kartchava, E. (eds.), Cambridge Handbook on Corrective Feedback.
Cambridge: Cambridge University Press.
Ellis, R., Basturkmen, H. & Loewen, S. (2001). Pre-emptive focus on form in the
ESL classroom. TESOL Quarterly 35, 407–32.
Ellis, R., Basturkmen, H. & Loewen, S. (2002). Doing focus-on-form. System 30,
419–32.
Ellis, R. & He, X. (1999). The roles of modiﬁed input and output in the incidental
acquisition of word meanings. Studies in Second Language Acquisition 21,
285–301.
Ellis, R. & Heimbach, R. (1997). Bugs and birds: Children’s acquisition of second
language vocabulary through interaction. System 25, 247–59.
Ellis, R., Li, S. & Zhu, Y. (2018). The effects of pre-task explicit instruction on the
performance of a focused task. System 80, 38–47.
Ellis, R., Loewen, S. & R. Erlam. (2006). Implicit and explicit corrective feedback
and the acquisition of L2 grammar. Studies in Second Language Acquisition
28, 339–68.
Ellis, R. & Mifka-Provozic, N. (2013). Recasts, uptake and noticing. In Bergsleit-
ner, J., Frota, S. & Yoshioka, J. (eds.), Noticing and Second Language Acqui-
sition: Studies in Honor of Richard Schmidt (pp. 61–79). Honolulu, HI:
University of Hawai'i at Manoa, National Foreign Language Center.
Ellis, R. & Shintani, N. (2014). Exploring Language Pedagogy through Second
Language Acquisition Research. New York: Routledge.
Ellis, R., Tanaka, Y. & Yamazaki, A. (1994). Classroom interaction, comprehen-
sion, and the acquisition of L2 word meanings. Language Learning 44, 449–91.
Ellis, R. & Yuan, Y. (2004). The effects of planning on ﬂuency, complexity, and
accuracy in second language narrative writing. Studies in Second Language
Acquisition 26, 59–84.
Engeström, Y. (1999). Activity theory and individual and social transformation.
In Engeström, Y., Miettinen, R. & Punmäki, R.-L. (eds.), Perspectives on
Activity Theory (pp. 19–38). Cambridge: Cambridge University Press.
Erlam, R. (2006). Elicited imitation as a measure of L2 implicit knowledge: An
empirical validation study. Applied Linguistics 27, 464–91.
Erlam, R. (2016). I’m still not sure what a task is: Teachers designing language
tasks. Language Teaching Research 20, 275–99.
Erlam, R. & Ellis, R. (2018). Task-based language teaching for beginner-level
learners of L2 French: An exploratory study. Canadian Modern Language
Review 74(1), 1–26.
Erlam, R., Ellis, R. & Batstone, R. (2013). Oral corrective feedback on L2 writing:
Two approaches compared. System 41, 257–68.
Estaire, S. & Zanon. J. (1994). Planning Classwork: A Task-Based Approach.
Oxford: Heinemann.
Ewald, J. D. (2007). Foreign language learning anxiety in upper-level classes:
Involving students as researchers. Foreign Language Annals 40(1), 122–42.
Feuerstein, R., Falik, L. & Rynders, J. E. (1988). Don’t Accept Me as I Am.
Helping Retarded Performers Excel. New York: Plenum.

Foster, P. (1998). A classroom perspective on the negotiation of meaning. Applied

Linguistics 19, 1–23.
Foster, P. (2001). Rules and routines: A consideration of their role in the task-
based language production of native and non-native speakers. In Bygate, M.,
Skehan, P. & Swain, M. (eds.), Researching Pedagogic Tasks: Second Lan-
guage Learning, Teaching, and Testing (pp. 75–93). Harlow: Longman.
Foster, P. & Skehan, P. (1996). The influence of planning on performance in task-
based learning. Studies in Second Language Acquisition 18, 299–324.
Foster, P. & Skehan, P. (1999). The influence of source of planning and focus
of planning on task-based performance. Language Teaching Research 3,
185–214.
Foster, P. & Skehan, P. (2012). Complexity, accuracy, fluency and lexis in task-
based performance: A synthesis of the Ealing research. In Housen, A., Kuiken,
F. & Vedder, I. (eds.), Dimensions of L2 Performance and Proficiency: Com-
plexity, Accuracy, and Fluency in SLA (pp. 199–220). Amsterdam: John
Benjamins.
Foster, P. & Skehan, P. (2013). The effects of post-task activities on the accuracy
of language during task performance. Canadian Modern Language Review 69,
249–73.
Foster, P. & Tavakoli, P. (2009). Lexical diversity and lexical selection:
A comparison of native and non-native speaker performance. Language Learn-
ing 59, 866–96.
Foster, P. & Wigglesworth, G. (2016). Capturing accuracy in second language
performance: The case for a weighted clause ratio. In Mackey, A. (ed.), Annual
Review of Applied Linguistics 36, 98–116.
Foster, P., Tonkyn, A. & Wigglesworth, G. (2000). Measuring spoken language:
A unit for all reasons. Applied Linguistics 21, 354–75.
Fotos, S. & R. Ellis. (1991). Communicating about grammar: A task-based
approach. TESOL Quarterly 25, 605–28.
Fredricks, J., Blumenfeld, P. & Paris, A. (2004). School engagement: Potential of
the concept, state of evidence. Review of Educational Research 74, 59–105.
Fried-Booth, D. (1986). Project Work. Oxford: Oxford University Press.
Fried-Booth, D. (1989). COLLINS COBUILD English Course Tests. London:
Collins.
Fu, M. & Li, S. (2017). The associations between the cognitive processes of task
performance and working memory. Modern Foreign Languages 40, 114–24.
Fulcher, G. (1996). Does thick description lead to smart tests? A data-based
approach to rating scale construction. Language Testing 13, 208–28.
Gal’perin, P. I. (1989). Mental actions as a basis for the formation of thoughts and
images. Soviet Psychology 27, 45–64.
Galaczi, E. (2008). Peer–peer interaction in a speaking test: The case of the
first certificate in English examination. Language Assessment Quarterly 5,
89–119.
Galaczi, E. & ffrench, A. (2010). Context validity. In Taylor, L. (ed.), Examining
Speaking: Studies in Language Testing 30 (pp. 112–70). Cambridge:
Cambridge University Press.

Gardner, R. C. (1985). Social Psychology and Second Language Learning: The

Role of Attitudes and Motivation. London: Edward Arnold.
Gardner, R. C. & Lambert, W. E. (1965). Language aptitude, intelligence, and
second-language achievement. Journal of Educational Psychology 56, 191–9.
Gass, S. (1997). Input, Interaction, and the Development of Second Languages.
Mahwah, NJ: Erlbaum.
Gass, S. (1998). Apples and oranges: Or why apples are not oranges and don’t
need to be. A response to Firth and Wagner. Modern Language Journal 82,
83–90.
Gass, S. & Mackey, A. (2000). Stimulated Recall Methodology in Second
Language Research. Mahwah, NJ: Erlbaum
Gass, S. & Mackey, A. (2007). Input, interaction, and output in second language
acquisition. In VanPatten, B. & Williams, J. (eds.), Theories in Second
Language Acquisition (pp. 175–200). London: LEA.
Gass, S. & Mackey, A. (2015). Input, interaction, and output in second language
acquisition. In VanPatten, B. & Williams, J. (eds.), Theories in Second
Language Acquisition (pp. 180–206). New York: Routledge.
Gass, S. & Varonis, E. (1985). Task variation and nonnative/nonnative negoti-
ation of meaning. In Gass, S. & Madden, C. (eds.), Input in Second Language
Acquisition. Rowley, MA: Newbury House.
Gass, S. & Varonis, E. (1994). Input, interaction and second language production.
Studies in Second Language Acquisition 16, 283–302.
Gass, S. & Varonis, M. (1984). The effect of familiarity on the comprehension of
nonnative speech. Language Learning 34, 65–89.
Gass, S., Mackey, A., Alvarez-Torres, M. & Fernandez-Garcıa, M. (1999).
The effects of task repetition on linguistic output. Language Learning 49,
549–81.
Gilabert, R. & Barón, J. (2013). The impact of increasing task complexity on L2
pragmatic moves. In Mackey, A. & McDonough, K. (eds.), Second Language
Interaction in Diverse Educational Settings (pp. 45–69). Amsterdam: John
Benjamins.
Gilabert, R. & Munoz, C. (2010). Differences in attainment and performance in a
foreign language: The role of working memory capacity. The International
Journal of Engineering and Science 10, 19–42.
Gilabert, R., Baron, J. & Levkina, M. (2011). Manipulating task complexity
across task types and modes. In Robinson, P. (ed.), Second Language Task
Complexity (pp. 105–40). Amsterdam: John Benjamins.
Gilabert, R., Baron, J. & Llanes, A. (2009). Manipulating cognitive complexity
across task types and its impact on learners’ interaction during oral perform-
ance. IRAL 47, 367–95.
Gilabert, R., Manchón, R. & Vasylets, O. (2016). Mode in theoretical and
empirical TBLT research: Advancing research agendas. Annual Review of
Applied Linguistics 36, 117–35.
Givon, T. (1985). Function, structure, and language acquisition. In Slobin, D.
(ed.), The Crosslinguistic Study of Language Acquisition, Vol. 1 (pp.
1008–25). Hillsdale, NJ: Lawrence Erlbaum Associates.

Gong, Y. (2014). New Notion English. Beijing: Foreign Language Teaching and
Research Press.
Gong, Y. (2015). Reconceptualising English Education: A Multi-Goal Approach
to English Curriculum Design for School-Age Learners. Beijing: CIP.
González-Lloret, M. & Nielson, K. (2015). Evaluating TBLT: The case of a task-
based Spanish program. Language Teaching Research 19, 525–49.
González-Lloret, M. & Ortega, L. (eds). (2015). Technology-Mediated TBLT:
Researching Technology and Tasks. Amsterdam: John Benjamins.
Goo, J. (2012). Corrective feedback and working memory capacity in interaction-
driven learning. Studies in Second Language Acquisition 31, 445–74.
Goo, J., Granema, G., Novella, M. & Yilmaz, Y. (2015). Implicit and
explicit instruction in L2 learning; Norris & Ortega (2000) revisited and
updated. In Rebuschat, P. (ed.), Implicit and Explicit Learning of Languages
(pp. 443–82). Amsterdam: John Benjamins.
Goo, J. & Mackey, A. (2013). The case against the case against recasts. Studies in
Second Language Acquisition 35, 127–65.
Granena, G. (2013). Individual differences in sequence learning ability and second
language acquisition in early childhood and adulthood. Language Learning
63, 665–704.
Granena, G. (2015). Cognitive aptitudes for implicit and explicit learning and
information-processing styles: An individual differences study. Applied Psy-
cholinguistics 37(3), 577–600.
Granena, G. (2016). Individual versus interactive task-based performance through
voice-based computer mediated interaction. Language Learning and Technol-
ogy 30, 40–59.
Grice, P. (1975). Logic and conversation. In Cole, P. & Morgan, J. (eds.), Syntax
and Semantics. 3: Speech Acts (pp. 41–58). New York: Academic Press.
Guará Tavares, M. (2011). Pre-task planning, working memory capacity, and L2
speech performance. Organon 51, 245–266.
Gurzynski-Weiss, L. & Baralt, M. (2014). Exploring learner perception and use of
task-based interactional feedback in FTF and CMC modes. Studies in Second
Language Acquisition 35, 1–28.
Gutierrez, G. A. G. (2014). The third dimension: A sociocultural theory approach
to the design and evaluation of 3D virtual worlds tasks. In González-Lloret, M.
& Ortega, L. (eds.), Technology-Mediated TBLT: Researching Technology
and Tasks (pp. 213–238). Amsterdam: John Benjamins.
Hadley. G. (2013). Review of task-based language teaching from the teacher’s
perspective. System 41, 194–6.
Hall, G. & Cook, G. (2012). Own-language use in language teaching and learn-
ing. Language Teaching 45, 271–308.
Halliday, M. (1973). Explorations in the Functions of Language. London: Arnold.
Han, Z.-H., Park, E. & Combs, C. (2008). Textual enhancement of input: Issues
and possibilities. Applied Linguistics 29, 597–618.
Harley, B. (1989). Functional grammar in French immersion: A classroom experi-
ment. Applied Linguistics 19, 331–59.

Harley, B. & Hart, D. (1997). Language aptitude and second language proficiency
in classroom learners of different starting ages. Studies in Second Language
Acquisition 19, 379–400.
Harley, B. & Hart, D. (2002). Age, aptitude and second language learning on a
bilingual exchange. In Robinson, P. (ed.), Individual Differences and
Instructed Language Learning. Amsterdam: John Benjamins.
Harley, B., Allen, J. P. B., Cummins, J. & Swain, M. (1990). The Development of
Second Language Proficiency. Cambridge: Cambridge University Press.
Harris, J. & Leeming, P. (2016). On Task 2. Tokyo: Abax.
Harrison, A. (1982). Student-centered testing: Assessing communication in pro-
gress. Levende Talen 372, 401–10.
Hedge, T. (2000). Teaching and Learning in the Language Classroom. Oxford:
Oxford University Press.
Henzl, V. (1979). Foreigner talk in the classroom. International Review of
Applied Linguistics 17, 159–67.
Hoey, M. (1983). On the Surface of Discourse. London: George Allen and Unwin.
Horwitz, E., Horwitz, M. & Cope, J. (1986). Foreign language classroom anxiety.
Modern Language Journal 70(2), 125–32.
Housen, A. & Kuiken, F. (2009). Complexity, accuracy, and fluency in second
language acquisition. Applied Linguistics 30(4), 461–73.
Howatt, A. P. R. (1984). A History of English Language Teaching. Oxford:
Oxford University Press.
Hu, R. (2013). Task-based language teaching: Responses from Chinese
teachers of English. The Electronic Journal for English as a Second Language
16(4), n.p.
Hulstijn, J. (2015). Language Proficiency in Native and Non-Native Speakers.
Amsterdam: John Benjamins.
Hwu, F., Wei, P. & Sun, S. (2014). Aptitude-treatment interaction effects on
explicit rule learning: A latent growth curve analysis. Language Teaching
Research 18, 294–319.
Hyland, F. (2003). Focusing on form: Student engagement with teacher feedback.
System 31, 217–30.
Hymes, D. (1971). On Communicative Competence. Philadelphia: University of
Pennsylvania Press.
Inoue, C. (2013). Task Equivalence in Speaking Tests. Bern: Peter Lang.
Inoue, C. (2016). A comparative study of the variables used to measure syntactic
complexity and accuracy in task-based research. The Language Learning
Journal 1, 1–18.
Ishikawa, T. (2007). The effect of manipulating task complexity along the +/-
here-and-now dimension on L2 written narrative discourse. In Garcia
Mayo, M. del P. (ed.), Investigating Tasks in Formal Language Learning
(pp. 136–56). Clevedon: Multilingual Matters.
Iwashita, N., McNamara. T. & Elder, C. (2001). Can we predict task difficulty in
an oral proficiency test? Exploring the potential of an information-processing
approach to task design. Language Learning 51, 401–36.

Jackson, D. O. & Suethanapornkul, S. (2013). The cognition hypothesis:

A synthesis and meta-analysis of research on second language task complexity.
Language Learning 63, 330–67.
Jauregi, K., de Graaff, R., van den Bergh, H. & Kriz, M. (2012).
Native/non-native speaker interactions through video-web communication:
A clue for enhancing motivation? Computer Assisted Language Learning
25(1), 1–19.
Jeon, I.-J. & Hahn, J.-W. (2006). Exploring EFL teachers’ perceptions of task-
based language teaching: A case study of Korean secondary school classroom
practice. Asian EFL Journal 8, 123–43.
Jepson, K. (2005). Conversation and negotiated interactions in text and voice chat
rooms. Learning & Technology 9, 79–98.
Jobard G., Crivello F. & Tzourio-Mazoyer N. (2003). Evaluation of the dual
route theory of reading: A meta-analysis of 35 neuroimaging studies. Neuro-
image 20, 693–712.
Johnson, K. (1982). Communicative Syllabus Design and Methodology. Oxford:
Pergamon.
Johnson, K. (2000). What task designers do. Language Teaching Research 4,
301–21.
Johnson, K. (2005). Fighting fossilization: Language at different stages in the
task cycle. In Edwards, C. & Willis, J. (eds.), Teachers Exploring Tasks in
English Language Teaching (pp. 191–200). Basingstoke: Palgrave
Macmillan.
Juffs, A. & Harrington, M. (2012). Aspects of working memory in L2 learning.
Language Teaching 44, 137–66.
Kahnemann, D. (2011). Thinking, Fast and Slow. London: Penguin.
Kahng, J. (2014). Exploring utterance and cognitive ﬂuency of L1 and L2 English
speakers: Temporal measures and stimulated recall. Language Learning 64,
809–54.
Keenan, J., MacWhinney, B. & Mayhew, D. (1977). Pragmatics in memory:
A study of natural conversation. Journal of Verbal Learning and Verbal
Behavior 16, 549–60.
Kelly, C. & Kelly, E. (1996). The Snoop Detective School Conversation Book.
Tokyo: Macmillan Language House.
Khabbazbashi, N. (2017). Topic and background knowledge effects on perform-
ance in speaking assessment. Language Testing 34, 23–48.
Kiele, R. & Rea-Dickens, P. (2005). Program Evaluation in Language Education.
Basingstoke: Palgrave Macmillan.
Kim, Y. (2009). The effects of task complexity on learner-learner interactions.
System 37, 254–68.
Kim, Y. (2012). Task complexity, learning opportunities, and Korean EFL
learners’ question development. Studies in Second Language Acquisition 34,
627–58.
Kim, Y. (2013). Effects of pre-task modelling on attention to form and question
development. TESOL Quarterly 47, 8–35.

Kim, J. & Lantolf, J. (2018). Developing understanding of sarcasm in

L2 English through explicit instruction. Language Teaching Research 22,
209–29.
Kim, Y., Payant, C. & Pearson, P. (2015). The intersection of task-based inter-
action, task complexity, and working memory. Studies in Second Language
Acquisition 37, 549–81.
Kim, Y. & Taguchi, N. (2015). Promoting task-based pragmatics instruction
in EFL classroom context: The role of task complexity. Modern Language
Journal 99, 656–77.
Kim, Y. & Tracy-Ventura, N. (2011). Task complexity, language anxiety, and the
development of the simple past. In Robinson, P. (ed.), Second Language Task
Complexity: Researching the Cognition Hypothesis of Language Learning and
Performance (pp. 287–306). Amsterdam: John Benjamins.
Kim, Y. & Tracy-Ventura, N. (2013). The role of task repetition in L2 perform-
ance development: What needs to be repeated during task-based interaction?
System 41, 829–40.
Klein, W. & Perdue, C. (1997). The basic variety (or: couldn’t natural languages
be much simpler?). Second Language Research 13, 301–47.
Kormos, J. (2006). Speech Production and Second Language Acquisition. Mah-
wah, NJ: Lawrence Erlbaum.
Kormos, J. (2011). Speech production and the cognition hypothesis. In Robinson,
P. (ed.), L2 Task Complexity: Researching the Cognition Hypothesis of Lan-
guage Learning and Performance (pp. 39–60). Amsterdam: John Benjamins.
Kormos, J. & Dörnyei Z. (2004). The interaction of linguistic and motivational
variables in second language task performance. Zeitschrift für interkulturellen
Fremdsprachenunterricht 9, 1–19.
Kormos, J., Kiddle, T. & Csizer, K. (2011). Systems of goals, attitudes, and self-
related beliefs in second-language learning motivation. Applied Linguistics 32,
495–516.
Kormos, J. & Trebits, A. (2011). Working memory capacity and narrative task
performance. In Robinson, P. (ed.), Second Language Task Complexity:
Researching the Cognition Hypothesis of Language Learning and Performance
(pp. 267–86). Amsterdam: John Benjamins.
Krashen, S. (1981a). Aptitude and attitude in relation to second language acquisi-
tion and learning. In Diller, K. (ed.), Individual Differences and Universals in
Language Learning Aptitude. Rowley, MA: Newbury House.
Krashen, S. (1981b). Second Language Acquisition and Second Language Learn-
ing. Oxford: Pergamon.
Krashen, S. (1985). The Input Hypothesis: Issues and Implications. London:
Longman.
Krashen, S. & Terrell, T. (1983). The Natural Approach: Language Acquisition in
the Classroom. Oxford: Pergamon.
Kuiken, F. & Vedder, I. (2008). Cognitive tasks complexity and written output in
Italian and French as a foreign language. Journal of Second Language Writing
17, 48–60.

Kumaravadivelu, B. (2001). Toward a postmethod pedagogy. TESOL Quarterly

35, 537–60.
Lai, C. (2015). Task-based language teaching in the Asian context: Where are
we now and where are we going? In Thomas, M. & Reinders, H. (eds.),
Contemporary Task-Based Teaching in Asia (pp. 12–29). London:
Bloomsbury.
Lai, C. & Li, G. (2011). Technology and task–based teaching: A critical review.
CALICO Journal 28, 498–521.
Lam, D. M. K. (2018). What counts as ‘responding’? Contingency on previous
speaker contribution as a feature of interactional competence. Language
Testing 35, 377–401.
Lambert, C. (1997). Motivation and personal investment in the learning process.
Journal of Nanzan Junior College 24, 55–88.
Lambert, C. (1998). The role of the learner in classroom task performance.
Journal of Nanzan Junior College 26, 85–101.
Lambert, C. (2002). Task sequencing and affective performance variables. Kita-
kyushu University Faculty of Foreign Studies Bulletin 103, 97–175.
Lambert, C. (2004). Reverse-engineering communication tasks. ELT Journal 58,
18–27.
Lambert, C. (2010). Task-based needs analysis: Putting principles into practice.
Language Teaching Research 14, 99–112.
Lambert, C. (2017). Tasks, affect and second language performance. Language
Teaching Research 21(6), 657–64.
Lambert, C., Gong, Q. & Zhang, G. (in press). Learner-generated content and the
lexical recall of beginning-level learners of Chinese as a Foreign Language.
Available at: www.academia.edu/39149761/_2019_Learner-generated_con
tent_and_the_lexical_recall_of_beginning-level_learners_of_Chinese_as_a_For
eign_Language.
Lambert, C. & Hailes, A. (2002). Simulations: A Task-Based Approach to Con-
versational English. Kitakyushu: The University of Kitakyushu, Faculty of
Foreign Studies.
Lambert, C., Kormos, J. & Minn, D. (2016). Task repetition and second language
speech processing. Studies in Second Language Acquisition 38, 1–30.
Lambert, C. & Minn, D. (2007). Personal investment in L2 task design and
learning: A case study of two Japanese learners of English. ELIA: Estudios
de Lingüística Inglesa Aplicada 7, 127–48.
Lambert, C., Philp, J. & Nakamura, S. (2017). Learner-generated content
and engagement in L2 task performance. Language Teaching Research 21,
665–80.
Lambert, C. & Robinson, P. (2014). Learning to perform narrative task:
A semester long study of task sequencing effects. In Baralt, M., Gilabert, R.
& Robinson, P. (eds.), Task Sequencing and Instructed Second Language
Learning (pp. 207–30). London: Bloomsbury.
Lambert, C. & Zhang, G. (2019). Engagement in the use of English and Chinese
as second languages: The roles of learner-generated and teacher-generated
content. Modern Language Journal 103(2), 391–411.

Lamendella, J. T. (1977). The limbic system in human communication. In

Whitaker, H. & Whitaker, H. A. (eds.), Studies in Neurolinguistics, Volume
3 (pp. 154–222). New York: Academic Press.
Lantolf, J. (2000). Introducing sociocultural theory. In Lantolf, J. (ed.), Sociocul-
tural Theory and Second Language Learning (pp. 1–26). Oxford: Oxford
University Press.
Lantolf, J. (2007). Conceptual knowledge and instructed second language
learning: A sociocultural perspective. In Fotos, S. & Nassaji, H. (eds.), For-
m-Focused Instruction and Teacher Education: Studies in Honour of Rod Ellis
(pp. 35–54). Oxford: Oxford University Press.
Lantolf, J. (2009). Dynamic assessment: The dialectic integration of instruction
and assessment. Language Teaching 42, 355–68.
Lantolf, J., Kurtz, L. & Kisselev, O. (2016). Understanding the revolutionary
character of L2 development in the ZPD: Why levels of mediation matter.
Language and Social Cultural Theory 3, 153–71.
Lantolf, J. P. & Poehner, M. E. (2014). Sociocultural Theory and the Pedagogical
Imperative in L2 Education: Vygotskian Praxis and the Theory/Practice
Divide. New York: Routledge.
Lantolf, J. & Thorne, S. (2006). Sociocultural Theory and the Genesis of Second
Language Development. Oxford: Oxford University Press.
Lantolf, J., Thorne, S. & Poehner, M. (2015). Sociocultural theory and second
language acquisition. In VanPatten, B. & Williams, J. (eds.), Theories in
Second Language Acquisition (pp. 207–26). New York: Routledge.
Lantolf, J. & Zhang, X. (2017). Concept-based language instruction. In Loewen,
S. & Sato, M. (eds.), The Routledge Handbook of Instructed Language Acqui-
sition (pp. 146–65). New York: Routledge.
Larsen-Freeman, D. (2000). Techniques and Principles in Language Teaching,
second edition. New York: Oxford University Press.
Laufer, B. & Hulstijn, J. (2001). Incidental vocabulary acquisition in a second
language: The construct of task-induced involvement. Applied Linguistics 22,
1–26.
Leaver, B. & J. Willis. (2004). Task-Based Instruction in Foreign Language
Education. Washington, DC: Georgetown University Press.
Lee, A. & Lyster, R. (2016). The effects of corrective feedback on instructed L2
speech production. Studies in Second Language Acquisition 38, 35–64.
Lee, L. (2005). Using web-based instruction to promote active learning: Learners’
perspectives. CALICO Journal 23, 139–56.
Lee, S. & Huang, S. (2008). Visual input enhancement and grammar learning.
A meta-analytic review. Studies in Second Language Acquisition 30, 307–31.
Leontiev, A. (1978). Activity, Consciousness, and Personality. Englewood Cliffs,
NJ: Prentice Hall.
Leontiev, A. (1981). Psychology and the Language Learning Process. London:
Pergamon.
Leow, R. (2015). Explicit and implicit learning in the L2 classroom: Processes and
products. In Rebuschat, P. (ed.), Implicit and Explicit Learning of Languages
(pp. 47–68). Amderdam: John Benjamins.

Leung, J. H. C. & Williams, J. N. (2011). The implicit learning of mappings

between forms and contextually derived meanings. Studies in Second Lan-
guage Acquisition 33, 33–55.
Levelt, W. J. (1989). Speaking: From Intention to Articulation. Cambridge, MA:
The MIT Press.
Levelt, W. J. (1999). Producing spoken language: A blueprint of the speaker. In
Brown, C. & Hagoort, P. (eds.), Neurocognition of Language (pp. 83–122).
Oxford: Oxford University Press.
Lewis, M. (1993). The Lexical Approach. Hove: Language Teaching Publications.
Li, D. (1998). It’s always more difficult than you planned: Teachers’ perceived
difficulties in introducing the communicative approach in South Korea.
TESOL Quarterly 32, 677–703.
Li, L., Chen, J. & L. Sun. (2015). The effects of different lengths of pre-task planning
time on L2 learners’ oral test performance. TESOL Quarterly 49, 38–66.
Li, Q. (2014). Get it right in the end: The effects of post-task transcribing on
learners’ oral performance. In Skehan, P. (ed.), Processing Perspectives on Task
Performance (pp. 129–54). Amsterdam: John Benjamins.
Li, S. (2010). The effectiveness of corrective feedback in SLA: A meta-analysis.
Language Learning 60, 309–65.
Li, S. (2013a). The interactions between the effects of implicit and explicit feed-
back and individual differences in language analytic ability and working
memory. Modern Language Journal 97, 634–54.
Li, S. (2013b). The differential roles of language analytic ability and working
memory in mediating the effects of two types of feedback on the acquisition of
an opaque linguistic structure. In Sanz, C. & Lado, B. (eds.), Individual
Differences, L2 Development & Language Program Administration: From
Theory to Application. Boston: Cengage Learning.
Li, S. (2015). The associations between language aptitude and second language
grammar acquisition: A meta-analytic review of five decades of research.
Applied Linguistics 36, 385–408.
Li, S. (2016). The construct validity of language aptitude. Studies in Second
Language Acquisition 38, 801–42.
Li, S. (2017). Teacher and learner beliefs about corrective feedback. In Nassaji, H.
& Kartchava, E. (eds.), Corrective Feedback in Second Language Teaching and
Learning (pp. 143–57). New York: Routledge.
Li, S., Ellis, R. & Zhu, Y. (2016). Task-based versus task-supported language
instruction: An experimental study. Annual Review of Applied Linguistics 36,
205–29.
Li, S., Ellis, R. & Zhu, Y. (2017). The influence of pre-task grammar instruction
on task performance and L2 learning in task-supported language instruction: A
process-product study.
Li, S. & Fu, M. (2017). Strategic and unpressured within-task planning and their
associations with working memory. Language Teaching Research 22, 230–53.
Li, S., Zhu, Y. & Ellis, R. (2016). The effects of the timing of corrective feedback
on the acquisition of a new linguistic structure. Modern Language Journal 100,
276–95.

Lin, T.-B. & Wu, C.-W. (2012). Teachers’ perceptions of task-based language teaching
in English classrooms in Taiwanese high schools. TESOL Journal 3, 586–609.
Linck, J., Osthus, P., Koeth, J. & Bunting, M. (2014). Working memory and
second language comprehension and production: A meta-analysis. Psycho-
nomic Bulletin & Review 21, 861–83.
Littlewood, W. (2007). Communicative and task-based language teaching in East
Asian classrooms. Language Teaching 40, 243–9.
Littlewood, W. (2014). Communication-oriented teaching: Where are we now?
Where do we go from here? Language Teaching 47, 249–362.
Loewen, S. (2005). Incidental focus on form and language learning. Studies in
Second Language Acquisition 27, 361–86.
Loewen, S., Erlam, R. & Ellis, R. (2009). The incidental acquisition of 3rd person – s as
implicit and explicit knowledge. In Ellis, R., Loewen, S., Elder, C., Erlam, R., Philp,
J. & Reinders, H. (eds.), Implicit and Explicit Knowledge in Second Language
Learning, Testing and Teaching (pp. 262–80). Bristol: Multilingual Matters.
Loewen, S. & Nabei, T. (2007). The effect of oral corrective feedback on implicit
and explicit L2 knowledge. In Mackey, A. (ed.), Conversational Interaction
and Second Language Acquisition: A Series of Empirical Studies (pp. 361–78).
Oxford: Oxford University Press.
Long, M. (1980). Input, interaction and second language acquisition. Unpub-
lished PhD dissertation, University of California at Los Angeles.
Long, M. (1983). Native-speaker/non-native speaker conversation and the nego-
tiation of comprehensible input. Applied Linguistics 4, 126–41.
Long, M. (1985). A role for instruction in second language acquisition:
Task-based language teaching. In Hyltenstam, K. & Pienemann, M. (eds.),
Modelling and Assessing Second Language Acquisition. Clevedon: Multilin-
gual Matters.
Long, M. (1989). Task, group, and task-group interactions. University of Hawaii
Working Papers in ESL 8, 1–26.
Long, M. (1990). Maturation constraints on language development. Studies in
Second Language Acquisition 12, 251–85.
Long, M. (1991a). Focus on form in task-based language teaching. In Lambert, R.
& Shohamy, E. (eds.), Language Policy and Pedagogy: Essays in Honour of
A. Ronald Walton (pp. 179–92). Amsterdam: John Benjamins.
Long, M. (1991b). Focus on form: A design feature in language teaching method-
ology. In de Bot, K., Ginsberg, R. & Kramsch, C. (eds.), Foreign Language
Research in Cross-Cultural Perspective (pp. 39–52). Amsterdam/Philadelphia:
John Benjamins.
Long, M. (1996). The role of the linguistic environment in second language
acquisition. In Ritchie, W. & Bhatia, T. (eds.), Handbook of Second Language
Acquisition (pp. 121–58).) San Diego, CA: Academic Press.
Long, M. (1998). SLA breaking the siege. University of Hawaii Working Papers in
ESL 17, 79–129.
Long, M. (2005). Methodological issues in learner needs analysis. In Long, M.
(ed.), Second Language Needs Analysis (pp. 19–76). Cambridge: Cambridge
University Press.

Long, M. (2006). Problems in SLA. Mahwah, NJ: Lawrence Erlbaum.

Long, M. (2015). Second Language Acquisition and Task-Based Language
Teaching. Oxford: Wiley-Blackwell.
Long, M. (2016). In defence of tasks and TBLT: Nonissues and real issues. Annual
Review of Applied Linguistics 36, 5–33.
Long, M. H., Adams, L., McLean, M. & Castaños, F. (1976). Doing things with
words: Verbal interaction in lockstep and small group classroom situations. In
Fanselow, J. & Crymes, R. (eds.), On TESOL ’76 (pp. 137–53). Washington,
DC: TESOL.
Long, M. & Norris, J. (2000). Task-based teaching and assessment. In Byram, M.
(ed.), Encyclopedia of Language Teaching (pp. 597–603). London: Routledge.
Long, M. & Porter, P. (1985). Group work, interlanguage talk, and second
language acquisition. TESOL Quarterly 19, 207–28.
Long, M. & Ross, S. (1993). Modiﬁcations that preserve language and content.
In Tickoo, M. (ed.), Simpliﬁcation: Theory and Application (pp. 29–52).
Singapore: SEAMEO Regional Language Centre.
Loschky, L. (1994). Comprehensible input and second language acquisition: What
is the relationship. Studies in Second Language Acquisition 16, 303–23.
Loschky, L. & Bley-Vroman, R. (1993). Grammar and task-based methodology.
In Crookes, G. & Gass, S. (eds.), Tasks and Language Learning: Integrating
Theory and Practice (pp. 123–67). Clevedon: Multilingual Matters.
Loumpourdi, L. (2005). Developing from PPP to TBL: A focused grammar task.
In Edwards, C. & Willis, J. (eds.), Teachers Exploring Tasks in English
Language Teaching (pp. 33–9). Basingstoke: Palgrave Macmillan.
Lovblad, K., Schaller, K. & Vargas, M. (2013). The fornix and limbic system.
Seminars in Ultrasound, CT and MRI 35, 459–73.
Lynch, T. (2009). Responding to learners’ perceptions of feedback: The use of
comparators in second language speaking courses. Innovation in Language
Learning and Teaching 3, 191–203.
Lyster, R. (1998). Recasts, repetition, and ambiguity in L2 classroom discourse.
Studies in Second Language Acquisition 20, 51–81.
Lyster, R. (2001). Negotiation of form, recasts, and explicit correction in relation
to error types and learner repair in immersion classrooms. Language Learning
51, 265–301.
Lyster, R. (2004). Differential effects of prompts and recasts in form-focused
instruction. Studies in Second Language Acquisition 26, 399–432.
Lyster, R. (2007). Learning and Teaching Languages through Content:
A Counterbalanced Approach. Amsterdam/Philadelphia: John Benjamins.
Lyster, R. & Mori, H. (2006). Interactional feedback and instructional counter-
balance. Studies in Second Language Acquisition 28, 269–300.
Lyster, R. & Ranta, L. (1997). Corrective feedback and learner uptake. Studies in
Second Language Acquisition 19, 37–66.
Lyster, R. & Ranta, L. (2013). Counterpoint piece: The case for variety in corrective
feedback research. Studies in Second Language Acquisition 35, 167–84.
Lyster, R. & Saito, K. (2010). Oral feedback in classroom SLA: A meta-analysis.
Studies in Second Language Acquisition 32, 265–302.

Macaro, E. (2001). Analysing student teachers’ codeswitching in foreign language

classrooms: Theories and decision making. The Modern Language Journal 85,
531–48.
MacIntyre, P. D. & Gardner, R. C. (1994). The subtle effects of language anxiety
on cognitive processing in the second language. Language Learning 44(2),
283–305.
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.
Mackey, A. (2006). Feedback, noticing and second language learning. Applied
Linguistics 27, 405–430.
Mackey, A. & Philp, J. (1998). Conversational interaction and second language
development: Recasts, responses and red herrings. The Modern Language
Journal 82, 338–56.
Mackey, A., Philp, J., Egi, T., Fujii, A. & Tatsumi, T. (2002). Individual differ-
ences in working memory, noticing of interactional feedback, and L2 develop-
ment. In Robinson. P. (ed.), Individual Differences and Instructed Language
Learning. Philadelphia: John Benjamins.
MacWhinney, B. (2000). The CHILDES Project: Tools for Analysing Talk:
Volume 1: Transcription Format and Programs, third edition. Mahwah, NJ:
Lawrence Erlbaum.
MacWhinney, B., Keenan, J. & Reinke, P. (1982). The role of arousal in memory
for conversation. Memory and Cognition 10, 308–17.
Maehr, M. (1984). Meaning and motivation: Toward a theory of personal invest-
ment. In Ames, R. & Ames, C. (eds.), Motivation in Education: Student
Motivation, vol. 1 (pp. 115–44). San Diego, CA: Academic Press.
Mak, B. (2011). An exploration of speaking-in-class anxiety with Chinese ESL
learners. System 39, 202–14.
Malicka, A. & Levkina, M. (2012). Measuring task complexity: Does L2 proﬁ-
ciency matter? In Shehadeh, A. & Coombe, C. (eds.), Task-Based Language
Teaching in Foreign Language Contexts: Research and Implementation
(pp. 43–66). Amsterdam: John Benjamins.
Malicka, A. & Sasayama, S. (2017). The importance of learning from the accu-
mulated knowledge: Findings from a research synthesis on task complexity.
Paper presented at the 7th Biennial International Conference on Task-Based
Language Teaching, Barcelona, Spain, April.
McDonough, K. (2015). Perceived beneﬁts and challenges with the use of collab-
orative tasks in EFL contexts. In Bygate, M. (ed.), Domains and Direction in
the Development of TBLT (pp. 225–45). Amsterdam: John Benjamins.
McDonough, K. & Chaikitmongkol, W. (2007). Teachers’ and learners’ reactions
to a task-based EFL course in Thailand. TESOL Quarterly 41, 107–32.
McNamara, T. (1996). Measuring Second Language Performance. London:
Longman.
Mehnert, U. (1998). The effects of different lengths of time for planning on
second language performance. Studies in Second Language Acquisition 20,
52–83.

Mercer, S., Ryan, S. & Williams, M. (2012). Introduction. In Mercer, S., Ryan, S. &
Williams, M. (eds.), Psychology for Language Learning: Insights from Research,
Theory, and Practice (pp. 10–25). New York: Palgrave.
Mesulam, M. (2000). Behavioural neuroanatomy: Large-scale networks,
association cortex, frontal syndromes, the limbic system, and the hemispheric
specializations. In Mesulam, M. (ed.), Principles of Behavioural and Cognitive
Neurology, second edition (pp. 1–120). Oxford: Oxford University Press.
Michel, M. (2011). Effects of task complexity and interaction on L2 performance.
In Robinson, P. (ed.), Second Language Task Complexity: Researching the
Cognition Hypothesis of Language Learning and Performance (pp. 141–74).
Amsterdam: John Benjamins.
Mifka-Profozic, N. (2013). The Effectiveness of Corrective Feedback and the Role
of Individual Differences in Language Learning. Frankfurt am Main: Peter
Lang.
Miyake, A. & Friedman, N. (1998). Individual differences in second language
proﬁciency: Working memory as language aptitude. In Healy, A. & Bourne, L.
(eds.), Foreign Language Learning: Psycholinguistic Studies on Training and
Retention. Mahwah, NJ: Erlbaum.
Mochizuki, N. (2017). Contingent need analysis for task implementation: An
activity systems analysis of group writing conferences. TESOL Quarterly 51,
607–31.
Mochizuki, N. & Ortega, L. (2008). Balancing communication and grammar in
beginning-level foreign language classrooms: A study of guided planning and
relativization. Language Teaching Research 12, 11–37.
Muller, T. (2005). Adding tasks to textbooks for beginning learners. In Edwards,
C. & Willis, J. (eds.), Teachers Exploring Tasks in English Language Teaching
(pp. 69–77). Basingstoke: Palgrave Macmillan.
Nabei, T. & Swain, M. (2001). Learner awareness of recasts in classroom
interaction: A case study of an adult EFL student. Language Awareness 11,
43–63.
Nakatsukasa, K. (2016). Efﬁcacy of requests and gestures on the acquisi-
tion of locative prepositions. Studies in Second Language Acquisition 38,
771–99.
Nassaji, H. (2009). Effects of recasts and elicitations in dyadic interaction and the
role of feedback explicitness. Language Learning 59, 411–52.
Nassaji, H. (2016). Interactional feedback in second language teaching and learn-
ing: A synthesis and analysis of current research. Language Teaching Research
20, 535–62.
Nassaji, H. (2017). The effectiveness of extensive versus intensive recasts for L2
learning of grammar. Modern Language Review 101, 353–68.
Nassaji, H. & Fu, T. (2016). Corrective feedback, learner uptake, and feedback
perception in a Chinese as a foreign language classroom. Studies in Second
Language Learning and Teaching 1, 159–81.
Nassaji, H. & Swain, M. (2000). A Vygotskian perspective on corrective feedback
in L2: The effect of random versus negotiated help in the learning of English
articles. Language Awareness 9, 34–51.

Negueruela, E. J. & Lantolf, J. P. (2006). A concept-based approach to teaching

Spanish grammar. In Salaberry, R. & Lafford, B. (eds.), Spanish Second
Language Acquisition: State of the Art (pp. 79–102). Washington, DC:
Georgetown University Press.
Neumann, O. (1987). Beyond capacity: A functional view of attention. In Heuer,
H. & Sanders, A. (eds.), Perspectives on Perception and Action (pp. 361–94).
Berlin: Springer.
Nielson, K. (2014). Evaluation of online, task-based Chinese course. In González-
Lloret, M. & Ortega, L. (eds.), Technology-Mediated TBLT: Researching
Technology and Tasks (pp. 295–321). Amsterdam: John Benjamins.
Nitta, R. & Nakatsuhara, F. (2014). A multifaceted approach to investigating pre-
task planning effects on paired oral test performance. Language Testing 31,
147–75.
Niwa, Y. (2000). Reasoning demands of L2 tasks and L2 narrative production:
Effects of individual differences in working memory, intelligence, and aptitude.
Unpublished master’s thesis, Aoyama Gakuin University, Tokyo.
Norris, J. (2009a). Task-based teaching and testing. In Long M. & Doughty C.
(eds.), The Handbook of Language Teaching (pp. 578–94). Malden: MA:
Blackwell.
Norris, J. (2009b). Understanding and improving language education through
program evaluation: Introduction to the special issue. Language Teaching
Research 13, 7–13.
Norris, J. (2015). Thinking and acting programmatically in task-based language
teaching. In Bygate, M. (ed.), Domains and Directions in the Development of
TBLT (pp. 27–57). Amsterdam: John Benjamins.
Norris, J. (2016). Current uses for task-based language assessment. Annual
Review of Applied Linguistics 36, 230–44.
Norris, J., Brown, J. D., Hudson, T. & Yoshioka, J. (1998). Designing Second
Language Performance Assessments. Honolulu, HI: University of Hawai’i
Press.
Norris, J. & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthe-
sis and quantitative meta-analysis. Language Learning 50, 417–528.
Norris, J. & Ortega, L. (2009). Towards an organic approach to investigating
CAF in instructed SLA: The case of complexity. Applied Linguistics 30,
555–78.
Norton, J. (2013). Performing identities in speaking tests: Co-construction
revisited. Language Assessment Quarterly 10, 309–30.
Nunan, D. (1989). Designing Tasks for the Communicative Classroom. Cam-
bridge: Cambridge University Press.
Nunan, D. (1990). The teacher as researcher. In Brumﬁt, C. & Mitchell, R. (eds.),
Research in the Language Classroom. ELT Documents 133. Modern English
Publications.
O’Grady, S. (2019). The impact of pre-task planning on speaking test perform-
ance for English-medium university admission. Language Testing.
Oga-Baldwin, W. & Nakata, Y. (2017). Engagement, gender, and motivation:
A predictive model for Japanese young language learners. System 65, 151–63.

Ohta, A. (2001). Second Language Acquisition Processes in the Classroom:

Learning Japanese. Mahwah, NJ: Lawrence Erlbaum Associates.
Ortega, L. (1999). Planning and focus on form in L2 oral performance. Studies in
Second Language Acquisition 21, 109–48.
Ortega, L. (2005). What do learners plan? Learner-driven attention to form
during pre-task planning. In Ellis, R. (ed.), Planning and Task Performance
in a Second Language (pp. 77–110). Amsterdam: John Benjamins.
Ortega, L. (2009). Interaction and attention to form in L2 text-based computer-
mediated communication. In Mackey, A. & Polio, C. (eds.), Multiple Perspec-
tives on Interaction in SLA: Research in Honor of Susan M. Gass
(pp. 226–53). New York: Routledge.
Ortega, L. (2015). Researching CLIL and TBLT interfaces. System 54, 103–9.
Oskoz, A. & Elola, I. (2014). Promoting foreign language collaborative writing
through the use of Web 2.0 tools and tasks. In González-Lloret, M. & Ortega,
L. (eds.), Technology-Mediated TBLT: Researching Technology and Tasks
(pp. 115–48). Amsterdam: John Benjamins.
Pang, F. & Skehan, P. (2014). Self-reported planning behaviour and second
language performance in narrative retelling. In Skehan, P. (ed.), Processing
Perspectives on Task Performance (pp. 95–128). Amsterdam: John
Benjamins.
Paradis, M. (1994). Neurolinguistic aspects of implicit and explicit memory:
Implications for bilingualism and SLA. In Ellis, N. (ed.), Implicit and Explicit
Learning of Languages (pp. 393–419). New York: Academic Press.
Paradis, M. (2004). A Neurolinguistic Theory of Bilingualism. Amsterdam: John
Benjamins.
Park, M. (2015). Development and validation of virtual interactive tasks for an
aviation English assessment. Unpublished doctoral dissertation, Iowa State
University, Ames, IA.
Park, S. (2010). The inﬂuence of pre-task instructions and pre-task planning on
focus on form during Korean EFL task-based interaction. Language Teaching
Research 14, 9–26.
Patanasorn, C. (2010). Effects of procedural, content, and task repetition on
accuracy and ﬂuency in an EFL context. Unpublished PhD dissertation, North-
ern Arizona University.
Phillips, E. M. (1992). The effects of language anxiety on students’ oral test
performance and attitudes. The Modern Language Journal 76(1), 14–26.
Philp, J. & Duchesne, S. (2016). Exploring engagement in tasks in the language
classroom. Annual Review of Applied Linguistics 36, 50–72.
Philp, J., Oliver, R. & Mackey, A. (2006). The impact of planning time on
children’s task–based interactions. System 34, 547–65.
Phung, L. (2017). Task preference, affective response, and engagement in L2
use in a US university context. Language Teaching Research 21, 751–66.
Pica, T. (1987). Second language acquisition, social interaction, and the
classroom. Applied Linguistics 8, 3–21.
Pica, T. (1996). Second language learning through interaction: Multiple perspec-
tives. Working Papers in Educational Linguistics 12, 1–22.

Pica, T. (1997). Second language teaching and research relationships: A North

American view. Language Teaching Research 1, 48–72.
Pica, T. & C. Doughty. (1985a). Input and interaction in the communicative
language classroom: A comparison of teacher-fronted and group activities. In
Gass, S. & Madden, C. (eds.), Input in Second Language Acquisition. Rowley,
MA: Newbury House.
Pica, T. & C. Doughty. (1985b). The role of group work in classroom second
language acquisition. Studies in Second Language Acquisition 7, 233–48.
Pica, T., Kanagy, R. & Falodun, J. (1993). Choosing and using communication
tasks for second language research and instruction. In Crookes, G. & Gass, S.
(eds.), Task-Based Learning in a Second Language (pp. 9–34). Clevedon:
Multilingual Matters.
Pienemann, M. (1985). Learnability and syllabus construction. In Hyltenstam, K.
& Pienemann, M. (eds.), Modelling and Assessing Second Language Acquisi-
tion (pp. 23–75). Clevedon: Multilingual Matters.
Plonsky, L. & Gass, S. (2011). Quantitative research methods, study quality, and
outcomes: The case of interaction research. Language Learning 61, 325–66.
Plonsky, L. & Kim, J. (2016). Task-based learner production: A substantive
review. Annual Review of Applied Linguistics 36, 73–97.
Plough, I. & Gass, S. (1993). Interlocutor and task familiarity: Effects on inter-
actional structure. In Crookes, G. & Gass, S. (eds.), Tasks and Language
Learning: Integrating Theory and Practice. Clevedon: Multilingual Matters.
Poehner, M. (2008). Dynamic Assessment: A Vygostkian Approach to Under-
standing and Promoting L2 Development. Berlin: Springer.
Poehner, M. & Infante, P. (2017). Mediated development: A Vygotskian
approach to transforming second language learner abilities. TESOL Quarterly
51, 332–57.
Poehner, M. & Lantolf, J. (2005). Dynamic assessment in the language classroom.
Language Teaching Research 9, 233–65.
Pollitt, A. (1991). Giving students a sporting chance: Assessment by counting and
judging. In Alderson, J. C. & North, B. (eds.), Language Testing in the 1990s
(pp. 46–59). London: Macmillan.
Prabhu, N. (1987). Second Language Pedagogy. Oxford: Oxford University Press.
Prabhu, V. (1990). Comments on Alan Beretta’s ‘Attention to form or meaning?
Error treatment in the Bangalore project’. TESOL Quarterly 24, 112–15.
Pulido, D. (2007). The effects of topic familiarity and passage sight vocabulary
on L2 lexical inferencing and retention through reading. Applied Linguistics
28, 66–86.
Qin, T. & van Compernolle, R. (forthcoming). Computerized dynamic assessment
of implicature comprehension in L2 Chinese. Language Learning and
Technology.
Quinn, P. (2014). Delayed versus immediate corrective feedback on orally
produced past passive errors in English. Unpublished PhD thesis, University
of Toronto, Canada.
Rainey, S. (2000). Action research and the English as a foreign language practi-
tioner: Time to take stock. Educational Action Research 8, 165–91.

Ranta, L. (2002). The role of learners’ language analytic ability in the communi-
cative classroom. In Robinson, P. (ed.), Individual Differences and Instructed
Language Learning. Amsterdam: John Benjamins.
Rassaei, E. (2015). Oral corrective feedback, foreign language anxiety and L2
development. System 49, 98–109.
Rebuschat, P. & Williams, J. N. (2009). Implicit learning of word order. In
Taatgen, N. & van Rijn, H. (eds.), Proceedings of the 31st Annual Conference
of the Cognitive Science Society (pp. 425–30). Austin, TX: Cognitive Science
Society.
Reese, C. & Wells, T. (2007). Teaching academic discussion skills with a card
game. Simulation and Gaming 38, 546–55.
Reeve, J. & Lee, W. (2014). Students’ classroom engagement produces longitu-
dinal changes in classroom motivation. Journal of Educational Psychology
106, 527–40.
Révész, A. (2009). Task complexity, focus on form, and second language devel-
opment. Studies in Second Language Acquisition 31, 437–70.
Révész, A. (2011). Task complexity, focus on L2 constructions, and individual
differences: A classroom-based study. Modern Language Journal 95, 162–81.
Révész, A. (2012). Working memory and the observed effectiveness of recasts on
different L2 outcome measures. Language Learning 62, 93–132.
Révész, A. (2017). Replication in task-based language teaching research: Kim
(2012) and Shintani (2012). First view: Language Teaching, 1–11.
Révész, A., Ekiert, M. & Torgersen, E. N. (2016). The effects of complexity,
accuracy, and fluency on communicative adequacy in oral task performance.
Applied Linguistics 37(6), 828–48.
Révész, A., Michel, M. & Gilabert, R. (2016). Measuring cognitive task demands
using dual-task methodology, subjective self-ratings, and expert judgments:
A validation study. Studies in Second Language Acquisition 38, 703–37.
Richards, J. (1984). Language curriculum development. RELC Journal 15, 1–29.
Richards, J. & Rodgers, T. (1986). Approaches and Methods in Language Teach-
ing, second edition. Cambridge: Cambridge University Press.
Richards, J. & Rodgers, T. (2014). Approaches and Methods in Language Teach-
ing, third edition. Cambridge: Cambridge University Press.
Robinson, P. (2001). Task complexity, cognitive resources, and syllabus design:
A triadic framework for examining task influences on SLA. In Robinson, P.
(ed.), Cognition and Second Language Instruction (pp. 287–318). Cambridge:
Cambridge University Press.
Robinson, P. (2002). Learning conditions, aptitude complexes and SLA:
A framework for research and pedagogy. In Robinson, P. (ed.), Individual
Differences and Instructed Language Learning. Amsterdam: John Benjamins.
Robinson, P. (2007a). Aptitudes, abilities, contexts, and practice. In DeKeyser, R.
M. (ed.), Practice in Second Language: Perspectives from Applied Linguistics
and Cognitive Psychology (pp. 256–86). New York/Cambridge: Cambridge
University Press.
Robinson, P. (2007b). Re-thinking-for-speaking and L2 task demands: The Cog-
nition Hypothesis, task classification, and sequencing. Plenary address at the

Second International Conference on Task-based Language Teaching, Hawaii,

20–22 September.
Robinson, P. (2007c). Task complexity, theory of mind, and intentional
reasoning: Effects on L2 speech production, interaction, uptake and percep-
tions of task Difﬁculty. International Review of Applied Linguistics 45,
193–214.
Robinson, P. (2010). Situating and distributing cognition across task demands:
The SSARC model of pedagogic task sequencing. In Putz, M. & Sicola, L.
(eds.), Cognitive Processing in Second Language Acquisition: Inside the
Learner’s Mind (pp. 243–68). Amsterdam: John Benjamins.
Robinson, P. (2011). Second language task complexity, the cognition hypothesis,
language learning, and performance. In Robinson P. (ed.), Second Language
Task Complexity: Researching the Cognition Hypothesis of Language Learn-
ing and Performance (pp. 3–38). Amsterdam: John Benjamins.
Robinson, P. (2015). The cognition hypothesis, second language task demands,
and the SSARC model of pedagogic task sequencing. In Bygate, M. (ed.),
Domains and Directions in the Development of TBLT (pp. 87–122). Amster-
dam: John Benjamins.
Roever, C. (2011). Testing of second language pragmatics: Past and future.
Language Testing 28, 463–81.
Roever, C. & Kasper, G. (2018). Speaking in turns and sequences: Interactional
competence as a target construct in testing speaking. Language Testing 35,
331–55.
Rolin-Ianziti, J. (2010). The organization of delayed second language correction.
Language Teaching Research 14, 183–206.
Romanova, N. (2010). Planning, recasts, and learning of L2 morphology. The
Canadian Modern Language Review 66, 843–75.
Rulon, K. & McCreary, J. (1986). Negotiation of content: Teacher-fronted
and small-group interaction. In Day, R. R. (ed.), Talking to Learn: Conversa-
tion in Second Language Acquisition (pp. 182–99). Rowley, MA: Newbury
House.
Sabet, P. & Zhang, G. (2015). Communicating through Vague Language:
A Comparative Study of L1 and L2 Speakers. Basingstoke: Palgrave
Macmillan.
Sachs, R. (2010). Individual differences and the effectiveness of visual feedback of
reﬂexive binding in L2 Japanese. PhD dissertation, Georgetown University.
Saito, K. & Akiyama, Y. (2017). Video-based interaction, negotiation for com-
prehensibility, and second language learning: A longitudinal study. Language
Learning 67, 43–74.
Saito, Y., Horwitz, E. K. & Garza, T. J. (1999). Foreign language reading anxiety.
The Modern Language Journal 83(2), 202–18.
Samimy, K. & Kobayashi, C. (2004). Toward the development of intercultural
competence: Theoretical and pedagogical implications for Japanese English
teachers. JALT Journal 26, 245–61.
Samuda, V. (2001). Guiding relationships between form and meaning during task
performance: The role of the teacher. In Bygate, M., Skehan, P. & Swain, M.

(eds.), Researching Pedagogic Tasks: Second Language Learning, Teaching

and Testing (pp. 119–40). Harlow: Longman.
Samuda, V. (2015). Tasks, design, and the architecture of pedagogical spaces. In
Bygate, M. (ed.), Domains and Directions in the Development of TBLT (pp.
271–301). Amsterdam: John Benjamins.
Samuda, V. & Bygate, M. (2008). Tasks in Second Language Learning. Basing-
stoke: Palgrave Macmillan.
Sanders, A. (1998). Elements of Human Performance. Mahwah, NJ: Lawrence
Erlbaum.
Sangarun, J. (2005). The effects of focusing on meaning and form in strategic
planning. In Ellis, R. (ed.), Planning and Task Performance in a Second
Language. Philadelphia: John Benjamins.
Sasayama, S. (2015). Validating the assumed relationship between task design,
cognitive complexity, and second language task performance. Unpublished
PhD dissertation, Georgetown University.
Sasayama, S. (2016). Is a ‘complex’ task really complex? Validating the assump-
tion of task complexity. Modern Language Journal 100, 231–54.
Satar, H. & Ozdener, N. (2008). The effects of synchronous CMC on speaking
proficiency and anxiety: Text versus voice chat. The Modern Language Journal
92, 595–613.
Sato, M. (2017). Interaction mindsets, interactional behaviors, and L2
development: An affective-social-cognitive model. Language Learning 67,
249–83.
Sato, M. & Loewen, S. (2018). Metacognitive instruction enhances the effective-
ness of corrective feedback: Variable effects of feedback types and linguistic
targets. Language Learning 68, 507–45.
Sato, M. & Lyster, R. (2012). Peer interaction and corrective feedback for
accuracy and fluency development: Monitoring, practice, and proceduraliza-
tion. Studies in Second Language Acquisition 34, 591–626.
Schmidt, R. (1990). The role of consciousness in second language learning.
Applied Linguistics 11, 129–58.
Schmidt, R. (1994). Deconstructing consciousness: In search of useful definitions
for applied linguistics. AILA Review 11, 11–26.
Schmidt, R. (2001). Attention. In Robinson, P. (ed.), Cognition and Second
Language Instruction (pp. 3–32). Cambridge: Cambridge University Press.
Schmidt, R. & Frota, S. (1986). Developing basic conversation ability in a second
language: A case-study of an adult learner. In Day, R. (ed.), Talking to Learn:
Conversation in Second Language Acquisition. Rowley, MA: Newbury House.
Schumann, J. (2001). Appraisal psychology, neurobiology, and language. Annual
Review of Applied Linguistics 21, 23–42.
Schwartz, B. D. (1993). On explicit and negative data effecting and affecting
competence and linguistic behavior. Studies in Second Language Acquisi-
tion 15, 147–63.
Seedhouse, P. (1999). Task-based interaction. ELT Journal 53, 149–56.
Seedhouse, P. (2005a). ‘Task’ as research construct. Language Learning 55,
533–70.

Seedhouse, P. (2005b). Conversation analysis and language learning. Language

Teaching 38, 165–87.
Seedhouse, P. (ed.). (2017). Task-Based Language Learning in a Real-World
Digital Environment: The European Digital Kitchen. London: Bloomsbury.
Seipp, B. (1991). Anxiety and academic performance: A meta-analysis of ﬁndings.
Anxiety Research 4(1), 27–41.
Sfard, A. (1998). On two metaphors for learning and the dangers of choosing just
one. Educational Researcher 27, 4–13.
Sheen, R. (1994). A critical analysis of the advocacy of the task-based syllabus.
TESOL Quarterly 28, 127–51.
Sheen, R. (2003). Focus on form–a myth in the making? ELT journal 57, 225–33.
Sheen, R. (2004). Corrective feedback and learner uptake in communicative class-
rooms across instructional settings. Language Teaching Research 8, 263–300.
Sheen, R. (2005). Focus on forms as a means of improving accurate oral produc-
tion. In Housen, A. & Pierrard, M. (eds.), Investigations in Instructed Lan-
guage Acquisition (pp. 270–310). Berlin: Mouton de Gruyter.
Sheen, R. (2006). Focus on forms as a means of improving accurate oral produc-
tion. In Housen, A. & Pierrard, M. (eds.), Investigations in Instructed Second
Language Acquisition (pp. 271–310). Berlin: Mouton de Gruyter.
Sheen, R. (2006). Exploring the relationship between characteristics of recasts and
learner uptake. Language Teaching Research 10, 361–92.
Sheen, R. (2007). The effects of corrective feedback, language aptitude, and
learner attitudes on the acquisition of English articles. In Mackey, A. (ed.),
Conversational Interaction in Second Language Acquisition. Oxford: Oxford
University Press.
Sheen, R. (2008). Recasts, language anxiety, modiﬁed output, and L2 learning.
Language Learning 58(4), 835–74.
Sheen, R. (2010). Differential effects of oral and written corrective feedback in the
ESL classroom. Studies in Second Language Acquisition 32, 203–34.
Shehadeh, A. (2005). Task-based learning and teaching: Theories and applica-
tions. In Edwards, C. & Willis, L. (eds.), Teachers Exploring Tasks (pp.
13–30). Basingstoke: Palgrave Macmillan.
Shehadeh, A. (2012). Broadening the perspective of task-based language teaching
scholarship: The contribution of research in foreign language contexts. In
Shehadeh A. & Coombe, C. A. (eds.), Task-Based Language Teaching in
Foreign Language Contexts: Research and Implementation (pp. 1–20). Phila-
delphia: John Benjamins.
Shintani, N. (2011). A comparison of the effects of comprehension-based and
production-based instruction on the acquisition of vocabulary and grammar
by young Japanese learners of English. Unpublished PhD thesis, the University
of Auckland.
Shintani, N. (2012). Repeating input-based tasks with young beginner learners.
RELC Journal 43, 39–51.
Shintani, N. (2013). The effect of focus on form and focus on forms instruction on
the acquisition of productive knowledge of L2 vocabulary by young beginning-
level learners. TESOL Quarterly 47, 36–62.

Shintani, N. (2015). The incidental grammar acquisition in focus on form and

focus on forms instruction for young beginner Learners. TESOL Quarterly 49,
115–40.
Shintani, N. (2016). Input-Based Tasks in Foreign Language Instruction for
Young Learners. Amsterdam: John Benjamins.
Shintani, N. & Ellis, R. (2010). The incidental acquisition of English plural –s by
Japanese children in comprehension-based lessons: A process–product study.
Studies in Second Language Acquisition 32, 607–37.
Skehan, P. (1984). Issues in the testing of English for specific purposes. Language
Testing 1/2, 202–20.
Skehan, P. (1996). A framework for the implementation of task-based learning.
Applied Linguistics 17, 38–62.
Skehan, P. (1998). A Cognitive Approach to Language Learning. Oxford: Oxford
University Press.
Skehan, P. (2001). Tasks and language performance assessment. In
Bygate, M., Skehan, P. & Swain, M. (eds.), Researching Pedagogic Tasks,
Second Language Learning, Teaching and Testing (pp. 167–85). Harlow:
Longman.
Skehan, P. (2002). Theorizing and updating aptitude. In Robinson, P. (ed.),
Individual Differences and Instructed Language Learning (pp. 69–94).
Amsterdam: John Benjamins.
Skehan, P. (2003). Task-based instruction. Language Teaching 36, 1–14.
Skehan, P. (2007). Task research and language teaching: Reciprocal relationships.
In Fotos, S. (ed.), Form-Meaning Relationships in Language Pedagogy: Essays
in Honour of Rod Ellis. Oxford: Oxford University Press.
Skehan, P. (2009a). Modelling second language performance: Integrating com-
plexity, accuracy, fluency and lexis. Applied Linguistics 30, 510–32.
Skehan, P. (2009b). Lexical performance by native and non-native speakers on
language-learning tasks. In Richards, B., Daller, H., Malvern, D. D. & Meara,
P. (eds.), Vocabulary Studies in First and Second Language Acquisition: The
Interface Between Theory and Application (pp. 107–24). London: Palgrave
Macmillan.
Skehan, P. (2009c). Models of speaking and the assessment of second language
proficiency. In Benati, A. G. (ed.), Issues in Second Language Proficiency
(pp. 202–15). London: Continuum International Pub Group.
Skehan, P. (2012). Researching Tasks: Performance, Assessment, Pedagogy. Shang-
hai/Amsterdam: Shanghai Foreign Language Education Press/De Gruyter.
Skehan, P. (2013). Nurturing noticing. In Bergsleithner, J. M., Frota, S. N. &
Yoshioka, J. K. (eds.), Noticing and Second Language Acquisition: Studies in
Honor of Richard Schmidt (pp. 169–80). Honolulu, HI: University of Hawai‘i,
National Foreign Language Resource Center.
Skehan, P. (2014a). Processing Perspectives on Task Performance. Amsterdam:
John Benjamins.
Skehan, P. (2014b). Introduction: Coordinated research into task-based perform-
ance. In Skehan, P. (ed.), Processing Perspectives on Task Performance
(pp. 1–26). Amsterdam: John Benjamins.

Skehan, P. (2014c). Synthesising and applying task research. In Skehan, P. (ed.),

Processing Perspectives on Task Performance (pp. 211–60). Amsterdam: John
Benjamins.
Skehan, P. (2015). Limited attentional capacity and cognition: Two hypotheses
regarding second language performance on tasks. In Bygate, M. (ed.), Domains
and Directions in the Development of TBLT: A Decade of Plenaries from the
International Conference (pp. 123–55). Amsterdam: John Benjamins.
Skehan, P. (2016). Tasks vs. conditions: Two perspectives on task research and its
implications for pedagogy. Annual Review of Applied Linguistics 36, 34–49.
Skehan, P. (2018). Second Language Task-Based Performance: Theory, Research,
and Assessment. New York: Routledge.
Skehan, P. & Foster, P. (1997). The influence of planning and post-task activities
on accuracy and complexity in task-based learning. Language Teaching
Research 1, 185–211.
Skehan, P. & Foster, P. (1999). Task structure and processing conditions in
narrative retellings. Language Learning 49(1), 93–120.
Skehan, P. & Foster, P. (2001). Cognition and tasks. In Robinson P. (ed.),
Cognition and Second Language Learning (pp. 183–205). New York: Cam-
bridge University Press.
Skehan, P. & Foster, P. (2005). Strategic and on-line planning: The influence of
surprise information and task time on second language performance. In Ellis,
R. (ed.), Planning and Task Performance in a Second Language (pp. 193–216).
Philadelphia: John Benjamins Publishing Company
Skehan, P. & Foster, P. (2007). Complexity, accuracy, fluency and lexis in task-
based performance: A meta-analysis of the Ealing research. In Van Daele, S.,
Housen, A., Kuiken, F., Pierrard, M. & Vedder I. (eds.), Complexity, Accur-
acy, and Fluency in Second Language Use, Learning, and Teaching (pp.
207–26). Brussels: University of Brussels Press.
Skehan, P. & Shum, S. (2017). What influences performance? Personal style or the
task being done? In Wong, L. L. C. & Hyland, K. (eds.), Faces of English
Education: Students, Teachers, and Pedagogy (pp. 28–43). London: Routledge.
Skehan, P., Xiaoyue, B., Qian, L. & Wang, Z. (2012). The task is not enough:
Processing approaches to task-based performance. Language Teaching
Research 16, 170–87.
Slimani-Rolls, A. (2005). Rethinking task-based language learning: What we can
learn from the learners. Language Teaching Research 9, 195–218.
Smith, B. (2003). Computer-mediated negotiated interaction: An expanded
model. Modern Language Journal 26, 365–98.
Smith, B. (2012). Eye-tracking as a measure of noticing: A study of explicit recasts
in SCMC. Language Learning and Technology 16, 53–18.
Snow, R. E. (1991). Aptitude-treatment interaction as a framework for research
on individual differences in psychotherapy. Journal of Consulting and Clinical
Psychology 59, 205–16.
Solon, M., Long, A. & Gurzynska-Weiss, L. (2017). Task complexity, language
related episodes, and production of L2 Spanish vowels. Studies in Second
Language Acquisition 39, 347–80.

Spada, N., Jessop, L., Tomita, Y., Suzuki, W. & Valeo, A. (2014). Isolated and
integrated form-focused instruction: Effects on different types of L2 know-
ledge. Language Teaching Research 18, 453–73.
Spada, N. & Lightbown, P. (1999). Instruction, ﬁrst language inﬂuence, and
developmental readiness in second language acquisition. Modern Language
Journal 83, 1–22.
Sparks, R. L. & Patton, J. (2013). Relationship of L1 skills and L2 aptitude to L2
anxiety on the foreign language classroom anxiety scale. Language Learning
63(4), 870–95.
Storch, N. (2002). Patterns of interaction in ESL pair work. Language Learning
52, 119–58.
Storch, N. (2007). Investigating the merits of pair work on a text-editing task in
ESL classes. Language Teaching Research 11, 143–59.
Storch, N. (2008). Metatalk in a pair work activity: Level of engagement
and implications for language development. Language Awareness 17,
95–114.
Storch, N. (2017). Sociocultural theory in the L2 classroom. In Loewen, S. & Sato,
M. (eds.), The Routledge Handbook of Instructed Second Language Acquisi-
tion (pp. 69–84). New York: Routledge.
Storch, N. & Wigglesworth, G. (2003). Is there a role for the use of the L1 in an
L2 setting? TESOL Quarterly 37, 760–70.
Stroud, R. (2017). The impact of task performance scoring and tracking on
second language engagement. System 69, 121–32.
Svalberg, A. (2009). Engagement with language: Developing a construct. Lan-
guage Awareness 18, 242–58.
Swain, M. (1995). Three functions of output in second language Learning. In
Cook, G. & Seidhofer, B. (eds.), Principles and Practice in the Study of
Language: Studies in Honour of H. G. Widdowson (pp. 125–44). Oxford:
Oxford University Press.
Swain, M. (2000). The output hypothesis and beyond: Mediating acquisition
through collaborative dialogue. In Lantolf, J. (ed.), Sociocultural Theory
and Second Language Learning (pp. 97–114). Oxford: Oxford University
Press.
Swain, M. (2006). Languaging, agency and collaboration in advanced second
language learning. In Byrnes, H. (ed.), Advanced Language Learning: The
Contributions of Halliday and Vygotsky (pp. 95–108). London: Continuum.
Swain, M. (2013). The inseparability of cognition and emotion in second language
learning. Language Teaching 46, 195–207.
Swain, M., Kinnear, P. & Steinman, L. (2011). Sociocultural Theory in Second
Language Education: An Introduction through Narratives. Bristol: Multilin-
gual Matters.
Swain, M. & Lapkin, S. (1998). Interaction and second language learning: Two
adolescent French immersion students working together. Modern Language
Journal 82, 320–37.
Swain, M. & Lapkin, S. (2001). Focus on form through collaborative dialogue:
Exploring task effects. In Bygate, M., Skehan, P. & Swain, M. (eds.),

Researching Pedagogic Tasks, Second Language Learning, Teaching and

Testing (pp. 99–118). Harlow: Longman.
Swain, M. & Lapkin, S. (2002). Talking it through: Two French immersion
learners’ response to reformulation. International Journal of Educational
Research 37, 285–304.
Swain, M. & Lapkin, S. (2007). The distributed nature of second language
learning: A case study. In Fotos, S. & Nassaji, H. (eds.), Focus on Form and
Teacher Education: Studies in Honour of Rod Ellis (pp. 73–86). Oxford:
Oxford University Press.
Swain, M., Lapkin, S., Knouzi, I., Suzuki, W. & Brooks, L. (2009). Languaging:
University students learn the grammatical concept of voice in French. Modern
Language Journal 93, 5–29.
Swan, M. (2005a). Legislation by hypothesis: The case of task-based instruction.
Applied Linguistics 26, 376–401.
Swan, M. (2005b). Review of Rod Ellis’ task-based language learning and teach-
ing. IJAL 15, 251–9.
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning.
Cognitive Science 12, 257–85.
Sweller, J. (1994). Cognitive load theory, learning difficulty, and instructional
design. Learning and Instruction 4, 295–312.
Sykes, J. M. (2014). TBLT and synthetic immersive environments: What can in-
game task restarts tell us about design and implementation? In González-
Lloret, M. & Ortega, L. (eds.), Technology-Mediated TBLT: Researching
Technology and Tasks (pp. 149–82). Amsterdam: John Benjamins.
Tavakoli, P. (2018). A detailed analysis of oral fluency at different levels of
proficiency. Paper presented at the L-SLARF Annual Conference. Birkbeck
College, London, 2 June.
Tavakoli, P. & Foster, P. (2008). Task design and second language performance:
The effect of narrative type on learner output. Language Learning 58(2),
439–73.
Tavakoli, P. & Skehan, P. (2005). Planning, task structure, and performance
testing. In Ellis, R. (ed.), Planning and Task Performance in a Second Language
(pp. 239–76). Amsterdam: John Benjamins.
Thai, C. & Boers, F. (2015). Repeating a monologue under increasing time
pressure: Effects on fluency, complexity, and accuracy. TESOL Quarterly
50, 369–93.
Thomas, M. & Reinders, H. (eds.). (2010). Task-Based Learning and Teaching
with Technology. London: Continuum.
Thomas, M. & Reinders, H. (eds.). (2015). Contemporary Task-Based Teaching
in Asia. London: Continuum.
Timpe-Laughlin, V. (2018). Pragmatics in task-based language assessment:
Opportunities and challenges. In Taguchi, N. & Kim, Y. (eds.), Task-Based
Approaches to Teaching and Assessing Pragmatics (pp. 287–304). Amsterdam:
John Benjamins.
Tobias, S. (1985). Test anxiety: Interference, defective skills, and cognitive cap-
acity. Educational Psychologist 20(3), 135–42.

Tomlin, R. & Villa, V. (1994). Attention in cognitive science and second language
acquisition. Studies in Second Language Acquisition 16, 183–203.
Tomlinson, B. (2015). TBLT and materials and curricula: From theory to practice.
In Thomas, M. & Reinders, H. (eds.), Contemporary Task-Based Teaching in
Asia (pp. 328–339). London: Bloomsbury.
Toth, P. (2008). Teacher- and learner-led discourse in task-based grammar
instruction: Providing procedural assistance for L2 morphosyntactic develop-
ment. Language Learning 58, 237–83.
Trebits, A. (2014). Sources of individual differences in L2 narrative production:
The contribution of input, processing, and output anxiety. Applied Linguistics
37(2), 155–74.
Troﬁmovich, P., Ammar, A. & Gatbonton, E. (2007). How effective are recasts?
The role of attention, memory, and analytical ability. In Mackey, A. (ed.),
Conversational Interaction in Second Language Acquisition. Oxford: Oxford
University Press.
Trondheim, L. (2007). Mr. I. New York: Nantier, Beall & Minoustchine.
Van Avermmaat, P. & Gysen, S. (2006). From needs to tasks: Language learning
needs in a task-based approach. In Van den Branden, K. (ed.), Task-Based
Language Education: From Theory to Practice (pp. 17–46). Cambridge: Cam-
bridge University Press.
Van Compernolle, R. & Zhang, H. (2014). Dynamic assessment of elicited imita-
tion: A case analysis of an advanced L2 English speaker. Language Testing 31,
395–412.
Van Compernolle, R., Weber, A. & Gomez-Laich, M. (2016). Teaching L2
Spanish sociopragmatics through concepts: A classroom-based study. Modern
Language Journal 100, 341–61.
Van de Guchte, M., Rijlaarsdam, G., Braaksma, M. & Bimmel, P. (2015). Learn-
ing new grammatical structures in task-based language learning: The effects of
recasts and prompts. The Modern Language Journal 99, 246–62.
Van de Guchte, M., Rijlaarsdam, G., Braaksma, M. & Bimmel, P. (2017). Focus
on language versus content in the pre-task: Effects of guided peer-video model
observations on task performance. Language Teaching Research 23(3),
310–329.
Van den Branden, K. (2009). Diffusion and implementation of innovations. In
Long, M. & Doughty, C. (eds.), The Handbook of Language Teaching (pp.
659–72). Malden, MA: Wiley-Blackwell.
Van den Branden, K. (ed.). (2006). Task-Based Language Education: From
Theory to Practice. Cambridge: Cambridge University Press.
Van den Branden, K., Bygate, M. & Norris, J. M. (2009). Task-Based Language
Teaching: A Reader. Amsterdam: John Benjamins.
Van den Branden, K., Van Gorp, K. & Verhelst, M. (2007). Introduction. In Van den
Branden, K. et al. (eds.), Tasks in Action: Task-Based Language Education from a
Classroom-Based Perspective (pp. 1–6). Newcastle: Cambridge Scholars Publishing.
Van Gorp, K. & Deygers, B. (2014). Task-based language assessment. In Kunnan,
A. (ed.), The Companion to Language Assessment, vol. 2 (pp. 271–87).
Malden, MA: John Wiley & Son.

Van Lier, L. (1996). Interaction in the Language Curriculum: Awareness, Auton-

omy and Authenticity. London: Longman.
VanPatten, B. (2015). Input processing in adult SLA. In VanPatten, B. & Williams,
J. (eds.), Theories in Second Language Acquisition. New York: Routledge.
Varnosfadrani, A. & Basturkmen, H. (2009). The effectiveness of implicit and
explicit error correction on learners’ performance. System 37, 82–98.
Varonis, E. & Gass, S. (1985). Non-native/non-native conversations: A model for
negotiation of meaning. Applied Linguistics 6, 71–90.
Vygotsky, L. (1978). Mind in Society: The Development of Higher Psychological
Processes, Cole, M., John-Steiner, V., Scribner, S. & Souberman, E. (eds.).
Cambridge, MA: Harvard University Press.
Vygotsky, L. (1986). Thought and Language. Cambridge, MA: MIT Press.
Vygotsky, L. (1987). The Collected Works of L. S. Vygotsky, Volume 1: The
Problems of General Psychology. R. W. Rieber & A. S. Carton (eds.). New
York: Plenum Press.
Vygotsky, L. (2000). Thought and Language. Cambridge, MA: MIT Press.
Wang Q. (2007). The national curriculum changes and their effects on english
language teaching in the People’s Republic of China. In Cummins, J. & Davison
C. (eds.), International Handbook of English Language Teaching. Springer
International Handbooks of Education, vol. 15 (pp. 87–105). Boston: Springer.
Wang, Z. (2009). Modelling speech production and performance: Evidence
from five types of planning and two task structures. Unpublished PhD thesis,
Chinese University of Hong Kong.
Wang, Z. (2014). On-line time pressure manipulations: L2 speaking performance
under five types of planning and repetition conditions. In Skehan, P. (ed.),
Processing Perspectives on Task Performance (pp. 27–62). Amsterdam: John
Benjamins.
Wang, Z. & Skehan P. (2014). Task structure, time perspective and lexical demands
during video-based narrative retellings. In Skehan, P. (ed.), Processing Perspec-
tives on Task Performance (pp. 155–86). Amsterdam: John Benjamins.
Watanabe, Y. & Swain, M. (2007). Effects of proficiency differences and patterns
of pair interaction on second language learning: Collaborative dialogue
between adult ESL learners. Language Teaching Research 11, 121–42.
Watson-Todd, R. (2006). Continuing change after innovation. System 34, 1–14.
Weaver, C. (2012). Incorporating a formative assessment cycle into task-based
language teaching in a university setting in Japan. In Shehadeh, A. & Coombe,
C. (eds.), Task-Based Language Teaching in Foreign Language Contexts
(pp. 287–312). Amsterdam: John Benjamins.
Weir, C. (2005). Language Testing and Validation: An Evidence-Based
Approach. Basingstoke: Palgrave Macmillan.
Weir, C. & Roberts, J. (1994). Evaluation in ELT. Oxford: Blackwell.
Weir, C., Vidakovic, I. & Galaczi, E. (2014). Measured Constructs: A History of
Cambridge English Language Examinations, 1913–2012: Studies in Language
Testing 37. Cambridge: Cambridge University Press.
Wen, Z. (2016). Working Memory and Second Language Learning: Towards an
Integrated Approach. Clevedon, Avon: Multilingual Matters.

Wen, Z., Skehan, P., Biedroń, A., Li, S. & Sparks, R. (eds.) (in press). Rethinking
Language Aptitude: Multiple Perspectives and Emerging Trends. New York:
Routledge.
Wertsch, J. V. (1985). Vygotsky and the Social Formation of Mind. Cambridge,
MA: Harvard University.
Wertsch, J. V., Minick, N. & Arns, F. J. (1984). The creation of context in joint problem-
solving. In Rogoff, B. & Lave, J. (eds.), Everyday Cognition: Its Development in
Social Context (pp. 151–71). Cambridge, MA: Harvard University Press.
White, L. (1987). Against comprehensible input: The input hypothesis and the
development of L2 competence. Applied Linguistics 8, 95–110.
Wickens, C. D. (2007). Attention to the second language. International Review of
Applied Linguistics 45, 177–91.
Widdowson, H. (2003). Defining Issues in English Language Teaching. Oxford:
Oxford University Press.
Wigglesworth, G. & Elder, C. (2010). An investigation of the effectiveness
and validity of planning time in speaking test tasks. Language Assessment
Quarterly 7(1), 1–24.
Wigglesworth, G. & Frost, K. (2017). Task and performance-based assessment.
In Shohamy E., Or, I. & May, S. (eds.), Language Testing and Assess-
ment: Encyclopedia of Language and Education, third edition. Cham:
Springer.
Wilkins, D. (1976). Notional Syllabuses. Oxford: Oxford University Press.
Williams, J. (1999). Learner-generated attention to form. Language Learning 49,
583–625.
Williams, J. (2013). Attention, awareness, and noticing in language processing
and learning. In Bergsleithner, J., Frota, M., Nagem, S. & Yoshioka, Jim Kei
(eds.), Noticing and Second Language Acquisition: Studies in Honor of Rich-
ard Schmidt (pp. 39–57). Honolulu, HI: University of Hawai‘i, National
Foreign Language Resource Center
Williams, L. (2018). Task-based teaching and concept-based instruction. In
Ahmadian, M. & Mayo, M. (eds.), Recent Perspectives on Task-Based Lan-
guage Teaching (pp. 121–41). Boston/Berlin: Mouton de Gruyter.
Willis, D. & Willis, J. (2007). Doing Task-Based Teaching. Oxford: Oxford
University Press.
Willis, J. (1996). A Framework for Task-Based Language Teaching. New York:
Longman.
Willis, J. & Willis, D. (1988). Collins COBUILD English Course: Book 1.
London: Collins.
Winke, P. (2014). Formative, task-based oral assessments in an advanced Chinese
language class. In González-Lloret, M. & Ortega, L. (eds.), Technology-Medi-
ated TBLT: Researching Technology and Tasks (pp. 264–93). Amsterdam:
John Benjamins.
Winter, E. (1976). Fundamentals of Information Structure: A Pilot Manual for
Further Development According to Student Need. Hatfield: Hatfield Polytech-
nic, English Department.

Xi, X. (2005). Do visual chunks and planning impact the overall quality of oral
descriptions of graphs? Language Testing 22(4), 463–508.
Yang, Y. & Lyster, R. (2010). Effects of form-focused practice and feedback on
Chinese EFL learners’ acquisition of regular and irregular past-tense forms.
Studies in Second Language Acquisition 32, 235–63.
Yilmaz, Y. (2013a). Relative effects of explicit and implicit feedback: The role of
working memory capacity and language analytic ability. Applied Linguistics
34, 344–68.
Yilmaz, Y. (2013b). The relative effectiveness of mixed, explicit and implicit
feedback in the acquisition of English articles. System 41, 691–705.
Youn, S. J. (2018). Task design and validity evidence for assessment of L2
pragmatics in interaction. In Taguchi, N. & Kim, Y. (eds.), Task-Based
Approaches to Teaching and Assessing Pragmatics (pp. 217–46). Amsterdam:
John Benjamins.
Yuan, F. & R. Ellis. (2003). The effects of pre-task and on-line planning on
fluency, complexity and accuracy in L2 monologic oral production. Applied
Linguistics 24, 1–27.
Yuksal, D. & Inan, B. (2014). The effects of communication mode on negotiation
and its noticing. ReCALL 26, 333–54.
Yule, G. (1997). Referential Communication Tasks. Hillsdale, NJ: Lawrence
Erlbaum.
Yule, G. & McDonald, D. (1990). Resolving referential conflicts in L2 interaction:
The effect of proficiency and interactive role. Language Learning 40, 539–56.
Zhang, J. & Rahimi, M. (2014). EFL learners’ anxiety level and their beliefs about
corrective feedback in oral communication classes. System 42, 429–39.
Zhang, X. & Lantolf, J. (2015). Natural or artificial: Is the route of L2 develop-
ment teachable? Language Learning 65, 152–80.
Zhao, Y. (2015). The effects of explicit and implicit recasts on the acquisition of
two grammatical structures and the mediating role of working memory.
Unpublished PhD thesis, University of Auckland.
Ziegler, N. (2016). Taking technology to task: Technology-mediated TBLT, per-
formance and production. Annual Review of Applied Linguistics 36, 136–63.
Zuengler, J. & Miller, E. (2006). Cognitive and social perspectives: Two parallel
SLA worlds. TESOL Quarterly 40, 35–58.

ability analysis, 190 Bangalore Project (Communicational

ability for use, 241, 245–6, 249, 254–6, Language Teaching Project), 175, 184,
259, 261, 265, 274 221, 285, 310
accountability, 181, 186, 189, 199, 223,
303, 307 co-construction, 105, 114, 118, 251, 253,
accuracy, 16, 23, 28, 53, 58, 60, 65–6, 265
68–9, 72–5, 78, 80, 82–3, 85, 87–8, Cognition Hypothesis, 41, 64, 69, 73–4, 82,
90–3, 98–100, 113, 117, 120, 122, 87, 93, 129–30, 148, 190, 212, 247,
125, 136, 138, 140, 144, 149, 191, 348, 359–60
194, 210, 212, 214, 218, 220–1, 226, ability analysis, 190
231, 247–9, 256, 260, 262–4, 288, behavioral analysis, 190
296, 299, 301, 321, 328, 334, 336, information-theoretic analysis, 190, 192
338–9, 346, 351, 361 resource-dispersing task factors, 194
achievement testing, 259, 266–8, 368 cognitive load, 99, 113, 206, 216, 292, 348,
acquisition poor environments, 344, 356, 355, 359
365 cognitive-interactionist perspective, 29, 105,
acquisition/acquisitional sequence, 338 109
action research, 308, 317–18, 373 collaborative dialogue, 116–21
Activity Theory, 109–11, 113, 115–16 common European framework of reference
affective profiles, 192 (CEFR), 257, 271, 273
anxiety, 81, 129–30, 147–54, 156, 161, communicative language teaching (CLT),
193, 212, 217, 225 3–4, 9, 18, 21, 24, 134–5, 208, 325,
aptitude, 70, 81, 129, 131–5, 148, 153, 156, 328–9, 338
190, 192, 196, 198, 204–6, 362 comparative method studies, 23, 281, 299,
language analytic ability, 131, 133–5 302, 330
MLAT, 131–2, 134 compensation ability, 255
working memory, 46, 51, 70, 76, 79, 81, complexity, 7, 14, 28, 39–40, 42, 47, 58,
87, 90, 116, 129, 135–41, 149, 153, 62, 65–6, 69, 73–5, 78, 80–3, 85, 87–9,
192–3, 227, 255, 261, 265, 362 91–2, 97–9, 113–14, 136, 143, 168,
aptitude profiles, 192, 196 188, 192, 194, 211–14, 220–1, 226,
articulation, 68, 71, 97, 136 231, 247–9, 252, 257, 262–4, 269,
attention, 9, 13, 15–17, 29, 38–9, 43, 45, 273, 288, 296, 299, 301, 305, 336,
48–9, 52, 54, 56, 63, 70, 72, 76–8, 80, 339, 346–7, 350, 354, 359, 361, 372
82, 87, 94, 98, 109, 111, 129, 136, complexity-accuracy-fluency (CAF), 58,
139, 147, 150, 152, 156, 158, 160, 65, 79, 87, 91, 214, 216, 218, 220,
162, 176, 185, 193, 200, 202, 206, 227, 231, 263
212, 217–19, 221–2, 224, 226, 228, complexity-accuracy-lexis-fluency
247, 250, 255, 285, 289, 303, 314, (CALF), 65, 70–2, 74, 87–8, 90, 247–9,
339, 343–4, 349, 357, 359–60 264, 361
attentional limitation, 71–2, 77, 89 lexical complexity, 65, 67, 72, 84, 98,
automatize, 82, 112, 121, 194, 200, 202, 101, 138, 214, 288, 296
204, 218, 363 lexical density, 83

412

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.023
Index 413
lexical sophistication, 67, 102, 258 definition of task, 38, 190, 196
lexical variety, 213 delayed CF, 55
syntactic complexity, 148–9, 213–14, discourse competence, 243
288, 296 discourse repair, 34
comprehensible input, 33 dynamic assessment, 108, 114, 116, 121–4,
comprehensible output, 32 127, 358, 361, 372
pushed output, 32
computer-mediated (CM) tasks, 42, 46 engagement, 63, 80, 88–9, 119, 143–4, 146,
concept-based language teaching, 112, 347 153, 160, 165, 171, 182, 295, 322, 326
consciousness-raising tasks, 321, 346, 355 engagement in language use (ELU), 171,
content-based language teaching/CLIL, 182
17–18 everyday concepts, 105, 107, 112, 114
contingency, 253, 361 experiential learning, 155, 157, 170, 172,
corrective feedback, 34–6, 39, 45, 50–1, 55, 325
60, 62, 116–17, 132–3, 137, 140, 148, explicit correction, 37, 54, 61, 117, 133,
151, 153, 171, 208, 215, 222–3, 232, 140, 224, 372
235, 237, 239, 290, 338–9, 341, 349 corrective feedback, 34–6, 39, 45, 50–1,
clarification request, 36, 46, 53, 57, 224, 55, 60, 62, 116, 132, 171, 208, 290,
322, 372 338–9, 341, 349
delayed feedback, 55, 232 explicit instruction, 31, 44, 57–8, 62–3,
elicitation, 37, 53, 224, 241, 372 111–12, 125, 223, 234, 291, 303, 311,
explicit correction, 37, 54, 61, 117, 133, 325, 328, 341, 349–50, 352
141, 224, 372 explicit learning, 30, 111, 133, 153, 334,
explicit feedback, 54, 60, 62, 116, 135, 372
140, 153, 224, 226 exploratory practice, 308, 317–19
focus on form, 16–17, 32, 35, 39, 44, 47, extensive CF, 52
51, 60, 62, 98, 133–4, 153, 176, 211, Extensive CF, 52
217, 219, 222–3, 232, 235, 237, 286,
293, 314, 334–5, 338 familiarity, 8, 40, 75, 80, 92, 113, 163–4,
immediate feedback, 56, 232 191, 194, 197, 215, 221, 248, 252, 317
implicit feedback, 54, 62, 116, 133, 135, filled pauses, 68, 258
140, 226, 237 flow theory, 171
intensive corrective feedback, 51 fluency, 4, 16, 23, 50, 58, 65, 68, 72–5, 78,
metalinguistic clue/feedback, 36–7, 54, 83–4, 88, 92–3, 97, 101–2, 113, 136,
133, 140, 152, 224, 372 138, 144, 168, 191–2, 199, 201, 211,
prompts, 45, 52, 55–7, 61, 117, 123, 143, 213–14, 220–1, 227, 231, 247–8, 250,
224, 226, 228 257, 263, 287, 296, 299, 301, 316,
recasts, 16, 35, 44, 46, 52, 54, 56, 61, 336, 338–9, 346, 350, 361, 372
117, 133, 140, 144, 152–3, 224, focus on form (FonF), 16–17, 32, 35, 39,
226–7, 371 44, 47, 51, 60, 98, 111, 133–4, 153,
repetition, 36, 40, 49, 68, 73–4, 91–3, 176, 189, 211, 217, 219, 222–3, 232,
96–7, 102, 140, 176, 224, 228, 230, 235, 237, 286, 293, 314, 334–5, 338,
232, 237, 247–8, 264, 362, 372 349
uptake, 21, 23, 38, 44, 46–7, 62, 87, 304, corrective feedback, 34–6, 39, 45, 50–1,
332, 372 55, 60, 62, 116–17, 132–3, 137, 140,
course book, 16, 267–8, 275–6, 278, 364 148, 151, 153, 171, 208, 215, 222–3,
criterion-referencing, 242, 279 232, 235, 237, 239, 290, 338–9, 341,
critics of TBLT, 335, 344 349
critiques/criticisms of TBLT, 23–4, 331, focus on forms (FonFs), 21, 111, 127, 229,
335, 350, 352 232, 286, 338

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.023
414 Index
formative, 21, 25, 176, 242, 279, 309, 318, working memory, 46, 51, 70, 76, 79, 81,
357, 368 87, 90, 116, 129, 135–41, 149, 153,
formative evluation, 308, 318 192–3, 227, 255, 261, 265, 362
language related episodes (LREs), 108, 219,
general programme comparisons, 281, 361
285–6, 295 languaging, 105–9
goal-tracking, 169, 205 language related epispdes (LREs)
Grading and sequencing tasks, 7–8, 14, 197, languaging, 105, 108, 120–1, 125, 355,
348, 371 357
grammar in TBLT, 338, 344, 352 learner-generated content, 162–3
lemma, 71, 89, 97, 255
here-and-now, 42, 66, 76, 80, 83, 85, 88, 95 lexical variety, 213
lexis, 65, 75, 84, 91–2, 97, 179, 247–8, 344
IATBLT, 364, 370 lexical complexity, 65, 67, 72, 84, 98,
imitation, 56, 105–9, 111 101, 138, 214, 288, 296
immediate corrective feedback (CF), 56 lexical density, 83
immediate feedback, 56, 232 lexical sophistication, 67, 102, 258
immersion, 45, 55, 132, 134–5 limbic system, 170
implementation variables, 40, 43, 345 Limited Attention Capacity Hypothesis, 136
implementing TBLT, 22, 146, 282, 304, listen-and-do task, 48, 340
306, 318, 324–30 longitudinal research, 360, 362
implicit language learning, 51, 111
incidental language learning, 170, 337 macro evaluation, 307–9, 316
individual differences, 90, 130, 133, 156, macrostructure, 78, 98
169, 190, 192, 195–6, 361 main task, 15, 143, 161, 176, 200, 206,
information-gap task, 11, 40, 339, 341, 371 209, 212–13, 221, 228, 234, 237, 239,
information-theoretic analysis, 190, 192 322, 339, 349, 357
initiation-response-feedback (IRF), 292 measurement, 8, 64, 67–8, 91, 122, 153,
innovation, 22, 282, 285, 303–7, 314, 324, 249, 256, 259, 267, 287, 296, 298,
329, 357, 366, 369 301, 361, 370
input-enhancement, 43 measuring task-based performance, 64–9
insider critics, 336, 344–51 mediation, 105–7, 109, 111, 114, 116–17,
intensive corrective feedback, 51 120, 122, 126–7
interaction approach, 29, 35, 40 meta-analyses/meta-analysis, 39, 43, 51–2,
Interaction Hypothesis, 29, 82, 136, 360 83–5, 90–3, 99, 131, 147, 225, 247–8,
interactional authenticity, 12, 21, 244, 250, 253, 263, 338, 349
260 metacognition, 256
interactionally modiﬁed input, 43, 48 metalinguistic feedback, 54, 133, 140, 152,
interactive ability, 243–4, 259, 261, 266, 372
268, 274, 280 corrective feedback, 34–6, 39, 45, 50–1,
interactive tasks, 43, 63, 75, 98, 156, 215, 55, 60, 62, 116, 132, 171, 208, 290,
338 338–9, 341, 349
International Association of Task-Based micro processes, 98
Language Teaching (IATBLT), 359 micro-evaluation, 282, 307–8, 316–23, 328,
330
LAC, 64, 72, 74, 76–8, 83, 86, 89, 98, 100 modality, 89, 91, 148, 150–2, 230
language aptitude, 129, 131–5 Modern Language Aptitude Test (MLAT),
language analytic ability, 131, 133–5 131–2, 134
MLAT, 131–2, 134 aptitude, 81, 129, 131–5, 148, 153, 362
psychological perspectives, 129 language aptitude, 129, 131–5

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.023
Index 415
modular curriculum/syllabus, 25, 33, 199, planning, 17, 19, 58, 69, 72–3, 75–7, 80, 85,
335, 355 90–1, 93, 96–7, 101, 130, 137, 139, 141,
modular syllabus, 204 150, 154, 176, 202–3, 211–13, 215–16,
motivation, 81, 90, 94, 129, 131, 142–7, 223, 226, 236, 244, 247–50, 254–5,
153, 156, 160, 162, 170–1, 181, 192, 263–4, 269, 273, 326–7, 334, 336, 339,
196, 199, 203, 205–6, 208, 210, 221, 343, 348, 354–5, 357, 362, 367
223, 225, 308, 320–2, 362 pre-task planning, 138, 141, 211
strategic planning, 73, 90, 97, 137, 212,
narrative tasks, 148, 152, 165 227, 247, 255
native-speaker, 66, 101–2, 316 within-task planning, 137, 139, 141, 213,
needs analysis, 20, 115, 163–4, 186, 188, 223, 226
190, 197, 200, 244, 269, 312, 316, post-task activities, 228, 322, 334, 357
318, 345–6, 354 post-task condition, 73, 75–6, 93, 247
negotiation of form, 35, 47, 61, 168–9 post-task feedback, 229, 232
negotiation of meaning, 35, 39–40, 42, 47, delayed feedback, 55, 232
60–1, 82, 168, 191, 294, 319, 322 post-task options, 161, 228–35, 237
non-understanding routines, 34–5 practitioner research, 235–6, 373
norm-referencing, 242 pragmatic ability/abilities, 251–3, 265
noticing, 29–30, 32, 39, 43–4, 46–7, 51, 61, pragmatic competence, 42, 196, 244
63, 80, 82, 87–8, 108, 140–1, 144, pre-emptive focus on form, 44, 222
153, 169, 227, 233, 334–5, 345 pre-modified input, 43, 48, 62
pre-task focus on form, 217
on-line planning, 73, 90, 97, 247, 255, 264, pre-task grammar instruction, 211,
357 217–18, 239
opinion gap task, 40 pre-task modelling, 211, 217, 234, 237
outcome, 4, 10, 12, 21, 40, 48, 80, 109, pre-task options/work, 161, 209
113–14, 133–4, 139–40, 142, 144, pre-task planning, 17, 137, 139, 141, 176,
147, 151–2, 157, 164, 169, 172, 175, 212–13, 215, 217, 227, 236, 250, 264,
182, 184, 189, 191, 196, 198, 203, 327, 339, 343, 348
205, 207, 210, 222, 228, 231, 235, planning, 17, 19, 58, 69, 72–3, 75–7, 80,
245, 262, 273, 284, 287, 289–90, 295, 85, 90–1, 93, 96–7, 102, 130, 137,
300, 308, 314, 317, 319, 322, 339, 139, 141, 150, 154, 176, 202–3,
342–3, 345, 353, 361 211–13, 215–16, 223, 226, 236, 244,
output prompting corrective feedback (CF), 247–50, 254–5, 263–4, 269, 273,
52 326–7, 334, 336, 339, 343, 348,
output-based task, 33, 47, 50, 169, 338, 354–5, 357, 362, 367
341 pre-verbal message, 71, 254
output-prompting corrective feedback (CF) private speech, 105–6, 108, 111, 120
prompts, 45, 52, 55–7, 61, 117, 123, 143, problem-solution structure, 76, 85
224, 226, 228 procedural syllabus, 198
outsider critics, 335–44, 350, 352 proficiency level, 57, 66, 79–80, 89, 138,
156, 169, 191, 252, 271, 275, 354, 364
participation variables, 79–80 proficiency tests, 260, 266, 304
participatory structure, 59–60, 146, 209, project work, 365
224, 235–7, 239, 335, 341 prompts, 45, 52, 55–7, 61, 117, 123, 143,
pedagogic tasks, 7, 12–13, 18, 89, 159, 164, 224, 226, 228
167, 172, 175, 182, 186–7, 191, 195, corrective feedback, 34–6, 39, 45, 50–1,
197, 200, 206, 336, 345–6, 354 55, 60, 62, 116, 132, 171, 208, 290,
performance measurement, 370 338–9, 341, 349
personal investment theory, 160–2, 169, 172 output-prompting feedback, 61–2, 137
phonological short-term memory, 139 pushed output, 32, 171

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.023
416 Index
qualitative methods/research, 85, 248, 355, resource pools, 79, 87
360 resource-directing task factors, 193
quantitative methods/research, 320 resource-dispersing task factors, 193
restructure, 82, 86, 188, 202–3
range, 23, 52, 60, 65, 70, 73, 77, 84, 89–91,
93, 95, 99–100, 119, 169, 206, 209, scaffolding, 105–9, 117, 119, 211, 221,
242, 245, 247, 250–1, 254–5, 257, 317, 358
260, 262, 267, 269, 271, 274, 276–8, scientific concepts, 105, 107, 109, 111, 114,
344, 354, 357, 360, 368 124, 127, 373
raters, 245–6, 252, 256–8 second language mental lexicon, 71, 255
rating scales, 245–6, 256, 258, 271 sequencing, 7–8, 11, 14, 79, 81, 89, 113, 175,
reactive focus on form, 36, 39, 44, 51, 222, 179, 187, 191, 322, 337, 371
339 sequencing tasks, 322
corrective feedback, 34–6, 39, 45, 50–1, skill-learning theory, 30, 328, 349
55, 60, 62, 116–17, 132–3, 137, 140, SLML, 72, 74, 78
148, 151, 153, 171, 208, 215, 222–3, small group work, 15, 29, 59–60, 120, 327,
232, 235, 237, 239, 290, 338, 341, 349 341
real-world tasks/real-life tasks, 7, 12, 20, sociocultural theory, 56, 103, 105, 108–9,
184–6, 190, 192, 197, 259–60, 269, 113, 126, 221, 322, 336, 361, 367
276, 345 scaffolding, 105–9, 117, 119, 211, 221,
reasoning, 8, 41–2, 80–1, 83, 85, 88, 91–2, 317, 358
113, 129, 139, 148, 175, 183–4, zone of proximal development, 106–7
192–3, 197–8, 263, 274, 348 specific purposes, 287, 371
reasoning demands, 84 SSARC, 64, 69, 79, 81–2, 84, 86, 88–9, 98,
reasoning gap tasks, 183 100, 188, 194, 200
recasts, 16, 35, 44, 46, 52, 54, 56, 61, 117, stabilise, 83, 86
133, 140, 144, 152–3, 224, 226–7, 372 stimulated recall, 45–6
corrective feedback, 34–6, 39, 45, 50–1, strategic competence, 243–4, 254, 260
55, 60, 62, 116, 132, 171, 208, 290, strategic planning, 73, 90, 97, 137, 212,
338–9, 341, 349 227, 247, 255
focus on form, 16–17, 32, 35, 39, 44, 47, pre-task planning, 17, 138, 141, 176,
51, 133–4, 153, 176, 189, 211, 217, 211–13, 215, 217, 227, 236, 250, 264,
219, 222–3, 232, 235, 237, 286, 293, 327, 339, 343, 348
314, 334–6, 338 structural complexity, 65, 67–9, 74, 78, 83,
reflection, 77, 80, 171, 186, 229, 235, 250, 91, 102
309, 317, 322 Structural Oral-Situational Method/Model,
reliability, 56, 241, 266–7, 270–1, 278, 329, 6, 285–6
368 structural syllabus, 3, 24, 180, 199, 204,
repair fluency/repair-linked fluency, 68 303, 328, 351
fluency, 50, 58, 65, 68, 72–5, 78, 83, 88, structured/unstructured tasks, 66, 73, 77,
91–3, 97, 101–2, 113, 136, 138, 144, 98, 138, 247, 249
211, 213–14, 218, 220–1, 227, 231, subordination, 65, 94, 102, 138, 201, 213,
247–8, 250, 257, 263, 287, 296, 299, 220, 257
301, 336, 339, 346, 350, 361, 372 syntactic complexity, 148–9, 213–14,
repetition, 36, 40, 49, 68, 73–4, 91–3, 96–7, 288, 296
102, 140, 176, 224, 228, 230–1, summative assessment, 21
237–8, 247–8, 264, 341, 343, 362, 372 summative evaluation, 308
corrective feedback, 34–6, 39, 45, 50–1, 55, support material, 269, 357
60, 62, 116–17, 132–3, 137, 140, 148, syllabus design, 180–4, 195–6
151, 153, 171, 208, 215, 222–3, 232, System 1, 94
235, 237, 239, 290, 338–9, 341, 349 System 2, 94

Downloaded from https://www.cambridge.org/core. National University of Singapore (NUS), on 15 Mar 2021 at 06:14:32, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108643689.023
Index 417
task characteristics, 70–2, 75, 78, 89, 93, TBLT as innovation, 304–7
96, 192, 211, 247–8, 250, 254, 259, technology mediated tasks/TBLT, 18–20
261, 263–4, 271, 355 there-and-then, 66, 76, 80, 83, 85, 88, 91, 95
task complexity, 14, 41, 79–83, 87–8, 90, time perspective, 76, 80, 84, 86, 88, 91, 96,
98–101, 115, 130, 137, 139, 146, 247
148–9, 152, 168, 217, 263–4, 347, time pressure, 74, 160, 202, 204, 215, 226,
354, 359, 370, 372 256, 264, 272, 278, 357
task conditions, 70–2, 75, 78, 80, 89, 93, topic preparation, 211, 356
96, 101, 130, 139, 141, 149–50, 152, trade-off, 69, 75, 77, 79
154, 191, 247–8, 250, 253–4, 261,
263–4, 348, 354–5, 360 underlying competence, 243–5, 249, 254–5,
task design, 11, 39–41, 43, 58, 62, 73, 94, 260–1
96, 113, 116, 127, 173, 188, 193, 319, unplanned, 83, 101, 250
345, 354 uptake, 21, 23, 38, 44, 46–7, 62, 87, 194,
task difficulty, 14, 74, 79, 81, 90, 304, 307, 372
99–100, 130, 259, 262, 270, 278, corrective feedback, 34–6, 39, 45, 50–1,
354, 361 55, 60, 62, 116–17, 132–3, 137, 140,
task fulfilment, 257, 262 148, 151, 153, 176, 208, 215, 222, 232,
task implementation, 40, 116–17, 127, 169, 235, 237, 239, 290, 338–9, 341, 349
173, 175, 208 use of first language (L1), 310
task repetition, 40, 73, 228, 231–2, 237, 343
posttask options, 229 validity, 67, 100, 124, 131, 151, 208, 242,
task selection, 13, 42, 113–15, 163, 175, 266–7, 270–1, 273, 276, 278, 284–5,
179, 187, 191–2, 197 300, 329, 337, 368
task sequencing, 79, 81, 180, 182–3, 188, virtual worlds, 367
191, 193–5, 197, 200, 206, 348
syllabus design, 79, 82, 87, 98, 179, 363, 365 washback, 176, 242, 266, 268, 275–6, 278
task support, 93, 248 task-based assessment, 20–1, 74, 169,
task types, 7, 11, 42, 70, 77, 151, 163, 165, 266, 277–9, 329
186–7, 204, 247, 316, 319, 364 within-task planning, 137, 139, 141, 213,
task-as-process, 7, 10, 196, 198 223, 226
task-as-workplan, 7, 10, 115, 143, 196, words per clause, 66
198, 336, 355 syntactic complexity, 148–9, 213–14,
task-based (language) assessment, 20–1, 288, 296
169, 266, 278–9, 329, 332 working memory, 46, 51, 70, 76, 79, 81, 87,
task-based lesson, 15, 25, 55, 175, 209, 90, 116, 129, 135–41, 149, 153,
219, 236, 309, 326, 334, 356–7 192–3, 227, 255, 261, 265, 362
task-based research, 39, 47, 66, 97, 139, www.tblt.org, 359
212, 279, 308, 319, 331, 343–4, 352,
357 Zone of Proximal Development, 106–7, 127

Grabe - Reading in A Second Language - Moving From Theory To Practice, Second Edition
100% (1)
Grabe - Reading in A Second Language - Moving From Theory To Practice, Second Edition
603 pages
Task-Based Language Teaching Theory and Practice
100% (5)
Task-Based Language Teaching Theory and Practice
434 pages
1 Key Questions in Language Teaching An Introduction PDF
100% (6)
1 Key Questions in Language Teaching An Introduction PDF
430 pages
Preview Key Issues in Language Teaching
No ratings yet
Preview Key Issues in Language Teaching
53 pages
Educating Second Language Teachers
91% (11)
Educating Second Language Teachers
355 pages
Quick Guide To Mexican Spanish Preview 10212013
No ratings yet
Quick Guide To Mexican Spanish Preview 10212013
22 pages
Task Based Language Teaching
100% (23)
Task Based Language Teaching
84 pages
(Applied Linguistics in Action) Michael Rost-Teaching and Researching Listening-Pearson Education ESL (2011) PDF
100% (3)
(Applied Linguistics in Action) Michael Rost-Teaching and Researching Listening-Pearson Education ESL (2011) PDF
425 pages
Test 1
No ratings yet
Test 1
16 pages
Comparative Stylistics of French and English. A Methodology For Translation
No ratings yet
Comparative Stylistics of French and English. A Methodology For Translation
33 pages
02.0 PP II II The Cambridge Applied Linguistics Series
No ratings yet
02.0 PP II II The Cambridge Applied Linguistics Series
4 pages
Task-based Language Teaching - Rod Ellis Et.al
No ratings yet
Task-based Language Teaching - Rod Ellis Et.al
435 pages
2020 Жылгы Кытап Task-based
No ratings yet
2020 Жылгы Кытап Task-based
44 pages
Frontmatter
No ratings yet
Frontmatter
15 pages
Ellis R. Skehan P. Li S. Shintani N. and
100% (3)
Ellis R. Skehan P. Li S. Shintani N. and
14 pages
Studying Speaking to Inform Second Language Learning
From Everand
Studying Speaking to Inform Second Language Learning
Diana Boxer
No ratings yet
Task-Based Language Teaching
100% (10)
Task-Based Language Teaching
238 pages
Morphosyntactic Issues in Second Language Acquisition
From Everand
Morphosyntactic Issues in Second Language Acquisition
Danuta Gabryś-Barker
No ratings yet
Michael Thomas_ Hayo Reinders (editors) - Contemporary Task-Based Language Teaching in Asia_ Contemporary Studies in Linguistics (2015, Bloomsbury Academic) [10.5040_9781474219877] - libgen.li
No ratings yet
Michael Thomas_ Hayo Reinders (editors) - Contemporary Task-Based Language Teaching in Asia_ Contemporary Studies in Linguistics (2015, Bloomsbury Academic) [10.5040_9781474219877] - libgen.li
417 pages
Task-Based Language Teaching in Foreign Language Contexts - Research and Implementation
100% (2)
Task-Based Language Teaching in Foreign Language Contexts - Research and Implementation
384 pages
Investigating Tasks in Formal Language Learning
From Everand
Investigating Tasks in Formal Language Learning
María del Pilar García Mayo
No ratings yet
01.0 PP I VI Frontmatter
No ratings yet
01.0 PP I VI Frontmatter
6 pages
Task-Based Language Teaching: Definition, Characteristics, Purpose and Scope
No ratings yet
Task-Based Language Teaching: Definition, Characteristics, Purpose and Scope
10 pages
Cross-linguistic Similarity in Foreign Language Learning
From Everand
Cross-linguistic Similarity in Foreign Language Learning
Håkan Ringbom
No ratings yet
Vocabulary Learning Strategies and Foreign Language Acquisition
From Everand
Vocabulary Learning Strategies and Foreign Language Acquisition
Višnja Pavičić Takač
2/5 (8)
Task-Based Teaching and Learning
No ratings yet
Task-Based Teaching and Learning
10 pages
Book+Review+1
No ratings yet
Book+Review+1
5 pages
Input for Instructed L2 Learners: The Relevance of Relevance
From Everand
Input for Instructed L2 Learners: The Relevance of Relevance
Anna Niżegorodcew
No ratings yet
Fossilization in Adult Second Language Acquisition
From Everand
Fossilization in Adult Second Language Acquisition
ZhaoHong Han
No ratings yet
Moore 2018
No ratings yet
Moore 2018
6 pages
01.0 PP I VI Frontmatter
No ratings yet
01.0 PP I VI Frontmatter
6 pages
Awanda Novia N. /138/Writing-IV
No ratings yet
Awanda Novia N. /138/Writing-IV
4 pages
Task_Based_Second_Language_Learning
No ratings yet
Task_Based_Second_Language_Learning
14 pages
Van Den Braden 2016 TBLT
No ratings yet
Van Den Braden 2016 TBLT
14 pages
The Language Teaching Matrix
No ratings yet
The Language Teaching Matrix
356 pages
EJ1311226
No ratings yet
EJ1311226
10 pages
Li Xueping - The Applications and Reflection of TBLT (Task-Based Language Teaching) in College English Reading Courses
No ratings yet
Li Xueping - The Applications and Reflection of TBLT (Task-Based Language Teaching) in College English Reading Courses
10 pages
9 TBLT
No ratings yet
9 TBLT
12 pages
Pro Elt Assignment 3 Action Research
No ratings yet
Pro Elt Assignment 3 Action Research
21 pages
TEFL
100% (1)
TEFL
116 pages
Candlin Taking The Curriculum To Task
No ratings yet
Candlin Taking The Curriculum To Task
16 pages
Task Based Language Teaching in Asia
No ratings yet
Task Based Language Teaching in Asia
298 pages
TBL and Young Learners
No ratings yet
TBL and Young Learners
8 pages
Perspectives On Pedagogical Grammar 0521445302 0521449901 Compress
No ratings yet
Perspectives On Pedagogical Grammar 0521445302 0521449901 Compress
171 pages
What Teachers Need to Know About Language
From Everand
What Teachers Need to Know About Language
Carolyn Temple Adger
No ratings yet
Chapter Ii
No ratings yet
Chapter Ii
40 pages
Studies in Second Language Acquisition of Chinese
From Everand
Studies in Second Language Acquisition of Chinese
ZhaoHong Han
4/5 (1)
460 1355 2 RV
No ratings yet
460 1355 2 RV
11 pages
9781003227267_previewpdf
No ratings yet
9781003227267_previewpdf
76 pages
The Cambridge Handbook of Task Based Lan
No ratings yet
The Cambridge Handbook of Task Based Lan
4 pages
Teaching Grammar Through Task-Based Language Teaching To Young EFL Learners
No ratings yet
Teaching Grammar Through Task-Based Language Teaching To Young EFL Learners
15 pages
Referência teórica para as discussões no 1º semestre
No ratings yet
Referência teórica para as discussões no 1º semestre
214 pages
Carte Nunan Designing Tasks For The Communicativ
100% (2)
Carte Nunan Designing Tasks For The Communicativ
110 pages
(1989) Livro Designing - Tasks - For - The - Communicative - Classroom - Nunan - D PDF
No ratings yet
(1989) Livro Designing - Tasks - For - The - Communicative - Classroom - Nunan - D PDF
110 pages
In Defense of Tasks and TBLT Nonissues and Real Issues
No ratings yet
In Defense of Tasks and TBLT Nonissues and Real Issues
29 pages
A Bird's Eye View On Teachers' Perception of Task-Based Language Teaching (TBLT)
No ratings yet
A Bird's Eye View On Teachers' Perception of Task-Based Language Teaching (TBLT)
17 pages
Tilfarlioglu e Basaran (2007) - Enhancing Reading Comprehension Through Task-based Writing Acativities
No ratings yet
Tilfarlioglu e Basaran (2007) - Enhancing Reading Comprehension Through Task-based Writing Acativities
19 pages
First Language Use in Second and Foreign Language Learning
From Everand
First Language Use in Second and Foreign Language Learning
Miles Turnbull
No ratings yet
Language Learners with Special Needs: An International Perspective
From Everand
Language Learners with Special Needs: An International Perspective
Judit Kormos
No ratings yet
Nunan
No ratings yet
Nunan
18 pages
T - B L T: ASK Ased Anguage Eaching
No ratings yet
T - B L T: ASK Ased Anguage Eaching
15 pages
Task-Based Language Teaching: What Every EFL Teacher Should Do
No ratings yet
Task-Based Language Teaching: What Every EFL Teacher Should Do
7 pages
Developing Writing Competence in L2 Chinese Classrooms: Research and Application
From Everand
Developing Writing Competence in L2 Chinese Classrooms: Research and Application
Li Yang
No ratings yet
Academic English Syllabus: Week 1
No ratings yet
Academic English Syllabus: Week 1
4 pages
Academic English Syllabus
No ratings yet
Academic English Syllabus
2 pages
Time
No ratings yet
Time
1 page
Comic Strip Vacation Activity
No ratings yet
Comic Strip Vacation Activity
3 pages
Factual Description
50% (2)
Factual Description
4 pages
Sociolinguistics - Language and Culture
No ratings yet
Sociolinguistics - Language and Culture
3 pages
Evolution of Semantic Similarity - A Survey
No ratings yet
Evolution of Semantic Similarity - A Survey
35 pages
Sample Written Output
No ratings yet
Sample Written Output
23 pages
Airstreammechanism 160507072326
No ratings yet
Airstreammechanism 160507072326
19 pages
Elmt Score Report: English Course
No ratings yet
Elmt Score Report: English Course
2 pages
Grade-8, English, WS
No ratings yet
Grade-8, English, WS
21 pages
Speech Context Quiz
No ratings yet
Speech Context Quiz
26 pages
1 Unit 1 Lesson 1
No ratings yet
1 Unit 1 Lesson 1
5 pages
APAIn-TextCitations Formula and Style
No ratings yet
APAIn-TextCitations Formula and Style
2 pages
POVERBS
No ratings yet
POVERBS
4 pages
Dzexams 3as Anglais Al - E3 20161 498870
No ratings yet
Dzexams 3as Anglais Al - E3 20161 498870
6 pages
Reduplication in CIA-CIA Language
No ratings yet
Reduplication in CIA-CIA Language
40 pages
Sabado Marzo 24 Inglés 1
No ratings yet
Sabado Marzo 24 Inglés 1
5 pages
AP Style Grammar Powerpoint
100% (1)
AP Style Grammar Powerpoint
99 pages
Taller de Ingles Jose
No ratings yet
Taller de Ingles Jose
4 pages
Test Result Analysis MTB DIVISION
No ratings yet
Test Result Analysis MTB DIVISION
7 pages
Language Myths
No ratings yet
Language Myths
9 pages
Matrix Pre Intermediate Tests
No ratings yet
Matrix Pre Intermediate Tests
31 pages
Chapter 10 Teaching Pronunciation
No ratings yet
Chapter 10 Teaching Pronunciation
20 pages
GREX3 U5RTG Chart5.1-5.6
No ratings yet
GREX3 U5RTG Chart5.1-5.6
6 pages
Expressive Language (Using Words and Language) : Print PDF
No ratings yet
Expressive Language (Using Words and Language) : Print PDF
8 pages
Homework 9 20184
No ratings yet
Homework 9 20184
5 pages
Business Communication Chapter 2-Communications Across Culture
No ratings yet
Business Communication Chapter 2-Communications Across Culture
14 pages
Dynamic, Summative Evaluation (UNIT V) Variant I
No ratings yet
Dynamic, Summative Evaluation (UNIT V) Variant I
3 pages
Book-Introduction To Linguistics 2019
No ratings yet
Book-Introduction To Linguistics 2019
153 pages
New TE Starter Students-E
No ratings yet
New TE Starter Students-E
143 pages