Applied Language Studies

edited by David Crystal and Keith Johnson

Instructed
Second
Language
Acquisition

Rod Ellis
 And Keith Johnson

.. deal with key topics within the main branches of

ianv,-<age studies — initially in the fields of foreign language
teazling and learning, child language acquisition with clinical or remedial
language studies. The series will provide students with a research
perspective in a particular topic, at the same time containing an original
slant which will make each volume a genuine contribution to the
development of ideas in the subject.

Series List

Communication Strategies in Second Language Acquisition

Ellen Bialystok

Computers and Written Texts

Christopher S. Butler

Chomsky’s Universal Grammar

An Introduction
V/. J. Cook

Principles of Language Testing

Alan Davies

Instructed Second Language Acquisition

Learning in the Classroom
Rod Ellis

Teaching Oral Communication

A Methodological Framework
William Littlewood

The ELT Curriculum

Design, Innovation and Management
Ronald V. White
INSTRUCTED SECOND
LANGUAGE ACQUISITION
Learning in the Classroom
i

Rod Ellis

B
BLACKWELL
Oxford UK fr Cambridge USA
 Copyright © Rod Ellis 1990

First published 1990

Reprinted 1991, 1992, 1994, 1995, 1997, 1999, 2000

Blackwell Publishers Ltd

108 Cowley Road
Oxford OX4 1JF, UK

Blackwell Publishers Inc

350 Main Street
Malden, Massachusetts 02148, USA

All rights reserved. Except for the quotation of short passages for the purposes
of criticism and review, no part of this publication may be reproduced, stored
in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior permission
of the publisher.

Except in the United States of America, this book is sold subject to the condition
that it shall not, by way of trade or otherwise, be lent, re-sold, hired out, or
otherwise circulated without the publisher’s prior consent in any form of binding
or cover other than that in which it is published and without a similar condition
including this condition being imposed on the subsequent purchaser.

British Library Cataloguing in Publication Data

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

Library of Congress Cataloging in Publication Data

Ellis, Rod
Instructed second language acquisition: learning in the classroom/Rod Ellis,
p. cm. — (Applied language studies)
Includes bibliographical references (p. ).
ISBN 0-631-16202-X
1. Language and languages — Study and teaching. 2. Second language
acquisition. I. Title. II. Series.
P51.E47 1990 89^19728
418’.007—dc20 CIP

Typeset in 10 on 12pt Ehrhardt

by Setrite Typesetters Ltd
Printed and bound in Great Britain
by TJ International Ltd., Padstow, Cornwall

This book is printed on acid-free paper

Contents

Preface vii

List of Abbreviations ix

1 Investigating Classroom Language Learning 1

2 Behaviourist Learning Theory and Classroom Language Learning 19

3 Naturalistic Language Acquisition and Classroom Language Learning 32

4 Classroom Process Research 64

5 The Relationship between Classroom Interaction and Language

Learning 93

6 Formal Instruction and Language Learning 130

7 An Integrated Theory of Instructed Second Language Learning 174

8 Final Comments 199

Bibliography 205

Index 225
For Lwindi and Emma
Preface

I have been interested for a long time now in how an understanding of second
language acquisition can contribute to language pedagogy. In an earlier book,
Understanding Second Language Acquisition, I tried to provide an overview of the
whole field. But Second Language Acquisition (SLA); as a field of enquiry,
has grown enormously over the past few years. It has spawned countless
empirical studies and is in danger of splintering into relatively isolated sub¬
fields. It has become almost impossible for one person to keep abreast of
publications in all areas of SLA. I have chosen, therefore, to adopt a more
focused approach in this book, by concentrating on the research that has
addressed how classroom second language acquisition takes place.
This book is aimed at teachers and applied linguists who wish to develop
their theoretical understanding of how learners learn a second language (L2)
through instruction and to inform themselves about the research that has
contributed to this understanding.
The book has the following aims:

1 To provide a historical context for the empirical study of classroom

language acquisition by examining early attempts to formulate a theory
through extrapolation from general learning theory and from the study
of naturalistic language acquisition.
2 To review a wide range of research which has investigated classroom
language acquisition.
3 To develop a theory of instructed second language acquisition which is
compatible with the results of this research.

In my earlier book, I argued that teachers would benefit from trying to

make their own theories of how acquisition takes place explicit. This is a view
I still hold. By focusing specifically on classroom language acquisition I have
tried to make my current position clear, so that readers can subject it to
critical scrutiny and, by so doing, make explicit their own position.
I am very aware that there are many perspectives on how learners leam a
second language in a classroom. A full understanding will need to make use of
insights provided by sociological, sociolinguistic and educational enquiry. In
viii Preface
choosing to operate from a relatively narrow base — the issues that preoccupy
L2 acquisition researchers — I do not intend to suggest that this is the only or
the most profitable line of enquiry. It is simply one possible way of obtaining
insights into the processes that contribute to language learning. Other com¬
mentators will doubtlessly follow other routes. We need to guard against
dogmatists who try to argue that one approach — whether linguistic, psycho-
linguistic, sociolinguistic or educational — provides us with a better view of
the truth than another. I believe that we build up our understanding of the
complex nature of classroom language learning bit by bit, drawing on different
approaches and the insights they afford. The picture this book affords is a
partial one. It will need to be filled out with the help of other types of enquiry.
I am indebted to a number of people who have assisted me in one way or
another to write this book. I am grateful to Keith Johnson for helping me
refine my own thinking without ever imposing his own. Ros Mitchell’s com¬
ments on the first draft were also very valuable. I am also grateful to fellow
lecturers and students at Ealing College of Higher Education who have
responded to the ideas expressed in this book and helped me to refine them.

A Note on Pronouns
In this book I have used ‘she’ and ‘her’ to refer generically to learners,
teachers, researchers etc. This is a departure from my previous practice, but I
have come to realize that the choice of pronouns is an important issue to many
women and that, overall, less offence is likely to be caused by the choice of the
female gender.
Abbreviations

ESL English as a Second Language

FL Foreign Language
IRF Initiate — Respond — Feedback
LAD Language-Acquisition Device
LI First Language (i.e. mother tongue)
L2 Second Language
MLU Mean Length of Utterance
NNS Non-native Speaker
NS Native Speaker
SLA Second Language Acquisition
TESOL Teaching English as a Second or Other Language
.
1 Investigating
Classroom Language
Learning

Introduction

The question which this book tries to answer is ‘How does second language
(L2) learning take place in a classroom?’ The primary aim of the book is to
explore learning rather than teaching. It is concerned with how classroom
learners construct the mental grammar that underlies their use of an L2, not
with how teachers teach. This focus on learning has been motivated by the
belief that it is necessary to understand as fully as possible the processes by
which learners internalize a knowledge of a L2. Such an understanding will
contribute to L2 acquisition research and will also serve as a basis for
pedagogic recommendations. The focus on learning is additionally motivated
by the conviction that a theory of classroom learning needs to be explicitly
formulated so that statements about how learners learn and how teachers
ought to teach can be subjected to critical scrutiny.
Of course, a book which sets out to examine how classroom L2 learning
takes place cannot ignore teaching — for the obvious reason that what dis¬
tinguishes the classroom from a naturalistic setting is the attempt to teach the
L2. It is this which makes the study of classroom L2 learning such a fascinating
subject of enquiry.

Classroom Language Learning

A good starting point, then, is to try to define ‘classroom language learning’.

The term exists in opposition to ‘naturalistic language learning’. The difference
between the two types of learning can be examined from a sociolinguistic, a
pyscholinguistic and an educational viewpoint.
Sociolinguistically, the distinction between classroom and naturalistic L2
learning can be viewed as one of domain (Fishman, 1964). Domains are
constellations of factors that affect the way language is used. The domains of
classroom and naturalistic learning can be distinguished with reference to
such factors as location, participants, topics and purposes. The differences on
2 Instructed Second Language Acquisition
each of these dimensions is fairly self-evident and needs little comment. In
general, the domain of classroom language learning is circumscribed in com¬
parison to that of naturalistic language learning. The latter is likely to be
characterized by a greater range of settings, participants, topics and purposes,
although it is possible to envisage exceptional cases of each type of learning
where the reverse applies — the classroom can provide the richer, more
diverse learning experiences.
Psycholinguistically, the key distinction is between formal and informal
language learning (Krashen and Seliger, 1976; d’Anglejan, 1978). Formal
learning involves some kind of studial activity on the part of the learner — for
example an attempt to learn about the language by obtaining information
about explicit rules of grammar. Informal learning takes place through obser¬
vation and direct participation in communication — learning is a process of
discovery which takes place spontaneously and automatically providing certain
conditions have been met. It would be a mistake to equate classroom and
formal learning on the one hand and naturalistic and informal learning on the
other. The fact is that classroom learning can and does involve informal
learning — for instance when learners have the opportunity to engage in
meaning-focused communication. Advocates of the communicative method
(e.g. Brumfit, 1984) emphasize the importance of teaching activities designed
to encourage informal learning. Similarly, naturalistic learning can involve
formal learning, for example when a learner asks a question about a linguistic
form in the middle of a conversation. However, it is probably true to say that
the classroom setting affords more opportunities for formal learning and
naturalistic settings more opportunities for informal learning. The psycho-
linguistic difference between classroom and naturalistic learning may not be
absolute but it is nevertheless significant.
Educationalists often distinguish the idea of formal training and apprentice¬
ship.1 Formal training typically occurs in classrooms. It involves some deliberate
attempt to shape the learning experiences in the belief that by so doing the
learner will be able to acquire knowledge more efficiently. Thus Stem (1983:19)
defines classroom language learning as ‘learning which has been induced or
influenced by some form of deliberately planned social intervention’.2 The
intervention need not be designed to cater for formal language learning.
Instruction that seeks to provide opportunities for learning through natural
communication is also an attempt at social intervention. After all, it is the
teacher who supplies the materials for an information-gap activity and who
instructs the learners to get into groups. Apprenticeship involves learning by
doing. No deliberate attempt is made to shape the learning environment by
devising a syllabus or providing special activities. The learner-apprentice
works side by side with the master-craftsman and ‘picks-up’ skills through
observation and practice. It should be noted, however, that the learning
environment provided by apprenticeship is possible in the classroom, and
formal training can occur in a naturalistic setting. But classrooms are ideally
suited to formal training, while naturalistic settings tend to give rise to
apprenticeship.
 Investigating Classroom Language Learning 3
In many respects the distinction between classroom and naturalistic language
learning is a crude one. It presupposes that it is possible to generalize about
the characteristics of each domain and the types of learning and teaching that
take place. In reality, there are many different types of classrooms and natural
settings are multifarious. Nevertheless, providing that the inherent variety of
both settings is borne in mind, the distinction is an important one. In studying
classroom language learning, we are trying to discover how a typical constellation
of social factors leads to attempts on the part of the teacher to control the
environment in order to provide opportunities for language learning. The
central questions are ‘Does intervention promote L2 learning?’ and ‘What
kind of intervention is most effective?’ These are important question to ask
both for theory building in L2 acquisition research and for language pedagogy.

Building a Theory of Classroom Language Learning:

Three Approaches

How can we develop our understanding of classroom language learning? How

can we build a theory of the way in which classroom learners acquire an L2?
There are three basic approaches that have been used.
One way is to assume that classroom language learning is just tike any other
kind of learning and can be explained with reference to a general theory of
learning. According to this view it is not necessary to build a separate theory
of classroom language learning. All that is required is to demonstrate the
applicability of a general theory to instructed L2 learning and to show how the
hypotheses that comprise the theory can be applied to it. Such an approach
has proved popular in the past and continues to do so today. It is, however,
contentious, for not everyone agrees that language learning is the same as
other kinds of learning. Chomsky (1965; 1986), for instance, has argued
strongly that language constitutes a separate mental faculty and is not acquired
in the same way as other knowledge systems.3 Irrespective of whether Chomsky is
right or wrong, it can be argued that extrapolating from a general theory is a
hazardous undertaking on the grounds that important facts relating to both
the nature of language and the classroom setting can be overlooked. It is
significant that those applied linguists who have turned to a general theory of
learning to explain classroom learning have rarely bothered to collect data
from inside the classroom to test the claims of the theory. The process is one¬
way only; extrapolation occurs, but there is no attempt at falsification.4
A second way of building a theory of classroom language learning is to
assume that instructed L2 learning proceeds in the same way as naturalistic
language learning. This approach has also proved popular. It takes two forms,
depending on whether the starting point is child LI acquisition or naturalistic
L2 acquisition. Clearly the degree of extrapolation to the classroom setting is
greater in the case of the former than the latter. The claim that classroom
language learning is identical (or similar) to LI acquisition involves an accept-
4 Instructed Second Language Acguisltlon
ance of the L2 = LI hypothesis. It also implies a belief that adults can
acquire an L2 in the same way as children. It is possible to question both
assumptions (cf. Bley-Vroman, 1988 and Long, 1988a) and very difficult to
prove either. The claim that naturalistic L2 learning and classroom language
learning have many features in common is a much safer one. It is also a claim
that can be subjected more readily to empirical enquiry.
Neither of the first two approaches involves going inside the classroom to
try to discover what actually happens when teachers ‘intervene’ in the learning
process. This is one reason why they have proved popular. It is much easier
to draw on the work of other researchers who have investigated the nature of
learning in general or who have studied naturalistic acquisition than to enter
the ‘black box’ and begin the messy business of trying to find out how learners
learn a language there. The reluctance to engage in classroom research can be
explained by the natural inclination not to undertake unnecessary work. Why
study die classroom if it can be argued that classroom language learning is like
other kinds of learning or like naturalistic language learning?
One reason why it proved necessary to research the language classroom
was precisely because there was radical disagreement about the theories of
classroom language learning which were derived respectively from a general
learning theory and from comparisons with naturalistic language learning.
This research — like all research — has been of two broad kinds: (1)
exploratory—interpretative research and (2) hypothesis-testing research. The
former makes use of a non-experimental design, collects qualitative data and
provides an interpretative analysis (Grotjahn, 1987). The latter makes use of
an experimental or quasi-experimental design, employs quantitative data and
offers statistical analysis. There are also various mixed forms of research.
There tends to be a certain tension between the advocates of the two
research paradigms. A number of researchers (e.g. Mitchell, 1988a; Van Lier,
1988) argue that an exploratory—interpretative approach is required in order
to unravel the complexities of behaviour in the language classroom. They
argue that in order to understand these behaviours it is necessary to study
them in depth and in context. They emphasize the social nature of classroom
activity and see the study of the interactions that occur between the classroom
participants as the principal object of enquiry. Such an approach, they suggest,
also has the advantage that it presents results in a form that makes them easily
understood by teachers. They are advocates of a ‘research-then-theory’
approach (Reynolds, 1971). That is, although they are interested in the 017-
building, they do not feel the need to base their research on a strong and
explicit theory. Rather they feel that the theory will evolve as understanding
becomes more complete. The principal tool of this branch of research is
ethnography. The emphasis is on describing and understanding classroom
processes rather than on testing what has been learnt.
Other researchers (e.g. Long, 1985a; 1988b) have argued in favour of a
‘theory-then-research’ approach. That is, they believe that ‘true experiments’
should be designed to test hypotheses based on a well-grounded theory.
 Investigating Classroom Language Learning 5
Hypothesis-testing research, it is claimed, is the principal way of understanding
cause—effect relationships. Thus experimental research is necessary to investi¬
gate whether instruction actually results in learning. Long (1985a: 391) writes:
‘The theory-then-research strategy and the related causal-process form of
theory allow for more efficient research. This is because the theory governing
the research at any point in time tells the researcher which the relevant data
are, which is the crucial experiment to run.’ Researchers like Long talk
frequently about explanation. They point out that the qualitative study of
classroom processes may help us to understand how events take place and
what motivates them but they are unable to explain how the events contrib¬
ute to language learning. To do this it is necessary to set up a controlled
experiment. *
The debate over the two research paradigms is an age-old one. It is also a
sterile one. For a start much of the research that has taken place is of a mixed
form. An experimental design does not preclude the collection of qualitative
data and interpretative analysis. Ethnographic research can be combined with
attempts to manipulate the learning environment in specific ways in order to
observe what happens. Second, given the complexity of the task facing the
researcher it would seem wise to employ as many and as varied strategies of
research as are available. Third, it would seem to me that the real issue is not
so much how the research is carried out as what is done with the results of the
research. The danger of much classroom research is to try to apply the results
directly to teaching by advocating specific methods or techniques. This danger
exists quite independently of whether the research is ethnographic or scientific.
Classroom research should be directed at building a theory of language
learning — a goal which both types of research can help to meet. The theory
can then serve as a basis for pedagogical advice. The piecemeal application of
the results of classroom research should be avoided.
Figure 1.1 shows the three ways of building a theory of classroom L2
learning which have now been discussed in broad outline. Two points are

Figure 1.1 Three approaches for developing a theory of classroom second language
learning
6 Instructed Second Language Acquisition
worth emphasizing. The first is that a general distinction can be drawn
between theory-building that proceeds without direct reference to classroom
behaviour — as occurs when extrapolation from a general learning theory or
from naturalistic language learning occurs — and theory-building that results
from attempts to research classroom activities. The second point is that the
theories that have resulted from the different approaches have not always been
coherent or even fully explicit. This is because the teaching-learning re¬
lationship has been viewed predominandy from the perspective of how to
teach rather than from how learning takes place. One of the aims of this book
is to remedy this by adopting a learning perspective.

Extrapolating from General Learning Theory

Now that the three ways of building a theory of classroom language learning
have been outlined, we can begin to examine each one in a little more detail.
We start with the attempts that have been made to extrapolate from general
learning theory. This makes a convenient starting point as this was the initial
approach adopted.5
In the fifties and sixties there was no field of enquiry that could be labelled
‘second language acquisition’. In order to determine the views about classroom
language learning which prevailed during that period it is necessary to examine
what language teaching methodologists had to say. There was no attempt to
develop an explicit theory of classroom language learning; rather views of
learning were invoked to lend support to a set of claims regarding how
language teaching should take place.
Ideas about language teaching during this period were derived in part from
linguistic theory and in part from a general theory of learning. The linguistic
theory was that propounded by structuralist linguists. A language was seen as
a set of formal patterns that could be described rigorously without reference
to meaning. The learning theory was that propounded by behaviourist psycho¬
logists. Learning was treated as a process of habit-formation that could be
described in a terms of stimulus-response associations, often linked together in
complex chains. Methodologists such as Brooks (1960) and Lado (1964)
drew extensively on both structuralist and behaviourist theories in developing
an approach to language teaching that became known as audiolingualism.
According to audiolingual principles the goal of classroom learning was the
acquisition of the habits that comprised the target language. A habit consisted
of the ability to perform a particular linguistic feature (a sound, a word, a
grammatical pattern) automatically, i.e. without having to pay conscious attention
to it. This ability entailed being able to link a particular response to a
particular stimulus. It was acquired through massive practice of a mechanical
nature. The teacher supplied the stimulus, the learner supplied the response.
The teacher then reinforced correct responses or corrected erroneous re-
 Investigating Classroom Language Learning 7
sponses. Ideally, errors were to be avoided because they were believed to have
a negative effect on learning. A major source of errors was the learner’s first
language (LI). This interfered with the acquisition of the L2 because the
learner tended to transfer the habits of her native language into the target
language. These views on classroom language learning — derived entirely
from a dieory of general learning — were articulated with conviction and came
to be accepted as received opinion by a large number of teachers.
It is easy to overemphasize the role played by behaviourist learning theory
in audiolingualism. As Howatt (1984) points out, the progenitors of the
method — Fries and Bloomfield — made little mention of behaviourist psy¬
chology but drew instead on structuralist descriptions of language as a basis
for pattern-prkctice. However, there can be little doubt that behaviourist
theories were used to underpin recommended instructional techniques. There
was, for instance, a clear link between Skinner’s theory of operant conditioning,
which described how complex behaviours can be systematically shaped, and
programmed language learning.
It is important to understand audiolingual learning theory and to subject its
assumptions to careful scrutiny. There are two good reasons why. First, the
theory addressed some key issues — issues which need to be addressed in any
theory of classroom language learning. It dealt with the difference between
explicit knowledge of a L2 (i.e. knowledge about the L2) and implicit
knowledge (i.e. the knowledge that underlies the ability to use the L2) and
made statements about which kind of knowledge should be the goal of
learning. It considered the cause of learner errors and the role they play in
learning. It articulated in some detail the kind of classroom behaviours which
were needed to ensure successful L2 learning. Second, the theory has had a
tremendous impact on teachers’ popular conceptions about how L2 learning
takes place in a classroom. This impact is still evident today, many years after
the theory has been rejected as an adequate account of classroom language
learning. One reason for this, perhaps, is that in audiolingual learning theory
L2 learning is treated like any other learning and, therefore, the task of
teaching a L2 can be seen as essentially the same as that of teaching any other
school subject.
Audiolingual learning theory is not the only classroom language learning
theory to have been based on a general theory of learning — although it has,
perhaps, been the most influential. Currendy considerable attention is being
paid to cognitive theory (J. Anderson, 1980; McLaughlin, 1978b and 1987).6
This exists in direct opposition to behaviourist learning theory as it emphasizes
the role of internal mental processing rather than external behaviour. Cognitive
theory seeks to explain three main aspects of learning: (1) how knowledge is
established, (2) how knowledge becomes automatic and (3) how new knowledge
is integrated into the learner’s existing cognitive system. It draws extensively
on research into information processing. A key distinction is that between
declarative and procedural knowledge. Applied to language learning, the
theory claims that the process by which new linguistic knowledge is internalized
8 Instructed Second Language Acquisition
is different from the process by which control over this knowledge is achieved.
New knowledge is ‘declarative’ (i.e. it involves ‘knowing that’). Automated
knowledge is ‘procedural’ (i.e. it involves ‘knowing how’). Learners typically
progress from declarative to procedural knowledge as they develop control.
Many of the errors that learners produce are not the result of a lack of
declarative knowledge but rather of procedural knowledge. The solution is to
provide conditions of learning that enable them to practise using their knowl¬
edge in authentic communicative situations (Johnson, 1988).
Cognitive learning theory provides a much more convincing account of
classroom language learning than audiolingual learning theory. This is because it
does not seek to explain L2 learning solely in terms of observable behaviours
but gives full recognition to the contribution of the learner’s internal mental
processing. The theory is particularly helpful in enabling us to understand
what learners need to do in order to obtain full control over L2 knowledge.
But, as we shall see, it is unable to account satisfactorily for a number of
aspects of classroom language learning — in particular for the fact that there
are remarkable regularities in the sequence in which L2 knowledge is acquired.
Extrapolation from general theories of learning is inevitable and desirable.
Learning an L2 in the classroom must share a number of characteristics with
the learning of other kinds of knowledge. After all, the conditions that prevail
in a language lesson are not so very different from those that prevail in the
history or science lesson. The fact that an L2 is acquirable in a classroom
context suggests that at least some of the processes involved must be the same
as in other kinds of learning. But there are dangers in extrapolating from a
general theory. It is one thing to claim that classroom language learning is like
other forms of learning, entirely another to assume that it is the same. There is
now ample evidence to suggest that in some respects at least classroom
language learning is special. Extrapolators are not likely to bother to go inside
the classroom to test their hypotheses and, sadly, they are likely to ignore the
evidence of those researchers that have done so.
Extrapolation from behaviourist learning theory is considered in chapter 2.
Cognitive learning theory is examined in chapter 7.

Extrapolating from Naturalistic Language Learning

In the sixties behaviourist theories of learning began to buckle under the

weight of Chomsky’s (1959) attack on the general principles of associationist
psychology. Mentalist theories of language learning began to assume ascend¬
ancy. These emphasized the importance of innate knowledge and the learner’s
contribution to the process of language learning. The learner was credited
with a mental grammar that comprised her competence and which underlay
her actual language behaviour. At the same time a number of important
studies of LI acquisition were undertaken (e.g. Bloom, 1970; Brown, 1973).
 Investigating Classroom Language Learning 9

The rejection of behaviourism on theoretical grounds and on the basis of

empirical studies had a considerable impact on applied linguists interested in
language teaching. Articles applying the results of LI acquisition research and
theory to the classroom began to appear (e.g. Corder, 1967; Cook, 1969).
In the late sixties teachers-tumed-researchers began to take an interest in
L2 learning. This period saw the beginnings of Second Language Acquisition
(SLA) as an area of empirical enquiry. There was sudden burst of research.
Many of the early studies employed the methodology of error analysis (e.g.
George, 1972); that is, corpora of learner utterances were inspected and
deviations from target language norms identified, described and explanations
for them sought. Other studies made use of performance analysis (e.g. Dulay
and Burt, 1973); that is, the degree of correct usage of a number of different
grammatical features was compared in order to establish an order of accuracy.
There were also a number of longitudinal case studies of individual L2
learners (e.g. Ravem, 1968; Huang and Hatch, 978) which examined how
grammatical sub-systems such as negatives and interrogatives were acquired.
There were two significant findings of this research. It was shown that
many of the errors produced by learners were developmental in nature. That
is, learners appeared to construct their own rules, which were independent of
both their native language and the target language. This finding was used as
evidence against the claim of audiolingual learning theory that the major
source of learner error was LI interference. The second finding was that
there appeared to be a natural sequence of acquisition for many grammatical
features. Learners from different language backgrounds displayed remarkable
regularity in the rank ordering of grammatical morphemes or in the stages of
development of negatives and interrogatives. It was argued that L2 acquisition
was a process of creative construction not dissimilar to that found in LI
acquisition.7
All this provided a basis for some very different views of classroom language
learning from that advanced by audiolingual learning theory. These views
were cognitivist in nature. Cognitive code learning theory (Chastain, 1971)
drew heavily on the learning theory that underpinned the study of generative
grammar. One of its central tenets was that the perception and awareness of
L2 rules preceded the actual use of these rules (i.e. competence preceded
performance). Considerable importance, therefore, was attached to metalingual
knowledge - knowledge about language. The cognitive code learning theory
was directly opposed to audiolingual learning theory in this respect. Also
opposed were the views of classroom language learning propagated by Newmark
and Reibel (1968). These emphasized the classroom learner’s innate capacity
for language learning. It was argued that classroom language learning - like
naturalistic language learning - would take place effortlessly and automatically
providing there was sufficient exposure to the target language and the learner
was sufficiently motivated. Errors were a natural concomitant of learning and
were to be tolerated rather than corrected. Learning could best be fostered if
there is no attempt to ‘interfere’ with it. Very similar views were propagated by
10 Instructed Second Language Acquisition
applied linguists who drew on the early research into L2 acquisition (e.g.
Corder, 1975).
A familiar feature of articles that appeared in journals such as Language
Learning, TESOL Quarterly and the International Review of Applied Linguistics
was a final section in which theoretical claims or the findings of empirical
research were ‘applied’ to the classroom. This approach found its arch¬
disciple in Stephen Krashen, whose Monitor Model, developed in the middle
seventies and popularized in the late seventies and early eighties, was specially
devised to take account of classroom language learning. The Model itself,
however, was based almost entirely on studies of naturalistic acquisition - LI
and L2. It represented the most complete attempt yet at extrapolation.8
Extrapolation from naturalistic LI and L2 acquisition research is discussed
in chapter 3.

Researching Classroom Language Learning

Although it is not entirely true to say that there was no classroom L2 research
before the seventies, it is certainly true that the decade saw a remarkable
growth of such studies. The research conducted in the sixties and earlier was
typically experimental in nature and did not involve the observation ol actual
classroom behaviour. Scherer and Wertheimer (1964), for instance, sought
to compare the efficacy of two different approaches to language teaching
(grammar—translation vs. audiolingualism) and did so by measuring classroom
learners’ L2 proficiency by means of pre- and post-tests. The ‘treatment’
consisted of instruction based on the two approaches, but no attempt was
made to investigate what actually happened when it took place. Another
frequendy cited comparative method study of the sixties was the Pennsylvania
Project (Smith, 1970). This investigated the relative efficacy of three different
methods: (1) traditional (grammar—translation), (2) inductive (the audiolingual
method) and (3) deductive (the cognitive code method). It employed a similar
pre/post-test design to the Scherer and Wertheimer study.
As Clark (1969) pointed out, there are inherent problems in studies that try
to treat the real-life classroom situation as if it were a psychologist’s laboratory.
One of the main problems is that there can be no certainty that the instructions
given to teachers are actually carried out by them. In other words there is no
guarantee that the ‘treatments’ really are different. Methods that are distinct
on paper may not be so in practice. Some effort was made in the Pennsylvania
Project to carry out classroom observation in order to establish whether there
were real differences in instruction, but this was of limited validity as the
observation schedules employed were flawed in a number of respects.
Neither the Scherer and Wertheimer not the Smith study was able to
demonstrate that one teaching method was significantly more effective in
promoting L2 learning than another. One of the effects of these disappointing
 Investigating Classroom Language Learning 11
results was that researchers began to question whether large-scale comparative
method studies were the right way to go about investigating teaching—learning.
Allwright (1988a: ch. 1) describes in detail how the rejection of‘method’ as an
appropriate goal of enquiry set the scene for the detailed, small-scale obser¬
vational studies of classroom behaviour that began to appear in the seventies.
The empirical studies of L2 classrooms which have taken place in the
seventies and eighties and which provide a basis for building a theory of
classroom L2 learning can be grouped under two headings: (1) classroom
process research, (2) the study of formal instruction and L2 acquisition. A
brief account of each category follows.

Classroom Process Research

Classroom process research has consisted typically of small-scale studies
aimed at documenting the events that take place in L2 classrooms. A variety
of methods have been employed — the use of observational schedules for
interaction analysis, educational ethnography, the analysis of classroom dis¬
course, case studies, interviews, action research (i.e. attempts to investigate
teaching—learning while carrying out normal teaching) and introspection
(e.g. asking learners to comment on how they set about performing different
classroom tasks). Van Lier (1988: ch. 3) provides a useful survey of these
methods. The general aim of classroom process research is to describe
classroom behaviour in detail in order to build up an accurate record of what
actually happens. Careful description serves as a basis for understanding and
explaining what happens:

What we are doing when we examine an utterance in context is basically

of course explaining the occurrence of that utterance as the same time as
describing. We therefore engage in explanation, and in that sense a dis¬
tinction between descriptive and explanatory research is simplistic; it is
not possible to do one without the other. (Van Lier, 1988: 10—11)

As this quotation makes clear, classroom process research is usually

(although not exclusively) sociological in orientation. The ‘explanation’ it
provides is of a social nature. That is, it does not provide an explanation
of how L2 learning takes place in the mind of the learner, although it can
provide some helpful clues about this.
Classroom process research has examined a number of different aspects
of the teacher’s and the learner’s language, as well as exploring the
nature of interaction between the participants. Some of the issues dealt
with are:

1 Error treatment, i.e. how the teacher deals with learner errors.
2 Teacher talk, i.e. the formal and functional characteristics of the ways
in which teachers talk to L2 learners.
12 Instructed Second Language Acquisition
3 Learners’ language, e.g. the communication strategies which learners
use to overcome communication problems in the classroom and the use
of code-switching.
4 The differences between pedagogic and natural discourse.
5 The different types of classroom discourse.

These issues are considered in some depth in chapter 4.

The Study of Classroom Interaction and L2 Acquisition

Classroom process research has helped us to understand what happens in
teaching—learning. It provides us with a clearer picture about the way in
which teachers and learners go about their business. It has also afforded a
number of insights about the relationship between overt classroom behaviours
and language learning and produced some valuable speculation on this re¬
lationship. The research has tended to be piecemeal, however, and has not
been informed by an explicit theory of classroom L2 learning. This comment
is not intended as a criticism of classroom process research. Indeed, the
strength of this kind of research rests in the detailed attention paid to specific
aspects of classroom activity. There is danger, however, that investigators
working in this tradition will be seduced by the attractiveness of manageable
research questions that can be answered by means of easily collected classroom
data but which add little to our understanding of how learning takes place. As
a complement to classroom process research, therefore, there is a need for
theory-led research, which addresses more directly the nature of the relation¬
ship between classroom interaction and L2 learning.
A number of applied linguists have argued for viewing teaching—learning as
‘interaction’. Allwright (1984a: 157), for instance, argues that interaction
should be viewed not just as an aspect of communicative language teaching
but as ‘the fundamental fact of classroom pedagogy’. According to Allwright it
is through the joint management of interaction by the teacher and the learners
that learning takes place. ‘Interaction’ in this sense refers not just to those
exchanges involving authentic communication but to every oral exchange that
occurs in the classroom, including those that arise in the course of formal
drilling.
A number of different hypotheses about the relationship between interaction
and learning have attracted researchers’ attention. One that has motivated a
number of recent studies is the interactional hypothesis. This states that L2
acquisition occurs most efficiendy when learners have plentiful opportunities
to negotiate meaning whenever there is some kind of communication difficulty.
Such negotiation, it is claimed, brings learners into contact with L2 data
which they are likely to attend to and so incorporate into their L2 mental
grammars. There are, however, considerable problems involved in testing this
hypothesis empirically, not least that of determining what is actually learnt as a
result of engaging in an interactional exchange where there is opportunity to
 Investigating Classroom Language Learning 13
negotiate meaning. The causal relationship between meaning-negotiation and
acquisition has not been conclusively demonstrated.9 However, the theory has
served as a basis for conducting a number of classroom studies aimed at
answering such questions as ‘What kinds of pedagogic task provide the best
opportunities for negotiation?’ and ‘Does small-group work provide greater
opportunity for the negotiation of meaning than teacher—class interaction?
These studies are, of course, only as good as the theory that motivates them.
It is probably true to say that we still know very little about the relationship
between interaction and learning. Chapter 5 reviews a number of hypotheses
(including the interactional hypothesis) which address the relationship, but the
evidence to support each is often indirect and meagre.
4

The Study of Formal Instruction and L2 Acquisition

A different branch of classroom research has examined the effect that formal
instruction has on L2 acquisition. ‘Formal instruction’ refers to the attempt to
teach some specific feature of the L2 code — usually a grammatical feature —
in one way or another. The studies belonging to this type of research fall into
two categories, depending on whether they examine the effect of formal
instruction on the rate/success of L2 learning or on the sequence/process of
acquisition.
Recent surveys of research which fall into the first category (Long, 1983b,
1988a; Ellis, 1985a) indicate that learners who receive formal instruction
generally outperform those who do not. The methodological basis of this
research is a comparison between classroom and naturalistic learners. The
assumption is made that if it can be shown that classroom learners learn more
rapidly and/or achieve higher levels of ultimate success than naturalistic
learners, then this must be the result of the essential difference in the two
learning environments — the focus on form that occurs in the classroom.
Studies which have investigated the effect of formal instruction on the
sequence/process of L2 acquisition have produced mixed results. Some re¬
searchers have compared the acquisitional sequences of classroom learners
with those of naturalistic learners (e.g. Lightbown, 1983; Pica, 1983; Ellis,
1984a). In general the sequences appear to be very similar, but some interesting
differences have also been observed. Other researchers have examined the
effect of formal instruction more directly by carrying out classroom experiments
(e.g. Pienemann, 1984; Eckman et al., 1988). That is, attempts are made to
control the learning of specific grammatical features through formal instruction.
The results show that some grammatical structures do not appear to be
teachable unless the learner is developmentally ‘ready’. Other grammatical
structures, however, appear more amenable to instruction. There is also some
evidence to suggest that formal instruction directed at one linguistic feature
can not only result in the learning of that structure but also trigger the
acquisition of other ‘implicated’ structures.
Much of this research has been motivated by theoretical questions concerning
14 Instructed Second Language Acquisition
SLA in general. The classroom serves as a convenient arena in which to carry
out controlled experiments. Nevertheless, the research has helped us to form
a clearer idea about the constraints that govern the acquisition of new linguistic
forms and whether formal language teaching enables the learner to ‘beat’ the
natural sequence. One of the strengths of this branch of classroom research is
that it has been able to examine the relationship between teaching and
learning directly because it utilizes measures of both.10 It is examined in detail
in chapter 6.

Summary
The empirical study of L2 classrooms came about largely as a result of
dissatisfaction with global method comparisons. The different categories of
research are summarized in table 1.1. It is not the purpose of this book to
advance the claims of one approach over the others. No one approach is
capable of providing the data needed to build a complete theory of L2
learning. The three approaches together, however, provide a picture of how
learning takes place in the classroom.

Theoretical Issues

In this section we will consider a central distinction that any theory of class¬
room L2 learning will need to address: the role of form-focused and meaning-
focused instruction. A number of questions will be raised but no attempt will
be made to answer them at this juncture — that must wait until chapter 7.

The Role of Form-focused Instruction

We have already seen that classroom language learning can be distinguished
from naturalistic language learning on the grounds that it involves planned
attempts to intervene in the learning process. One of the key questions,
therefore, is what form this planning should take. Broadly speaking, there are
three options. The intervention can involve the provision of form-focused
instruction where the learners are encouraged to focus their attention on
specific properties of the linguistic code. Alternatively, intervention can take
the form of specially contrived meaning-focused activities designed to promote
authentic communication in the classroom. The third option consists of some
kind of combination of form and meaning-focused instruction.
The choice of the form-focused option involves an acceptance of one of two
assumptions. The strong assumption is that attention to the code is necessary
for L2 learning in a classroom context. The weaker assumption is that it is not
necessary but is desirable as an aid to learning. We will examine both
assumptions in the following chapters.
 Investigating Classroom Language Learning 15
Table 1.1 Empirical research of L2 classrooms

Category Goal Principal research methods

Classroom The understanding of The detailed, ethnographic

process how the ‘social events’ of observation of classroom
research the language classroom behaviours
are enacted
The study of To test a number of Controlled experimental
classroom hypotheses relating to how studies; ethnographic
interaction interacting in the studies of interaction
and L2 classroom contributes to
acquisition L2 acquisition and to
4 explore which types of
interaction best facilitate
acquisition
t he study of To discover whether Linguistic comparisons of
instruction formal instruction results L2 acquisition by classroom
and L2 in the acquisition of new and naturalistic learners;
acquisition L2 knowledge and the experimental studies of the
constraints that govern effects of formal instruction
whether formal instruction
is successful

Both the strong and the weaker assumption imply that it is possible to
influence the speed and/or course of L2 learning by directing the learner’s
attention to the formal properties of the code. In other words, it is accepted
that learners are able to learn what they are taught. It is precisely this
acceptance which has been challenged in recent years, notably in the publi¬
cations of Stephen Krashen (1981; 1982; 1985). Krashen has argued that
grammar-teaching is powerless to alter the natural route of L2 acquisition and
that learners should be left to follow their own internal syllabus. This view has
been strongly criticized by a number of applied linguists (McLaughlin, 1978b;
Sharwood-Smith, 1981; Ellis, 1984a).
If it is accepted that attention to the code contributes to L2 acquisition, a
further question arises. How should attention to the code be organized? The
answer to this question depends on whether it is believed that L2 learning
involves the incremental mastery of discrete items. Rutherford (1987) refers to
this as the ‘accumulated entities’ view of language learning. Such a view
underpins both audiolingual and cognitive code learning theories. Imparting
the necessary information about the items that comprise the code can be
attempted inductively (as in audiolingualism) or deductively (as in the cognitive
code method). That is, the instruction may simply provide the learners with
plentiful opportunities to produce utterances containing the target item or it
can provide explicit information about the properties of the item. An alternative
to the ‘accumulated entities’ view of learning is consciousness-raising. This
differs from traditional grammar-teaching in that it sees form-focused in¬
struction as a means to the attainment of grammatical competence not as an
attempt to instil it. Consciousness-raising aims to facilitate acquisition, not to
16 Instructed Second Language Acquisition
bring it about directly. It recognizes that the learner will contribute to and
shape the process of acquisition herself.
Form-focused instruction can be considered from the point of view of how
the input to the learner is planned (i.e. syllabus-design and lesson-planning)
and also from the point of view of the processes that occur in the course of
teaching (i.e. classroom methodology). The choice of syllabus type is based on
a particular view of classroom language learning — be this overt or covert.
Methodological choices are similarly based. For instance, the teacher who
believes that it is necessary to correct learners’ errors does so because she
believes that this will contribute, in one way or another, to learning. The
construction of a theory of language learning that addresses the role of formal
instruction is, therefore, of importance not just for curriculum planning but
also for curriculum implementation.

The Role of Meaning-focused Instruction

Meaning-focused instruction consists of the provision of classroom activities
that encourage learners to communicate using whatever resources, linguistic
and non-linguistic, they have at their disposal. Two pedagogical arguments
have been advanced in its favour (Ellis, 1986). The first is that learners need
the opportunity to communicate in order to develop fluency. Through trying
to communicate they develop the strategic competence they need to deal with
communication problems and at the same time they automatize their existing
L2 knowledge. The second argument is that learners are able to acquire new
L2 knowledge as a result of taking part in communication. They ‘pick up’
knowledge from the input they are exposed to through interaction.
One the issues addressed by classroom process research is the extent to
which classrooms are able to afford opportunities for authentic communication
(e.g. Riley, 1977; Edmondson, 1985). The study of the teacher’s and the
learner’s language and of the discourse they jointly construct have shown that
the communication that takes place in the classroom is usually very different
that which takes place outside. The asymmetrical role relationships between
teacher and learners together with an educational ideology that views the
process of teaching—learning as one of transmission result in a distorted form
of communication. A number of questions arise. One concerns whether the
distortion has a negative effect on L2 learning. Another concerns how the
distortion can be overcome (e.g. through talk in small-group work).
The key question regarding the role of meaning-focused instruction concerns
how interaction contributes to the acquisition of new linguistic knowledge.
According to one view comprehensible input is necessary. Krashen (1985) has
emphasized the importance of the teacher simplifying input to ensure that
learners are able to understand. Long (1983a) has argued the importance of
negotiation of meaning when there is a communication difficulty. This helps
to make input which contains new linguistic material comprehensible and so
facilitates its acquisition. According to another view learner output is important.
 Investigating Classroom Language Learning 17
Swain (1985), for instance, argues that learners need the opportunity for
‘pushed output’ (i.e. output that stretches their linguistic capacity) in order to
avoid stopping learning some way short of native-speaker competence. A
theory of classroom language learning needs to consider the role of meaning-
focused instruction. It needs to explain the respective contributions of form-
focused and meaning-focused instruction and to provide a principled basis for
how the two can be combined.

Conclusion
4

The goal of this book is the construction of a comprehensive theory of

classroom L2 learning. This chapter has had three main purposes: to define
the term ‘classroom language learning’, to identify and describe the types of
information to be drawn on in the process of building a theory and to outline
some of the major issues which the theory will have to address.
The study of classroom language learning can proceed in different ways. It
can draw on the theories and methods of different disciplines — linguistics,
sociology, psychology and education. This book does not aim to promote the
claims of any one discipline over those of the others. If it gives greater
attention to psycholinguistic enquiry, this is not because this is considered the
‘right’ way to go about building a theory but because it reflects the personal
interests of the author. Nor is it intended to suggest that a psycholinguistic
theory is the ‘best’ kind of theory upon which to base pedagogic practice. The
relationship between research and theory on the one hand and pedagogy on
the other, however, does need careful consideration. It will be examined in
chapter 8.

NOTES

1 The distinction between formal training and apprenticeship in language teaching

is mirrored in Stem’s (1981) pedagogic distinction between an ‘L’ approach
(characterized as linguistic and analytical) and a ‘P’ approach (characterized as
psychological and experiential). Howatt (1984) distinguishes a ‘rational’ and a
‘natural’ approach.
2 Klein (1986) distinguishes classroom and naturalistic acquisition in a similar way.
He describes the former as ‘spontaneous’ and the latter as ‘guided’. ‘Guidance’ is
evident in both the choice of what to teach (syllabus) and how to teach it
(methodology).
3 Chomsky continues to maintain that language constitutes a separate faculty which
cannot be explained with reference to a general cognitive system (cf. Chomsky,
1986). He rejects, therefore, explanations of language learning within a Piagetian
framework.
4 The fifties and sixties were not devoid of empirical research, of course. In
18 Instructed Second Language Acquisition
particular, there were numerous studies of language aptitude (e.g. Pimsleur, 1960;
Carroll, 1963). However, these were experimental in nature and extrinsic to the
classroom. No attempt was made to enter the classroom and observe what
happened there.
5 I am, of course, ignoring the history of language teaching prior to the advent of
audiolingualism. My reason for this is that before andiolingualism there was no
explicit attempt to justify pedagogic techniques by reference to the way a L2 was
learnt. Grammar—translation, for instance, made no mention of how languages
are learned. The learning theory that underlay the oral approach was ill-defined,
to say the least (cf. Richards and Rodgers, 1986: 36—7).
6 There is, in fact, no one ‘cognitive theory’. Rather there are a number of separate
theories which are broadly similar in that they recognize similar distinctions —
such as that between declarative and procedural knowledge — and draw extensively
on research into information processing.
7 It is important to note, however, that the idea of a ‘natural order hypothesis’, in
the L2 acquisition of grammatical structures is a controversial one. Not all studies
lend support to this. The hypothesis and the research are considered in chapter 3.
8 Krashen’s later work (e.g. Krashen, 1985) draws on studies of both naturalistic
and classroom learning. The Monitor Model, however, was constructed on evidence
supplied predominantly from research into naturalistic acquisition.
9 Long (1985a) argues that it is not necessary to demonstrate a direct relationship
between interaction and the acquisition of specific linguistic features. He suggests
that the link can be demonstrated indirectly by showing that the negotiation of
meaning promotes comprehension which in turn promotes acquisition. This pro¬
posal is discussed in detail in ch. 5.
10 In fact, though, many of the studies of formal instruction are not really classroom
research in the strict sense of this term, i.e. they did not involve going inside the
classroom. Comparative studies of naturalistic and classroom learning, for example,
have been conducted by collecting data outside the classroom.
2 Behaviourist
Learning Theory and
Classroom Language
Learning

Introduction

The dominant language-teaching method of the fifties and sixties was audio-
lingualism. This method originated from the ‘scientific’ descriptions of language
provided by American linguists such as Bloomfield and Fries, both of whom
belonged to the structuralist school, which sought to identify and describe the
formal patterns of a language in an explicit and rigorous manner. Subsequently,
language-teaching methodologists^such as Brooks and Lado drew extensively
on behaviourist psychology as a means of justifying the prescriptions of the
method they advocated. Thus, audiolingualism drew on linguistics in the first
place, extrapolating from the theoretical claims of general psychology only at a
later date.1 This chapter begins with a brief account of the emergence of
audiolingualism.
Several behaviourist accounts of learning were drawn on as a warrant for
audiolingual principles of teaching. Methodologists drew liberally and eclec¬
tically on available theories to justify their own views about teaching. However,
it is probably true to say that the views on learning advanced by Watson and
more especially Skinner constituted the primary source for extrapolation. No
attempt will be made here to describe in detail particular behaviourist theories
of learning. This is not necessary, as the main purpose of this chapter is to
identify the views of L2 learning that underpinned audiolingual teaching. The
chapter, therefore, will examine the assumptions about L2 learning found in
audiolingualism, borrowing from and extending a list which Rivers (1964)
drew up. Where some background theoretical knowledge is needed in order to
make sense of a particular assumption this will be provided.
The assumptions of audiolingualism were subjected to a critical evaluation
in the early sixties on the grounds that language learning is a process to which
the learner actively contributes. The central idea that learning could be
directed from the outside by manipulating the behaviour of the learner was
rejected by nativists, who argued that learners themselves control the course
of language learning. This chapter examines the objections which have been
levelled against audiolingual learning theory. In so doing it stands back from
20 Instructed Second Language Acquisition
the preoccupations of the day to ask whether audiolingual learning theory
offers any insights of value for an understanding of classroom L2 learning.

The Emergence of Audiolingualism

Audiolingualism was a method of teaching that grew out of the structural

approach developed by a number of American linguists — notably Fries. The
method originated in the attempts to provide training to army personnel
during the Second World War. It was designed to develop oral fluency in a L2
in nine months. An outline of the approach was later described by Fries
(1948) in an article written for the first issue of Language Learning. First,
linguists drew up descriptions of the patterns of the L2, using native speakers
of the language as informants. Second, applied linguists carried out contrastive
descriptions of the learners’ LI and the target language. These served as a
basis for the selection and grading of the patterns which were to be taught.
The applied linguists then prepared teaching materials to practise the patterns.
Finally, the materials were implemented. Howatt (1984:267) noted that the
structural approach lacked a set of clearly defined methodological principles;
it was based on ‘the cult of materials’ derived from ‘sound linguistic principles.’
The structural approach served as the basis for the Foreign Language in
the Elementary School (FLES) programmes that began throughout the United
States in the fifties. The new techniques held out the promise that children
could quickly and easily master a L2. Whereas the grammar—translation
method, which the audiolingual method replaced, had emphasized intellectual
and literary study, the new approach rested on techniques of mimicry and
memorization (‘mim—mem’) designed to develop the ability for oral com¬
munication. As such, it was seen as appropriate for teaching young children
and of greater practical relevance to society.2 McLaughlin (1985) reports that
by 1960 there were over a million children in the FLES programmes.
The late fifties and early sixties in the States saw a number of books on
applied linguistics and language-teaching methodology, stimulated by the
rebirth of interest in foreign language teaching in American education. In
1957 Lado published Language across Cultures. This provided a lucid account
of the principles and methods involved in contrastive descriptions of two
languages. In 1960 Language and Language Learning by Brooks was published,
followed in 1964 by Lado’s Language Teaching. In 1961 the staff of the
Modern Language Materials Development Center published teaching manuals
for the major foreign languages. A feature of all these - and other -
publications was reference to the principles of learning propounded by be¬
haviourist psychologists such as Watson and Skinner. It is perhaps significant
that, as attention was given to actual classroom practice, as opposed to the
specification and ordering of teaching content, reference was made to
psychological principles.
 Behaviourist Theory and Classroom Learning 21

Audiolingualism was very much an American method. In its purist form it

was never very popular in Britain and Europe, where less attention was paid to
teaching the formal patterns of the L2 and more to the situational uses of the
L2.3 In general, approaches to language teaching in Europe have treated
theories of learning provided by psychologists or second language acquisition
researchers with suspicion. Thus Catford (1959) in a review of the teaching
of English as a foreign language in Britain paid lip-service to the need for
language teaching to seek guidance from both linguistics and psychology but,
in fact, made scarcely any reference to learning. However, many of the
audiolingual assumptions regarding the way language is learnt can be found in
the pedagogical prescriptions of British and European methodologists writing
at this time (cf.*early issues of English Language Teaching).

Assumptions about L2 Learning in the Audiolingual

Approach
In this section the term foreign language (FL) learning will be used throughout
as audiolingual methodologists held that foreign or classroom language learning
was different in kind from naturalistic or second language learning and were
concerned only with the former.

Assumption 1: Foreign language is the same as any other kind learning

and can be explained by the same laws and principles This assumption
was basic to all the other assumptions. Skinner (1957:10) made his position on
this point quite clear: ‘We have no reason to assume ... that verbal behavior
differs in any fundamental respect from non-verbal behavior, or that any new
principles must be invoked to account for it.’ This claim was the basis for
extrapolating from a general theory of learning. Skinner, in fact, went further
and argued that laboratory experiments designed to investigate how animals
such as rats could be trained to perform specific sequences of behaviour
served as a basis for making claims about the way language learning took
place.

Assumption 2: All learning is the result of experience and is evident in

changes in behaviour All bevahiourist theories were based on observable
behaviour. For Watson (1924) there was no such thing as ‘mind’ or ‘con¬
sciousness’. The sole criterion he used for determining whether something
existed was whether it could be observed. Other behaviourists — such as
Skinner — accepted the existence of mind, but rejected any explanation of
human behaviour in terms of feelings or mental processes.
Learning, then, consisted of changes in behaviour that were brought about
through experience. It followed that learning could be effected by manipulating
the environment to provide the required experiences. This is exactly what
Skinner attempted to do in his animal experiments. It is also what audiolingual
techniques set out to do with the language learner. One of the great attractions
22 Instructed Second Language Acquisition
of audiolingualism was the conviction that FL learning could be manipulated
and induced by ensuring that the learners produced the appropriate behaviours.4
If learning was behaviour it could be taught by inducing the correct behaviour
(Politzer, 1961:2).
One of the results of treating learning as behaviour was that ‘meaning’ was
excluded from consideration, as Politzer (1961:19) indicates: ‘The other
possibility . . . that has been proved feasible ... is to eliminate meaning
almost totally from the initial phase of language instruction. It is entirely
possible to teach the major patterns of a foreign language without letting the
student know what he is saying.’ The exclusion of meaning was considered
particularly important in the early stages of learning.

Assumption 3: Foreign language learning is different from first language

learning According to behaviourist theories of learning, the mechanism by
which all language learning took place was the same. All language was ‘verbal
behaviour’, so it followed that both LI and foreign language (FL) learning
consisted of developing the correct behavioural responses.
However, language teaching methodologists were often at pains to point out
the differences between LI learning and FL learning. They were well aware
that FL learning posed difficulties which the child learning an LI did not
face. As Mackey (1965:107) put it: ‘The learning of one language in childhood is
an inevitable process; the learning of a second language is a special ac¬
complishment.’ Brooks (1960) also emphasizes the differences. Despite an
otherwise strong adherence to behaviourist orthodoxy, he speaks mentalistically
of the ‘vital force’ which characterizes the newborn baby and which enables it
to ‘break the code’ of the LI.5
However, despite the recognition that the two types of learning differed, the
relevancy of a general theory of learning to L2 learning was not seriously
questioned. Comparisons between LI and FL learning were undertaken with
two purposes in mind. The first was to emphasize the relative ease of LI
learning and to make out a case for commencing FL instruction as early as
possible in order to tap the natural language-learning capacity of the young
child. Brooks recommended starting at eight years old. Arguments such as
this were used to justify setting up FLES programmes. The second purpose
was to dispute the view that the classroom should try to imitate ‘real life’. LI
learning took place successfully in ‘real life’, but FL learning required setting
up the ‘optimal conditions’ for learning in order to overcome the various
problems.

Assumption 4: Foreign language learning is a mechanical process of

habit formation This assumption lay at the centre of audiolingualism. Brooks
(1960:49) wrote: ‘The single paramount fact about language learning is that
it concerns, not problem solving, but the formation and performance of
habits.’ Methodologists recognized two different ways in which habits could
be formed. According to the theory of classical conditioning, a stimulus
 Behaviourist Theory and Classroom Learning 23
present in the environment comes to be closely associated with a particular
response such that whenever the stimulus occurs the response is automatically
produced. Instrumental learning, as described by Skinner (1957), worked
somewhat differently. Responses were emitted rather than elicited. Operants
were produced randomly in the first place as a result of some ill-determined
inner drive. Selected responses, however, were then subjected to reinforcement.
Through the process of rewarding responses produced on a trial-and-error
basis new habits were formed.6
Both theories of habit formation are evident in the writings on audio-
lingualism. Classical conditioning fitted in nicely with the structuralists’ view
of language as a set of patterns. Each pattern, once identified, could be
practised through stimulatus—response drills until it became a habit. Be¬
haviourist psychologists, such as Thorndike, had carried out experiments to
determine the optimal conditions for habit-formation. Lado (1964) and others
refer to a number of ‘laws of learning’ which had been derived from these
experiments. For example, the law of effect stated that if an act was followed
by a satisfying state of affairs, the probability of its occurrence in a similar
situation increased. In the case of the pattern-drills the ‘satisfying state of
affairs’ was brought about by the learner knowing that he had produced the
correct response. Another key law was the law of intensity, according to
which the more intensely a response was practised, the better it was leamt and
the longer it was remembered. The audiolingual emphasis on repetition,
imitation and simple substitution on the part of the learner and systematic
reinforcement on the part of the teacher was intended to reflect these laws of
learning.
For learning to be effective habits had to become automatic. This is what
Bloomfield (1942) had in mind when he claimed: ‘Language learning is over¬
learning. Anything else is of no use.’ Only if the patterns of the L2 had been
‘over-leamt’ was it possible for the learner to produce them correcdy in real
communication. Stimulus—response associations had to be leamt so that they
could be accessed when the attention was on content and this required
developing ‘bundles of habits below the threshold of awareness’ (Lado,
1964:34).
Skinner’s account of instrumental learning was used to provide a principled
explanation of how learning could be shaped by dividing the learning target
into a series of minimal steps and arranging these in a carefully ordered
sequence. The idea here was to effect efficient learning by making sure that
each step in the learning process lay within the capabilities of the learner. In
language teaching this was reflected in designing drills that exposed learners
to minimal differences between patterns and, especially, in programmed
instruction designed to lead the learner towards some predetermined criterion
in an optimal fashion.
Audiolingual methodologists rejected ‘total language experiences’ such as
those provided by normal communication on the grounds that these were
beyond the capacity of the learner. Forcing the learner to adopt ‘a normal
24 Instructed Second Language Acguisitlon
language set’ would result in imperfect experiences and, therefore, faulty or
non-learning. Audiolingual teaching, therefore, was based on providing ‘partial
experiences’. Success in a partial experience provided a platform for more
complicated behaviour. Thus Skinnerian principles of operant conditioning
lay behind the audiolingual conviction that learners should not engage in
meaningful communication before they were ready.

Assumption 5: Language learning proceeds by means of analogy rather

than analysis The grammar—translation method emphasized the value of
deductive learning through formal explanations of L2 rules. In contrast, the
audiolingual approach emphasized inductive learning through pattern-practice.
This contrast is often referred to as that of analysis vs. analogy (cf. Brooks,
1960). Analysis involved ‘problem solving’7 rather than ‘habit-formation’.
Metalingual knowledge was no guarantee of the ability to use a grammatical
feature correctly. Indeed, metalingual knowledge was worse than useless as it
was a source of potential interference in the process of developing FL habits:
‘The learner who has been made to see only how language works has not
learned any language; on the contrary, he has learned something he will have
to forget before he can make any progress in that area of language’ (Brooks,
1960:49). Thus, as Valdman (1966a) puts it, the use of grammatical rules as
‘predictors’ of linguistic behaviour was outlawed.
Two uses of analysis were allowed, however. First, Brooks thought that
giving an analytical explanation could help to ‘clear the track’ with learners
who were used to deductive teaching. Brooks emphasized, however, that
‘clearing the track’ was not the same as advancing along the track. Second,
Politzer (1961) suggested that rules could be provided as ‘summaries of
behaviour’. Politzer seems to accept that learners must undergo some kind
of conscious realization of what a pattern consisted of during pattern practice
and that this could be consolidated by providing a grammatical explanation.
The central process in forming habits was analogy. This involved dis¬
crimination and generalization. The learner needed to identify the underlying
structure of a pattern by perceiving its similarities and differences with other
patterns. Also the learner was expected to reproduce a pattern in similar but
not identical situations, using different vocabulary. Substitution and cue drills
were intended to foster both discrimination and generalization.

Assumption 6: Errors are the result of LI interference and are to be

avoided or corrected if they do occur The patterns of the learners’ LI
and the FL would be the same in some cases and different in others. Where
they were the same, it was assumed that the learning of the FL would be
faciliated because all the learner had to do was to transfer LI habits. Flowever,
where they were different, learning difficulties arose as a result of proactive
inhibition - the inhibition of new habits by previous learning. The learner’s
LI interfered with the acquisition of new, FL habits. As a result errors
appeared in the learner’s responses which were directly traceable to the LI.
 Behaviourist Theory and Classroom Learning 25
Audiolingualists recognized other sources of error (e.g. random responses or
overgeneralization of a pattern resulting from incomplete learning’8) but con¬
sidered LI interference by far the most serious.
The goal of language teaching was to develop the FL as a coordinate
system, independent of the learner’s LI rather than a compound system, in
which the mother tongue accompanied and dominated attempted behaviour in
the L2. To achieve a coordinate system it was necessary to prevent learners
from speaking their LI in class — although they might occasionally be allowed
to hear it. More important still was the need to prevent LI interference
occurring. The avoidance of error was one of the central precepts of
audiolingualism. Statements like the following from Brooks (1960:58) were
common in the literature: ‘Like sin, error is to be avoided and its influence
overcome, but its presence to be expected.’ Error prevention necessitated
massive pattern-practice by means of mechanical drills which had a low
probability of error and which could ensure use of the L2.
Errors would be prevented more easily if they could be predicted. To this
end, the contrastive analysis hypothesis was formulated. In its strong form
(cf. Wardhaugh, 1970) this stated that all L2 errors could be predicted by
identifying the differences between the target language and the learner’s LI.
On the grounds that linguistic difference constituted learning difficulty a
number of contrastive studies between English and the major European
languages were carried out in the United States at the newly formed Center
for Applied Linguistics (e.g. Stockwell and Bowen, 1965; Stockwell, Bowen
and Martin, 1965). These studies were designed to identify the patterns
which were most likely to cause difficulty so that materials writers could give
precedence to them.
Despite all the preventive measures some errors were bound to occur. They
had to be dealt with. Students were not to be allowed to discover and correct
their own mistakes (Modem Language Materials Development Center,
1961:28). The recommended method was immediate correction by the
teacher followed by further opportunity to produce the correct response. Such
an approach to error treatment was compatible with the central tenet of
operant conditioning, namely that correct responses received positive
reinforcement and negative responses negative reinforcement.

An Evaluation of Audiolingual Learning Theory

These six assumptions constitute the theory of FL learning that underpinned

audiolingualism. Of course, the learning theory was not monolithic. Politzer
(1965:14), for instance, was not convinced that learning was simply a matter of
method: ‘It is not the method that performs the trick, it is the learner.’
Valdman (1966a) queried the general assumption that learning was over¬
learning, referring to such a view as a ‘simplistic model’. Even Lado (1964),
26 Instructed Second Language Acquisition
one of the main proponents of audiolingualism, was cautious in the claims he
made for his theory of FL learning. ‘The theory is in the nature of a proposed
explanation whose status is still partly conjectural’ (p. 39). Nevertheless, the
six assumptions reflected the confidence placed in behaviourist accounts of
learning. They also reflected a preparedness to apply theory to practice or,
more accurately perhaps, to justify practice with reference to theory.
Attacks on audiolingual learning theory came from two directions. First,
applied linguists such as Rivers (1964) and Carroll (1966:101) took a close
look at some of the principal assumptions from the perspective of develop¬
ments in general learning theory. Carroll, in particular, was sceptical of the
kinds of claims about FL learning advanced by Brooks and Lado: ‘we do not
yet have either a good general theory concerning the conditions under which
learning takes place or a general theory of language behavior that would
enable us to select optimal components of a foreign language teaching system
for any given case.’ He argued that the audiolingual habit theory was not
linked to any contemporary theory of learning and had, in fact, only a ‘vague
resemblance’ to an early version of association theory. He also noted the
absence of any relevant research that could lend support to the theory. Rivers
adopted a similar approach. She drew on the neo-behaviourist theories of
Osgood and Mowrer which gave recognition to the mediational and emotional
processes inside the learner. Her main point was that any approach to teaching
which failed to recognize the contribution of the learner was inadequate. The
basis of the first kind of attack on audiolingual learning theory, then, was to
argue that the theory was no longer abreast of recent developments.
The second source of attack came from the paradigm shift in the psychology of
language brought about by Chomsky’s (1959) attack on Skinner’s Verbal
Behavior. Chomsky challenged the very basis of behaviourism by arguing that
language was not a set of habits but a set of abstract rules and that language
was not acquired through imitation and repetition but rather was an innate
inheritance triggered off by minimal environmental exposure. Chomsky’s
arguments were lent further support by Lenneberg (1967), who provided
experimental and clinical evidence that language development was biologically
rather than environmentally determined. This position, which contradicted
the assumption that all learning was the result of experience (see Assumption
2), soon began to be reflected in discussions of language-teaching methodology
(cf. Newmark, 1963; Corder, 1967; Chastain, 1969). It is dealt with in detail
in the next chapter.
The pedagogic proposals that derived from these two sources of attack were
very similar. However, the sources were different in a significant way. Carroll
and Rivers still sought to extrapolate from general theories of learning to FL
learning. No consideration was given to the possibility that FL learning was
different in kind from other types of learning. Chomsky and Lenneberg,
however, took up just such a position. They argued that language was a
specifically human faculty and that, therefore, it was not possible to extrapolate
from general theories of learning.9 A language-specific theory was needed.
 Behaviourist Theory and Classroom Learning 27
This constituted a direct attack on the principal assumption (i.e. Assumption
1) of audiolingual learning theory.
Here is an account of the main objections raised against some of the other
assumptions of audiolingual learning theory.

FL Learning vs. LI Learning

According to audiolingual learning theory FL and LI learning were both
similar and different. Interest in the similarities and differences has continued
since (cf. Cook, 1969; Ervin-Tripp, 1974; Ellis, 1985b). The differences
which audiolingual theory assumed to exist as a result of LI interference and
age have been challenged, although not conclusively. The basis of the assumed
similarity (habit-formation) is no longer accepted, but the idea of similarity
itself lives on. The L2 = LI hypothesis (Dulay and Burt, 1974a) was pro¬
posed on the grounds that the process of L2 development was essentially the
same in both types of learning (see chapter 3).
The assumption that differences existed was largely motivated by the rec¬
ognition that whereas success in LI learning was guaranteed, this was not the
case in FL learning. It is probably true to say that the conditions that govern
FL learning are still seen as less favourable than those governing LI learning.
The search for the ‘optimum conditions, for classroom language learning —
the main goal of the audiolingual approach — continues (see chapter 5).

Habit-formation
Language learning cannot be convincingly treated as a process of mechanical
habit-formation. Rivers (1964) disputes the ‘mechanical’ part of the assumption,
arguing that a habit is developed only when the learner has a communicative
need and is in a relaxed state. Chomsky (1959) disputes the concept of‘habit’
itself. He argues that the concepts of ‘stimulus’ and ‘response’ are vacuous
where language behaviour is concerned. We do not use language in response
to clearly delineated behavioural stimuli. Also no behaviourist theory of learning
can account for the speaker-hearer’s ability to understand and produce
sentences which are not merely novel in terms of lexical choice but also in
their underlying pattern.
With these kind of arguments the whole edifice of audiolingualism began to
collapse. In particular, the belief that learning could be controlled from the
outside was eroded. As Dakin (1973:16) put it, ‘though the teacher may
control the experiences the learner is exposed to, it is the learner who selects
what is learnt from them.’ Implicit in audiolingual learning theory was the
belief that patterns and items could be learnt in the order they were practised.
The criterion used to determine the order of presentation was that of linguistic
difficulty. No thought was given to the inherent leamability of specific linguistic
features. Indeed such a notion was unthinkable in a theory that equated
learning with behaviourally induced habits. The essence of both neo-
28 Instructed Second Language Acguisitlon
behaviourist and nativist theories of learning was that the learner played an
active role in determining both what was learnt and when it was learnt.
Leamability became a central issue.
Can a case still be made for language learning as habit-formation? In so far
as memorization has a part to play, the methods by which ‘habits’ are formed
may contribute to at least some aspects of classroom L2 learning (e.g. vocabu¬
lary or ready-made phrases). They may also enable the learner to automatize
new knowledge. Current cognitive theories of learning distinguish declarative
and procedural knowledge (see chapters 1 and 8). Pattern-practice may be a
means of helping the learner to proceduralize declarative knowledge. This is,
in fact, the usual justification offered today for the continued use of pattern-
practice in language teaching (Gower and Walters, 1983). Such a view is
controversial, however. Johnson (1988:91), for instance, argues that mistake-
correction can only take place if there is ‘an opportunity to practise in real
conditions’. Pattern-practice does not afford such conditions. How L2 knowl¬
edge is acquired and how it is automatized have become central questions in
the study of classroom L2 learning. It is doubtful whether the notion of habit-
formation can provide adequate answers.
There was one theory of language learning — originating from Russia —
which integrated mentalist and behaviourist perspectives. Belyayev (1963)
distinguished ‘knowledge’, ‘skills’, and ‘habits’. In terms of language, ‘knowl¬
edge’ consisted of declarative facts about linguistic rules. ‘Habits’ were actions
that can be performed without conscious thought, such as the pronunciation
of a particular sound or the choice of lexical items. ‘Skills’ were communicative
actions carried out consciously, such as the stylistic use of language for
particular effects. Belyayev was critical of both teaching based on explanations
of grammatical rules and teaching that required endless repetition of correct
sentences. Although ‘knowledge’ and ‘habits’ could provide a basis for learning,
the goal was to develop ‘skills’ in relation to communicative language use. He
argued that teaching could promote learning if it included activities designed
to develop all three aspects. Belyayev’s theory with its recommended balance
between accuracy and fluency teaching was not so very dissimilar to that
proposed in England by Palmer (1921) or, more recently, by Brumfit (1984).

Analogy vs. Analysis

A number of arguments were advanced to refute the assumption that analysis
was of no value in language learning. One of the most compelling was that
voiced by Dakin (1973:19): ‘if no public explanation is given, the learners will
often form their own private explanation.’ Rivers (1964) argued that explanations
served to prevent practice becoming blind. The explanation helped the learner
to recognize and focus on those elements in the pattern that were the goal of
the lesson. The cognitive code method, which came to rival audiolingualism
in the late sixties and seventies, gave emphasis to the importance of metalingual
knowledge and the intellectual activity needed to acquire it. However, the
 Behaviourist Theory and Classroom Learning 29
status of metalingual knowledge has continued to be uncertain in theories of
classroom L2 learning and there is continuing debate about its usefulness.

Error Avoidance and LI Interference

Nativist theories of language learning have led to a complete revision in
thinking regarding the role of error. Errors have come to be seen as part of
the learning process. They are inevitable and provide evidence that acquisition
is in fact taking place. The next chapter considers these views in some detail.
The theory of transfer was closely tied to the belief that FL learning
consisted of habit-formation. Attacks on the latter were also attacks on the
former. If proactive inhibition was discounted, LI interference could no
longer be treated as a major source of error. A number of attacks were
mounted on the contrastive analysis hypothesis. It was pointed out that
linguistic difference could not be equated with learning difficulty in any
simple, straightforward way. Wardhaugh (1970) argued that the strong form
of the hypothesis (which stated that errors could be predicted by identifying
the linguistic differences between the learners’ LI and the target language)
was untenable, although the weak form (which stated that contrastive analysis
might be used to explain some of the errors that were seen to occur) was still
possible.
Transfer has continued to be a problematic area in L2 acquisition research,
not least because of its associations with behaviourist learning theory. There
can be little doubt, however, that the learner’s LI does play a significant part
in L2 learning and, in particular perhaps, FL learning (Marton, 1981). One of
the major problems in determining exactly what this role is lies in the
difficulty of distinguishing transfer that occurs in conscious translation from
transfer that results from subconscious processes. Corder (1978a) talks of the
former as a communication strategy and the latter as a learning strategy.
The ‘transfer debate’ continues in full flow today (cf. Gass and Selinker,
1983; Kellerman and Sharwood-Smith, 1986). Discussion now takes place
within the framework of cognitive or linguistic rather than behaviourist learning
theories. The term crosslinguistic influence is often preferred to that of
transfer.

Conclusion

Audiolingualism rapidly lost popularity in the United States, partly as a result

of the strong theoretical arguments that were advanced against it, but also
because the results obtained from classroom practice were disappointing. The
method required highly skilled teachers who were not always available. Many
learners found pattern-practice boring and lost interest in FL learning. Even
learners who were ‘motivated’ to persevere found that memorizing patterns
30 Instructed Second Language Acquisition
did not lead to fluent and effective communication in real-life situations.
Older learners found little appeal in a method that denied them opportunity to
use their cognitive skills in a more rule-governed approach to language
learning.
Nevertheless, the premises of behaviourist psychology continue to influence
many teachers thinking about how a L2 is learnt. Brumfit (1979) contrasted a
traditional language teaching methodology (present —> controlled practice —»
free practice) with a communicative methodology (communicate —» present —»
controlled practice). There was a role still for controlled practice in a
communicative methodology — presumably as a way of enabling the learner to
acquire and use linguistic features which are not yet part of her repertoire.
The theoretical justification for controlled practice — implicit or explicit —
remains behaviourist ideas of habit-formation. These are rarely acknowledged
these days, however.
Audiolingual learning theory is an inadequate explanation of classroom L2
learning. Its value is that it offered a number of explicit claims about the
nature of learning which were open to challenge. The theory, therefore,
instigated the discussion of a number of questions of central importance for
understanding instructed language learning:

1 To what extent is it possible to control the course of L2 development

by the external manipulation of the linguistic environment?
2 What role does controlled production of L2 forms play in their
acquisition?
3 Does classroom L2 learning consist of the incremental acquisition of
discrete linguistic features?
4 To what extent are the errors produced by instructed learners the
result of LI interference?
5 What role do errors play in the process of classroom L2 learning?
6 What role does metalingual knowledge play in classroom L2 learning?

Audiolingual learning theory offered definite answers to these questions. It

soon became clear that they were the wrong answers but they provided clear
reference points for subsequent thinking and research. What was missing
from the theory was a recognition that language learning is a developmental
process. Only a theory that paid due attention to the contribution of the
learner - as opposed to the environment - was capable of addressing this
major lacuna.

NOTES

1 It is probably true to say that most approaches to language teaching — in Britain at

least - have owed more to linguistics than to psychology. The so-called com¬
municative approach, for example, derives primarily from functional descriptions
 Behaviourist Theory and Classroom Learning 31
of English (Wilkins, 1976). Language teaching in the United States, however, has
paid more attention to psychology. The humanistic approach (cf. Stevick, 1976;
1980) owes much to a particular view about the nature of learning.
2 In fact, the methodological procedures remained the same irrespective of the age of
the learner. None of the language-teaching handbooks of the time saw any need to
modify the approach to take account of the cognitive differences between young
children and adolescents/adults.
3 Structuralist linguistics — the origins of audiolingualism — were never as popular
in Britain as the United States. Firthian linguistics, which emphasized the relationship
between context and language, encouraged a different view of language, one that
gave greater importance to meaning.
4 The ideal learner was an ‘automaton’. Consider what Morton (1964:3) had to say
about learning'L2 pronunciation: ‘It can be accomplished most quickly if the
student makes a conscious effort to ‘automatize’ himself — turn himself into a non¬
thinking ‘machine’ whose only responsibility is to respond immediately, mechanically,
AUTOMATICALLY to an external, physical stimulus.’
5 Brooks also makes the point that classroom L2 learners will approach the task in a
formal way as a result of their previous experiences of what it means to learn in a
classroom. When students have their first taste of FL learning they will have
already developed a learning ‘set’ as a result of their earlier experiences of studying
other school subjects. They bring this ‘set’ to the task of FL learning.
6 Skinner’s account of operant conditioning was developed to explain LI learning.
No language-teaching methodologist advocated a ‘trial-and-error’ approach. Lado
specifically dismissed such a possibility. In order to ensure that there were optimal
conditions for learning it was necessary to elicit responses. Thus, mainstream
audiolingualism conformed more to theories of classical conditioning, although
operant conditioning was drawn on as a justification for ‘shaping’ learners’ responses
in stages.
7 The term ‘problem solving’ appears to be used to refer to any learning activity
that required intellectual effort (e.g. understanding a grammatical explanation or
translating).
8 ‘Incomplete learning’ occurred when the teaching was unsatisfactory in some way.
Implicit in the audiolingual method was the belief that as long as the teaching
provided optimal conditions, complete learning would take place and errors be
avoided.
9 Chomsky also argued that language was separate from other cognitive systems and
therefore could not be explained by means of a general theory of human learning
such as that developed by Piagetian psychologists. A number of applied linguists
have found such a theory attractive, however. Dakin (1973), for example, provides
a cognitive account of learning, the main points of which are: (1) learners need
real-life experience, (2) learning involves assimilation and accommodation and (3),
as a result of (2), learning takes place in stages.
3 Naturalistic
Language
Acquisition and
Classroom Language
Learning

Introduction

In the last chapter we saw that one of the factors contributing to the rejection
of the audioiingual theory of foreign-language learning was the emergence of
a powerful theory that addressed head-on how language is learnt. This theory
was derived from a particular view of language, one that emphasized the
mentalistic nature of linguistic knowledge and was, therefore, diametrically
opposed to the structuralist linguists’ view of language as a set of habits. The
theory initially addressed how children acquired their mother tongue (LI),
generating a number of studies of LI acquisition. The theory and the empirical
research provided a basis which was used by applied linguists to support two
different views about the nature of classroom L2 learning. One view held that
classroom language learning was essentially the same as LI acquisition and
would proceed most effectively if no attempt was made to interfere with the
natural processes of learning. This view of learning was used to justify a set of
pedagogical proposals which have become known as the cognitive anti¬
method. The other view drew on Chomsky’s distinction between competence
and performance. A number of applied linguists argued that as the real goal of
classroom language learning was competence, learners should be encouraged
to engage in the conscious ‘analysis’ of linguistic forms. This view led to the
cognitive code method. This chapter begins with an account of these two
views of learning. This leads into a fuller consideration of the L2 = LI
hypothesis, which we touched on briefly in the last chapter.
Towards the end of the sixties interest in the nature of second language
(L2) acquisition began to gather steam. This interest manifested itself in three
major directions. First, researchers began to examine the errors which L2
learners produced, collecting corpora of learner language, identifying the
errors and then seeking to describe and explain them. Second, a number of
studies of naturalistic L2 acquisition were undertaken. These were of two
kinds: longitudinal case studies of individual learners and cross-sectional
 Naturalistic and Classroom Learning 33
studies of large groups of learners. Both the error-analysis and naturalistic
studies were motivated initially by the need to determine to what extent L2
learning was different from LI learning, in particular as a result of transfer
and age. Third, theories of L2 acquisition began to appear. This period —
from the late sixties up to the end of the seventies — provided a rich source of
descriptive information about the nature of L2 learner language and the
processes of naturalistic acquisition together with a theoretical framework,
which has become known as interlanguage. Although much of the research
was speculative, incomplete and did not investigate classroom language learning
directly, it was used to justify a number of radical proposals about how
teaching should proceed.
This chapter 'also considers Krashen’s Monitor Model. This was first
described in a series of papers that appeared in the late seventies, subsequently
brought together in Second Language Acquisition and Second Language Learning,
published at the beginning of the eighties. Krashen drew extensively on
studies of naturalistic acquisition (both LI and L2) to advance a theory that
has had considerable impact on language teaching. Krashen himself has never
been shy to apply his model to the classroom. Together with Terrell (1983) he
has proposed a method (the ‘Natural Method’) which he claims is based on a
thoroughly tested set of principles of L2 acquisition. Krashen — more than
any other researcher — has been prepared to stick his neck out and commit
himself to a highly controversial view of the teaching—learning relationship.
Krashen’s influential work represents the attractions and the dangers of an
approach that rests on the application of non-classroom based research.
Finally, in the conclusion to the chapter, the legitimacy of an approach to
classroom L2 learning based on the application of the results of research into
naturalistic acquisition is questioned. This is an issue that has been hotly
debated with strongly divergent views.

The Applications of LI Acquisition Research

The impetus for the LI acquisition research which took place in the sixties
and early seventies came from Chomskyan linguistics. This provided a radically
different view of language from that of the structuralist school. Whereas the
latter saw language in terms of the surface patterns that make up speech and
emphasized the differences between languages, Chomsky emphasized the
abstract nature of the rules that constitute the individual speaker-hearer’s
underlying competence and the universal nature of these rules. These differ¬
ences in the way language was viewed were matched by corresponding differ¬
ences in how language acquisition was thought to take place. As we saw in the
last chapter, behaviourist psychologists maintained that language was learnt as
a set of habits in which particular stimuli were associated with particular
responses through reinforcement. This position was challenged by a generative
34 Instructed Second Language Acquisition
theory of language which emphasized the abstract nature of linguistic knowledge.
Chomsky has consistently argued the impossibility of a child arriving at the
rules of the target language grammar solely on the basis of primary linguistic
data. In other words, competence could not be achieved simply by attending
to performance because the available input data were simply insufficient to
enable the child to discover the ‘hidden’ rules.1 The child’s task — as seen by
Chomsky — was to devise an appropriate grammar given imperfect primary
linguistic data. This was possible only if the child was credited with a language-
acquisition device consisting of innate knowledge of grammatical principles:
‘As a precondition for language learning, he [the child] must possess, first, a
linguistic theory that specifies the form of the grammar of a possible human
language, and, second, a strategy for selecting a grammar of the appropriate
form that is compatible with the primary linguistic data’ (Chomsky, 1965:25).
The task of the linguist was to specify the nature of the ‘linguistic theory’
that enabled the child to learn language. Although the details of the linguistic
theory have changed over the years, Chomsky’s initial premise has not. The
search for an adequate description of the Universal Grammar which makes
the learning of a particular language possible continues today (cf. Cook,
1988).
Empirical studies of LI acquisition were instigated as a means of testing
Chomsky’s claims about language and language learning. One of the important
issues was whether the speech produced by children provided evidence of
habit-formation or of innate linguistic knowledge. Another important issue
was whether there was any evidence of developmental progression in the way
children’s competence evolved — and, in particular, whether this progression
conformed to the specific proposals of generative grammarians such as
Chomsky. For example, Chomsky (1957; 1965) distinguished phrase structure
rules which generated abstract base or deep structures and transformational
rules which converted these into the surface structures observable in actual
speech or writing. Researchers were interested in whether children began the
process of acquiring language with a ‘simple’ grammar composed only of base
structures and then added the transformational component later (cf. McNeill,
1970). The sixties and early seventies saw a number of child acquisition
projects involving both longitudinal studies (e.g. Brown, 1973; Bloom, 1970)
and cross-sectional studies (e.g. Villiers and Villiers, 1973). These provided
evidence that children passed through a series of stages en route to adult
competence.2 The progression was evident both in the gradual increase in the
mean length of utterance in children’s speech and also in the order of
acquisition of grammatical features such as progressive (-ing, plurals), copula
(‘be’, and 3rd person -s). Thus, children’s utterances grew increasingly longer
as they acquired first one grammatical feature, then another and so on.
Evidence was also forthcoming that the child revised grammatical rules as
acquisition took place; for example, the past tense of irregular verbs, such as
‘went’, showed a U-shaped pattern of development, with ‘goed’ replacing the
correct ‘went’ in an intemiediate stage. These findings provided support for
 Naturalistic and Classroom Learning 35
Chomsky’s arguments that language acquisition was internally rather than
environmentally driven.
Applied linguists responded rapidly both to Chomsky’s view of language
and to LI acquisition research. The teaching manuals that now appeared (e.g.
Rivers, 1968; Jakobovits, 1970; Chastain, 1971) reflected the rejection of
audiolingual theory and argued the case for a teaching approach that gave
recognition to the learner and the abstract nature of linguistic competence.
There was, however, far less conviction regarding both how an L2 was learnt
and how it should be taught. This was understandable, given the feeling that
all the old certainties had gone: ‘The theoretical foundations of our assumptions
about the nature of language and how it is acquired have been badly shaken if
not actually destroyed’ (Donaldson, 1971:123) and that nothing clear-cut had
taken their place. Applied linguists spoke of ‘psycholinguistic upheaval’, ‘un¬
certainty’ and ‘current ferment’ and they were aware that the new proposals
they put forward were tentative. As Wardhaugh (1971:16) put it: ‘Perhaps a
new method will develop which will achieve the same kind of general approval
as the Audiolingual method, but at the moment there is no consensus as to
what it would be like.’
The general feeling was that the new linguistics and language acquisition
research could not be used to justify confident prescriptions about pedagogic
procedures but could only serve to provide ‘insights’ about the teaching-
learning relationship. These insights took the form of the cognitive anti¬
method and the cognitive code method, both drawing on different aspects of
Chomsky’s ‘cognitivism’. It is probably true to say that the theories of classroom
language learning upon which these methods were based lacked the cohesive¬
ness of audiolingual learning theory. The cognitive code method had little
impact on classrooms. Cognitive anti-method was also peripheral, although its
principal assumptions were incorporated into subsequent theories of classroom
language learning derived from L2 acquisition research. At the time definitive
new practical directions were lacking (Smith, 1971). It is perhaps for this
reason that Richards and Rogers (1986) include neither method in their
admirable survey of language-teaching approaches and methods.

The Cognitive Anti-method

The cognitive anti-method was articulated in a series of articles published by
Newmark and Reibel in the sixties. Very similar views were also expressed in
Corder (1967) and Jakobovits (1970). The theory of classroom learning which
supported the method is outlined below.

Assumption 1: Second-language learning is controlled by the learner

rather than the teacher This assumption is central to the theory. Newmark
and Reibel (1968), for instance, talk disparagingly of ‘the excessive preoccu¬
pation with the contribution of the teacher’ and argue that this has ‘distracted
the theorists from considering the role of the learner’. Kennedy (1973:76)
36 Instructed Second Language Acquisition
gives expression to the new belief that it is the learner who is in charge: ‘in the
final analysis we can never completely control what the learner does, for he
selects and organises, whatever the input.’ The learner is seen as a problem-
solver who contributes hugely and actively to the learning process.

Assumption 2: Human beings possess an innate capacity for learning

language Chomsky’s arguments that children made use of a biologically
endowed language-acquisition device for acquiring language were widely
accepted. Newmark (1966) saw LI acquisition as a model of successful
learning and argued that L2 learning in the classroom could proceed in much
the same way as long as the teacher avoided ‘interfering’. Newmark and
Reibel (1968) saw the adult as ‘a potentially magnificent language learner’.
They went on to argue that the foreign language (FL) learner’s capability was
‘qualitatively’ the same as the child’s — although, possibly, ‘quantitatively’
different. In other words, they credited both the child and adult L2 learner
with continued access to the child’s language-acquisition device. It should be
noted, however, that this was contentious; other applied linguists were less
sure that adults in particular could learn a L2 in the same way as children
acquired their L2.

Assumption 3: It is not necessary to attend to linguistic form in order to

acquire a L2 The cognitive anti-method shared one assumption with audio-
lingual learning theory — namely that linguistic analysis was not necessary.
Newmark (1963: 217) argued: ‘Systematic attention to the grammatical form
of utterances is neither a necessary condition nor a sufficient one for successful
language learning.’ Newmark pointed out that LI acquisition took place very
successfully without any analysis and he noted that many classroom learners
were unsuccessful despite it. He advised special caution about applying the
descriptions from transformational generative grammar to the classroom,
arguing instead that there was a need to ‘liberate language teaching from
grammatical theory’.

Assumption 4: Classroom language learning is not an additive process

Perhaps the most radical departure from audiolingual learning theory was
Newmark’s (1966) claim that classroom language learning, like child LI
acquisition, was not an additive process, i.e. the learner did not acquire
linguistic features incrementally but rather learnt ‘whole chunks’ at a go.
Newmark considered the isolation of parts from wholes ‘artificial’ and rec¬
ommended the provision of ‘natural contexts’ where no attempt was made to
select and grade the input for the learner.

Assumption 5: Errors are a concomitant of the learning process and are

therefore, inevitable Empirical studies of LI acquisition showed that many
of the utterances produced by children were unique. Utterances such as the
 Naturalistic and Classroom Learning 37
following could not be found in the input data to which the children were
exposed (Brown, 1970):

Mommy sock
Allgone rattle
Sweater chair
Walk street

This finding was important because it provided strong evidence against be¬
haviourist claims that language learning involved imitation of parental speech.
Another finding was that the child’s unique utterances were highly systematic,
reflecting the existence of underlying rules. This led researchers such as
McNeill (1966) tto propose that the child derived utterances from an internally
consistent grammar which was constructed in the process of acquisition.
When the child produced deviant utterances — such as those above — she was
testing out hypotheses about the rule system of the language. In McNeill’s
(1966:2) terms the child was trying ‘to discover how a more or less fixed
concept of a sentence is expressed in the language to which he has, by
accident, been exposed’.
This view of LI acquisition led to a re-evaluation of the role of error in L2
learning in what was, perhaps, the most significant advance in thinking about
classroom language learning. Jakobovits (1970) argued that learners should
not be forced to produce well-formed sentences. He pointed out that the
production of deviant sentences in LI acquisition did not prevent the child
from going on to produce well-formed sentences later. Corder (1967) suggested
that it was much more important that L2 learners be allowed to discover their
own errors rather than be corrected by the teacher.

Assumption 6: LI interference is the result of ignorance Newmark and

Reibel (1968) accepted that LI interference did take place. However, they did
not consider it a significant factor in classroom language learning. They
argued that interference was more likely to occur if the learner’s attention was
drawn to the contrasts between her LI and the L2 — as was the case in both
grammar—translation and audiolingualism. They suggested that the conditions
that produced interference should be minimized by reproducing the learning
environment of LI acquisition in the classroom. In short, they viewed the
‘problem of interference’ as a ‘problem of ignorance’; the solution was simply
‘more and better training in the foreign language’. This could not be provided
by drilling the points of difference but only by ‘observation and exercise of
particular instances of the language in use’.

These six assumptions constituted a radical alternative to audiolingual learning

theory. Whereas the latter emphasized the importance of controlling the
course of L2 development by manipulating the linguistic environment, the
cognitive anti-method rested on the belief that the innate language-learning
38 Instructed Second Language Acquisition
capability of the student was sufficient. This would work best if the conditions
of learning found in LI acquisition were reproduced in the classroom.
Newmark and Reibel’s proposals were before their time but, as we shall see,
they bear a striking resemblance to those advanced a decade later by Stephen
Krashen.

The Cognitive Code Method

The cognitive code method was far less radical. Carroll (1966:102) describes
it as ‘a modified, up-to-date grammar translation theory’. It was for this
reason, perhaps, that it was taken more seriously than the cognitive anti¬
method by the teaching profession. As we saw earlier, it motivated a major
comparative method study (Smith, 1970) designed to establish whether it led
to more successful language learning than audiolingualism. The method is
described most fully in the work of Chastain (1971).
The principal assumption of cognitive code learning theory was that per¬
ception and awareness of second language rules preceded the use of these
rules. It was in this respect that it differed most strongly from audiolingualism.
Audiolingual learning theory discounted the contribution of metalingual
knowledge, except as ‘summaries’ of acquired behaviour; cognitive code learning
theory saw conscious grammatical knowledge as essential to the learning
process. This assumption derived in part from the competence/performance
distinction. Donaldson (1971:123) commented: ‘We have discovered rather
late that we have been teaching speech and not language.’
Thus competence (i.e. ‘language’) was equated with explicit knowledge of
grammatical rules, while performance (‘speech’) was equated with practising
their use. The learners’ first task was to lay the foundations of their linguistic
knowledge; only when this had been achieved could they ‘perform’. Thus
according to theorists such as Chastain (1971) learning proceeded from
competence to performance. He argued that the presentation of grammatical
rules should precede the provision of opportunities for practice. Whereas
audiolingual learning theory emphasized inductive learning, the cognitive code
learning theory placed considerable store on deductive learning, at least as a
basis for practice.
Although applied linguists lent their support to the explicit teaching of
grammar, they were careful to emphasize that they were not advocating a
return to the bad old days of the grammar—translation method. They argued
that learners needed to understand what they were learning and that they
ought to be allowed to make use of their cognitive skills to help them achieve
comprehension of grammatical structure. ‘Rules can obviously provide a
short-cut to learning,’ argued Prator (1969).
The other major difference from audiolingualism lay in the recognition that
language learners needed opportunities to use the language innovatively
and creatively. One of the principal objections to the audiolingual learning
theory was that language use was innovative and stimulus-free. As Chomsky
 Naturalistic and Classroom Learning 39
(1966:153) stated: ‘Ordinary linguistic behavior characteristically involves
innovation, formation of new sentences and new patterns in accordance with
rules of great abstractness and intricacy.’ Language learning, therefore, could
not be accounted for in terms of the memorization of a fixed set of patterns.
‘The use of language resembles more writing a play than performing one,’
McNeill (1965) claimed.
The rejection of stimulus—response association as the basis of language
learning led to a reconsideration of the role of practice. Jakobovits (1968) and
Chastain (1971) rejected mechanical practice as a means of forming gram¬
matical rules. According to them new grammatical forms were not acquired
through imitation and not stamped in through practice. The starting point was
‘concept attainrpent’, which resulted from verbalizing grammatical relations.
Practice was useful, however, once a grammatical concept had been acquired.
Jakobovits recommended transformation exercises on the grounds that these
dealt with deep structure and, except for pronunciation, rejected pattern drills
which only dealt with surface structure. Chastain emphasized the importance
of providing opportunities for the learner to create language in ‘language¬
demanding situations’. These opportunities needed to be ‘linguistically unique’.
These, then, were the principal views of learning upon which the cognitive
code method was based. The emphasis on rule perception and conscious
analysis set cognitive code learning theory off against both audiolingual learning
theory and the position adopted by Newmark and Reibel. The recognition
that learners needed an opportunity to use the L2 meaningfully was shared
with cognitive anti-method but contrasted with audiolingualism. In many
respects cognitive code learning theory was a half-way house between the
received opinion of audiolingualism and the radical proposals of the cognitive
anti-method.
Cognitive code learning theory was problematic in a number of respects.
Applied linguists such as Chastain were too ready to extend terms like
‘competence’ and ‘performance’ beyond the theoretical boundaries in which
they were formulated. Equating an understanding of grammatical rules with
competence and speaking with performance was a distortion of Chomsky’s
position. In fact, Chomsky made it clear that both understanding and speaking
were aspects of performance since the same non-linguistic variables (memory
span, attention etc.) affected both. Also, competence for Chomsky was not
explicit knowledge but rather the abstract knowledge that underlies the
speaker-hearer’s use of language. It makes no sense, therefore, to talk of
competence (= explicit knowledge) preceding performance (= language use).
Nor, in fact, does it make much sense to insist that language instruction
should be concerned primarily with competence, as learners clearly need to be
able to perform successfully in a L2.
Another problem concerned the sequencing and grading of linguistic items.
In audiolingualism there were clear criteria for determining the order of
presentation (e.g. difficulty). Cognitive code learning theory continued to view
L2 learning as an incremental process but offered no rationale for deciding in
40 Instructed Second Language Acquisition
which order items should be dealt with. One proposal was that the teaching
syllabus should follow the natural sequence of acquisition observed in LI
acquisition. McNeill (1970) hypothesized that children begin with a ‘base’, not
dissimilar from the base proposed in early models of generative grammar, and
then go on to acquire various ‘transformations’.
This led Chastain (1971:91) to suggest that L2 learners might do the same:
‘One would suspect that, since first-language learners first produce base
structure, second-language learners should also begin with base structure.’
However, he was not certain that a L2 should be learnt in the same maimer as
the LI, admitting that the best way of estabfishing the relationship between
base structure and surface structure was still unclear. Jakobovits (1970:23) put
forward the proposition that the learner should be exposed to utterances
which are ‘grammatically progressive at each stage but which fall short of
having the complexity of well-formed sentences’. In other words, an attempt
should be made to follow the LI learner’s developmental route by teaching
appropriate learner-rules at each stage. McNeill (1965) also felt that it might
be worth while teaching adult L2 learners child grammar so that they first
learn base structure. Jakobovits, however, recognized that this kind of proposal
was likely to be resisted by most teachers. He acknowledged that not enough
was yet known about whether such finely tuned input facilitated acquisition or
not. These various suggestions testify to the uncertainties about what to teach
and when to teach it, which were so evident in accounts of the cognitive code
method.
The cognitive code learning theory represents an uneasy attempt to incor¬
porate the new ideas provided by generative linguists and LI acquisition
researchers. It maintained the view that classroom learning could be externally
manipulated — ‘interfered with’, as Newmark and Reibel would have it — by
presenting and practising discrete linguistic features. But it did introduce the
important idea that there were constraints to do with the nature of language
learning which had to be taken into account if instruction directed at specific
items was to work.
Neither the cognitive anti-method nor the cognitive code method made
much of an impact on language teachers. The former was probably too radical
and the latter insufficiently distinct from audiolinguafism. The methods are of
historical interest because they reveal the initial attempts of applied linguists
to attend to the way language is acquired when they formulated pedagogic
proposals. They reflect the attempt to obtain insights from an understanding
of how LI acquisition took place. Newmark and Reibel accepted the equation
of LI and L2 acquisition with confidence. Jakobovits (1968:275), however,
voiced the kind of doubt that most applied linguists felt at the time when he
said that ‘the fastest method of acquiring a second language need not be one
that replicates the conditions existing under “natural” language acquisition.’
It is now time to consider in more detail the basis of the L2 = LI
hypothesis.
 Naturalistic and Classroom Learning 41

The L2 = LI Hypothesis

Applied linguists of this period found themselves in something of a quandary.

On the one hand they wished to use the evidence of LI learning to support
arguments for a more cognitive approach to language teaching. On the other
hand, they recognized the obvious differences that existed between LI and
classroom L2 learning. Most applied linguists rejected the ‘natural approach’
advocated by Newmark and Reibel. They wished to maintain their advocacy
of a ‘structured content’ (Prator, 1969) for L2 learning (i.e. selection and
grading).
The arguments in favour of the non-equivalence of L2 and LI learning
seemed overwhelming. They were frequendy rehearsed in a number of publi¬
cations (e.g. Prator, 1969; Cook, 1969; Kennedy, 1973). They are summarized
in table 3.1. It should be noted that the differences all related to the conditions
of learning or to the learner and concerned the physical, cognitive and
affective domains of learning. However, there was no consideration of whether
the process of learning in the linguistic domain was the same in both types,
despite the fact that this was really the central issue.
It became common to distinguish acquisition and learning (cf. Corder,
1973; Wilkins, 1974). These terms were defined with specific reference to the
conditions of learning:

The term acquisition is used here for the process where language is
acquired as a result of natural and largely random exposure to language,
the term language learning where the exposure is structured through
language teaching. (Wilkins, 1974:26)

The basis of this definition contrasts with that of Krashen’s (1981) later
definition. Krashen emphasized the differences in the way linguistic
knowledge is internalized and stored, i.e. process differences.3
There were two responses to the differences in learning conditions. Some
applied linguists assumed that they were inevitable. Prator (1969:143)
commented: ‘There is actually no way whereby the circumstances under
which a child learned his mother tongue can ever be reduplicated for learning
a second language.’
Others, however, felt that L2 conditions were not immutable and that much
more could be done to make the classroom environment resemble that of LI
learning. Kennedy (1973), for instance, suggested that a structured content
could result in an artificial learning context which was counterproductive.
Both Cook (1969) and Kennedy felt that avoidance of errors might inhibit
hypothesis testing and communication. Thus, although the presence of differ¬
ences in the conditions of learning were widely acknowledged, their inevitability
was also beginning to be challenged.
42 Instructed Second Language Acquisition
Table 3.1 Differences between LI and classroom L2 learning

Aspect Points of difference Comments

A. Conditions 1 Amount of time LI learner has much more time

at disposal; classroom L2 learner
has very restricted time.
2 Structured LI learner is exposed to
content naturally occurring language; L2
learner is presented with
carefully selected and graded
input.
3 Avoidance of In LI learning errors permitted
errors and, if they do occur, they are
usually not corrected; in
classroom L2 learning errors are
avoided and corrected.
B. Learner 1 Age LI learner has innate capacity
for learning language; ‘critical
period’ for this capacity may
have passed for L2 learner, who
is also cognitively more mature.
2 Motivation Child has strong motivation to
learn LI because of importance
of communication for satisfying
basic needs; L2 learner’s
motivation is necessarily weaker.
3 Linguistic LI learner has no previous
knowledge knowledge of language; L2
learner approaches task already
knowing a language and
potentially transfers this
knowledge.
4 Cultural The LI learner acquired cultural
knowledge norms in the process of
acquiring language; the L2
learner has already acquired a
set of cultural values and may
experience anomie as a result of
learning L2.

Source: based on Prator (1969); Cook (1969); Kennedy (1973)

Applied linguists of this period were clearly reluctant to claim that the
child’s language-acquisition device was still available to the adult learner.
Even those who leaned in this direction did so uncertainly. Cook (1969:109),
for example, concluded with the warning that the soundness of the analogy
between LI and L2 learning still had to be proved. There were two main
reasons for this hesitancy; LI transfer and age.
Applied linguists were doubtful whether LI and L2 learning involved
similar processes because they remained convinced that language transfer was
 Naturalistic and Classroom Learning 43
a major factor in L2 learning. Prator (1969:151) believed that linguistic
interference was ‘so obvious that it hardly seems necessary to consider it
further’. Other commentators such as Chastain (1971) continued to recognize
the role of the learner’s LI but saw it in a more positive light, arguing that
learning could be facilitated by concentrating on areas of structural similarity
between mother tongue and target language. Jakobovits (1970) thought that
transfer might be more a feature of performance than competence and also
speculated that different performance factors (understanding, speaking, reading,
writing) were affected to different extents. But it was not clear how transfer
could be reconciled with the idea of an innate capacity for learning languages.
Only in recent years has the idea of language transfer been satisfactorily
incorporated ir^to a nativist model of L2 learning (cf. Kellerman, 1984). For
example, it has been suggested that the learner resorts to her LI in cases of
indeterminancy, i.e. where the language-acquisition device fails to specify
narrowly the syntactic rule of the L2 (Zobl, 1983).
The other major impediment to grasping the innate capacity nettle was age.
There were both folk beliefs and scientific arguments to support the idea that
the language-learning device atrophied with age. It was a well-known fact that
children picked up their LI effortlessly, whereas adults had to struggle
ineffectively with a new language. Neurophysiological evidence seemed to
support such a belief. Penfield and Roberts (1959) advanced the critical period
hypothesis according to which the ability to learn a language naturally and
effortlessly was linked to cerebral plasticity, which terminated around the age
of ten years when puberty set in.4 This occurred as a result of lateralization of
the language function in the left hemisphere of the brain. Lenneberg (1967)
provided clinical evidence to support the hypothesis. For example, children
who underwent surgery on the left hemisphere recovered total language
control, whereas adults did not.
The critical period hypothesis continues to be hotly debated today. Seliger
(1978) has put forward the view that there might be different critical periods
for different aspects of language. Long (1988c) suggests that the critical
period for acquiring a native-speaker pronunciation is around six years, while
puberty is the key age for the acquisition of L2 grammar. In general, it is
accepted that whereas children are able to access their language-acquisition
device in L2 learning, adults make use of strategies of learning of the kind
employed in general problem-solving. Bley-Vroman (1988), for example,
disputes the L2 = LI hypothesis on the grounds that adult language learning
is more like general skill-learning than LI acquisition. However, it has also
been argued that the adult’s use of problem-solving strategies does not
preclude the possibility that L2 acquisition proceeds in the same way as LI
acquisition. Other researchers (e.g. White, 1985) maintain that adults have
continued access to the language-acquisition device and that the overall
process of acquisition is the same in LI and L2 acquisition, irrespective of
age.
In the sixties LI transfer and the critical period hypothesis provided powerful
44 Instructed Second Language Acquisition
reasons for believing that the process of L2 acquisition must be different from
that of LI acquisition. However, as we have seen, there were those who began
to argue otherwise. Newmark (1966), for instance, put forward the ignorance
hypothesis to account for LI interference (see p. 00). Newmark’s view was
that ‘the cure for interference is simply the cure for ignorance dearning’. The
ignorance hypothesis was compatible with an innate acquisition device in L2
learning. Macnamara (1973) queried the critical period hypothesis. He pointed
out that there was no evidence that a language could not be learnt just as well
after puberty as before. Nevertheless, the received opinion was that knowledge
of a previous language and age made L2 learning qualitatively different from
LI learning.
The uncertainty of the L2 = LI hypothesis prevented a quantum leap away
from the well-tried assumptions of the audiolingual learning theory. It lay
behind the tentativeness and incompleteness of the cognitive code learning
theory and the peripheral status of the cognitive anti-method. The uncertainty
could only be resolved if a number of questions were answered. What role did
the learner’s LI play? Was the child’s language-acquisition device available
for use in L2 learning? Could the classroom replicate the conditions of LI
learning? These were the questions that motivated the first empirical studies
of L2 acquisition.

The Applications of Naturalistic L2 Acquisition Research

It would be incorrect to say that there were no empirical studies of L2

acquisition until the late sixties. McLaughlin (1978a) in a review of early
research into child L2 acquisition mentions a number of early studies. But it
is true to say that there was a sudden explosion of studies at this time. This
was stimulated by the theoretical questioning of behaviourist theories of
language learning and the need to obtain answers to the questions listed at the
end of the previous section. It was made possible by the methodological
techniques developed by LI acquisition researchers.
The vast majority of the studies examined either pure naturalistic or mixed
L2 acquisition. There were very few studies of pure classroom learners. This
might seem surprising given that most of the research was frequendy motivated
by the desire to find out how classroom learning could be facilitated and was
carried out by researchers-cum-teachers. There were, however, a number of
reasons. The main one was the desire to explore to what extent L2 acquisition
was different from LI acquisition. It made better sense if the comparison was
made between learners who experienced similar learning conditions. Another
was that much of the research took place in second rather than foreign
language settings (e.g. learners of L2 English in the United States or learners
of L2 German in Germany). Also the methodological techniques of LI
acquisition research transferred easily to naturalistic L2, acquisition research.
 Naturalistic and Classroom Learning 45
It was several years before these techniques were applied to classroom language
learning. Allwright (1980) lamented that the case-study approach — fore¬
most in early studies of naturalistic acquisition — was spumed in classroom
research.1 2 3 4 5
The purpose of this section is not to review the findings of the research. A
number of reviews are already available (e.g. McLaughlin, 1978a and 1985;
Ellis, 1985a; Long and Larsen-Freeman, forthcoming). Instead, there will be
a brief account of the different types of enquiry and an oudine of the theory
associated with it — interlanguage theory. This will be followed by a discussion of
the main conclusions that were reached regarding the applications of the
research to the classroom.

Types of Enquiry
There were three distinct types of enquiry: error analysis, performance analysis
and form—function analysis. These are described below. Also some of the
methodological problems of each technique are considered so that the limi¬
tations of the research are clearly understood before the applications are
considered.

Error analysis Error analysis was not new. There had been a number of
earlier studies of errors in L2 learner-language (e.g. French, 1949). These,
however, consisted of lists of typical errors made by learners from different
language backgrounds. Error analysis in the late sixties and seventies was used
for different purposes — in particular to investigate the contrastive analysis
hypothesis. If it could be shown that many L2 errors were not traceable to the
LI this would provide strong evidence against the audiolingual learning theory.
Also, if it could be shown that many of the errors were developmental (i.e.
reflected the stage of development of the learner) this would constitute evidence
that the process of L2 and LI acquisition were similar.
The error analysis movement never lost sight of the classroom. Corder
(1967) claimed that errors were significant in three different ways and the first
of these was of value to the teacher — they indicated how far the learner had
progressed towards the final goal. The other two concerned the researcher
(errors provided evidence of how language is learned) and the learner (errors
were used to test hypotheses). The steps involved in error analysis also
reflected the close relationship between this technique and pedagogy. Els et
al. (1984) identify the steps as:

1 identification of errors
2 description
3 explanation
4 evaluation
5 prevention/correction
46 Instructed Second Language Acquisition
The last two steps make a direct link between analysis and teaching. Evaluation
involved considerations of error gravity (James, 1974) and communicative
efficiency. It was concerned with the attempt to identify which errors most
warrant pedagogic attention. Prevention and correction were explicitly
pedagogic.
The results of error-analysis studies were important because they provided
empirical support for the some of the theoretically derived claims. The majority
of learner errors were intralingual (i.e. caused by the structure of the L2
itself) rather than interlingual (i.e. the result of LI transfer). Dulay and Burt
(1974a) claimed that only 3 per cent of the errors produced by their subjects
could be explained by interference. Studies of errors, such as that carried out
by George (1972), showed that the learner herself made a substantial cognitive
contribution to the learning process — just like the child in LI acquisition.
However, the technique was open to a number of criticisms. Long and Sato
(1984) list six. Error analysis ignored what learners were doing correctly. The
classification of errors was subjective and unreliable. Analyses were unquan-
tified. Explanations of errors were ‘impressionistic and vague’. The samples
they were based on were often biased so that valid generalizations were not
possible. No account was made of the learner’s avoidance of certain structures
(Schachter, 1974). Some of these criticisms could be overcome by improving
the methodology but others could not; they indicated the need for an alternative
technique.

Performance analysis Performance analysis started a little later than error

analysis and came to dominate L2 acquisition research in North America in
the seventies. Performance analysis differed from error analysis in that it
aimed to provide a description of the L2 learner’s language development and,
therefore, looked not just at deviant but also at well-formed utterances. As
such it provided a much more powerful analysis.
The analytical tools were borrowed from LI acquisition research. Brown
(1973) used the concept of an obligatory occasion to determine to what
extent a child had acquired a specific grammatical feature. An obligatory
occasion consisted of any context for the use of a specific grammatical feature
which a learner incidentally created in the course of using the language. The
procedure consisted of first identifying all such contexts for a functor and then
establishing whether the feature was supplied in each one. The percentage of
accurate suppliance was then calculated. In order to determine whether the
feature was actually acquired it was necessary to decide on a criterion level
of suppliance. The actual level chosen varied from researcher to researcher,
but approximated to 90 per cent. Thus a functor was considered ‘acquired’ if
it was supplied in 90 per cent or more of the contexts requiring it.
Performance analysis was used in both cross-sectional and longitudinal L2
studies. In the case of the former, the accuracy order of a number of different
grammatical features was calculated and this order was then equated with the
order of learning. Thus, there was an a priori assumption that the accuracy
 Naturalistic and Classroom Learning 47
order corresponded to the acquisition order. The seventies saw a number of
cross-sectional studies that followed this procedure. These became known as
the morpheme studies. They investigated a common set of English gram¬
matical morphemes. The results were used to claim that there was a more or
less invariant order of acquisition which was independent of LI background
and age (see figure 3.1). Although this order was slightly different from that
found for the same morphemes in LI acquisition research, it provided evidence
against the contrastive analysis hypothesis and in favour of the existence of
universal cognitive mechanisms which enabled learners to discover the structure
of a particular language. This view of L2 acquisition became known as
creative construction (Dulay and Burt, 1975). However, the evidence provided
by the morpheme studies needs to be treated circumspectly as the method¬
ological procedure was seriously flawed. Also, despite some of the claims
made for the consistency of the findings, a number of studies produced different
accuracy orders (e.g. Hakuta, 1976). Lightbown (1984:248) comments at the

Figure 3.1 L2 acquisition hierarchy for thirteen English grammatical morphemes

Note: The hierarchy consists of groups of morphemes representing different
overall levels of accuracy and, therefore, of acquisition. The
morphemes were grouped in this way in recognition of the fact that the
difference in accuracy between some morphemes was too small to
claim that they were acquired at different times.
Source: Based on Dulay and Burt (1975)
48 Instructed Second Language Acquisition
end of a review of performance-analysis studies which provided evidence of
an invariant order of accuracy/acquisition:

This research has been very persuasive, but it can be shown to be

limited in generalizability. For every one of the examples given above,
there are other studies which provide counter-evidence, usually related
to the influence of the first language on SLA, but also to variation in
input or even in the degree of social integration.

In other words, the so-called ‘natural’ order of acquisition of L2 English

morphemes was not as fixed and definite as some researchers claimed.
Performance analysis was also used in longitudinal studies of individual L2
learners. Some of these studies also examined grammatical morphemes (e.g.
Rosansky, 1976), but others looked at syntactical sub-systems such as English
negatives and interrogatives (e.g. Cazden et al. 1975; Wode, 1976). The
results of these latter studies indicated that learners passed through a series of
overlapping stages en route to the full target language system (see table 3.2).
Minor variations in the sequence of development did occur as a result of the
learners’ LI background, but in general the route they followed was very
much the same. In the late seventies and eighties work on the L2 acquisition
of German word order in naturalistic learners (Meisel, Clahsen and Pienemann,
1981) provided strong evidence for the presence of highly regular acquisitional
sequences. These longitudinal studies constitute the strongest evidence to
date in support of a nativist theory of L2 acquisition.
Performance analysis research was a major breakthrough in the study of L2
learning. It gave us — for the first time — evidence of what happened when
learners set about acquiring an L2. The case-study data, in particular, enabled
us to see the process at work. There were a number of problems identified
with this technique, however (Hatch, 1983a; Long and Sato, 1984). Some of
the main ones will be mentioned here. Obligatory context analysis ignored
occasions when the learner overgeneralized. It was possible that a learner
could supply a feature like verb-ing on every occasion that called for it and

Table 3.2 Summary of stages of acquisition for auxiliary verbs in L2 English

Stage Description

1 Verbless utterances (e.g. ‘This kite’)

2 Base form of verb with no auxiliary; some learners use ‘dummy’ verbs
such as ‘get’, and ‘wanna’.
Almost concurrently V-ing appears without ‘be’.
3 ‘be’ auxiliary appears — first as ‘is’ and later as °m’ and ‘’re’.
4 ‘do’ auxiliary appears in negative form only (‘don’t’).
Modal verbs begin to appear but order varies from learner to learner.
5 ‘do’ auxiliary' is acquired in its various forms and functions.
6 ‘have’ auxiliary' is acquired late - rarely seen in first 2 years.

Source: Based on Hatch (1974)

 Naturalistic and Classroom Learning 49
thus be credited with having acquired it. However, the same learner might
also use verb-ing in contexts that did not require it. To say that this learner
had ‘acquired’ verb-ing would clearly be misleading. A feature is not fully
acquired until the learner has mastered the particular grammatical functions
that it serves. Performance analysis is also based on the assumption that the
‘rules’ which a learner internalizes correspond to the ‘rules’ of a fully competent
native speaker. This need not be the case, however; the learner may construct
highly idiosyncractic systems consisting of unique rules not found in the native
speaker’s competence. Bley-Vroman (1983) talks of the comparative fallacy;
he argues that instead of measuring L2 acquisition using the yardstick of the
target language, we need to illuminate the inner logic of learner grammars by
examining them in their own right.

Form—function analysis More recently form—function analyses of L2

learner language have been undertaken in order to overcome the kinds of
problems in performance analysis referred to above. The basis of this approach is
the study of the different functions which a specific form performs at different
stages of development.
Researchers have often commented on the high degree of variability evident
in learner language. Cancino et al, for example, (1978) talk about ‘the constant
development and concomitant variation’ in their subjects’ speech. They found
that their subjects alternated in the use of different negative forms (no + V,
don’t + V, aux.—neg. and analysed don’t) with two or more of these forms
apparent in their speech at any one time. Wagner-Gough (1975) studied the
L2 acquisition of Homer, a six-year-old Iranian boy. She found variation in
his form—function mapping of V-ing. In addition to marking progressive
aspect, V-ing was used to refer to immediate intention, distant future, past
time and process-state. Homer also used simple V to perform the same range
of functions. In other words, V-ing and V were used in apparent free variation.
Variability is, in part, a reflection of inconsistent form—function mappings.
The learner frequently uses two or more forms to perform the same set of
functions. The learner needs to discover the functional restrictions that apply
to each form and only when this has been achieved can acquisition of the
forms be said to have taken place. The study of form—function relationships,
therefore, has been seen as methodologically necessary by a number of
researchers (e.g. Huebner, 1979; Ellis, 1985d).
It is important to study these relationships for another reason. The studies
by Cancino et al. and Wagner-Gough suggested that the variation is often
random. However, a careful form—function analysis can often reveal hidden
systematicity. An example will make this clearer. Let us imagine that a learner
produces the following negative utterances:

No play football today.

Me no see Mariana.
Don’t look my card.
Don’t talk to that one.
50 Instructed Second Language Acquisition
The first two utterances make use of the ‘no + V’ pattern, while the second
two make use of ‘don’t + V’. An obligatory occasion analysis of these data
would show a 50 per cent level of accuracy for negatives. However, a closer
look reveals that ‘no + V’ is used in statements, while ‘don’t + V’ is used in
commands. The two negative forms are being used systematically to perform
different functions.
Form—function analyses, then, help to show the inner logic of the learner’s
mental grammar. Studies by Huebner (1979; 1983) and Schachter (1986)
have revealed the hidden system of learners’ grammar/, once a form—function
analysis is applied. For example, Huebner shows hoy an uninstructed Hmong-
speaking adult uses an article system at different stages of development of L2
English. Initially, the learner used ‘da’ to mark nouns as specific and known
to the hearer. Nouns that constituted the topic of an utterance were not
marked, however. Next, the learner used ‘da’ more widely with all nouns,
while later still, he used ‘da’ in conformity with native-speaker usage. Huebner
was able to show how the system was gradually revised so that it accorded
with the form—function distribution of English. He refers to form—function
analysis as ‘the dynamic paradigm’ and observes: ‘Although an approach
which looks at only those morphemes found in Standard English obligatory
contexts can tell us when morphemes are acquired with respect to one another, it
may not be the most insightful approach to the question of how they are
acquired’ (Huebner, 1979:22). Huebner is right. Form—function analyses
provide us with windows for viewing the process by which learners build their
grammars. Arguably, however, a full understanding of interlanguage develop¬
ment requires both the order-of-acquisition approach afforded by performance
analysis and Huebner’s dynamic paradigm.
The empirical research carried out using these techniques helped to provide
answers to important questions. It showed that LI transfer was not necessarily
the dominant factor in L2 acquisition and it also showed that the process by
which learners construct their knowledge of a L2 was not substantially affected
by age. In other words, the main objections to the L2 = LI hypothesis
discussed above were removed.

Interlanguage Theory

The theory that motivated and fed off the empirical research become known
as interlanguage theory, after the term coined by Selinker (1972). It is a
constantly evolving theory, having changed considerably since its initial for¬
mulation. It is, therefore, not any easy task to produce an accurate account of
the theory. The aim here is to provide a brief and composite account of the
central premises of early interlanguage theory. The reader who wants a more
detailed description together with an account of differences in the positions
adopted by individual researchers and the way these have evolved is referred
to McLaughlin (1987).6
 Naturalistic and Classroom Learning 51
The main premises of interlanguage theory are:

(1) The learner constructs a system of abstract linguistic rules which underlies
comprehension and production.
The system of rules is referred to as ‘an interlanguage’. The learner draws
on these rules in much the same way as the native speaker draws on linguistic
competence. The rules enable the learner to produce novel sentences. They
are also responsible for the systematicity evident in L2 learner language. An
interlanguage is ‘a linguistic system ... in its own right’ (Adjemian, 1976).
As such it is a natural language and is entirely functional.

(2) The learner’s grammar is permeable.

The grammar' that the learner builds is incomplete and unstable. It is
amenable to penetration by new linguistic forms and rules, which may be
derived internally (i.e. by means of transfer from the LI or overgeneralization
of an interlanguage rule) or externally (i.e. through exposure to target language
input).

(3) The learner’s competence is transitional.

As a result of the permeability of an interlanguage system learners rapidly
revise it. They pass through a number of stages in the process of acquiring the
target language. Each stage constitutes ‘an interlanguage’ — or, in Corder’s
(1967) terms a ‘transitional competence’. The series of stages together com¬
prise the ‘interlanguage continuum’. Of course the stages are not discrete
but overlap because every part of an interlanguage is subject to constant
revision. _ .
Interlanguage can be seen as a restructuring or a recreating continuum
(Corder, 1978b). According to the former the starting point of L2 acquisition
is the LI; the learner gradually substitutes target language for mother-tongue
rules. According to the latter, the learner starts from some basic simple
grammar which is independent of the LI. Most interlanguage theorists see
interlanguage as a recreating continuum, although they do not eliminate the
possibility of transfer.

(4) The learner’s competence is variable.

At any one stage of development the language produced by-learners will
display systematic variability. This variability reflects the particular form-
function correlations which comprise the rules of the learner’s grammar at
that stage of development. The learner’s competence must be viewed as
heterogeneous rather than homogeneous (cf. Tarone, 1983).
This principle is more contentious than the previous ones. There is in fact
considerable disagreement about how best to characterize the nature of an
interlanguage system. Adjemian (1976) and, more recently, Bialystok and
Sharwood-Smith (1985) and Gregg (forthcoming) reject a variable competence
account of interlanguage. Nevertheless, this principle is able to account for
the insights provided by form-function analyses.
52 Instructed Second Language Acquisition
(5) Interlanguage development reflects the operation of cognitive learning
strategies.
The process by which interlanguages are constructed has been explained in
various ways. One type of explanation identifies a number of cognitive learning
processes such as LI transfer, overgeneralization and simplification (Cancino
et al., 1974). According to this view, the L2 learner does not necessarily
utilize the same language-acquisition device as the child. The similarity between
LI and L2 acquisition lies in the process of hypothesis-formation and testing.
Hypothetical rules, formulated on the basis of learning strategies, are tested
out in comprehension and production and amended if understanding is defec¬
tive or if the utterances produced fail to communicate.
An alternative account credits the L2 learner with a language-acquisition
device. According to this view, L2 learning, like LI learning, involves the
learner discovering how the general principles that constitute any learner’s
innate knowledge of language are realized in the target language (cf. Cook,
1985).

(6) Interlanguage use can also reflect the operation of communication

strategies.
When learners are faced with having to communicate messages for which
the necessary linguistic resources are not available, they resort to a variety of
communication strategies (Tarone et al., 1976). These enable them to com¬
pensate for their lack of knowledge. Typical communication strategies are
paraphrase, code-switching and appeals-for-assistance.

(7) Interlanguage systems may fossilize.

Selinker used the term fossilization to refer to the tendency of many
learners to stop developing their interlanguage grammar in the direction of the
target language. Instead they reach a plateau beyond which they do not
progress. This may be because there is no communicative need for further
development. Alternatively it may be because full competence in a L2 is
neurolinguistically impossible for most learners (Selinker and Lamendella,
1978). Fossilization is a unique feature of interlanguage systems.
Interlanguage theory provided an explanation for how both children and
adults acquired a L2. Together with the empirical research which it supported
and was supported by, it led to a number of specific proposals that had a
considerable impact on language pedagogy.

The Applications
Most of the applications proposed on the basis of the results of interlanguage
studies were not revolutionary. They corresponded quite closely to the kinds
of implications drawn from LI acquisition research found in Newmark and
Reibel’s proposals. This was not surprising, given that the central premises of
interlanguage theory (i.e. that L2 learner language is rule-governed and that
 Naturalistic and Classroom Learning 53
learners pass through a series of developmental stages as they test out hypoth¬
eses about the target language) paralleled the premises of LI acquisition
theory. However, the proposals based on interlanguage theory had a far
profounder effect on language teaching than did those of Newmark and
Reibel.
The basis lor the application of interlanguage theory was the conviction that
it was necessary and possible to conduct teaching in such a way that it
corresponded to the way L2 learners learn. Corder (1976) commented: ‘Ef¬
ficient language teaching must work with rather than against natural processes,
facilitate and expedite rather then impede learning. Teachers and teaching
materials must adapt to the learner rather than vice-versa.’ In other words,
there was a strong conviction that classroom learning would be more successful if
it more closely resembled naturalistic L2 learning. The main proposals related
to three aspects of language teaching; (1) remedial procedures, (2) error
treatment and (3) the organization of the syllabus. These are discussed in turn
below.

Remedial procedures The use of error analysis as a basis for remedial

language teaching was nothing new. Language teachers had for a long time
carried out informal error analyses as a means of assessing what their students
had learnt and of diagnosing areas of continuing difficulty as a basis for later
remedial activity. This generally consisted of re-teaching the problem areas
using the same methods and even the same materials as in the initial teaching.
As Corder (1974) pointed out, there was no obvious reason why the re¬
teaching should work when the first teaching had failed. What was missing
from this informal approach was any understanding of the nature of the
learners’ difficulties. Corder argued that it was insufficient simply to describe
and classify the errors. It was also essential to identify the psycholinguistic
causes of the errors.
The procedure that Corder advocated, however, was not so very different
from existing practice. Error analysis was to be used to provide a detailed
description of the learner’s interlanguage at a particular developmental point.
This description could then be compared with a standard description of the
target language and the differences identified. Corder claimed: ‘It is the
account of the precise nature of these differences which gives us the information
which enables us to ‘correct’ the language learner’s errors in a systematic
fashion in our remedial teaching’ (1974:54).
Corder admitted that suqh a procedure was limited because it provided no
indication of the gaps in the learner’s communicative competence, as error
analysis only dealt with the language code. There were other major problems,
however, which Corder’s proposal did not consider. How was the teacher to
undertake the re-teaching of problem areas in a way that took account of the
psycholinguistic causes of the learners’ errors? How was the teacher to take
account of the inter-learner variation that would almost certainly be revealed
by the error analysis? These were major issues, which the remedial teaching
54 Instructed Second Language Acguisitlon
movement never really came to grips with.7 It became apparent that remedial
language teaching would make no headway until the more general problem of
how to treat learner errors was dealt with.

Error treatment Whereas remedial language teaching involves a post-hoc

attempt at dealing with errors, error treatment — as the term is generally used —
refers to attempts to handle errors concurrent with their appearance in learner
language.
Traditional treatments of learner error consisted of the immediate correction
of any utterance containing a deviant form by the teacher (see chapter 4).
Interlanguage researchers argued that it was poindess correcting errors which
were the result of the learner’s attempts to communicate beyond her existing
resources i.e. those errors that were the result of communication strategies.
They also argued that errors were inevitable and an integral part of the
process of L2 acquisition, reflecting the active way in which the learner tested
out hypotheses about the nature of L2 rules. As such they should not just be
tolerated, but welcomed in order to encourage learners to take risks.8
One view, therefore, was that errors should be ignored. Another, more
commonly held view, however, was there should be selective correction of
errors. A number of proposals were made regarding which errors ought to be
corrected:

(1) Distinguish mistakes and errors and treat differently.

Corder (1967) distinguished mistakes, accidental lapses in performance
resulting from inattention from errors, deviations from the target language
norms that occurred as a result of a lack of knowledge. Learners should be
encouraged to correct their own mistakes. The teacher should take responsi¬
bility for dealing with errors — through further teaching.

(2) Correct global, not local errors.

Burt and Kiparsky (1974) distinguished global and local errors. The
former violated rules involving the overall structure of a sentence, while the
latter involved mistakes in a particular constituent of a sentence. James (1974)
made a similar proposal. He developed the concept of error gravity based on
the number and the nature of the rules that were transgressed. The graver an
error, the more it warranted correction.

(3) Correct errors that affect the overall comprehensibility of an utterance.

This suggestion was closely related to (2), in so far as the motivation for
recommending correction of global and grave errors was the belief that they
interfered with communication more seriously. The criterion of ‘comprehen¬
sibility’, however, is broader. Even local and non-grave errors can cause
communication breakdown in some contexts.

(4) Correct stigmatized errors.

Native speakers respond differently to different errors. Certain kinds of
 Naturalistic and Classroom Learning 55
errors can be considered stigmatized because they invoke highly negative
responses from negative speakers (e.g. double negatives).

(5) Correct errors relating to the learner’s next stage of development.

The assumption underlying this proposal was that the learner would be able
to respond to such corrections because she was developmentally ready to
discover the target language rule. There was, of course, the problem of
determining whether a particular learner was ready or not.

In addition to these proposals regarding what errors to correct, there were

others relating to when and how to correct (cf. Hendrickson, 1978, for an
excellent overview of error treatment).
Researchers wno emphasized the variable nature of interlanguage took a
special interest in error treatment. Dickerson (1975), for example, applied the
results of her study of the systematic variability of the phonological rules
acquired by Japanese learners of L2 English by suggesting that teachers
needed to recognize that learners’ production of specific sounds will vary
according to the linguistic context of the sound and also according to the task
they are performing. She encouraged teachers to expect variability, as this was
the norm, not the exception. Teachers were also advised to abandon the
notion that an utterance was ‘right’ or ‘wrong’ and to acknowledge that there
were degrees of attainment.

Syllabus organization The most radical proposals based on interlanguage

theory concerned the organization of the teaching content. The recognition
that the learner passed through a series of developmental stages brought into
question the traditional principles of selection, grading and sequencing. There
seemed little sense in ordering the teaching content according to external
criteria of linguistic difficulty if the resulting order did not conform to the
built-in syllabus of the learner. A more sensible approach was to design a
teaching syllabus that was compatible with the learner’s syllabus.
One proposal for achieving this was to sequence the content of the syllabus
so that it mirrored the learner’s developmental sequence. Nickel (1973) was
one of the first to suggest that the syllabus should accommodate the progressive
complication evident in interlanguage by teaching the sequence of approximative
systems. Bailey et al. (1974) and Valdman (1978) put forward similar ideas. In
order to implement this proposal, however, it would be necessary actually to
teach erroneous forms, as the early stages of interlanguage involve the con¬
struction of simplified rules not found in the target language. Most applied
linguists considered this proposal pedagogically unworkable, because it was
contrary to both the teacher’s and the learner’s concept of teaching — learning.
Widdowson (1975) — drawing on Labov’s studies of stylistic variation -
pointed out that it was natural for teachers and learners to refer to norms of
correctness in the classroom as this constituted a formal situation.
The second proposal was far more workable and, in fact, has had consider¬
able impact on language pedagogy. It was argued that classroom teaching
56 Instructed Second Language Acquisition
should be focused exclusively on communication and that no attempt should
be made to teach specific parts of the code. The syllabus should attempt to
control the level of difficulty of the communicative demands placed on the
learner but the learning of grammar was to be left to the learner. Corder
(1976:78) commented: ‘Instead, then, of grading the linguistic material that
we expose the learner to, we should consider grading the communicative
demands we make on him, thereby gently leading him to elaborate his approxi¬
mate system.’ This was in agreement with Newmark and Reibel’s earlier
proposals that language teachers should stop ‘interfering’ with the process of
language learning. As Corder frequently pointed out, the proposal was also
compatible with the educational arguments that were being advanced for
group work and discovery learning. Teaching based on communication activities
in contrast to language exercises seemed especially suited to learner-centred
instruction.
Corder was, in fact, recommending a return to a ‘natural method’
i.e. the use of language in real situations to perform authentic communicative
functions. This was justified on other grounds quite apart from interlanguage
theory. It reflected the growing conviction among applied linguists that
Chomsky’s view of language was too narrow and that the real goal of language
teaching should be communicative competence, not just linguistic
competence. That is, language teaching needed to ensure that learners could
not only use English correctly but also appropriately and that they could
perform in a variety of situations requiring different types of discourse. By
claiming that a knowledge of grammar grew in response to communicative
needs, Corder was simply helping to push the teaching profession where it
was going in any case. When provided with opportunities for communication,
the learner would not only learn how to communicate but, in the process,
would also acquire a knowledge of the linguistic system.
It is difficult to assess the contribution of interlanguage studies to language
teaching. Language pedagogy was changing in any case, in response to theories
of language that emphasized language use over linguistic form. It is probably
true to say, however, that the interlanguage research of this period functioned
as a catalyst to the process of change. It also contributed to the construction of
a theory of L2 acquisition - the Monitor Model — which was much more
cohesive and complete than early interlanguage theory and which was to have
a direct and substantial effect on language teaching. We turn now to the
contribution of Stephen Krashen.

The Monitor Model

The Monitor Model is undoubtedly the best known theory of L2 acquisition.

It has been popularized in a number of articles in various journals - both
 Naturalistic and Classroom Learning 57

academic and educational — and in four books (Krashen, 1981; 1982; 1985:
Krashen and Terrell, 1984). For a while the theory dominated the field of L2
acquisition to such an extent that researchers felt compelled to measure their
results and theoretical positions against those covered by the Monitor Model.
This is a testimony to the lucidity, simplicity and explanatory power of
Krashen’s theory.

Table 3.3 An overview of the Monitor Model

Hypothesis Description

1 Acquisition—learhing Acquisition is the subconscious process by which

linguistic competence is developed as a result of
using language for real communication. ‘Learning’
is the conscious process by which metalingual
knowledge of a language is developed through
study. ‘Acquired’ and ‘learnt’ knowledge are stored
separately; ‘learnt’ knowledge cannot be converted
into ‘acquired’ knowledge.
2 Natural order Grammatical structures are ‘acquired’ in a
predictable order, which is the same for adults and
children and for learners with different Lis. It is
evident whenever the focus is on communication.
However, it is not evident in language activities such
as grammar tests when the focus is on form.
3 Monitor ‘Learnt’ knowledge can be used to edit utterances
generated by means of ‘acquired’ knowledge either
before or after they occur in the output. Monitoring
can only take place when there is sufficient time, the
focus is on form and the rule is known. Monitoring
is restricted to relatively simple rules such as regular
past tense or 3rd person -s.
4 Input This explains how ‘acquisition’ takes place. When
learners are exposed to input that contains
grammatical features a little beyond their current
level, those features are ‘acquired’. Krashen
emphasizes that ‘acquisition’ is the result of
comprehensible input and not production. Input
is made comprehensible because of the help
provided by the context. Krashen also emphasizes
that the input only needs to be roughly and not
finely tuned (i.e. there is no need to try deliberately
to expose the learners to samples of a specific
grammatical feature).
5 Affective Filter This concerns attitudinal variables that affect
‘acquisition’. Learners with optimal attitudes have a
low affective filter. This means they try to get more
input and are more receptive to the input they get.
The affective filter controls the rate and ultimate
level of success of ‘acquisition’.
58 Instructed Second Language Acguisition
No attempt will be made here to provide a thorough account of the model
(cf. Ellis, 1985a; Mclaughlin, 1987), the details of which have changed over
the years in response to criticisms. A summary of the five hypotheses which
constitute the core of the theory is given in table 3.3 for those readers who are
not already familiar with the model. In accordance with the main aim of this
chapter, the focus will be on the view of classroom language learning and
teaching that Krashen derived from the theory and the theoretical objections
which the model aroused.

Krashen’s View of Classroom Language Learning and

Teaching
Krashen has always maintained a keen interest in classroom language learning
and teaching. The Monitor Model was constructed with the classroom in
mind, even though it was developed to account for the results of studies of
naturalistic L2 acquisition in the main. Krashen has expounded the applications
of the model to language teaching—learning in considerable detail (cf. in
particular Krashen, 1982 and Krashen and Terrell, 1984). His main proposals
are:

(1) The principal goal of language teaching is to supply comprehensible

input in order to facilitate ‘acquisition’.
Krashen (1982:64) writes that ‘the defining characteristic of a good teacher
is someone who can make input comprehensible to a non-native speaker,
regardless of his or her competence in the target language’.
Optimal input is supplied when the teacher engages the learner in real
communication which the learner finds interesting. Krashen draws on research
into the role of motherese in facilitating LI acquisition to describe the
characteristics of simplified input that aid comprehension and, through this,
‘acquisition’. The application of the input hypothesis to classroom learning
and teaching is considered in greater detail in chapter 5.

(2) Teaching should be seen as a preparation for ‘acquisition’ in the wider

world.
Krashen argues that it is doubtful if the classroom can supply sufficient
comprehensible input to ensure successful L2 acquisition. The classroom is
more important to the beginner and the FL learner, who often are unable to
secure adequate comprehensible input in the outside world. One of the aims
of teaching must be to equip the learner to manage real-life conversations.
This can be done by teaching ‘conversational competence’ in the form of a
few well-chosen routines, discourse devices for getting the native speakers to
explain their meaning, back-channel cues and topic-changing devices.

(3) The teacher must ensure that learners do not feel anxious or are put
on the defensive.
 Naturalistic and Classroom Learning 59

The basis of this proposal is the affective filter hypothesis. The learner has
to feel relaxed and confident to ensure that the filter is down so that compre¬
hensible input gets in. Krashen argues that if teachers insist on learner
production too soon or if they correct errors in communicative activities, the
learner will be inhibited from learning.

(4) Grammar teaching should be restricted to simple forms and its goal is
to enable the learner to monitor.
Grammar teaching of any kind (inductive or deductive) is of limited value
because it can only contribute to ‘learning’ and never to ‘acquisition’. Krashen
argues that only a limited sub-set of the total rules of a language are ‘leamable’.
Complex rules such as WH questions or negatives cannot be learnt by most
students. Krashen also considers that monitoring has a limited role in language
production: ‘Not everyone Monitors. Those who do, only Monitor some of
the time and use the Monitor for only a sub-part of the grammar .. . the
effects of self-correction on accuracy is modest’ (Krashen, 1982:112). Grammar
can also be taught as ‘subject matter’, but this is not to be confused with the
main goal of language teaching.

(5) Errors should not be corrected when the goal is ‘acquisition’ but
should be corrected when the goal is ‘learning’.
Error correction has no role in ‘acquisition’ which only occurs as a result of
the learner processing comprehensible input. However, error correction can
help the learner to ‘learn’ a simple rule. Given that the main goal of teaching
is ‘acquisition’, error correction is generally to be avoided.
The ‘Natural Approach’ (Krashen and Terrell, 1984) was based on these
proposals. Its main principles (1984:58) can be summarized as:

1 The goal is communicative skills.

2 Comprehension precedes production.
3 Production emerges when the learner is ready.
4 Acquisition activities are central.
5 The affective filter needs to be kept low.

Krashen’s proposals represent a radical move away from audiolingualism. In

many respects they constitute a logical development of the pedagogic ideas
first voiced in the theory of L2 learning that supported the cognitive anti¬
method and then by applied linguists who drew on interlanguage theory. The
‘Natural Approach’ rejects any attempt to shape the main process of acquisition
through the systematic presentation and practice of the linguistic code.
The Monitor Model has been subjected to increasing criticism (McLaughlin,
1978b; Sharwood-Smith, 1981; Gregg, 1984; Barasch and James, forth¬
coming). One of the chief criticisms is that several of the hypotheses are not
falsifiable. In particular, it is not clear how Krashen’s claim that ‘learnt’
knowledge does not contribute to the development of ‘acquired’ knowledge
60 Instructed Second Language Acguisition
can be empirically tested. Similarly it is difficult to see how the input hypothesis
can be properly tested, given the vagueness of Krashen’s idea of‘input containing
structures just beyond those in the learner’s current grammar’ (see chapter 5
for a fuller discussion of this criticism). The Monitor hypothesis has been
criticized on the grounds that it is far too restricting; learners are capable of
learning and using metalingual knowledge to a far greater extent than Krashen
allows for. From the point of view of Krashen’s pedagogic proposals these
three hypotheses are central. If ‘learnt’ knowledge is convertible into ‘acquired’
knowledge, if production is important for development (contrary to Krashen’s
insistence that only comprehension counts) and if monitoring is widespread,
the theoretical foundations of the proposals are destroyed. As we will see
in subsequent chapters, there are strong grounds for questioning these
hypotheses.
The Monitor Model is inadequate in a number of ways and the pedagogic
proposals based on it cannot be fully supported. Nevertheless, Krashen has
done the teaching profession a service. He has provided a coherent set of
ideas firmly grounded in L2 acquisition research. His work has stimulated not
only a discussion of key issues in language pedagogy but, most important, has
contributed to the growth of the empirical study of classroom L2 learning
itself.

Summary and Conclusion

The aim of this chapter has been to examine a number of views about
classroom language learning which were derived from the study of naturalistic
language acquisition. A secondary aim was to provide a historical perspective
for understanding both why certain beliefs about classroom L2 learning came
about and what motivated much of the research into classroom L2 learning
discussed in the next section of this book. It was for this reason that attention
has been focused on the opinions current in the late sixties and seventies.
Naturalistic language learning, of course, has continued to serve as a basis for
theorizing about classroom language learning in the eighties and doubtlessly
will do so in the future. The main ideas derived from it, however, have not
significandy changed.
The chapter began with an account of the theory of LI acquisition that
resulted from Chomsky’s views on the nature of language and die empirical
studies which this theory stimulated. The cognitive anti-method drew directly
on these views in presenting a radical alternative to audiolingualism. It em¬
phasized the centrality of the learner and proposed that classroom language
learning would take place most efficiently if the conditions of LI acquisition
were reproduced. There was to be no direct attempt to teach the language
code. Cognitive code learning theory also drew on Chomsky’s ideas of language
but it maintained several of the key assumptions of audiolingualism - namely
that the L2 code needed to be explicitly taught and that the classroom should
 Naturalistic and Classroom Learning 61
not try to replicate the conditions of natural learning. The theory differed
from audiolingualism in that it gave primacy to the deductive learning of
linguistic knowledge and to opportunities for meaningful language practice.
Neither the cognitive anti-method nor the cognitive code method had much
impact on classroom teaching, however. They are of interest because they
represent the first attempts to challenge the certainties of audiolingual learning
theory and because, in so doing, they addressed two issues of central import¬
ance: the role of metalingual knowledge and the value of form-focused
instruction.
This chapter also reviewed the various kinds of empirical studies into L2
acquisition that were carried out in the sixties and seventies and the inter¬
language theory with which the research was associated. The research lent
support to the view that adult L2 learning was not so very different from child
LI acquisition in that transfer did not appear to be the major factor it was
previously assumed to have been and the sequence of acquisition was not
affected in any major way by age. It helped to reinforce the view of classroom
language learning articulated by the advocates of the cognitive anti-method.
Applied linguists drawing on L2 acquisition research emphasized the role of
the learner and began to make radical proposals regarding both syllabus
design and classroom methodology. In particular, the idea of attempting to
control the order in which learners acquired knowledge of the linguistic code
was further questioned. Instead it was suggested that learning should be
allowed to take place naturally in the course of using the L2 for communication.
Naturalistic L2 acquisition was seen as a model of successful learning; the
goal of language pedagogy was to reproduce the conditions that made it
successful. This led to questioning the value of a linguistic teaching syllabus
and also the efficacy of error correction.
Finally, this chapter considered the Monitor Model and its applications to
language teaching. Krashen’s proposals represented a complete break with
audiolingualism. The main goal of teaching was to provide comprehensible
input for ‘acquisition’ — the spontaneous process of picking up linguistic
knowledge from attempts at communicating in the L2. There was only a
limited role for productive practice of specific linguistic forms; there was no
role for explicit grammatical rules where ‘acquisition’ was concerned and only
a very limited role in the case of ‘learning’.
This chapter, then, has examined the various proposals about classroom
language learning which were derived from the study of naturalistic acquisition.
But how valid is this approach? This chapter concludes with a discussion of
the legitimacy of such extrapolation.
A number of researchers have expressed misgivings about applying the
results of language-acquisition studies to the classroom. Chomsky (1966),
himself, was doubtful whether either psychology or linguistics had achieved
sufficient understanding to support ‘a technology of language teaching’. For
Lightbown (1985a: 180) the position was no better after several years of
empirical research into L2 acquisition: ‘Many of the recommendations based
on second-language acquisition research have been (a) premature, (b) based
62 Instructed Second Language Acguisition
on research which was extremely narrow in scope, (c) based on overinterpret¬
ations of the data, (d) based largely on intuition or (e) all of the above.’ She
goes on to argue that the research should not be used to prescribe what
should be taught because the available evidence about acquisitional sequences
is incomplete and often contradictory. She agreed, however, that it could
inform about how not to teach.
Hughes (1983:1—2) is even more pessimistic about the utility of research:
‘It must be said at the outset that it is not at all certain that at the present time
there are any clear implications for language teaching to be drawn from the
study of second language learning.’ There is general agreement, however, that
language teachers need to keep informed about the insights and understanding
provided by the research and that language pedagogy needs to be consistent
with the results that have been obtained.
Other researchers — like Krashen — have taken a stronger stance and
argued that the study of naturalistic language acquisition provides a sound
basis for making recommendations. Corder (1980:1) believes that ‘we always
have an obligation to attempt to answer practical questions in the light of the
best available knowledge’, although he too is prepared to offer only ‘tentative
proposals’. Perhaps the best reason for going ahead with applications is that
there never will be a time when there is a generally agreed theory of L2
acquisition based on solid and enduring research. Theory constructions is a
never-ending affair. Applying the results of research, then, should take place,
although not with the expectancy that a single, incontrovertible account of the
teaching—learning relationship will arise.
The research discussed in this chapter has dealt only with naturalistic
acquisition. It resulted in ‘a more reciprocal relationship between psychology
and language pedagogy’ (Stem, 1983) - a distinct advance on the relationship
evidenced in the audiolingual learning theory where language learning was
treated as the same as any other form of learning. However, the situational
differences between naturalistic and classroom domains are bound to give rise
to doubts about whether the results of research in the former are fully
applicable to the study of the latter (doubts which received expression in the
cognitive code learning theory). It is true to say, however, that the seeds of
classroom research - the subject of the next few chapters - would never
have been sown but for the essential preparation of the ground by the
naturalistic research and theoretical perspectives discussed in this chapter.
These helped to raise questions about the nature of instructed language
learning which could only be answered by entering the classroom itself.

NOTES

1 The inability of the input to provide the child with the data needed to arrive at a
grammar of the language has become known as the poverty of the stimulus
argument. Initially Chomsky argued that the input was impoverished on the
 Naturalistic and Classroom Learning 63
grounds that parental speech to children was degenerate (i.e. consisted of sentence
fragments etc.). Subsequent research into motherese has shown that it is, in fact,
remarkably well formed (Cross, 1977). The poverty of the stimulus argument can
still be maintained, however, by claiming that discovery of certain highly abstract
rules is not possible without negative feedback, which is in fact only rarely
supplied by parents.
2 Chomsky himself was dubious about whether the study of child language acquisition
was of any great value. He distinguished development — the real-time learning
of language — and acquisition — language learning unaffected by other factors
such as memory and processing capacity. Acquisition in this sense is something
abstract, not amenable to direct enquiry through the analysis of performance data.
3 In other words, for Wilkins acquisition and learning were synonomous with informal
and formal language learning contexts. Krashen (cf. Krashen and Seliger, 1976),
however, argued that ‘acquisition’ and ‘learning’ (defined in process terms) could
both occur in both contexts.
4 The age at which cerebral plasticity is lost as a result of lateralization is controversial.
Krashen (1973) reviewed the research which has investigated this issue and arrived
at the conclusion that cerebral dominance was largely complete by the age of 5 yrs
— well before puberty.
5 There are still only a few case studies of classroom learners (e.g. Ellis, 1984a). One
of the inhibiting factors is the problem of obtaining adequate data in a classroom
setting. Individual tutored learners say very little and what they do say is often not
genuinely ‘communicative’.
6 McLaughlin (1987) points to two current developments in interlanguage theory.
One involves the comparison of the process of L2 acquisition with that of the
creolization of pidgins and emphasizes the universality of strategies of simplification
and complexification. The other follows developments in Universal Grammar in
arguing that the learner’s L2 knowledge derives from the process of setting the
parameters of abstract linguistic principles.
7 In fact, remedial teaching has fallen out of fashion. Current attempts at applying
L2 acquisition theories make no mention of it.
8 It was often pointed out that the term ‘error’ was misapplied, as the learners’
utterances were ‘grammatical’ with reference to their own transitional competences.
4 Classroom Process
Research

Introduction: A Historical Perspective

Classroom process research is concerned with the careful description of

the interpersonal events which take place in the classroom as a means of
developing understanding about how instruction and learning take place. As
Van Lier (1988:71) puts it: ‘Second-language classroom research, in studying
the processes and circumstances of second-language development, aims to
identify the phenomena that promote or hamper learning in the classroom.’
The principal method of studying L2 classroom activity is observation,
defined by Allwright (1988a:xvi) as ‘a procedure for keeping a record of
classroom events in such a way that it can later be studied’. Much of the
research (particularly the early research) reported in this chapter has not been
theory-led; rather it has been exploratory and illuminative in nature. However,
as Van Lier points out, observation cannot take place unless prior decisions
have been taken about what to look for. There is no such thing as atheoretical
research.
Classroom process research owes much to observational research in education
and the social sciences. Researchers interested in the L2 classroom drew
heavily on the instruments and techniques developed by educationalists and
sociologists. The rejection of‘method’ in favour of the detailed study of actual
classroom behaviour had taken place in general educational research some
time before it did in L2 classroom research. Grittner (1968, cited in Allwright,
1988a), for instance, pointed out that large-scale studies of the kind carried
out by Scherer and Wertheimer (1964) and Smith (1967) had lost favour in
educational circles. He summarizes their failure in this way: ‘In short, a half
century of such ‘research’ has told us almost ‘nothing’ about the relative
superiority of one educational strategy or system over another.’ Thus, as
Allwright (1983) has noted, the study of language teaching—learning has
tended to lag behind general educational research.
Developments in educational research involved a switch from a faith in
measurement to a faith in observation. This was motivated initially by a desire
to identify the characteristics of different teaching styles (often with a view to
prescribing those behaviours that constituted the ‘best’ style) and, increasingly,
 Classroom Process Research 65
by the recognition that very little was known about what actually happened
inside a classroom. As Stubbs (1976:69) put it: ‘There was an enormous
psychological literature on ‘learning theory’ based largely on experimental
situations, but very little is known about what and how children learn in
schools. The only way to discover this is to observe children in classrooms.’
Initially, observational research was carried out by means of interaction
analysis. Numerous observational instruments were designed for coding what
teachers and pupils said and did in the classroom. Simon and Boyer (1967)
refer to over seventy such coding schemes. Increasingly, this method was
abandoned as a research tool (but not as a resource in teacher training) in
favour of anthropological or ethnographic research (e.g. Mehan, 1974). The
emphasis here wqs on detailed fieldwork of the kind that anthropologists
undertook in documenting and understanding the beliefs and behaviours of
remote tribes. In another branch of classroom research, educationalists set out
to discover the structure of classroom dialogue (e.g. Bellack et al., 1966),
describing the various ‘moves’ that made up teaching ‘cycles’ and identifying
the hierarchical nature of classroom discourse.
Sociologists have also shown an interest in classrooms, focusing on the
different types of social organization found there and the ways in which social
control and personal relationships are enacted (e.g. Walker and Adelman,
1972). The difference between the norms of social interaction in home and
classroom situations have been identified and described and the potential
effects this difference has on educational success discussed (Labov, 1969).
The complex nature of classroom social settings is emphasized. Much of what
takes place can only be understood if the researcher has access to the same
‘shared context’ as the classroom participants themselves.
Classroom process research reflects all these various types of enquiry. As in
educational research, interaction analysis was the first method employed,
perhaps because it involved less of a break with measurement. But researchers
were quick to take advantage of the new tools provided by anthropologists,
discourse analysts and sociologists. The seventies and the eighties saw an
increasing array of research techniques directed at describing and understanding
the L2 classroom.
Although the techniques employed by L2 classroom researchers owe much
to those used in education and social science, the questions which researchers
sought to answer were more parochial; they were derived from views about
language teaching and learning. Much of the early research focused on
investigating specific pedagogic practices. For example, Politzer (1970) used
observation to try to establish whether behaviours such as ‘direct reference to
textbook’, ‘use of visual aids’ and ‘student/student interaction’ counted as
‘good’ or ‘bad’ teaching practices. Teacher trainers developed instruments to
help them assess teachers’ teaching and also tried to develop instruments that
would provide them with a technical language to ‘designate the teaching
behaviour in second language settings’ (Fanselow, 1977). The aim was to
facilitate ‘technical instruction’. Many of the initial descriptions of the language
66 Instructed Second Language Acquisition
classroom were made by teacher trainers rather than by researchers and the
tools they developed reflected their particular needs.
It would be wrong to suggest that the research in the early seventies was
uninformed by views of how L2 learning took place, but it is probably correct
to assert that learning took second place to teaching. Allwright (1988a: 197)
describes the classroom observation that took place during this period as ‘a
procedure looking for a purpose’. It was not until the late seventies and
eighties that researchers began to address more directly issues arising out of
work in L2 acquisition. Researchers became less interested in evaluating
pedagogic practice than in providing comprehensive and accurate accounts of
aspects of classroom behaviour which were deemed significant on the basis of
L2 acquisition theory. There were detailed studies of aspects of the teacher’s
language — for example, the treatment of learner error (e.g. Allwright, 1975)
and of teacher-talk (e.g. Henzl, 1973; Gaies, 1977). Studies of the student’s
language were less common — partly because the teacher was still seen as
central and partly because in many classrooms there was not much student-
talk to study — but one study often cited is Seliger’s (1977) investigation of
the role of language practice. In the late seventies and early eighties attention
switched to the study of discourse in L2 classrooms, motivated again by work
done in L2 acquisition. This suggested that interaction served as the matrix in
which grammatical competence was acquired. It should also be noted that L2
acquisition research provided not only the research topics but, increasingly,
the actual categories of analysis (Allwright, 1988a:239).
A great variety of classrooms have been investigated — foreign language
classrooms, ESL classrooms, immersion classrooms, bilingual classrooms,
mainstream classrooms containing limited-proficiency students. The learners
in these classrooms have come from widely differing backgrounds and the
teachers have been varied in experience. Amongst this diversity it is possible
to identify common ground with regard to both the general approach and the
classroom language behaviours that have been studied. The approach empha¬
sizes description but not as an end in itself, rather as a means to understanding
the factors that contribute to L2 learning in a classroom context.1
The rest of the chapter provides a review of a selection of classroom
process research. No further attempt to adopt an historical perspective will be
made. The chapter begins with a brief account of the principal research
procedures used and then examines, in turn, what has been discovered about
the teacher’s language, the learner’s language and classroom discourse.

The Methodology of Classroom Process Research

Long (1980) published a comprehensive review of the different methodological

approaches for investigating classroom processes. He distinguishes two broad
approaches: interaction analysis and anthropological observation. The former
 Classroom Process Research 67
typically leads to quantitative research involving the measurement of the
frequencies with which classroom events of various kinds occur. The latter
involves relatively unstructured observation of classrooms in the sense that
what is observed is not predetermined by the researcher but depends on the
observer’s developing understanding of what is significant.

Interaction Analysis
Interaction analysis involves the use of some form of schedule consisting of
sets of categories for coding classroom behaviours. In a category system
each event is coded each time it occurs. In a sign system each event is
recorded only onge within a fixed time-span irrespective of how many times it
actually occurs. A rating scale is used after the period of observation to
estimate how frequently a specific type of event occurred.
Initially, the schedules used were those that had been developed for content
classrooms. The best known was that of Flanders (1970). This included
categories of teacher behaviour (e.g. ‘Teacher asks questions’ and ‘Teacher
accepts feelings’) and also learner behaviour (e.g. ‘Pupil responds’). Shortly
afterwards, expanded interaction schedules for use in the language classroom
were developed (e.g. Moscowitz, 1970) in order to take account of the special
kinds of behaviours found in language instruction. The content of these early
interactional systems was not motivated by any strong theory that predicted
which behaviours were important for language learning. Instead, it reflected
opinion then current about what processes ought to occur. For example, the
Marburg form (reported in Freudenstein 1977) included categories such as
‘phases of instruction’ in which it was assumed that the basic pattern of a
lesson would consist of ‘warming up’, ‘presentation’, ‘learning’ and ‘using’.
Subsequent schedules (e.g. Fanselow, 1977; Allwright, 1980) were more
sophisticated. They aimed to provide a much more comprehensive description
of classroom interaction and were less tied to a particular view of how
teaching should take place. Observation schedules continued to become more
complex as they tried to take account of the findings of more theoretical
analyses of classroom discourse (e.g. Sinclair and Coulthard, 1975). Such
schedules attempted to preserve the discourse structure of a lesson, rather
than segmenting it into discrete units (e.g. Riley, 1977).2
The increasing complexity of observational schedules was the result of two
factors. First, one of the original motivations for observing classroom behaviours
was teacher training. The observational categories, therefore, derived from the
prevailing views about what should happen in a language lesson and were
designed to sensitize teacher trainees to the kinds of behaviours they should
incorporate into their own teaching. The other factor that led to an increase in
complexity was the use of audio and video recordings of lessons. The early
schedules were designed for use in ‘real time’. That is, the observers sat in on
lessons and coded each event as it occurred. It was essential that the schedules
were manageable under such conditions and, therefore, they contained a small
68 Instructed Second Language Acgulsitlon
number of ‘low inference’ categories. Later schedules were designed for use
with recorded or transcribed data. It became possible to code larger numbers
of categories and also to develop much more explicit descriptions of the
categories used. Systems that attempted some kind of discourse analysis (e.g.
Allwright, 1980; Bowers, 1980) necessitated retrospective interpretation.
Interaction analysis systems are useful in teacher training. Also the task of
designing a schedule has itself proved valuable in opening researchers’ eyes to
the nature of classroom processes. There has been a proliferation of systems,
however. Long (1980) lists twenty-two and there have been new ones since.
The status of many of the categories is questionable. Long (1980:10) comments:
‘The value of analytical systems must ultimately depend on the significance
for teaching and learning of the categories they contain.’ Each category
constitutes an untested claim about teaching — learning. Also the categories
tend to focus on the teacher’s behaviour and to neglect the learner’s. A more
serious criticism is voiced by McLaughlin (1985:149), who notes that because
the behaviour of the teacher and the learners is treated separately, information
is lost about ‘the sequential flow of classroom activities’ and how participants
provide each other with feedback. The product of interaction analysis is
disconnected tallies of behaviours without any general picture of how the
classroom process is negotiated.

Anthropological Observation
Anthropological observation provides a more qualitative approach to the de¬
scription of classroom processes. The general aim is to describe every aspect
of those experiences considered significant, in as much detail and as openly as
possible. Emphasis is placed on building up a holistic picture. The idea is to
observe classroom events in the same way as anthropologists observe primitive
societies. Anthropological observation typically involves the collection of massive
amounts of data which then have to be sifted and reported in a coherent
manner.
A number of procedures are to be found (cf. Long, 1980). Participant
ethnography involves the researcher taking a regular part in the activities
under study. This approach is evident in the journal records of classroom
learning experiences kept by learners (cf. Bailey, 1983). Non-participant
observation sets the researcher outside the classroom events being observed.
A variety of techniques can be used to obtain data - note-taking, interviewing,
administering questionnaires, eliciting personal opinions by means of ratings
and rankings and analysing relevant documents (e.g. students’ homework and
teachers’ handouts). This approach has been used by researchers in the
United States interested in examining the behaviours of ethnic minority
children in mixed classrooms (e.g. Phillips, 1972). Constitutive ethnography
aims to make explicit the way in which classroom participants succeed in
creating and managing the events in which they take part. Mehan (1979)
employed repeated viewings of videotaped lessons and also solicited the
 Classroom Process Research 69

participants’ reactions to the events that took place, in order to examine the
way behaviour is organized and how the interactions were accomplished.
Microethnography focuses on ‘particular cultural scenes in key institutional
settings’ (Erickson and Mohatt, 1982; quoted in Mitchell, 1985). It has a
similar aim and employs similar techniques — such as the detailed analysis of
audiovisual records of instructional events — to those used in constitutive
ethnography, but restricts the object of study to delimited contexts of activity
which are recognized as real by the participants (e.g. ‘getting ready for a
break’ or ‘setting homework’). Anthropological studies of these kinds generally
present their results in a discursive and illustrative manner.
Anthropological research involves more than observation. It also typically
includes collection of introspective and retrospective accounts of classroom
events. This is motivated by the desire to obtain different perspectives on the
same phenomena. Researchers in this tradition recognize that there is no
single ‘true’ interpretation of an event. The teacher’s, the learner’s and the
researcher’s views about what took place may all differ. Classroom processes,
therefore, can best be understood if the principle of triangulation is adhered
to. Anthropological research is distinguished from interaction analysis not only
in the manner in which observation is carried out but also by the use of verbal
report data (Cohen, 1987).
There are several advantages of anthropological research. Gaies (1983a)
lists three: (1) it can account for learners who do not participate actively in
class; (2) it can provide insights into the conscious thought processes of
participants; and (3) it helps to identify variables which have not been previously
acknowledged. However, the anthropological approach also has a number of
limitations (Long, 1980). It depends to a great extent on the skills of the
researcher, it is time-consuming and it is difficult to generalize the results
obtained because we can never be certain what is idiosyncratic and what is
common to many classrooms. Ogbu (1981; referred to in McLaughlin, 1985)
also criticizes anthropological studies of classroom interaction on the grounds
that they ignore the social status of the learners and thus fail to acknowledge
that minority children often develop survival strategies which are ‘more or less
incompatible with the demands of classroom teaching and learning’. Ogbu
was concerned with second- rather than foreign-language classrooms, but his
general point — that anthropological research neglects superordinate vari¬
ables — may also hold true for the latter.
A distinction is often drawn in classroom process research between objective
research based on observation schedules designed to provide an accurate and
reliable record of behaviours and subjective research that emphasizes the
interpretative, value-laden nature of all description and seeks to guard against
partial accounts by means of painstaking research from a number of different
perspectives. Objective research is characterized as hypothesis-testing and
subjective research as illuminative (Allwright, 1983). In a field of enquiry as
complex as classroom language processes, this distinction becomes blurred.
The research to be considered in this chapter - whether based on interaction
70 Instructed Second Language Acguisltion
systems or anthropological techniques — has often been ad hoc, unmotivated
by a well-defined theory of teaching—learning. It would be more honest,
therefore, to treat the information it has supplied as illuminative in nature,
irrespective of whether the research purports to be objective or subjective.

The Teacher’s Language

The teacher typically dominates classroom interaction. Holmes (1978) provides

a lucid account of how this takes place. The role relationship between the
teacher and the pupils is asymmetrical. Teachers have a superior status
because they are older and because they are presumed to know more about
the subject they are teaching. Teachers also have ultimate responsibility for
managing classroom interaction — ensuring that it proceeds in an orderly way
despite the large number of participants. Because of these factors, teachers
talk more than the other participants (around 70 per cent of the time). They
also typically control the turn-taking by adopting a role similar to that of a
chairperson in a formal meeting. They control the topic of the discourse by
means of questions. It is not surprising, therefore, that much of the research
into classroom language processes has focused on the teacher’s language.
This chapter will consider three aspects of the teacher’s language: error
treatment, ‘input’ features and ‘interactional’ features. The term ‘input’ is
used to refer to the formal adjustments that teachers make to their speech
when talking to L2 learners; ‘interactional’ features consist of the special
functional properties of teacher talk (cf. Long, 1981).

Error Treatment
The teacher has a traditional right to provide learners with feedback regarding
the correctness or appropriateness of their responses. In the case of the
language classroom, the focus of the treatment is likely to be on the formal
correctness of the learners’ productions. Error treatment in language classrooms
is often widespread. It is frequently not only welcomed but demanded by
learners (Cathcart and Olsen, 1976).
The study of error treatment requires an operational definition. Researchers
need to be able to identify incidences in a lesson. This has proved to be less
easy than expected. One problem is that teachers do not correct every error
that occurs. It is useful, therefore, to distinguish T-errors and U-errors
(Edmondson, 1986). A T-error is any discourse act which the teacher treats
explicitly or implicitly as erroneous, while a U-error is any learner utterance
which deviates from target language norms. Researchers usually elect to
examine T-errors (e.g. Allwright, 1975; Chaudron, 1977).
The search for an operational definition of error treatment is characteristic
of much of the research to date. Long (1977) proposes a distinction between
 Classroom Process Research 71
‘feedback’ and ‘correction’. He suggests that the term ‘feedback’ be used in
the case of teachers’ attempts to supply learners with information about the
correctness of their productions, while ‘correction’ should be used to refer to
the result of feedback (i.e. its effects on learning). Chaudron distinguishes
four types of ‘treatment’:

1 Treatment that results in the learners ‘autonomous ability to correct

herself on an item’.
2 Treatment that results in the elicitation of a correct response from the
learner.
3 Any reaction by the teacher that clearly transforms, disapprovingly
refers to or demands improvement.
4 Positive or negative reinforcement involving expressions of approval or
disapproval.

The broadest definition is provided by (3); (4) is very limiting as it restricts the
object of enquiry to occasions when the teacher draws explicit attention to a
learner error; (1) and (2) are concerned with the relationship between ‘feedback’
and ‘correction’. However, (1) cannot be determined within the context of a
single lesson as it involves investigating the effects of error treatment on long¬
term learning. Most researchers adopt (3) as a basis for research.
More recently, arguments have been put forward for widening the scope of
error-treatment studies to cover the whole concept of repair. Such an approach,
it is claimed, would provide ‘a process-centred approach to error’ in place of
one that concentrates on the discrete products of linguistic failure (Bruton
and Samuda, 1980). The study of repair derives from the work of Schegloff,
Jefferson and Sacks (1977), who investigated how native speakers are able to
sort out potential communication problems. They describe repairs in terms of
the production of the trouble source, the initiation of repair and its completion.
The trouble source can be located in the teacher or the learner, as can the
initiation and completion of repair. Kasper (1985) found that repair in the
language-centred phase of a lesson in a Danish gymnasium differed from that
in the message-focused stage. In the former, trouble sources in the learners’
utterances were identified by the teacher and were either corrected by the
teacher or by another learner to whom the teacher delegated the task. In the
latter, self-initiated and self-completed repair were preferred by both learners
and teachers, although teacher-initiated and completed repair were still strongly
represented. Van Lier (1988) points out that other repair (by the teacher) can
be of two types, depending on whether it takes place while the trouble-source
turn is in progress or in the next-turn position. Other repair that takes place
during the learner’s turn is a form of‘helping’ and may be a special characteristic
of the language classroom. It tends to focus on specific linguistic properties
involved in the execution of a complex utterance. Other repair that takes place
on the completion of a learner turn often involves some form of evaluation of
the learner’s performance on the part of the teacher.
72 Instructed Second Language Acgulsitlon
One way of describing error treatment is by listing the various options
available to the teacher (e.g. ‘To treat or ignore’, ‘To treat immediately or
delay’, ‘To transfer treatment or not’) and the various features of the error
treatment provided (e.g. ‘Fact of error indicated’, ‘Blame indicated’, ‘Location
indicated’). In other words, a kind of observation schedule specially designed
for those parts of a lesson in which error treatment occurs is developed.
Allwright (1975) — from whom the examples listed above come — and
Chaudron (1977) have developed comprehensive taxonomies of this kind.
In another approach, discourse systems to show how corrective discourse is
constructed have been developed. The emphasis here is not on isolated acts of
treatment but on whole exchanges involving one or more learners and the
teacher. Long (1977) provides a model of the decision-making teachers go
through in providing feedback. Chaudron (1977) adapts the model of classroom
discourse proposed by Sinclair and Coulthard (1975) and incorporates the
various options listed by Allwright (1975) in an explicit description of error
treatment. Van Lier (1988:205) provides a model of ‘repair decisions’. He
proposes that teachers respond initially to whether an utterance is ‘intelligible’,
then to whether it is ‘acceptable’ and finally to whether it is ‘correct’. Criteria
of correctness often override criteria of acceptability in the classroom context.
A number of characteristics of error treatment have been noted. Holley and
King (1975) found that graduate trainee teachers consistently overcorrected.
If a student hesitated in producing a response, the teacher immediately
jumped in to complete it. A student response that was deemed incorrect was
cut off before completion. Holley and King found that it was possible to
modify the trainees’ treatment through training by encouraging them to delay
filling in student pauses and to avoid correcting obviously incorrect answers.
Other researchers have found that experienced teachers vary enormously in
the extent to which they correct and, also, in their attitudes to error correction
(Nystrom, 1983; Chaudron, 1986a). For example, some teachers feel that it is
important to correct every linguistic error that occurs, while others feel that
linguistic errors should be ignored and only content errors corrected. It has
also been shown that native-speaker teachers tend to be more tolerant of error
than non-native-speaker teachers (Hughes and Lascaratou, 1982).
Teachers also display inconsistency and lack of clarity in their treatment of
error. Long (1977) notes that teachers often give more than one form of
feedback simultaneously and that many of their feedback moves go unnoticed
by pupils. Teachers use the same overt behaviour for more than one purpose.
For example, a teacher repetition can occur after a learner error and serve as
a model for imitation or it can function as reinforcement of a correct response.
Teachers often fail to indicate where or how an utterance is deviant. Positive
feedback is sometimes provided, even when the learner’s response is incorrect.
Teachers correct an error in one part of a lesson but ignore it in another.
They also correct an error produced by one student but fail to correct the
same error in another student’s speech. McTear (1975) notes that teachers
also frequently give up the task of correction. Nystrom (1983) sums it all up
when she writes: ‘the fact remains that each interactive situation is quite
 Classroom Process Research 73
complex, and teachers typically are unable to sort through the feedback
options available to them and arrive at the most appropriate response.’ However,
there is disagreement as to whether this constitutes a desirable or undesirable
state of affairs. Long argues that teachers need to be consistent. Allwright
(1975) suggests that teachers may have a duty to be inconsistent in order to
cater for individual differences amongst their students.
One explanation for the teacher’s inconsistency — apart from the complexity of
the task they face — may lie in differences in learner proficiency. Chaudron
(1986a) compared the corrections made by teachers in French immersion
classes in Canada on two different occasions, separated by a six-month
period. He found that the teachers corrected more morphological errors and
fewer discourse errors (e.g. errors involving the use of incomplete phrases or
speaking out of turn) at time two. Another factor accounting for inconsistency
is the nature of the teaching task. Teachers are less likely to correct linguistic
errors in tasks that call for real communication (Kasper, 1985).
There have been very few studies that have examined the effects on
learning. Chaudron (1977) investigated the relationship between different
types of corrective repetition and student uptake (i.e. the extent to which a
student incorporated the treatment into her next utterance). Reduced repetitions
containing emphasis on the key word were most strongly related to uptake.
Expanded repetitions without emphasis were the least successful. However,
students were often unable to produce the correct form after any kind of
treatment. Chaudron (1986a) found only a 39 per cent rate of success in the
three classes he studied.3
Even if the rate of success had been much higher, it is not clear that the
ability to uptake the teacher’s correction is the same as learning. In order to
demonstrate that error treatment has a real effect, it is necessary to show that
learners are able to produce the correct forms autonomously (i.e. in their
own-initiated utterances) when they were unable to do so before treatment.
From what is known about the way interlanguage develops, error treatment is
unlikely to achieve such an effect. As Long (1977) points out, errors disappear
slowly. It is possible, of course, that error treatment helps to speed up the
process. Ramirez and Stromquist (1979) found a positive correlation between
the correction of grammatical errors and general gains in linguistic proficiency. It
should be noted, however, that this study did not examine the relationship
between the correction of specific errors and the elimination of these errors in
the learners’ interlanguages. There is no definitive study showing that error
treatment promotes acquisition.
There are, however, plenty of strong opinions. Krashen (1982) emphasizes
the limitations of error correction. He argues that it puts learners on the
defensive and encourages them to avoid difficult structures and to focus on
form rather than on meaning, all of which are bad for ‘acquisition’. He
concludes (1982:119): ‘even under the best of conditions, with the most
learning-oriented students, teacher corrections will not produce results that
will live up to the expectations of many instructors.’
Van Lier (1988) also thinks that other-repair (by the teacher) may be less
74 Instructed Second Language Acguisition
conducive of acquisition than self-repair (by the learner). In particular, learners
may find classrooms where self-repair predominates more hospitable and less
threatening than ones where teacher-repair is the norm. Edmondson (1985),
however, argues that bringing errors to the learner’s attention helps learning.
He sees error treatment as contributing to the process of consciousness-
raising which he deems important for language acquisition. All these arguments
derive from theoretical positions rather than from empirical research.
Several researchers have noted that teachers can actually bring error into
existence. Stenson (1975) describes how errors can be elicited by the way a
teacher frames a question or an explanation of a grammatical rule. For
example, the learner utterance:

* In this class there are any students who speak German.

resulted from the teacher saying that ‘any’ is used in negative constructions,
which the learner understood as meaning that ‘any’ replaces the negative
marker. Edmondson (1986) illustrates how teachers actually elicit error in the
process of correcting. He observes that ‘some teaching methods .. . require
the occurrence of error as a teaching tactic (1986:118). Thus teachers exert
interactional pressure on learners to produce errors!
The research reported above has provided some valuable insights into the
nature of error treatment. The general picture that emerges is that error
treatment is not a manipulative process — as it was seen to be in audiolingual
learning theory. Rather it is a process of negotiation, one of several ways in
which the teacher and the learners collaborate in managing interactional tasks
in the classroom. This is why treatment is often inconsistent and why error
can be created in the very act of trying to correct. It is not clear yet to what
extent, if at all, error treatment facilitates learning. Even less clear is which
kinds of treatment are effective and which ones are not.

Input Features
The input features in teachers’ language which have been studied include
amount of talk, rate of speech, vocabulary, syntactic complexity and correctness.
These studies have been motivated by a particular view of the language
classroom — one that sees instruction in terms of the provision of input for
acquisition rather than as attempts to achieve specific linguistic aims. They
mirror studies of motherese in first language acquisition research.
Delamont (1976; quoted in Holmes, 1978) has noted that ‘research reported
from all over the world shows a similar pattern: in India, Belgium, Iraq, South
America and New Zealand the teacher keeps on talking.’ The dominance of
the teacher’s language is most evident in classrooms which are teacher-centred,
where instruction proceeds with the whole class in lockstep. However, even in
primary school classrooms, where a more child-centred methodology might be
anticipated, the teacher has been shown to take up the main share of the
talking time.
 Classroom Process Research 75
Also of interest is the amount of language directed at individual students.
This has been examined with reference to mixed classrooms (i.e. classrooms
containing both native-speaking children and L2 learners). Schinke-Llano
(1983) found that teachers in her study of 5th and 6th Grade content classrooms
in the United States interacted twice as much with the native speakers as with
the L2 learners. Nearly half of the latter were not addressed direcdy at all,
while this happened to only 14 per cent of the native-speaking children.
There was, however, considerable variation among the teachers. Kleifgen
(1985) found that kindergarten teachers varied the amount of speech they
addressed to individual learners according to the students’ needs in performing
various tasks. Which learners teachers talk to most is likely to vary enormously
according to the. teacher’s teaching style and the level and composition of the
class. For example, Ellis and Rathbone (1987) found that teachers directed
practice opportunities at those students who would be continuing their study
of L2 German beyond the first year at the expense of those students who did
not intend to continue.
Teachers typically reduce their rate of speaking when they are teaching L2
learners. Also they slow down in accordance with the learners’ levels of
proficiency. Chaudron (1985; 1988) has reviewed a number of studies which
have examined this aspect of teacher input in either simulated or real instruction.
He notes that ‘the absolute values of speech to beginning learners are around
100 w.p.m.’ while intermediate and advanced learners receive speech which is
30—40 w.p.m. faster. The importance of this research is that it shows that
beginner learners receive speech which may help them to process input.
Adjustments in vocabulary also occur. Henzl (1979) found that when teachers
were asked to tell a story in English to beginners, advanced learners and
native speakers, they simplified their lexis according to level. With the beginners,
lexical items with general meanings were substituted for items with narrow
semantic fields (e.g. ‘woman’ was preferred to ‘young gal’). The teachers also
avoided colloquial items with beginners and tended to use stylistically neutral
language. Both Henzl and Kleifgen (1985) found that the type—token ratio
was higher with less proficient learners (i.e. fewer different items were used).
There is also variation in the syntactic complexity of teachers’ speech.
Chaudron (1985; 1988) again provides an excellent review of the relevant
studies. He notes that the findings are contradictory. Whereas some researchers
(e.g. Pica and Long, 1986; Wesche and Ready, 1985) found no difference in
their chosen measures of syntactic complexity (e.g. tensed and non-tensed
forms), other researchers (e.g. Gaies, 1977; Hakansson, 1986) have found
quite marked differences. At the moment it is not clear why this contradiction
occurs. One reason might be the different syntactic measures used.
In naturalistic conversations between native and non-native speakers, adjust¬
ments can result in ungrammatical input under certain circumstances (Long,
1983c). Hakansson (1986) found that the input of teachers of L2 Swedish
contained a higher proportion of correct sentences than that found in either
formal or informal Swedish addressed to native speakers. The ungrammatical
input data consisted mainly of incompleted sentences and was the result of the
76 Instructed Second Language Acguisition
teachers’ attempts to prompt or prod the students into responding. She claims
that ungrammatical input in teacher-talk is rare. However, it has been found
to occur in communication sequences which are not overdy instructional.
Hatch, Shapira and Gough (1978) found that a teacher in a beginners’ class
in an adult community school in Los Angeles always used standard English in
drill practice, but resorted to ungrammatical simplifications in instructions and
discussions, as in these examples:

Is important.
The writing not important.
I say bye bye, I no want.

Ellis (1982) also found evidence of similar simplifications in the speech of

teachers of a London language unit.
The research outlined above indicates that teachers are generally sensitive
to their students’ levels of proficiency. Probably teachers are not aware of the
precise state of their L2 knowledge, but respond more generally according to
their ability to understand. It is also unlikely that teachers have a very exact
idea regarding gradations in the linguistic complexity of different structures
(cf. Chaudron, 1986b) but rely on a general intuitive feel for what makes input
simple or complex. An interesting question is what teachers use as a basis for
determining the level of input adjustments to strive for in classes containing
learners with mixed ability in the L2. Hakansson (1986) speculates that they
may aim at some hypothetical average learner. In this case, the input may not
be tuned very accurately to the level of either the less or more advanced
learners. It is reasonable to assume that in the one-to-many interactions that
occur in teacher-fronted lessons, the input which individual learners receive is
not so well adjusted to their level of proficiency as is the case in the one-to-
one interactions that are more common in naturalistic L2 acquisition. In
homogeneous classes, however, it is reasonable to assume that learners receive
input which is similar in complexity to that experienced by naturalistic learners
(Pica and Long, 1986), although, as noted above, it is likely to be more
grammatical.

Interactional Features
Interactional features of teacher-talk have attracted increasing attention from
classroom researchers. We will consider them in greater detail in the next
chapter when we will review the research that has derived from Long’s claims
regarding the role of negotiated input in L2 acquisition. Here we will sketch
some of the main characteristics.
The functional nature of teacher-talk is largely a product of the discourse
rights that are invested in the teacher. Gremmo, Holec and Riley (1978) list
these rights. The teacher has the right to:
 Classroom Process Research 71
1 participate in all exchanges;
2 initiate exchanges;
3 decide on the length of an exchange;
4 close exchanges;
5 include and exclude other participants in exchanges;
6 open all adjacency pairs;
7 be the only possible addressee of any exchange initiated by another
participant;
8 decide on the order of other participants’ turns;
9 decide on the number of turns to be attributed to each participant.

Such a list is the*testimony to the enormous power that the teacher can exert
in classroom discourse. Of course, the teacher can elect not to use this power.
Much depends on the nature of the role that the teacher chooses to adopt
(Corder, 1977). If the teacher chooses to act as ‘informant’ or ‘knower’, she
places the learner in a totally dependent position. In such a situation the
teacher is likely to make full use of her rights. However, if the teacher acts as
‘referee’ or ‘mentor’ and the student plays the role of ‘actor’ or ‘apprentice
member of society’, the learner has greater control over the discourse and,
Corder argues, over the conduct of her own learning.
The reality of most classrooms is that the teacher is in control of most of
the interactions that take place. This is reflected in the preponderance of
teacher acts over student acts. Gremmo, Holec and Riley (1977) find a 2:1
ratio. Also teachers typically perform a greater variety of acts and their nature
is different. Teachers open and close interactional exchanges, while students
are restricted to replying. Teachers elicit and students supply responses. The
nature of the acts found in teacher-talk also differs considerably from that
found in foreigner-talk (i.e. the language used by native speakers to address
non-native speakers in naturalistic contexts). Long and Sato (1983) investigated
the language used by six ESL teachers in classes containing beginners and
false beginners and compared it to the language used in native speaker/non¬
native speaker conversations. They found that statements accounted for most
of the teacher-talk (54 per cent), followed by questions (35 per cent) and
imperatives (11 per cent). In foreigner-talk conversations the percentages
were statements (33 per cent), questions (66 per cent) and imperatives (only
1 per cent). In other words, questions were more frequent in foreigner-talk,
although this might have been a reflection of the ‘interview-style’ conversations
of which their data consisted.
We can expect considerable variation from teacher to teacher and also from
class to class. Politzer, Ramirez and Lewis (1981) looked at classes where
standard English was being taught to speakers of Black American Vernacular.
They examined the proportion of teacher ‘moves’ that were eliciting, directing,
informing, evaluating, expressing (i.e. social, personalized comments) and
replying. Overall, they found that eliciting moves were the most frequent,
followed by informing and evaluating moves. Replying moves were very in-
78 Instructed Second Language Acquisition
frequent. This result contradicts, in part, Long and Sato’s findings.4 However,
they report that the frequencies of the different functions varied enormously
from teacher to teacher and also from lesson to lesson. Quantitative studies of
the kind reported above are likely to produce different results depending on a
variety of factors.
There can also be variation within a single teacher’s practice, according to
the nature of the activity engaged in. Mitchell (1988b) found that teachers of
elementary French in Scottish secondary schools varied in their use of English
(the LI) or French (the L2) in different kinds of classroom-management dis¬
course. The L2 was widely used in organizational instructions (i.e. utterances
which dealt with telling the pupils how to group themselves, what materials to
get out and what page to turn to etc.). The L2 was also employed in disciplinary
interventions. But the LI was generally preferred in activity instructions (i.e.
utterances in which the teacher explained how to perform a pedagogic activity).
The teachers in Mitchell’s study had a high level of commitment to communi¬
cative language teaching and, therefore, were much more likely to try to use
the L2 for management purposes than more traditional teachers.5
One of the interactional features that is a more or less universal characteristic
of teacher-talk is questions. Questions serve as the principal way in which the
teacher maintains control over the classroom discourse. They have attracted
considerable attention from researchers of both content and language class¬
rooms. Barnes (1969) reports on a study of teachers’ questions in secondary
school subject classrooms. He found a predominance of factual as opposed to
reasoning questions. Also open questions (i.e. questions which permit a
number of possible answers) were extremely rare, while closed questions
(i.e. questions designed to have only one acceptable answer) were very common.
The teachers used questions as a device for eliciting the information they
wanted to transmit. Long and Sato (1983) found that 79 per cent of the
questions which requested information from the students were closed (or, as
they call them, display questions). This contrasted with the use of questions
by native speakers in naturalistic discourse with non-native speakers, where
display questions occurred only rarely. Long and Sato concluded:

This result suggests that, contrary to the recommendations of many

writers on SL (second language) teaching methodology, communicative
use of the target language makes up only a minor part of typical
classroom activities. ‘Is the clock on the wall?’ and ‘Are you a student?’
are still the staple diet, at least for beginners. (1983:280)

Such is the prevalence of closed questions that learners can actually become
confused when the teacher asks an open question (McTear, 1975) and try to
interpret it as if it were a closed question.
Another interactional feature of teachers’ classroom speech derives from
the attempt to solve the comprehension problems the learners experience.
Mitchell (1988b) identifies a number of communication strategies which
 Classroom Process Research 79
the teachers in her study employed in order to consolidate the learners’ com¬
prehension of unfamiliar items. She divides these strategies into those that
maintained the L2 medium and those that involved use of the LI. Table 4.1
lists the various strategies employed and provides examples of each.
Teacher-talk in language classrooms is also typically context-embedded, at
least with beginners. Kleifgen (1985) reports that the L2 learners in the
kindergarten classroom she studied received more context-embedded language
than the native-speaking children. Long and Sato (1983) found that six ESL
teachers used verbs that were marked for present time significandy more than
verbs marked for past time (71 per cent versus 29 per cent). Henzl (1979)
found that teachers increased their use of explanatory gestures when talking to
low-level learners. It would seem that the ‘here-and-now’ principle is adhered
to fairly stricdy in the classroom, even with adult learners. In foreigner-talk
studies it has been found that this principle figures strongly in conversations
with children but less so with adults (Hatch, 1978b; Scarcella and Higa,
1981).
The extent to which the teacher’s language promotes or inhibits acquisition
of the L2 remains largely a matter of speculation. Gaies (1977) suggests that
teachers’ language is characterized by similar ‘training strategies’ to those
found in motherese (e.g. the use of repetitions, prompts, modelling and
expansions) and for this reason may help learning. It should be noted, however,
that the kind of modified language that characterizes teacher-talk is not
necessarily ideal for language learning. Chaudron (1983) argues that some
kinds of modification — in input and interaction — can result in either
ambiguous oversimplification or redundant over-elaboration. In such cases the
learner finds it more rather than less difficult to understand. Learners may
also react negatively to simplified teacher language (Lynch, 1988).
Other researchers (notably Long and Sato) see the teacher’s language — in
particular her use of closed questions — as restrictive. There is a general
conviction that teachers talk too much. Politzer, Ramirez and Lewis (1981)
suggest that the more successful teachers in their study dominated classroom
talk to a lesser extent than the more unsuccessful teachers, but they found no
overall relationship between the number of teacher moves and learning
outcomes.
However, not all researchers have taken such a bleak view of the teacher’s
contribution. Mitchell (1988b: 166) concludes her study of elementary French
teaching by drawing attention to the ‘critical importance of the teacher as the
central classroom resource’. She writes:

No functional syllabus, ‘authentic’ materials, or microcomputer program

can replace the capacity of the live, fluent speaker to hit upon the
follow-up topics of interest to particular individuals, continually adjust
his/her speech to an appropriate level of difficulty and solve unpredictable
communication difficulties from moment to moment ... In all this, the
teacher and his/her interactive skills are decisive.
80 Instructed Second Language Acguisition
Mitchell’s comments are salutary in reminding us that teachers’ efforts to
teach the L2 are not always in vain!

Table 4.1 Teacher communication strategies

Strategy Description Example

A. L2-medium strategies:

Repetition The teacher (T) simply repeats T: Si vous commencez la et

the problematic item vous continuez tout
droit . . . Vous comprenez?
Non? Ecoutez encore.
Alain, si tu commences la et
tu continues tout droit . . .
Substitution The teacher substitutes T: Alors les Tuileries, c’est un
another, roughly equivalent jardin — vous comprenez,
item un jardin? C’est comme un
pare.
Explanation The teacher explains the T: Patiner a roulettes, qa veut
meaning of the problematic dire patiner, pas sur la
item. glace, mais avec . . . mais
avec des roulettes comme
qa. (gestures)
Contrast The teacher contrasts the T: Tout le monde comprend,
problematic item with others les autobus? . . . alors, si
which in some way belong to a vous voulez faire un tour
similar set. de Paris, vous pouvez
prendre le metro, vous
pouvez prendre l’autobus
normal . . .
Exemplification The teacher exemplifies the T: Ecrivez dans votre cahier en
problematic item. anglais les mots que vous
allez entendre. Par exemple,
vous entendez ‘janvier’ et
vous ecrivez dans votre
cahier ‘January’.
Clue-giving The teacher suggests an T: Vous comprenez l’Egypte?
associated concept. Vous comprenez les
pyr amides?
B. LI-based strategies:

Pupil The teacher invites other pupils T: Alors voila un chat (pointing
interpretation to supply an LI equivalent of a to picture) ‘Un chat’,
problematic item. qu’est-ce que c’est, ‘un
chat’?
Teacher The teacher supplies a T: Choisis! Tu ne comprends
interpretation translation of the problematic pas ‘choisis’? Qa veut dire
item ‘to choose’.
 Classroom Process Research 81
Language The teacher speaks bilingually, T: Tu vas preparer le test.
switching repeating messages in the LI Prepare ... tu as fait le
that were first said in the L2? test? Have you done the
test?
Interpretation In LI-dominant exchanges the T: What does ‘Parle avec ton
quoted problem is the only L2 partenaire’ mean?
heard

Source: Mitchell (1988b)

The Learner’jS Language

We now turn to the learner’s contribution to classroom processes. This varies

greatly according to whether the learner is engaging in teacher-directed
discourse or in the ‘interlanguage talk’ that results from small group work.
The focus of this review will be on the former.
The restrictions that are placed on the learner’s contribution in lockstep
lessons as a result of the asymmetrical roles of teacher and students have
already been commented on. Thus learners typically talk far less than teachers
and perform a much narrower range of language functions. The crucial factor
appears to be the extent to which the learner or the teacher controls the
discourse. Barnes (1976) describes how a group of junior school children in
Britain produced language that was marked by a rich vocabulary, complexity
and a range of grammatical structures and long utterances. This occurred
when the teacher ceased to control the moment by moment progress of the
discussion. Barnes studied native-speaking children, but Cathcart (1986)
reached similar conclusions in her study of the different kinds of communicative
acts performed by eight Spanish-speaking children in a variety of school
settings (recess; seatwork; free play; ESL instruction; playhouse; interview;
storytelling). She found that situations in which the learners had control of the
talk were characterized by a wide variety of communicative acts and syntactic
structures, whereas the situations where the teacher had control seemed to
produce single-word utterances, short phrases and formulaic chunks.
Learners learn how to take part in classroom discourse. Their performance
in this type of discourse differs significantly from their performance in natural
discourse. House (1986) reports an interesting comparison of advanced German
learners of L2 English in a role-play situation and in a teacher-led discussion
of transcriptions of the role-plays. She found that the learners confined
themselves to an ‘interactional core’ in the discussion, omitting any non-
essentials such as Topic Introducers (e.g. ‘You know ...’), Grounders (i.e.
supportive moves used by a speaker to justify a speech act such as a request)
or Expanders (i.e. amplifications of a speech act). Instead they came to the
point almost immediately. In contrast, the role-play conversations sounded
much more natural because they did contain appropriate ‘discourse lubricants’.
82 Instructed Second Language Acquisition
House argues that the asymmetry of the classroom discussions, where the
teacher functioned as a manager of the talk, led to the learners adopting
relatively passive roles.
The most common communicative act performed by the learner is that of
replying to a teacher question. Politzer, Ramirez and Lewis (1981) report that
90 per cent of all student moves in their learner data were responses. The
answers students give to closed questions are typically short. Brock (1986)
found that learners’ responses were longer and syntactically more complex
when they followed referential (i.e. open) questions than when they followed
display questions. The responses of university-level ESL students averaged
10 words in the case of the former and only 4.23 words in the case of the
latter. Brock also found that referential questions resulted in responses with
more sentential connectives. However, Long et al. (1984) did not find any
significant differences in learner responses as a result of question types,
although there was a tendency for display questions to elicit more student
turns and for referential questions to elicit more and longer utterances. Any
question — even an open one — exerts a form of control as it constitutes the
first part of a question — answer pair. It is likely that learners’ contributions
will be more ample and syntactically complex if they have topic control (i.e.
are not answering questions), as has been argued above. In other words,
responses of any kind are likely to be sparser and leaner than speech acts that
initiate discourse.
Although replying to questions is the main communicative act, students do
perform other acts even in communication with the teacher. For example, in
classrooms where the target language functions not only as a pedagogic norm
but also as a means of conducting classroom business, learners have the
opportunity to perform a wide variety of speech acts (see table 4.2). Also,
when the teacher and a learner are joindy and collaboratively engaged in
completing some task, the learner has the opportunity to initiate and greater
variety of language results. Ellis (1981) illustrates how classroom learners can
be given the same opportunity to talk as children in conversation with their
mothers. He analyses a conversation between a group of L2 Asian children
and a teacher while they are having a meal together in the classroom. The
learners request permission, complain, make requests for goods, accept offers
and make comments. Not all classrooms afford such occasions, however. The
traditional foreign-language classroom is unlikely to provide opportunities for
this kind of learner language.
Another approach to the study of learner language in the classroom is to
examine communication strategies. These can be defined as attempts by
the learner to overcome communication problems by compensating for a lack
of linguistic knowledge (cf. Faerch and Kasper, 1980). There have been few
studies of communication strategies in a classroom setting. Rosing-Schow and
Haastrup (1982; as described by Holmen, 1985a) compared the performance
of four advanced Danish learners of English in the classroom and in two-way
interaction with a native speaker in a role-play activity. In the classroom, the
 Classroom Process Research 83
Table 4.2 Language functions performed by one L2 learner in speech
addressed to the teacher

Learner utterance Context Function

This one or this one? Pointing at two pages in Request for

textbook information
You milkman. It is near breaktime, when the Request for
teacher usually chooses a S to information
collect the milk.
Me milkman. As above. Offer
Me no blue. S is colouring but does not have Request for
i
a blue pencil. goods
Sharpening please. S has taken his pencil to the Request for
teacher. services
Look me do. Pointing at Christmas card Attracting
which he has made. attention
Me me. The Ss are playing a game of Demanding
rounders.

Source: Based on Ellis (1980)

learners either resorted to a reduction strategy (i.e. gave up on their original

communicative goal) or appealed to the teacher as a ‘walking dictionary’. The
authors concluded that the content and structure of most language classes do
not encourage the use of a wide variety of communication strategies and
therefore do not equip the learners for participation in ordinary face-to-face
interaction. This conclusion is very similar to one House reached.
Mitchell (1988b), however, did find evidence of the learners’ use of com¬
munication strategies in elementary French lessons in Scottish secondary
schools. Some of the teachers she observed made efforts to equip the learners
with the L2 means they needed to indicate that they were having comprehension
problems. They taught such phrases as ‘Je ne comprends pas’ and ‘Comment
est-ce qu’on dit xV The learners appeared to have mastered a number of such
routines and showed a willingness to use them in the classroom. Ellis (1982;
1984a) also reports successful attempts by a teacher to teach ESL learners
useful phrases such as ‘I don’t know’. Again, once they had learnt them, the
learners made regular use of them. It would appear, therefore, that in both
foreign and second language classrooms opportunities can be made for learners
to learn and use the linguistic formulas needed to perform communication
strategies. Whether these opportunities actually arise may depend on whether
the teacher is committed to using the L2 for communication.
So far we have been considering learner language from the point of view of
the learner herself. Gaies (1983b) examines ‘learner feedback’, which he
defines as ‘information provided by a learner to a teacher about the compre-
84 Instructed Second Language Acquisition
hensibility and usefulness of some prior teacher utterance(s)’. In other words
he looks at learner language from the perspective of the teacher. Gaies argues
that even in teacher-dominated lessons the learners negotiate the input they
receive by providing feedback. He investigated twelve ESL dyads (teacher-
student) and triads (teacher-student-student), collecting data from problem¬
solving tasks that required the teacher to describe graphic designs. The
learners were instructed to provide whatever feedback they considered
necessary in order to identify the designs being described. In the analysis
Gaies distinguished learner utterances which were elicited by the teacher as
opposed to those that were volunteered. The latter he divided into soliciting
(i.e. the learners tried to obtain information), reacting (i.e. the learners indicated
comprehension or non-comprehension) and structuring (i.e. the learners
attempted to reorient or redefine the basis for subsequent interaction). The
results revealed considerable variation from learner to learner in the amount
of feedback provided. Not all the learners used the full range of feedback
moves. Reacting moves were the most frequent and structuring moves the
least. In the case of the triads one of the two learners always provided more
feedback than the other. Gaies used the study to raise a number of interesting
questions to do with the effects of the size of the learning group, the activity,
learner characteristics and the teacher’s language on the nature of learner
feedback.
There are likely to be individual differences in learner language. One
source of variation is the ethnic background of the learner. Sato (1981)
compared nineteen Asian and twelve non-Asian university-level ESL students.
The Asian students initiated fewer turns and were also nominated to speak on
fewer occasions than the non-Asians. Irrespective of ethnicity learners vary in
the extent to which they are prepared to risk themselves in front of the rest of
the class. Diary studies (e.g. Ellis and Rathbone, 1987) show that some
learners develop considerable anxiety when called upon to answer a question
and avoid speaking if possible.
Holmen (1985b) suggests that learners can be placed on a continuum
between very active and very passive. The older learners evidenced greater
passivity than the younger learners she studied, a reflection of the different
kinds of schools they attended. Schools with teacher-centred classrooms
encourage more passive learners. However, Holmen recognizes that personality
is also a factor. Some learners ‘steal turns’ and grab the teacher’s attention
whenever they can. Other learners contribute frequently because they have
more teacher questions directed at them or because they have difficulty in
making themselves understood — like Allwright’s (1980) Igor.
Underlying much of the research that has been reported above is the
assumption that production is an aid to learning. It is also felt that learners
need the opportunity to perform a range of speech acts and communication
strategies. The classroom is often seen as affording only limited opportunities
for the kind of production that facilitates acquisition. It is, however an open
 Classroom Process Research 85
question whether production is centrally important in all aspects of L2 acqui¬
sition. It is possible, for instance, that the development of linguistic competence
owes more to comprehension than production (as Krashen argues). Production
may be more important for the development of oral fluency and strategic
competence. The next chapter takes a closer look at the role of learner
participation.

Classroom Discourse

Much is lost by considering the teacher’s and the learners’ language behaviour
separately. Classroom discourse is best seen as a cooperative enterprise. In
this section of the chapter, then, we examine classroom discourse as a whole.

The Teacher’s Paradox

The discourse that results from attempts to teach the target language is
necessarily different from the discourse that occurs naturally outside the
classroom. Edmondson (1985:162) refers to ‘the teacher’s paradox’, which
states:

We seek in the classroom to teach people how to talk when they are not
being taught.

This paradox results in a tension between discourse that is appropriate to

pedagogic goals and discourse that is appropriate to pedagogic settings. On
the one hand the teacher can create opportunities for discourse where the
learner behaves as a learner, while on the other she can also offer opportunities
for the learner to behave in other roles than that of the learner (i.e. by
providing tasks that stimulate authentic communication). The classroom affords
‘co-existing discourse worlds’, depending on whether the participants are
engaged in the act of trying to learn or trying to communicate.
This basic point about classroom discourse is reflected in the four types of
language use which McTear (1975) observed in EFL lessons with adult
Venezuelan students:

1 mechanical (i.e. no exchange of meaning involved);

2 meaningful (i.e. meaning is contextualized but there is still no new
information conveyed);
3 pseudo-communication (i.e. new information is conveyed but in a manner
that would be unlikely to occur outside the classroom);
4 real communication (e.g. spontaneous speech resulting from the exchange
of opinions, jokes, classroom management etc.).
86 Instructed Second Language Acgulsltlon
Table 4.3 Continuum of classroom interaction

Instructional discourse Natural discourse

Roles Fixed statuses Negotiated roles

Tasks Teacher-oriented, position- Group-oriented, person-

centred centred

Knowledge Focus on content, accuracy Focus on process, fluency of

of facts interaction

Source: Kramsch, (1985)

Pedagogic discourse is produced by (1) and (2), natural discourse by (4), while
(3) falls somewhere in between.
It is perhaps wiser to treat pedagogic and natural discourse as two poles of a
continuum rather than as alternatives. Kramsch (1985) discusses the factors
that induce the discourse to lean one way rather than the other (see table 4.3).
Instructional discourse occurs when the teacher and the students act out
institutional roles, the tasks are concerned with the transmission and reception
of information controlled by the teacher and there is a focus on knowledge as
a product and on accuracy. Natural discourse is characterized by much more
fluid roles established through interaction, tasks that encourage equal partici¬
pation in the negotiation of meaning and a focus on the interactional process
itself and on fluency.
Breen (1985a; 1985b) suggests how learning and the pedagogic discourse it
produces can be reconciled with communication and the natural discourse it
produces through metacommunication about the target language and the
problems of how to learn it. In other words, the discourse worlds can be
brought together through communication about learning.

Turn-taking
One aspect of classroom discourse which has received considerable atten¬
tion — largely because marked differences from conversational discourse are
often evident — is turn-taking. This refers to the systematic rules that govern
how the roles of speaker and listener change (Coulthard, 1977). Sacks et al.
(1974) identified a number of such rules in American English conversation
and these have served as a yardstick for the study of turn-taking in the
classroom. One of these rules is that only one speaker speaks at a time. Other
rules relate to how speaker-change is negotiated. A speaker can select the next
speaker by nominating her or performing the first part of an adjacency pair
(e.g. asking a question which requires an answer). Alternatively, the speaker
can allow the next speaker to self-select. Turns to speak are highly valued and
there is frequently competition to take the next turn. There are also rules
governing when a speaker can enter or re-enter the conversation. The general
 Classroom Process Research 87
characteristics of conversational turn-taking as described by Sacks et al. are
those of self-regulated competition and initiative.
Most researchers highlight the differences between turn-taking in natural
and classroom settings (e.g. McHoul, 1978; Sinclair and Brazil, 1982). In
classrooms there is a need to maintain centralized attention (usually directed
at the teacher), the content of lessons is pre-planned, the actual sequence of
utterances may be predetermined and the concern to maintain the ‘official’
topic inhibits small-talk. These factors can lead to a radically different pattern
of turn-taking. Classroom discourse is often organized so that there is a strict
allocation of turns to deal with potential transition and distribution problems.
Who speaks to whom at what time is stricdy controlled. This results in less
tum-by-tum negotiation; competition and individual learner initiative are
discouraged.
Lorscher (1986) is a good example of a study of turn-taking in the L2
classroom that highlights the differences between conversational and pedagogic
discourse. Lorscher examined openings, turn-taking and closings using video¬
tapes of English lessons involving pupils in the age range of eleven to eighteen
in different types of German schools. In the case of openings he found a
number of differences between classroom and natural discourse. The former had a
simpler structure, they were not negotiated or built up cooperatively, the topic
was determined by the teacher and the students never attempted to postpone
the topic. The turn-taking rules he identified were also different. In the
classroom turns were allocated by the teacher, the right to speak always
returned to the teacher when a student turn was completed and the teacher
had the right to stop or interrupt a student turn. In the case of closings, there
was no negotiation and ‘possible pre-closings’ — a feature of naturally occurring
conversations — were very rare. Lorscher argues that these classroom discourse
structures reflect the need for the communication partners to effect successful
and economical pedagogical outcomes. They are determined by the nature of
the school as a public institution and the teaching—learning process.
One of the possible effects of all this is that learners are robbed of ‘an
intrinsic motivation for listening’ with the result that they do not attend fully
to the input and thus L2 acquisition is impeded (Van Lier, 1988).
However, it is easy to overstate the lack of flexibility evident in L2 classroom
turn-taking. Van Lier (1988) reports that in his data learners frequently do
self-select and that ‘schismatic talk’ (i.e. talk that deviates from some predeter¬
mined plan) also occurs quite often. Teachers, too, perform many undirected
utterances, thus allowing learners to self-select. The extent of the difference
between turn-taking in conversation and classroom discourse is likely to be
influenced by the teacher’s perception of her role. If the teacher sees herself
as a ‘knower’, transmitting L2 knowledge to the learners, then turn-taking is
likely to be stricdy regulated. If, on the other hand, the teacher sees herself as
a ‘facilitator’ of self-directed L2 acquisition, the turn-taking is likely to
be negotiated in a manner not dissimilar to that which occurs in ordinary
conversation.
88 Instructed Second Language Acquisition

The Structure of Pedagogic Discourse

Pedagogic discourse manifests a particular structure. Sinclair and Coulthard
(1975) found that the predominant exchange structure in teacher-controlled
discourse was three-phase, involving a teacher elicitation, a student response
and teacher feedback — IRF:

Teacher: Ask Anan what his name is.

Student: What’s your name?
Teacher: Good.

As long as the teacher is controlling the discourse, this structure is evident no

matter whether the participants are focusing on linguistic form (as in the
example above) or on meaning. Mehan (1979) found a similar structure. He
distinguishes a ‘basic sequence’ which occurs when the learner provides the
response the teacher wants and an ‘extended sequence’ which is found when
the expected response does not occur. In this case ‘teacher—student interaction
continues until symmetry between initiation and reply acts is established’
(1979:55). The three-phase structure of classroom discourse contrasts with
the typical two-phase structure of natural discourse.
A series of exchanges form transactions (Sinclair and Coulthard, 1975). A
transaction constitutes an episode in the lesson with a unitary purpose. Trans¬
actions are less tighdy structured than exchanges. It is also possible to identify
discourse organization at the level of lesson. Mehan (1979) distinguished
three components: (1) an opening phase where the participants ‘inform each
other that they are, in fact, going to conduct a lesson as opposed to some
other activity’; (2) an instructional phase where information is exchanged
between teacher and students; and (3) a closing phase where participants are
reminded of what went on in the core of the lesson. However, not all language
lessons have such a pattern, perhaps because the content is ‘language’. Some
language lessons seem just to start and stop (i.e. consist entirely of an
instructional phase).

Types of Classroom Discourse

A basic distinction has already been drawn between pedagogic and natural
discourse — both of which can occur in the language classroom. A number of
researchers, however, have attempted to produce more detailed descriptions
of the types of interaction which can occur.
Ellis (1984a) describes classroom discourse in terms of two dimensions: (1)
the interactive goal and (2) address. Three broad types of interactive goal are
distinguished. ‘Core goals’ are reflected in the explicit pedagogical intentions
of the teacher. These can be medium-centred (i.e. concerned with linguistic
form), message-centred (i.e. concerned with the exchange of information) or
activity-centred (i.e. directed towards the completion of some non-verbal
task). ‘Framework goals’ are the interactional goals associated with the organ-
 Classroom Process Research 89
ization of classroom activity. ‘Social goals’ arise when the participants interact
on everyday social matters. ‘Address’ types are determined with reference to
one of four identities which a classroom participant can have (teacher, pupil,
class member or group member) and her interactive role (speaker, addressee
and hearer). The two dimensions form a matrix which can be used to classify
classroom interactions.
Van Lier (1982; 1988) identifies four basic types of classroom interaction
according to whether the teacher controls the topic (i.e. what is being talked
about) and the activity (i.e. the way the topic is talked about). Type 1
interaction occurs when the teacher controls neither the topic nor the activity.
Examples are small-talk with the teacher at the beginning of a lesson and
private talk between students. Type 2 interaction involves teacher control of
the topic but not of activity. This kind of interaction occurs when there is
some information that needs to be transmitted or some issue that needs to be
sorted out — for instance, when the teacher makes an announcement, gives
some instructions or delivers a lecture. Type 3 interaction occurs when the
teacher controls both topic and activity, as when some information needs to be
transmitted in a specific way (e.g. teacher elicitation in the context of a
language exercise). Type 4 involves teacher control of the activity but no
control of the topic, as when the teacher sets up small-group work with
prescribed procedural rules but allows freedom of choice of topic or when the
teacher nominates a learner to ask a question of another learner.
The systems developed by both Ellis and Van Lier involve two dimensions.
More complex descriptions have been developed involving several dimensions.
Fanselow (1977), for example, set out to provide ‘a technical, operationally
defined set of terms’ for discussing different kinds of teaching in terms of five
dimensions: (1) source (i.e. who communicates); (2) move type (i.e. whether
structuring, soliciting, responding or reacting); (3) mediums (i.e. linguistic,
non-linguistic or paralinguistic); (4) content (i.e. language, life, procedure or
subject matter) and (5) use (i.e. how the mediums are used to communicate
content). Allwright (1980) attempts to characterize classroom interaction with
reference to turns, topics and tasks.
Each system of classification represents the factors which its deviser considers
important for understanding what takes place in classroom discourse. As
Gaies (1983b:202) observes, the categories identified are not directly observable
in actual data but rather ‘reflect the particular order that a researcher chooses
to impose on the data’. As such, the value of such systems lies in the insights
they provide about potentially significant aspects' of classroom behaviour.
They should not be seen as comprehensive descriptions.

Interpreting the Research

Most of the studies reported above involved either a direct or an indirect

comparison between classroom and natural discourse. In summary, it can be
90 Instructed Second Language Acquisition
pointed out that the majority of studies emphasize the differences between the
two types of discourse. However, some studies also acknowledge that classroom
and pedagogic discourse are not entirely distinct, because the classroom can
afford some opportunities for natural discourse.
The general position taken by most researchers is that the absence of
natural discourse in the classroom inhibits language learning. Riley (1977), for
instance, talks about the ‘falsification of behaviour’ and ‘distortion’ that occurs
in pedagogic discourse. He argues that learners need to be exposed to
discourse that has undergone the minimum of adulteration and interference.
However, other researchers (e.g. Edmondson) acknowledge the inevitability of
pedagogical discourse. They see it as the product of attempts to learn the
target language and argue that such attempts are both logical and sensible. A
somewhat smaller group of researchers (e.g. Mitchell, 1988b) bring evidence
to show that teachers can provide opportunities for real communication and
argue that the teacher functions as a valuable resource of L2 input. All these
researchers share a common perspective: they see classroom discourse and
the interaction through which it is constructed as the matrix of language
learning—teaching — a radically different view from that discussed in the
earlier chapters.6

Summary and Conclusion

This chapter has examined a range of classroom process research. This

research was stimulated by the recognition that large-scale comparative method
studies were unlikely to throw fight on the factors that influenced L2 learning
and that this could be better achieved by small-scale studies of teaching-
learning behaviours. The principal aim of the research was to describe and
understand the various behaviours that take place in the language classroom.
Initially, the research was exploratory in nature,7 but as time passed, researchers
drew increasingly on the theoretical insights afforded by SLA studies. The
principal method of classroom process research is observation. This has been
carried out by means of interaction analysis, based on observation schedules
for coding different types of behaviour, by anthropological or ethnomethod-
ological studies of the social norms that govern classroom interaction and by
the study of the structure of classroom discourse. Classroom process research
has examined the nature of the teacher’s and the learners’ independent
contributions and also classroom discourse taken as a whole.
There can be little doubt that the research has helped to increase our
understanding of the activities from which learning derives. This understanding
relates both to the general characteristics of the language classroom and also
to the differences between classrooms - which should never be underestimated.
The understanding that has been achieved is valuable in two ways. First, it is
now clear that prescriptions regarding desirable teaching behaviours which
 Classroom Process Research 91
have been generated by a particular method may not correspond to the actual
behaviour that occurs in the classroom. Error treatment, for instance, has
turned out to be much more complicated than was ever envisaged by advocates of
audiolingualism. Second, a new way of conceiving the relationship between
teaching and learning has developed. This is, that any learning that takes
place must in some way result from the process of interaction the learner
takes part in. Allwright (1984b) suggests two ‘interaction hypotheses’ to
formalize this understanding. The first simply states that ‘the processes of
classroom interaction determine what learning opportunities become available
to be learned from’. The second states more strongly that ‘the process of
interaction is the learning process’. By this Allwright means that interaction
somehow produces linguistic development.
Classroom process research has done much to show us how learning
opportunities are made available to the learner, but it has not been able to
show how interaction results in L2 learning. As Van Lier (1988:91) points
out:

When observing an L2 classroom in action it is clear that no direct fink

can be made between observable behaviour and language development.
Learning is not generally direcdy and immediately observable. In the
first place, it is characterized by improved performance or increased
knowledge, and manifested by the learner’s behaviour at some time
(unspecified) after the learning has occurred. Secondly, the learning
itself may not be produced by one specifically identifiable event, but
rather by the cumulative effect of a number of events.

For this reason classroom process research has had difficulty in describing
what is ‘good’ or ‘bad’ for L2 learning. In order to establish a link between the
observation of teaching—learning behaviours and the internal process of L2
acquisition researchers have had recourse to various theories of SLA. In this
way they have felt able to make evaluative statements about what promotes
and hinders acquisition. The next chapter examines a number of these theories.

NOTES

1 As pointed out in ch. 1, classroom process researchers have tended to emphasize

social and sociolinguistic enquiry rather than psychological or psycholinguistic
enquiry. That is, they have viewed the classroom as a social context and treated
‘learning’ as an interpersonal activity rather than as an activity that takes place in
the learner’s mind.
2 Recent interaction analysis systems have also been informed by research into L2
proficiency and language acquisition. The Communicative Orientation of Language
Teaching (Allen, Frohlich and Spada, 1984), for instance, was derived ‘from
current theories of communicative competence, from the literature on communicative
92 Instructed Second Language Acquisition
language teaching, and from a review of recent research into first and second
language acquisition’ (p. 223).
3 The failure of learners to incorporate teacher corrections into their next utterances
is, of course, predicted by interlanguage theory. Learners will only be able to
handle corrections of their ‘mistakes’, not of their ‘errors’ (see p. 54). Of course,
they may be able to reproduce the teacher utterance containing the correction, but
studies of learners’ attempts at imitation (e.g. Naiman, 1974) suggest that in many
instances they can only reproduce items that are already part of their competence.
4 One possible explanation for the contradiction is that the two sets of researchers
used different operational definitions of such terms as ‘questions’, ‘statements’,
‘eliciting’ and ‘informing’. One of the difficulties of comparing the results obtained
from different studies rests in the lack of a widely accepted set of analytical
categories.
5 Teachers’ attitudes to the nature of the instructional task almost certainly act as a
major determinant of classroom behaviour. However, there have been surprisingly
few studies of the relationship between attitudes and behaviour in L2 classrooms.
6 There are two reasons for claiming that learners need the opportunity' to engage in
conversational discourse: (1) it enables them to develop discourse competence and
(2) it facilitates the development of linguistic competence. It has also been argued
(e.g. Ellis, 1984a; Van Lier, 1988) that the two are inextricably finked. That is,
through the process of developing discourse competence learners come to extend
their linguistic resources.
7 Ethnographic research is not only ‘exploratory’ and ‘illuminative’. It is also theory¬
building — designed to provide a valid interpretation of the L2 classroom. However, it
is probably true to say that much of the L2 classroom process research that has
taken place has not been theory-building in the strict sense of this term. It has not
produced a clearly articulated theory of L2 teaching—learning.
5 The Relationship
between Classroom
Interaction and
Language Learning

Introduction

In this chapter and the one that follows we will explore a number of theoretical
positions regarding classroom language learning. We will also examine some
of the research which these positions have generated and which has been used
to support them. It should be noted, however, that in many instances the
theoretical positions described in both chapters are speculative and contro¬
versial in nature. There is no consensus among applied linguists as to which
theory most adequately explains classroom language learning. No attempt will
be made to advance the claims of one theory over those of another. Rather the
approach adopted will be that of reviewing and evaluating each position and
the associated research. Each chapter, though, will conclude with a more
personal view of the teaching—learning relationship.
This chapter treats teaching as ‘interaction’ — the process by which samples
of the target language become available to the learner for interlanguage
construction through classroom interaction. In the next chapter teaching is
viewed as ‘formal instruction’ — the attempt to intervene directly in the
process of interlanguage construction by providing samples of specific linguistic
features for learning.
These two ways of viewing teaching are not to be treated as alternatives. It
is not intended to suggest that there are some classroom processes that can be
classified as ‘interaction’ and others as ‘formal instruction’. Everything that
happens in the classroom involves communication of one kind or another. It is
possible to view the same instructional sequence in two different ways — as
‘interaction’ and as ‘formal instruction’.
An example may help to make this basic point clearer. Consider the
following sequence:
94 Instructed Second Language Acguisition
T: R_, what is_?
LI: This is a book.
T: Wait. What is this?
LI: This is book.
T: ‘A’.
LI: This is a book.
T: Right. Very good. What are
these?
LI: What is _
L2: These are

LI: This - these are rubbers.

T: Rubbers.
LI: Rubbers.

As ‘interaction’, this sequence is characterized by tight topic control, the use

of display questions and evaluative and corrective feedback on the part of the
teacher and minimal responses from the learners. A linguistic feature is the
high frequency of third-person copula (‘is’ and ‘are’) in both the teacher’s
input and the learners’ output. As ‘formal instruction’, the sequence involves
an attempt to teach markers of plurality.
Not all instructional sequences can be viewed as both ‘interaction’ and
‘formal instruction’. Meaning-focused sequences (see chapter 1), where, by
definition, there is no attempt to teach specific linguistic items, can only be
viewed as ‘interaction’. It is form-focused sequences, such as the one above,
which can be treated as both ‘interaction’ and ‘formal instruction’. It should
also be noted that the interactional characteristics of sequences which are
meaning-focused are likely to vary greatly from those that are form-focused,
as we discovered in the previous chapter.
This way of looking at classroom processes differs from that currently
found in many teacher-training manuals, where ‘communication’ is contrasted
with ‘formal instruction’, with the assumption that all teaching is one or the
other. Such a contrast is misleading. We need to recognize that teaching
intended as formal instruction also serves as ‘interaction’. ‘Formal instruction’
does more than teach a specific item; it also exposes learners to features which
are not the focus of the lesson. For example, participation in the above
sequence may result in the acquisition of markers of plurality (the actual focus
of the lesson) but might also facilitate the acquisition of other linguistic
features such as third-person copula (which was not the focus of the lesson).
In other words, learners may respond to form-focused instruction as ‘interac¬
tion’ and acquire features not envisaged by the teacher.
These two perspectives lead to very different kinds of questions being asked
about the teaching—learning relationship. If we view teaching as ‘interaction’,
we will be interested in the extent to which the different input and interac¬
tional features contribute to acquisition. We might ask, for instance, whether
 Classroom Interaction and Language Learning 95
tight topic control helps L2 learning. If we view teaching as ‘formal instruction’,
we will be concerned with whether the attempt to teach specific linguistic
features results in their acquisition.
This chapter is concerned only with ‘interaction’ and the learning that
derives from it. Interaction can be hypothesized to contribute to learning in
two ways: (1) via the learner’s reception and comprehension of the L2 and (2)
via the learner’s attempts to produce samples of the L2. Reception-based
theories of L2 acquisition emphasize the importance of input, as opposed to
learner output. The frequency hypothesis states that learners acquire linguis¬
tic features according to their frequency in the input; features which occur
frequently are learnt before those which occur infrequently. The evidence for
this hypothesis lies in quantitative studies which have correlated the frequency
of input features with measures of acquisition. Krashen (1985) has advanced
the input hypothesis (see p. 57) to account for how learners’ interlanguages
develop as a result of comprehending input that contains linguistic features
one step beyond their current knowledge. Long (1983a) proposes the inter¬
action hypothesis, according to which input is made comprehensible as a
result of modification to the interactional structure of conversations when
communication problems arise. Both the input and the interaction hypotheses
have attracted considerable interest and have served as the basis for a number
of strongly argued pedagogical prescriptions. For this reason, they will be
considered in some detail.
There are also a number of production-based theories. Swain (1985) puts
forward the output hypothesis which claims that learners need the opportunity
to produce in order to develop native-speaker levels of grammatical proficiency.
Ellis (1984c) proposes the discourse hypothesis to account for the fink
between input and production. Like Swain, Ellis argues that output is essential
for learning. He suggests that the nature of the linguistic competence which
learners acquire depends on the nature of the discourse in which they partici¬
pate. The discourse hypothesis is closely linked to theories of L2 variability
(cf. Tarone, 1983). Next, the role of collaborative discourse in assisting
learners to produce new syntactic structures is considered. Studies which have
explored this aspect of interaction have made use of ‘discourse analysis
methodology’ (Hatch, 1978b; 1978c). Finally, the topicalization hypothesis
is considered. This states that learners are more likely to obtain intake when
they — rather than the teacher — initiate and control a topic.
This plethora of hypotheses (shown in figure 5.1) testifies to the importance
which researchers have attached to investigating ‘interaction’ as the matrix of
L2 acquisition. It also testifies to the lack of agreement regarding how class¬
room interaction contributes to the acquisition of new linguistic knowledge.
We are a long way from being able to answer the two central questions: (1)
What kind of communication is best for language learning? and (2) How much
communication (of whatever kind) is needed for learning? This chapter con¬
cludes with a consideration of what constitutes an optimal mode of classroom
communication for rapid and successful learning.
96 Instructed Second Language Acquisition
r- frequency hypothesis (1)

reception-based — input hypothesis (2)

theories

- interaction hypothesis (3)

instruction as _
interaction
[— output hypothesis (4)

- discourse hypothesis (5)

production-based_
theories
_ collaborative-discourse
hypothesis (6)

topicalization
hypothesis (7)

Figure 5.1 Theories of classroom learning based on instruction as interaction

Reception-based Theories

In this section we will consider a number of hypotheses based on the assump¬

tion that L2 acquisition occurs when the learner processes input data.
Reception-based theories are of two kinds: those that emphasize exposure,
such as the frequency hypothesis, and those that emphasize the need for
comprehension of the input data, such as the input and interactional hypotheses.
It is helpful to begin by defining a number of key terms. Input refers to the
target language samples to which the learner is exposed. It contains the raw
data which the learner has to work on in the process of interlanguage con¬
struction. Corder (1967) distinguishes input and intake, the latter consisting
of that portion of the input which the learner actually attends to and, therefore,
uses for acquisition. Not all input serves as intake as only a subset of the total
samples available are salient to the learner at any one stage of development.
Interaction refers to the process of interpersonal communication; it involves
the efforts of both die learner and the teacher. Input is supplied and made
comprehensible through interaction.

The Frequency Hypothesis

The frequency hypothesis states that the order in which learners acquire L2
grammatical features is determined by the frequency of those features in the
input; more frequent features are acquired before less frequent. The hypoth¬
esis makes no claims about the need for learners to comprehend the input.
Input simplifications and interactional adjustments of the kind found in teacher-
talk (see previous chapter) are important because these affect the frequency
with which specific linguistic features occur in the input at different stages of
the learner’s development.
The frequency hypothesis has been tested by means of correlational studies.
 Classroom Interaction and Language Learning 97
These relate measures of the frequency of different linguistic features in the
input to measures of acquisition of the same features, using statistical tech¬
niques. Table 5.1 provides a summary of the main classroom correlational
studies. It will be seen that the results are very mixed. Larsen-Freeman
(1976a & b), for instance, found a significant relationship between features in
the input and output, but Long and Sato (1983) did not. There may be a
number of reasons for this — for example, whether the learners are elementary or
advanced. The studies also have design problems. Larsen-Freeman and Long
and Sato correlated unrelated sets of input and output data.
It makes little sense to look for input—output relationships in data collected
at the same time. Any effect that input might have will only become apparent
at some point after the input has been supplied. In this respect, it is worthwhile
reporting Wells’s (1985) finding for LI acquisition. He observed a rapid
increase in adult caretakers’ use of specific linguistic forms, such as auxiliary
verbs, just before the emergence of the same features in the children’s speech.
Wells suggests that the frequency of more complex items might be kept low
until the child is ready. Evidence for a similar delayed relationship in classroom
L2 acquisition comes from Lightbown’s (1983) study. She found that learners
in Grade 6 ESL classrooms in Canada evidenced overuse of V-ing even
though this feature was relatively infrequent in the classroom language at that
time. However, they had received extensive exposure to this form during
Grade 5. Lightbown also notes that the use of V-ing fell away in Grade 7.
The possibility of instruction having a delayed effect is explored more fully in
the next chapter.
The non-authenticity of classroom discourse has already been discussed, in
chapter 4. It has implications for the frequency with which some linguistic
forms are either under- or over-represented in the input. Lightbown (1985b)
provides a number of examples of how this arises and the effects it has on
acquisition. For example, she found that Grade 6 French learners of L2
English underused ‘have’ in their free speech. They even failed to produce
the familiar error:

* I have twelve years old.

This could be explained by the absence of ‘have’ in the input in previous

years. However, once ‘have’ was introduced, the learners rapidly switched to
using it, including in sentences like the one above. V-ing is an example of a
feature that is commonly over-represented. Lightbown argues that low or high
frequency of a form in the input is a reflection of the distorted nature of
classroom dialects and can result in pseudo-acquisition of that form.
There are two major problems with the frequency hypothesis. The first is
that it is based on the assumption that linguistic items are acquired sequentially,
i.e. it draws heavily on the idea of a ‘morpheme order’, which, as we have seen
in chapter 3, is controversial. Second, the hypothesis does not really constitute
an explanation of any kind. It does not tell us why input frequency should
affect acquisition. Hatch and Wagner-Gough (1976) point out that there is an
Table 5.1 Correlational studies of classroom input and L2 learner production

Study Design Principal results

Larsen-Freeman, 1976a Accuracy orders of ESL The results reported are

learners correlated with generally significant.
input frequency orders of
same morphemes in the
classroom speech of 2 ESL
teachers.

Larsen-Freeman, 1976b L2 accuracy orders of Significant relationships

grammatical morphemes found for all accuracy
correlated with order of orders except that derived
frequency of the same from imitation task.
morphemes in parental
speech to children (based
on Brown, 1973). The
orders were calculated
using data from a variety of
tasks.

Hamayan and Tucker, The order of accurate Significant rank order

1980 production of nine French correlations reported for
syntactic structures in the both LI and L2. Also
elicited speech of children orders different for
speaking French as a LI different classrooms but
and learning French as a relationship between input
L2 was correlated with and production orders
teachers’ frequency of remained significant.
production of the same
structures in the
classrooms.

Snow and Hoefnagel- Achievement scores and No significant relationships

Hohle, 1982 gain scores of American reported.
children acquiring L2
Dutch in the Netherland
correlated with (1) quantity
of Dutch heard in the
classroom, (2) percentage
of speech directed
specifically at them and (3)
quantity of speech directed
at individual learners.

Long and Sato, 1983 Krashen’s (1977) average No significant relationship

order of acquisition of found. However, the
grammatical morphemes teachers’ production order
correlated with the order did correlate with native-
of frequency of the same speakers’ production order
morphemes in ESL in conversations and this
teachers’ speech. latter order correlated with
Krashen’s acquisition
order.
 Classroom Interaction and Language Learning 99
Lightbown, 1983 Frequency of production No direct relationship
of a number of between input frequency
grammatical morphemes and learners’ morpheme
(e.g. -s morphemes) by order found at the same
students in an oral time. But there was
communication game evidence that a high level
correlated with the of frequency of some
frequency of same morphemes (e.g. -ing)
morphemes in samples of resulted in subsequent
one teacher’s speech and overuse by the students.
in the textbook used by the Frequency may, therefore
students. The study was have a delayed effect.
longitudinal.

overlap between frequency, structural complexity and cognitive learning diffi¬

culty. The correlation between input and acquisition, therefore, is difficult to
interpret, because other factors are confounded. It may not be frequency per se
that counts but structural complexity. The hypothesis is descriptive rather
than explanatory. It is of value from the point of view of language pedagogy,
however, in that it highlights the need to consider instruction in terms of the
total input to which the learner is exposed.

The Input Hypothesis

The other two hypotheses which are reception-based are explanatory in
nature. They offer much more powerful accounts of how classroom input and
interaction contribute to L2 acquisition.
The input hypothesis has already been briefly considered as part of the
Monitor Model in chapter 3. It warrants a more detailed consideration,
however. Its importance is reflected in the fact that Krashen has devoted a
whole book to it (Krashen, 1985) and in the attention it has attracted from
language teachers. A detailed description of Krashen’s claims about the role
of input follows, together with a review of the research which has been used to
support them. Finally, the hypothesis is evaluated.

Description Krashen provides a number of accounts of the input hypothesis,

the most extensive of which is to be found in Krashen (1985). The hypothesis
takes the form of a principal statement and two corollaries (see Table 5.2).

Long (1983a) makes some equally strong claims:

1 Access to comprehensible input is characteristic of all cases of successful

acquisition, first or second.
2 Greater quantities of comprehensible input result in better (or at least
faster) acquisition.
3 Lack of access to comprehensible input (as distinct from incomprehen¬
sible, not any, input) results in little or no acquisition.
100 Instructed Second Language Acguisition
Table 5.2 The Input Hypothesis

Principal statement: Learners progress along the natural order by under¬

standing input that contains structures that are a little bit
beyond their current level of competence (i.e. i + 1).

Corollary 1: Speaking is a result of acquisition, not its cause; it emerges

as a result of building competence via comprehensible
input.

Corollary 2: If input is understood and there is enough of it, the

necessary grammar is automatically provided; the teacher
does not need to teach the next structure.

Source: Based on Krashen, (1985).

For Krashen the input hypothesis is the central hypothesis of the Monitor
Model. He calls it the Fundamental Principle. This is because he sees
comprehensible input as the necessary (although not sufficient) condition
for ‘acquisition’.1 It is not possible to control what the learner acquires nor the
order in which she acquires it. Both are determined by the internal language
processor or language-acquisition device (LAD) which regulates what features
in the input are attended to at each stage of development. However, the LAD
can function properly only if the learner receives input that contains structures
not yet learnt and if, despite this, the learner is able to understand the input.
Krashen describes two ways in which comprehension of input containing
new linguistic material is achieved: the utilization of context by the learner and
the provision of simplified input by the teacher. The learner makes use of
context to infer the meaning of an utterance when existing linguistic resources
are insufficient for immediate decoding. Three kinds of contextual information
are available: extra-linguistic information, the learner’s knowledge of the
world and the learner’s previously acquired linguistic competence. Krashen
refers to a number of studies which demonstrate the ‘dramatic effects’ that
contextual information can have on the comprehension of written texts; a
study by Adams (1982), for example, was able to show a sixfold improvement
in the comprehension of new lexical material when background information
was made available.
Krashen does not illustrate how the learner makes use of context, although
it is not difficult to think up examples. A passive sentence like ‘Raza was given
a pencil by Surjit’ can be understood even though the passive is not yet part of
a learner’s interlanguage if it is accompanied with non-verbal behaviour which
makes clear who the agent and patient are. A passive sentence like ‘Raza was
bitten by the dog’ can be understood correctly even without extra linguistic
information as the learner will know that the action of biting is performed by
dogs on humans; this is part of her world knowledge. A passive sentence like
‘Raza was hit’ can be understood even without extralinguistic information or
world knowledge if the learner’s existing linguistic competence includes knowl-
 Classroom Interaction and Language Learning 101
edge that ‘hit’ requires a patient; ‘Raza’ must perform the role of patient
because it is the only argument of the predicate ‘hit’ in the sentence.2
Krashen also places great store on the provision of simplified input, although
he acknowledges that input can be comprehensible without any simplification.
He emphasizes that simplification is designed to promote communication
rather than to teach and that it results in ‘rough’ rather than ‘fine tuning’. By
‘rough tuning’ he means that the input is not exactly related to the learner’s
developmental level so that it does not contain precisely the next rule the
learner is ready for. Thus the learner’s ‘next rule’ is not present in every
utterance. However, providing there is enough comprehensible input the
learner is assured of exposure to the necessary grammar at the right time.
Roughly tuned Input makes the task of comprehension by means of inference
easier. It ensures that the learner is not faced with input which overloads her
processing and inferencing capacity by removing excess rules that are way
beyond what is immediately acquirable. Also input to beginner learners typically
relates to ‘here-and-now’ topics which facilitate comprehension by means of
extralinguistic information. Krashen also argues that the simplified input
found in interlanguage talk (i.e. communication between L2 learners) can
promote learning.
Simplified input can be made available to the learner through one-way or
two-way interaction. Examples of the former include listening to a lecture,
watching television and reading. Two-way interaction occurs in conversations.
Krashen acknowledges that two-way interaction is a particularly good way of
providing comprehensible input because it enables the learner to obtain
additional contextual information and optimally adjusted input when meaning
has to be negotiated because of some communication problem. However, he
rejects claims (of the kind made by the interaction hypothesis — see below)
that two-way interaction is necessary for comprehensible input. According to
Krashen, input can be comprehensible without any active participation on the
part of the learner. Advanced learners can benefit from watching television.
Also there is plenty of evidence to support the facilitative effects of extensive
reading on acquisition.

Evidence Krashen offers no direct evidence in support of the input hypoth¬

esis. That is, he reports no studies which show that roughly tuned input
containing ‘i + T results in learners advancing to the next step in their
interlanguage development. There are, however, various kinds of indirect
evidence which lend support to the hypothesis. Some of the most important
are considered below.
The input hypothesis predicts that learners will be able to learn specific
grammatical features from classroom input irrespective of whether these are
formally taught. Terrell et al. (1980) were able to show that classroom
learners of L2 Spanish acquired the ability to produce interrogatives even
though no explicit teaching was provided. Classroom input is presumably rich
in interrogatives (see chapter 4) and these are often predictable and repetitive.
102 Instructed Second Language Acguisitlon
As a result they are readily comprehensible. Other studies have shown
that the acquisition of some linguistic features is delayed in classroom L2
acquisition, possibly because they are absent from the input. For example,
past tense forms occur infrequently in teachers’ language and are acquired
late by classroom learners (Ellis, 1982; Long and Sato, 1983).
A number of studies have shown a significant relationship between features
of modified input and learner comprehension. These studies have all involved
LI acquisition, however. Cross (1977), for instance, reported significant cor¬
relations between the mean lengths of the caretaker’s and child’s utterances (a
measure which gives an indication of general syntactic complexity) and an
even higher correlation between the caretaker’s Mean Length of Utterance
(MLU) and the child’s comprehension. Krashen — like Cross — interprets
these results as indicating that caretakers adjust their input in rough accordance
to the child’s ability to comprehend and produce. Cross’s study, however, was
cross-sectional and could not be used to make claims for any causative effect
of simplified input on learning. Other LI acquisition studies have failed to
show any relationship between formal characteristics of caretaker input and
learning (e.g. Gleitman et al., 1984). Even if the LI acquisition research was
uniformly supportive of the input hypothesis it would still need to be shown
that similar relationships occurred in L2 acquisition research. There has been
a singular lack of such studies. However, given that Krashen (1985:27)
accepts that learners may be able to learn from ‘non-tuned’ input (he refers to
the Gleitman et al. study) providing it is still comprehensible, this lacuna is
not very damaging to the hypothesis.
Long (1985a:378) has suggested that one way of demonstrating the
relationship between comprehensible input and acquisition is to break the task
down into three steps:

1 Show that (a) linguistic/conversational adjustments promote (b) com¬

prehension of input.
2 Show that (b) comprehensible input promotes (c) acquisition.
3 Deduce that (a) linguistic/conversational adjustments promote (c)
acquisition.

Evidence showing that simplified input aids comprehension can, therefore,

contribute support to the hypothesis. However, the evidence available is quite
limited. Chaudron (1985) in a review of the literature was able to find only
one study (Kelch, 1985) demonstrating that a slower rate of speech led to
increased comprehensibility. In this study slowing down from 200 words per
minute to 130 w.p.m. resulted in significandy improved dictation scores. Long
(1985a) investigated the effects of a number of modifications of lecture input
on L2 university students’ comprehension. The modifications included syntactic
complexity, rate of speech and the addition of rephrasings and restatements —
a mixture of input and interactional features. The problem with Long’s study
is that it is not possible to say whether it was the input or the interactional
 Classroom Interaction and Language Learning 103
modifications that assisted comprehension. Pica, Young and Doughty (1987)
carried out a study designed to investigate which type of modification was
most effective. They compared the effects of pre-modified input in a lecturette
with those of interactionally modified input that arose in two-way exchanges
during a similar lecturette with no pre-modified input. They found that
comprehension was assisted by the interactional modifications, whereas re¬
duction of linguistic complexity in the pre-modified input had less effect. The
importance of input modifications as an aid to comprehension cannot be
considered proven by any stretch of the imagination. Even if it is finally
demonstrated that they play their part, it is likely that other sources of
comprehensibility (i.e. interactional modifications and contextual knowledge)
will turn out t6 be more significant.
There have also been a number of studies which have investigated
the relationship between the frequency of specific linguistic features in the
input and in learner output. These are not considered here, however, as they
throw no fight on the input hypothesis. These studies demonstrate a role for
frequency but not for comprehensible input (see the earlier discussion of the
frequency hypothesis).
Krashen (1985) claims that the input hypothesis accounts for other research
findings. Learners typically go through a silent period before they begin to
produce in the L2; this serves as a means of building up competence through
listening. Language classes have been shown to be important in the early
rather than late stages of language learning; this is because the learner finds
difficulty in obtaining comprehensible input in naturalistic contexts in the
beginning. Methods that supply plenty of comprehensible input have been
shown to be more successful than methods that are based on formal language
study or practice.3 For example, Asher’s Total Physical Response — a method
that requires the learner to respond non-verbally to instructions for a lengthy
period before production is allowed (Asher, 1977) — produces consistently
successful results. Immersion classes based on ‘comprehensible subject-matter
teaching’ have met with tremendous success in Canada (Swain and Lapkin,
1982). Krashen argues that this is because they provide the learner with
plenty of comprehensible input. The skill of writing has been shown to be
fostered by large amounts of ‘self-motivated reading’ rather than by the
deliberate teaching of writing skills.4 Once again this is because reading
provides plenty of comprehensible input containing the subtle grammatical
and discourse features that characterize the written medium. For Krashen all
this constitutes ‘evidence’ in favour of the input hypothesis.

Evaluation Evaluation of the input hypothesis can proceed in two ways.

First, the quality of the evidence which Krashen cites in favour of it can be
considered. Second, the hypothesis can be subjected to theoretical appraisal.
The lack of direct supporting evidence has already been commented on. It
is unlikely that such evidence will ever become available. An empirical study
would need to show (1) that only comprehensible input promotes acquisition
104 Instructed Second Language Acguisition
and (2) that it promotes acquisition of only that linguistic feature which is next
in line along the natural route the learner is following. Krashen gives no
suggestions for how such a study could be carried out, nor is it easy to
envisage what it would look like.
The hypothesis can be tested, therefore, only indirectly — in the way that
Long (1985a) suggests. However, to date there have been only a few studies
which have shown that simplified input facilitates comprehension (see above)
and these are far from conclusive. There is some evidence to suggest that
simplified input can interfere with learning. Chaudron (1983), for example,
illustrates how the pressure to achieve comprehension can lead to either
ambiguous oversimplification or to confusing over-elaboration. For simplifi¬
cation to be effective the teacher needs to achieve a balance between these
two extremes — no easy task, given that a class is likely to contain mixed-
ability learners. There has also been very little research into the relationship
between comprehension and acquisition. Faerch and Kasper (1986) in an
article entitled ‘The Role of Comprehension in Second-language Learning’
fail to cite even one empirical study that has investigated this relationship.
Even the indirect empirical evidence is, therefore, slim.
The other kind of evidence cited by Krashen relates to various findings
about L2 acquisition which are compatible with the predictions of the input
hypothesis (e.g. the results of immersion studies). The argument is a circular
one, however. For example, Krashen (1985:17) states: ‘The Input Hypothesis
thus asserts that it is the comprehensible input factor that is responsible for
the success of immersion’ and at the same time argues that the success of
immersion is evidence in favour of the input hypothesis. The hypothesis is
used to interpret the research results which are then claimed as evidence in
favour of the hypothesis. In addition, it should be noted that the kind of
evidence provided by the immersion and other kinds of studies do not demon¬
strate that comprehensible input results in the acquisition of i + 1. At best,
they only constitute evidence that acquisition in general benefits from the kind
of linguistic environment provided in immersion classrooms. Later, we will
find that even Krashen’s claims regarding the success of immersion programmes
is questionable.
The input hypothesis has also been subjected to a number of theoretical
criticisms (cf. Gregg, 1984; Faerch and Kasper, 1986; White, 1987;
McLaughlin, 1987). These are summarized briefly below:

(1) The ‘i + 1’ construct is not operationally defined. This construct derives

from another of the hypotheses of the Monitor Model — the Natural Order
hypothesis - but not all researchers agree that L2 acquisition is characterized
by a rigid sequence of development. Also it is not clear how input containing
‘i + 1 ’ can result in transitional structures of the kind found in the acquisition
of negatives and interrogatives. These structures are deviant and yet constitute
an advance in acquisition. Krashen sees learning as an incremental process by
which one rule is added to another, but acquisition may be better viewed as a
 Classroom Interaction and Language Learning 105
process of constant restructuring of interim grammars involving die gain of
new features, the functional re-employment of some old features and the
eradication of other old features.5 It is not clear how the presence of some
unacquired feature in the input can generate such changes. White (1987)
comments that what is needed is ‘a theory of what precise aspect of the input
interacts with what aspect(s) of the learner’s existing system’. The input
hypothesis does not constitute such a theory.

(2) The relationship between comprehension and acquisition is not clearly

spelled out. Faerch and Kasper (1986) point out that there is a distinction to
be drawn between input functioning as intake for comprehension and input
functioning as'intake for acquisition. It is particularly difficult to see how
understanding input can result in the acquisition of third person -s or many
other grammatical morphemes, as these contribute no essential mean¬
ing to messages and are, therefore, redundant as far as comprehension is
concerned.

(3) Krashen’s account of how comprehension takes place is inadequate. Faerch

and Kasper (1986) provide an excellent summary of language-comprehension
processes.6 They emphasize that they consist of both ‘bottom—up processes’
and ‘top-down processes’. The former are based more or less exclusively on
the input and occur when the hearer is unable to activate contextual information
(as, for instance, when one turns on the radio in the middle of a broadcast).
The latter are based on the recipient’s existing knowledge; some messages are
so ritualized (as in service encounters) that they can be understood with
almost no processing of the actual input. This distinction is of importance for
the input hypothesis, as presumably acquisition of new linguistic material can
only take place when the learner attends to actual input (i.e. when some
degree of bottom—up processing occurs).

(4) Simplified input may impede rather than facilitate acquisition. Gregg
(1984:90—1) claims that ‘no-one — certainly not Krashen — has shown that
caretaker speech (CS) makes any significant contribution toward the acquisition
of a grammar by a child, and no one has shown that the existence of CS has
any bearing on second language acquisition’. White (1987) goes as far as to
suggest that simplified input may in fact be detrimental. Interlanguage talk, for
example, is likely to be lacking in just those features which the learner needs
to acquire. White also argues that simplified input from native speakers can
cause deprivation; for example, talking to learners in simple sentences deprives
them of crucial input. Gregg and White overstate their cases. There is
evidence to suggest that simplified input is involved in acquisition. Wells
(1985), for instance, reports a marked increase of linguistic features in care¬
taker speech shortly before they first appear in the speech of children during LI
acquisition. This, of course does not prove that input causes acquisition but it
is highly likely that it at least contributes to the process.
106 Instructed Second Language Acquisition
(5) Some acquisition is not based on input at all. Krashen gives passing
recognition to the fact that some changes in interlanguage are internally
driven, i.e. result from the learner’s current lexical and syntactical knowledge.
This aspect of L2 acquisition is played down, however. There are two aspects
of acquisition which are ‘input-free’. The first concerns features which are
triggered off by the acquisition of some other feature, which may be acquired
from the input. For example, there is some evidence to suggest that exposure
to input containing expletive pronouns (‘There/It is . .. ’) provides the necessary
evidence for learners to recognize that English does not permit pronoun drop
(as in ‘*Is very busy’). The acquisition of pronouns appears to trigger off the
acquisition of modal verbs, so that the learner does not require any input data
relating to this grammatical feature. Zobl (1983) suggests that learners must
possess a projection device to enable them to learn in this way.
The second aspect of acquisition which is input-free concerns the loss of
deviant interlanguage rules. Researchers working within the Chomskyan tra¬
dition of Universal Grammar argue that the input does not typically supply the
learner with the direct negative evidence (in the form of corrections, for
instance) that is needed to ‘unlearn’ overgeneralizations. No amount of com¬
paring such forms with the input will demonstrate to the learner that she is
wrong. For example, if the learner produces sentences such as:

* John donated the hospital some money,

by overgeneralizing from sentences like:

John gave the hospital some money.

comprehending input will not help the learner to see that she is wrong. To
unlearn this kind of overgeneralization the learner must either be equipped
with innate knowledge or must receive direct negative evidence in the form
of instruction. Neither has any role in acquisition according to the input
hypothesis.

There are two other theoretical objections to the input hypothesis: (6) concerns
Krashen’s rejection of formal instruction as a basis for promoting ‘acquisition’
(as opposed to ‘learning’) and (7) concerns his rejection of output as a
causative factor in acquisition. Both points are dealt with elsewhere. The next
chapter covers (6), while (7) is considered in a later part of this chapter.
The input hypothesis is a bucket full of holes. It is an example of the
‘tremendous leaps’ which researchers make from ‘low-level constructs about
input itself to high-level constructs about their effects on L2 acquisition
(Hatch, 1983b). As a basis for research the hypothesis is clearly inadequate,
although, as White (1987) points out, there is still a need for an input
hypothesis of a more precise kind. Krashen’s input hypothesis is not without
value for language pedagogy, however. It provides a statement of important
principle, namely that for successful classroom acquisition learners require
 Classroom Interaction and Language Learning 107
access to message-oriented communication that they can understand. It also
provides a rough explanation of why this might be so. The main problem with
Krashen’s hypothesis is that it is nothing like as ‘fundamental’ as he claims.
There is more to teaching than ‘comprehensible input’.

The Interaction Hypothesis

Like the input hypothesis, the interaction hypothesis emphasizes the import¬
ance of comprehensible input. It also seeks to explain how acquisition comes
about and makes claims regarding which kinds of interaction will best promote it.
The interactidn hypothesis has led to a substantial amount of classroom
research, although, as we will see, this research has not been directed at
falsifying the hypothesis. Rather, it has been based on the assumption that the
hypothesis is correct.

Description The interaction hypothesis differs from the input hypothesis in

the way it attaches importance to one particular method of making input
comprehensible. There are three ways of making input comprehensible (Long,
1983a):

1 by means of input simplifications;

2 through the use of linguistic and extra-linguistic context; and
3 through modification of the interactional structure of conversation.

Long (1981) refers to (1) as input features and (3) as interactional features
(see chapter 4, p. 70). In formulating the interaction hypothesis, Long argues
that acquisition is made possible and is primarily facilitated when interactional
adjustments are present.
The main interactional features which have been found to vary in foreigner-
talk are described and exemplified in table 5.3. The presence of such features
indicates that there is negotiation of meaning taking place. Long (1983c)
identities two kinds of negotiation: (1) negotiation aimed at avoiding conver¬
sational trouble and (2) negotiation aimed at repairing discourse when trouble
occurs. Interactional modifications that derive from (1) reflect long-term
planning by the native speaker and govern the way entire conversations are
conducted. Long describes these modifications as strategies. Examples are
‘Relinquish Topic Control’, ‘Select Salient Topics’ and ‘Treat Topics Briefly’.
Interactional modifications that derive from (2) are spontaneous and mainly
affect how topics are talked about. They are referred to as tactics. Examples
include ‘Request Clarification’, ‘Confirm own Comprehension’ and ‘Tolerate
Ambiguity’. Other researchers have tended to restrict the term ‘negotiation of
meaning’ to (2). Varonis and Gass (1985:39), for instance, define negotiation
exchanges as ‘those exchanges in which some overt indication that under¬
standing between participants has not been complete and there is a resultant
attempt to clarify nonunderstanding’.
Table 5.3 Interactional modifications involved in the negotiation of meaning

Interactional feature Definition Example

Clarification requests Any expression that elicits A: She is on welfare.

clarification of the B: What do you mean by
preceding utterance. welfare?

Confirmation checks Any expression immediately A: Mexican food have a

following the previous lot of ulcers?
speaker’s utterance B: Mexicans have a lot of
intended to confirm that the ulcers? Because of the
utterance was understood food?
or heard correctly.

Comprehension checks Any expression designed to A. There was no one

establish whether the there. Do you know
speaker’s own preceding what I mean?
utterance has been
understood by the
addressee

Self-repetitions:
(1) repairing The speaker repeats/ A: Maybe there would
paraphrases some part of be¬
her own utterance in order ll: Two?
to help the addressee A: Yes, because one
overcome a communication mother goes to work
problem and the other mother
stays home.

(2) preventive The speaker repeats/ A: Do you share his

paraphrases some part of feelings? Does anyone
her own utterance in order agree with Gustavo?
to prevent the addressee
experiencing a
communication problem.

(3) reacting The speaker repeats/ A: I think she has a lot of

paraphrases some part of money.
one of her previous B: But we don’t know
utterances to help establish that?
or develop the topic of A: But her husband is very
conversation. rich.

Other-repetitions:
(1) repairing The speaker repeats/ A: I think the fourth
paraphrases some part of family.
the other speakers utterance B: Not the fourth family,
in order to help overcome a the third family.
communication problem.
 Classroom Interaction and Language Learning 109
(2) reacting The speaker repeats/ A: I think she has three
paraphrases some part of children.
the other speaker’s B: This is the thing. She
utterance in order to help has three children.
establish or develop the
topic of conversation.

Source: Based on Pica and Doughty (1985a)

The interaction hypothesis also considers two-way communication of greater

importance for acquisition than one-way. This is because negotiation of
meaning is much more extensive when the learner is able to provide feedback
on her success‘in understanding the native speaker. Negotiation of meaning,
therefore, involves the learner as well as the native speaker/teacher. In this
respect, therefore, the interactional hypothesis takes learner production as
well as input into account. However, like the input hypothesis, it sees no
direct role for output in promoting acquisition.
Figure 5.2 summarizes how interaction leads to acquisition, according to
Long (1983a:214).

Figure 5.2 Model of the relationship between type of conversational adjustments and
language acquisition

Evidence There have been no L2 studies investigating whether interactional

adjustments promote acquisition. However, there have been more interesting
results obtained from LI acquisition research which has investigated the same
question. Wells (1985) reports on the relationship between various input
features and gains in acquisition by children. He found that children who
received larger amounts of speech addressed to them showed larger gains in
the mean length of their longest utterances, semantic complexity, noun phrase
structure and clause structure. Not surprisingly, sheer quantity of exposure
counts. The facilitative effect of a number of specific interactional features,
however, was independent of the general effect for amount of speech. For
example, both direct requests and ‘extending utterances’ (i.e. utterances which
either added new propositional content to the topic or encouraged the child to
do so) were related to progress in acquisition and were independent of the
relationship between amount of speech and acquisition gains. In contrast,
other interactional features (e.g. paraphrases and repetitions) tended to increase
in proportion to amount of speech. There are obvious dangers in trying to
apply the results of LI acquisition research to L2 acquisition (see chapter 3),
110 Instructed Second Language Acguisltlon
but it is somewhat surprising that research of the kind reported by Wells has
received no attention from proponents of the interaction hypothesis. This
research both gives general support to the claim that interaction facilitates
acquisition (at least as far as rate is concerned) and also raises a number of
interesting questions regarding which interactional features are the crucial
ones.
The evidence cited in support of the interaction hypothesis is indirect in
nature, consisting largely of theoretical argument. Two principal arguments
can be identified. The first is based on research evidence which shows that
foreigner-talk is characterized by interactional rather than input modifications.
For example, Long (1981) carried out a study involving a comparison of
conversations between sixteen native-speaker dyads and conversations between
sixteen native and non-native-speaker dyads and found that whereas there
were differences in only one out of the five input features he investigated,
differences in ten out of eleven interactional variables reached statistical
significance. In other words, there simply may not be much in the way of
input modifications. If foreigner-talk is necessary for acquisition (as Long
claims), then it must be because of the interactional features. The second
argument concerns the interactional modifications that occur when there is
some communication difficulty (i.e. tactics). For instance, in the example
below (from Long, 1983c) the native-speaker’s repetition (turn 3) is a response to
the learner’s apparent lack of comprehension, while the request for confirmation
(turn 5) occurs because the native speaker wishes to make certain she has
understood correctly:

1 NS: When did you finish?

2 NNS: Um?
3 NS: When did you finish?
4 NNS: Ten o’clock.
5 NS: Ten o’clock?
6 NNS: Yeah.

The modifications serve to make input which has proved problematic to the
learner comprehensible. Through them, it is argued, the learner’s attention is
drawn to new linguistic material, which then becomes intake. Thus, in the
above example the learner may be helped to focus attention on how past tense
interrogatives are formed.7
Indirect support for the claim that negotiation leading to interactional
adjustments results in more optimal input than do formal simplifications has
been adduced from a study carried out by Scarcella and Higa (1981). They
found that child learners received simpler input than adolescent learners.
However, the latter participated more actively and generated more interactional
adjustments from their native-speaker interlocutors. On the grounds that
research has shown that adolescents learn more rapidly than young children
(Snow and Hoefnagel-Hohle, 1978), it can be inferred that negotiation leading
 Classroom Interaction and Language Learning 111
to native-speaker assistance in sorting out misunderstandings is more beneficial
for learning than simplified input.
There is also some evidence to suggest that interactional adjustments
facilitate comprehension. Reference has already been made to a study carried
out by Pica, Young and Doughty (1987). They found that 73 per cent of the
repetitions in the native speaker’s two-way lecturette were triggered by inter¬
actional modification moves consisting of requests for clarification, confirmation
checks and comprehension checks. Only 27 per cent were not triggered by
some sort of interactional move. They consider it likely that comprehension
was assisted by these interactionally motivated repetitions. Other interactional
features have also been found beneficial. Chaudron and Richards (1986)
examined the effect of different kinds of discourse markers on the compre¬
hension of lectures by pre-university and university ESL students. They
found that macro-markers (i.e. signals or metastatements about the major
propositions in a lecture such as ‘What I’m going to talk about today ...’)
aided the learners’ retention of lecture content to a greater extent than micro¬
markers (i.e. signals of intersentential relations by means of connectors such
‘then’, ‘because’, ‘but’). Comprehension in conversations (as opposed to lec¬
tures) has been poorly researched, however, and doubts remain as to whether
modifications to the structure of interactions help with comprehension (see
below).

Classroom Research Based on the Interaction

Hypothesis
The lack of research lending direct support to the interaction hypothesis has
not inhibited researchers from basing classroom research on it. The general
assumption is that the greater the amount of interactional modification there
is, the more rapid and successful acquisition will be. A number of issues have
been investigated:

1: Negotiation in classroom discourse vs. negotiation in naturally oc¬

curring conversations The previous chapter reviewed a number of studies
which have compared classroom and naturalistic discourse so there remains
little to add here. One study — Pica and Long (1986) — has specifically
examined the nature of interactional modifications in the classroom. It compared
the classroom speech of ten ESL teachers with the speech of native speakers
conversing with non-native speakers whom they had met for the first time (cf.
Long, 1981). There was no difference in linguistic complexity but there were
functional differences. Teachers produced more questions and imperatives
and fewer statements. The teachers used more display questions and compre¬
hension checks but fewer referential questions, confirmation checks and clari¬
fication requests.8 Pica and Long conclude that there is less negotiation of
meaning in classroom settings and suggest that, as a result, there is less
comprehensible input.
112 Instructed Second Language Acquisition
2: The effect of task type on the amount of interactional adjustment that
takes place Crookes (1986) and Long (1988b) have both argued that L2
classroom research should be directed at ‘psycholinguistically relevant design
features’. Where tasks are concerned such a feature is ‘one which can be
shown to affect the nature of the language produced in performing a task in
ways that are relevant to SL [i.e. L2] processing and SL learning’ (Crookes,
1986:7). A number of such task characteristics have been investigated — in
particular the difference between one-way and two-way tasks. One-way infor¬
mation-gap tasks are tasks that do not require participants to pool information;
contributions from individual learners are a matter of choice. Two-way
information-gap tasks are tasks that equip all the participants with information
which they must share in order to solve a problem. Long (1983a) found
evidence of more interactional modification in two-way than in one-way
tasks in conversations between pairs of native speakers. However, not all
studies have produced similar results. Gass and Varonis (1985) were unable to
find any significant differences between the two types of task in interlanguage
talk involving intermediate-level university students studying L2 English. The
research is, therefore, not conclusive. This is a reflection of the difficulty of
identifying and controlling variables that affect the interaction resulting from
different tasks. For example, the intrinsic interest generated by a task may well
be more important than whether the task is one-way or two-way. Also the
distinction between one-way and two-way tasks is less clear-cut than Long’s
initial distinction recognized. As Gass and Varonis point out, the distinction
represents a continuous rather than dichotomous variable.
Other researchers have been more hesitant to suggest that one particular
task type is desirable and have argued that different task types contribute
to acquisition in different ways. Tong-Fredericks (1984) compared learner
output on three types of communication task: (1) a problem-solving task, (2) a
role-play task and (3) an ‘authentic interaction’ task, requiring learners to find
out from their partners what they had done the day before. Task (1) produced
the highest number of turns per minute, but the turn-taking on this task was
also the most disorderly, partly because the learners were thinking aloud.
Self-corrections were more frequent on tasks (2) and (3). On task (1) the
learners may have been too engrossed to bother with accuracy. No differences
in speaking speed were found. Tong-Fredericks argues that one activity type
is not necessarily better or more effective than another. Different tasks elicit
different kinds of responses, which can promote acquisition in different ways.
Duff (1986) investigated divergent and convergent problem-solving tasks.
The former requires learners to defend different viewpoints on an issue; the
latter requires them to reach a mutually acceptable solution. From the point of
view of negotiation of input, the convergent tasks worked better in pairs of
Chinese and Japanese learners of English. However, Duff noted that the
divergent tasks resulted in greater complexity of output. Like Tong-Fredericks,
she concludes that the two task types may be complementary in pedagogic and
psycholinguistic value.
 Classroom Interaction and Language Learning 113
In general, the results of research which has investigated the effect that
different tasks have on classroom interaction have not been very conclusive,
mainly because of the difficulty of identifying and controlling the task and
learner variables that affect interaction.

3: Changes in the teacher’s interactional modifications over time This

issue has attracted less interest, although it is of considerable theoretical
importance. If interactional modifications are the main source of comprehen¬
sible input, the quantity and perhaps also type of interactions ought to vary
over time according to the learner’s proficiency. Ellis (1985c) looked at the
interactional modifications in one teacher’s speech to two beginner learners
on two occasions separated by a period of six months. He found that the
teacher’s speech contained fewer self-repetitions but more expansions on the
second occasion. Also, the teacher switched from asking questions that required
simple object identification to questions that invited some kind of comment.
This study opens up the possibility that some interactional features may be
more important at one stage of development and others at later stages.

4: Differences in the nature of interaction found in teacher class lessons

-

and small-group work A number of researchers have sought to establish

whether small-group work results in more negotiation of meaning than teacher-
directed lessons (Pica and Doughty, 1985a and 1985b; Doughty and Pica,
1986; Duff, 1986; Rulon and Creary, 1986; Porter, 1986; Fillmore, 1985;
Ellis and Rathbone, 1987). The results obtained from these studies are
complex and difficult to interpret. For example, group work does not always
result in more modified interaction than lockstep teaching. Two-way infor¬
mation-gap tasks are much more likely to result in differences than one-way
tasks as in the latter many of the learners opt out, irrespective of the mode of
participation. Another mediating variable is the proficiency level of the learners.
There are also differing positions regarding the value of interlanguage talk.
Long and Porter (1985) argue that group work is potentially beneficial because
there is more negotiation of meaning. However, these claims are not based on
any research showing that interlanguage talk promotes acquisition. The only
two studies that have attempted to relate participation type to learning out¬
comes (Fillmore, 1982 and 1985; Ellis and Rathbone, 1987) conclude that
group work is not necessarily beneficial. The results of these studies certainly
do not constitute definitive support for interlanguage talk, although it should
be noted that there is a strong educational rationale for group work (Brumfit,
1984).

5: Differences in the amount of interactional adjustments undertaken by

experienced and inexperienced teachers It is generally acknowledged that
native speakers vary in the extent to which they modify their speech inter-
actionally in conversations with non-native speakers. Long (1983c) suggests
that one factor might be the previous experience native speakers have of
114 Instructed Second Language Acquisition
foreigner-talk. Pica and Long (1986) in the study referred to above set out to
investigate whether there were any classroom interactional differences between
experienced and inexperienced elementary ESL teachers. The experienced
teachers did use more other-repetitions and WH questions (as opposed to
Yes/No questions), but in general the similarities outweighed the differences.
Pica and Long suggest that this was because the interaction in all the class¬
rooms reflected a one-way flow of information from teacher to learner. The
absence of any negotiation of meaning was the ‘natural’ pattern of discourse in
all the classrooms.

6: The extent to which teachers can be trained to adjust the way they
interact in classrooms If there is little negotiation of meaning irrespective of
experience, it is worth finding out whether training can induce teachers to
change the ‘natural’ pattern. Long et al. (1984) provided six teachers with
training designed to encourage them to ask fewer display questions and more
referential questions and to extend the time they allowed for students to
respond to their questions. They found that the training had an immediate
effect and also that this effect carried over into later lessons.9 In a similar
study, Brock (1986) found that two teachers who received training in the use
of referential questions asked significantly more of these questions during a
reading and vocabulary lesson than two control-group teachers. It would be
interesting to find out if the kind of strategies and tactics which Long
considers important for acquisition are equally amenable to training.

All this research has been descriptive in nature. It differs from much of the
other process research reviewed in chapter 4 in that it has been based on a
theory of L2 acquisition which makes claims about what kind of input promotes
acquisition. Because of this, researchers have felt able to make statements
about what effect the processes they uncover have on acquisition. Such
statements are, of course, worth no more than the theory on which they are
based.

Evaluation
The evidence cited by Long (1983a) - like that provided by Krashen in
support of the input hypothesis - is essentially indirect. I know of no L2
acquisition study that has succeeded in demonstrating that interlanguage
changes are brought about by interactional modifications. This is not to say,
however, that such modifications are not important - indeed the weight of the
indirect evidence suggests that they do facilitate rapid language learning.
It does warrant caution. There are sound reasons for believing that the
negotiation of meaning is less central to the process of acquisition than
advocates of the interactional hypothesis would have us believe.
First, some research giving counter-evidence. One of the few studies to
have investigated the direct relationship between interaction and acquisition
 Classroom Interaction and Language Learning 115
has produced results that do not support the interaction hypothesis. Slimani
(1987) studied the effects of conversational adjustments on the acquisition of
English by thirteen highly motivated learners of L2 English in an Algerian
technical institution. She focused in particular on the learners’ own requests
for clarification. She concludes (1987:314): ‘The use of conversational adjust¬
ments does not often reveal itself to have a direct beneficial effect on those
who employ them.’ In other words, those students who sought clarification did
not gain from the ‘negotiation’ that ensued. It should be noted, however, that
the data for the study came from form-focused, teacher-centred language
lessons. Also Slimani’s measure of acquisition was unusual. She measured
uptake — the linguistic items which individual learners reported learning on
charts they wer„e asked to complete after each lesson. Slimani’s study is
interesting because it provides a methodology for investigating the relationship
between various aspects of classroom interaction and acquisition. Such
a methodology has been conspicuously missing. The results she obtained,
however, must be treated with circumspection given the novelty of the way she
measured acquisition.
Other doubts regarding the centrality of the negotiation for meaning have
been theoretical in nature. A number of these are considered below:

1 Conversational interaction may facilitate communicative performance

without facilitating acquisition of new linguistic features. Sato (1986), for
example, points out that because of the interactional support the learner
receives she may not experience any need to learn new features. In other
words, the negotiation of meaning may have the opposite effect from that
claimed by the interaction hypothesis; it may inhibit learning. A corollary of
this argument, however, is that, although linguistic competence is not developed,
strategic competence (i.e. the ability to deal with communication problems)
may be enhanced. Thus, even if negotiation does not benefit acquisition of the
L2 code it may help the learner to make better use of existing resources.
2 Negotiation of meaning may only be apparent and may not result in
comprehension. Hawkins (1985) carried out a study to determine whether
apparently appropriate responses made by learners actually signalled compre¬
hension. She collected retrospective data from learners in order to determine
whether what appeared appropriate responses in context did in fact represent
comprehension. She found that in 50 per cent of the responses for which
retrospective data from the learners were available comprehension had not
taken place. Learners, in fact, often feign comprehension for social reasons, as
Fillmore (1979) has pointed out. This is potentially very damaging for the
interaction hypothesis as it suggests that negotiation results in far less com¬
prehensible input than has been claimed.
3 The conversational adjustments which have been associated with the
negotiation of meaning are also used for other purposes (cf. Aston, 1986).
Varonis and Gass (1985) admit to difficulties in distinguishing exchanges
involving the negotiation of meaning from exchanges involving understanding.
116 Instructed Second Language Acguisition
Similar interactional acts are involved in ‘keeping the conversation going’ as in
the negotiation of meaning.
4 Negotiation of meaning only results in acquisition if there is a low
affective filter (i.e. the affective state of the learner which controls whether
the input becomes intake). Aston (1986) argues that difficult interactive class¬
room tasks designed to promote plenty of negotiation may in fact fail to do so
because they produce tension in the learner.

The interaction hypothesis is part of a psychological theory of language

acquisition. However, interaction in classroom settings, as elsewhere, is essen¬
tially a social phenomenon. Aston (1986:139) makes this point clear by suggest¬
ing that ‘trouble shooting routines’ often occur in order to allow the participants
‘to perform a ritual of understanding or agreement’. They go through the
motions of agreeing so that they can display solidarity. There is no reason why
such social/negotiation should contribute to L2 acquisition.

Review
In this section we have taken a critical look at three hypotheses which share
the view that acquisition is the result of the reception rather than the production
of L2 samples. The weakest of the three hypotheses is the frequency hypoth¬
esis, as this merely claims that the frequency of items in the input determines
the order in which they are acquired but does not specify why this should be
so. The results of classroom research which has investigated this hypothesis
are mixed. It is possible that input frequency has a delayed rather than an
immediate effect on acquisition.
The other two hypotheses aim to be explanatory. Both the input and the
interaction hypotheses emphasize the importance of meaning-focused com¬
munication as a source of comprehensible input, which is seen as the necessary
condition for acquisition to take place. The input hypothesis considers com¬
prehensible input from any source valuable. The interactional hypothesis
emphasizes the importance of conversational adjustments which occur in
attempts to negotiate meaning when there is a communication problem. As
yet, there is very little direct evidence for either hypothesis. The strong claim
that comprehensible input is necessary for acquisition is difficult to maintain.
However, the weaker claim that comprehensible input facilitates acquisition is
tenable, although it still remains to be shown exactly how this happens. It is
still not clear how comprehensible input, obtained by whatever means, results
in the acquisition of linguistic knowledge.

Production-based Theories

According to reception-based theories production plays no direct part in the

process of L2 acquisition. Such theories do allow for an indirect role, however.
 Classroom Interaction and Language Learning 117
The interaction hypothesis, for example, acknowledges that certain types of
learner utterance help to negotiate input which contains L2 features the
learner is likely to attend to. In this section we will consider a number of
hypotheses which claim production contributes directly to acquisition.
The output hypothesis considers learner production in isolation from input.
The discourse hypothesis and the collaborative discourse hypothesis treat
output as an integral part of classroom interaction. The topicalization hypoth¬
esis concerns the kind of learner participation which best promotes acquisition. It
should be noted that all these theories address how production contributes to
rapid and successful L2 learning rather than how new linguistic knowledge is
internalized.
4

The Output Hypothesis

The output hypothesis (Swain, 1985) was not proposed as an alternative to the
input/interaction hypotheses but as an addition. Swain argues that compre¬
hensible input may well be important for L2 learning, but is insufficient to
ensure that native-speaker levels of grammatical accuracy are attained. The
learner needs the opportunity for meaningful use of her linguistic resources to
achieve this. Swain (1985:248—9) attributes three roles to output:

1 The need to produce output in the process of negotiating meaning that

is precise, coherent and appropriate encourages the learner to develop
the necessary grammatical resources. Swain refers to this as ‘pushed
language use’.
2 Output provides the learner with the opportunity to try out hypotheses
to see if they work.
3 Production, as opposed to comprehension, may help to force the learner
to move from semantic to syntactic processing. It is possible to compre¬
hend a message without any syntactic analysis of the input it contains.
Production is the trigger that forces learners to pay attention to the
means of expression.

The main evidence for the output hypothesis comes from research carried out
as part of the Development of Bilingual Proficiency Project (cf. Swain et al.,
1989). One part of this project examined the nature of the language proficiency
achieved by 175 Grade 6 immersion students learning French as an L2 in the
Ottawa region of Canada. Proficiency was investigated using the modular ,
model of communicative competence developed by Canale and Swain (1980).
According to this model communicative competence consists of three separate
traits: grammar, discourse and sociolinguistic components. The subjects were
administered separate tests for each trait. The results indicated that whereas
they had achieved a level of discourse competence similar to that of native
speakers of French, they had failed to achieve native-speaker levels for the
grammar and sociolinguistic traits, despite several years’ exposure to French
in subject lessons. For example, in an oral interview they evidenced a low level
118 Instructed Second Language Acquisition
of accuracy in verb forms. Swain (1985) argues that the students’ failure to
achieve native-speaker grammatical competence was not because they lacked
comprehensible input as there was plenty of this available in the immersion
classrooms (cf. Krashen, 1985). She speculates that the real reason was
because they had limited opportunities for speaking in the classroom and were
not ‘pushed’ in the output they did produce. Swain et al. (1989) report that
student talk in teacher-fronted activities was indeed severely restricted.
Teacher-initiated talk dominated, allowing little opportunity for extended
student talk of a clause or more (less than 15 per cent of total student turns).
Swain (1985) suggests that in such an environment the students develop
strategies for getting their meaning across and consequently experience little
need to develop their interlanguages.10
It is important to recognize that the output hypothesis predicts that pro¬
duction will aid acquisition only when the learner is pushed. Opportunities to
speak may not in themselves be sufficient. Schmidt’s (1983) study of Wes, a
Japanese painter, acquiring English in first Japan and then naturalistically in
Hawaii testifies to this. Wes had plenty of opportunity to speak English. His
discourse and sociolinguistic skills developed substantially but his grammatical
competence progressed very little. One reason for this could be that he was
able to achieve his communicative goals without the need for grammatical
accuracy. In other words, he was not ‘pushed’.
One way of investigating ‘pushed output’ is to study those learner productions
that occur in response to a native speaker’s signals of comprehension difficulty.
Pica (1988) examined the interactions between a native speaker and ten non¬
native speakers of English in order to discover whether the latter would resort
to more target-like use of the L2 when the native speaker indicated difficulty
in understanding what they had said. She did find some evidence to support
the view that ‘pushed output’ of this kind became more grammatical, but she
also found that it did so in less than half of the total learner responses. She
suggested that, on the whole, the negotiated interaction offered the learners
models of what the comprehensible output could have been like, rather than
opportunities for producing comprehensible output themselves. In other words,
the study gave greater support to the role of comprehensible input rather than
comprehensible output.
A subsequent study by Pica et al. (1989), however, found that ‘comprehensible
output was alive and well and very much an outcome of linguistic demands
placed on the NNS by the NS in the name of negotiated interaction’. The
crucial factor uncovered by this study was the nature of the native speaker’s
signals of comprehension difficulty. The learners (in this case ten Japanese
adults) proved much more likely to produce output modifications in response
to clarification requests than to confirmation requests. Clarification requests
such as ‘Huh?’ leave it up to the learner to resolve comprehension problems,
whereas confirmation requests, which model what the native speaker thought
the learner meant, obviate the need for improved output. The same study also
showed that many of the modifications which learners made in their pushed
output were morphosyntactic, as opposed to semantic.
 Classroom Interaction and Language Learning 119
These studies lend some credence to the output hypothesis. But they need
to be treated with caution. Evidence that learners improve the grammaticality
of their utterances when pushed does not demonstrate that they acquire
knowledge of new L2 items. As with the input and interactional hypotheses,
there is, as yet, no direct evidence to support the claim that pushed output
promotes learning.
Research which has investigated the effectiveness of listening methods on
learning (e.g. Asher, 1977; Winitz, 1978) indicate that the beginning stage of
language learning does not require production on the part of the learner.
However, this does not constitute evidence against the output hypothesis,
which claims that pushed output is necessary in order to achieve advanced
levels of grammatical proficiency. The real contribution of pushed output may
be to encourage learners to make use of those variants in their current
interlanguage systems which are more target-like. It may enable them to
resolve variability in favour of target language norms. It is in this sense,
perhaps, that it contributes to acquisition.

The Discourse Hypothesis

The discourse hypothesis derives from work on the variability of language use
in general and of interlanguage systems in particular. Givon (1979) argues
that syntax is inextricably linked to discourse. He sees it as ‘a dependent,
functionally motivated entity’ with formal properties which reflect its communi¬
cative uses. Givon distinguishes two types of language: loose, paratactie
pragmatic discourse structures and tight, ‘grammaticalized’ syntactic structures.
This distinction is evident in the differences between informal/unplanned and
fonnal/planned speech. Table 5.4 details the main differences between the
two types of discourse.
Givon goes on to identify the communicative conditions which give rise to
the two types of language use. Informal communication takes place under
relaxed conditions where planning is not necessary and also when there is no
time for planned speech. Planning takes place ‘as you go’. Informal speech is
accompanied by constant face-to-face monitoring which affords the speaker
the opportunity for constant repair so that careful planning is not necessary. It
occurs in contexts where there is a shared pragmatic background involving
here-and-now topics, which obviate the need for a high level of explicitness.
In contrast, formal language use is characterized by careful planning involving
corrections and reformulations. There is often a lack of communicative stress
in the actual performance of formal discourse because it comes already
prepared. There is an absence of face-to-face monitoring and of shared
pragmatic background. Topics regularly derive from displaced activity or
abstract ideas. As a result meaning has to be encoded explicitly in the
discourse. The distinction between the two types of discourse reflects a
continuum rather than a dichotomy.11
A number of researchers have studied the variability inherent in learner
language (see Tarone, 1988 for a review of the research). Tarone (1983)
120 Instructed Second Language Acguisltlon
Table 5.4 Informal/planned discourse and formal/planned discourse contrasted

Informal/unplanned discourse Formal/planned discourse

Topicalized constructions Grammatical constructions

(e.g. My father, he’s sick) (e.g. It’s my father who is sick)
Loose coordination Tight subordination

Plentiful repetition and slow delivery Little repetition and fast delivery

Reduction and simplification of Full grammatical morphology

grammatical morphology

Short verbal clauses with few noun Long verbal clauses with several noun
phrases per verb phrases per verb
Prominent topic-comment structure Use of grammatical subjects

Source: Based on Givon (1979)

argues that interlanguage is best viewed as a capability continuum. That is,

the learner’s competence consists of variable rules which account for the
systematic way in which certain variants of a rule are used in formal/careful
language use and other variants in informal/vemacular language use. For
example, Ellis (1987a) found that L2 learners were more accurate in their use
of the regular past tense in a written narrative when they had ample time to
plan than in an oral narrative which they were required to perform spontaneously.
The discourse hypothesis states that learners will acquire those structures
associated with the type of language use in which they typically participate.
Thus, if a learner only has experience of informal/unplanned language use
she will only develop the capability of performing in this kind of discourse; she
will, therefore, develop what Givon refers to as the pragmatic mode. Conversely, if
a learner only has experience of formal/planned language use she will only be
able to take part in this kind of discourse and will develop the syntactic
mode. Ellis (1984c) comments: ‘Because different kinds of knowledge and
different processes of language use are involved in different discourse types, it
cannot be expected that the acquisition of one style will facilitate the use of
another style.’ The term style here refers to the learner’s internalized linguistic
competence.
There is an important corollary of the discourse hypothesis: in order to
acquire a full linguistic repertoire the learner needs exposure to formal/
planned language use. It is through participation in this type of language use
that the learner comes to ‘grammaticalize’ her language i.e. acquire the full
range of morphological and syntactic devices provided by the grammar of the
target language. The process of ‘grammaticalization’ requires both exposure
to formal language and also the opportunity to produce it. Learners who
experience only informal/unplanned language use will not be able to develop
the full range of grammatical features of the target language, both because
 Classroom Interaction and Language Learning 121
they will receive insufficient exposure to them and also because they will have
no communicative need to produce them.
Applied to the classroom, the discourse hypothesis predicts that different
learning outcomes will occur according to the type of language use the
learners experience (Ellis, 1987b). Formal classroom teaching with its emphasis
on linguistic accuracy will engage the learner in planned discourse and develop
the corresponding type of competence. Informal teaching will provide oppor¬
tunities for unplanned discourse and result in the learner developing the kind
of competence that enables her to perform in this kind of language use.
Relevant teaching is teaching that gives the learner access to the type of
language use which she needs to master. The discourse hypothesis, then,
leads to pedagogic proposals radically different to those based on the input or
interaction hypothesis.
There has been no controlled study to test the discourse hypothesis. The
hypothesis is, however, compatible with the results of a number of studies.
Swain (1985), in the study referred to earlier, reports that immersion students’
scores on literacy-based tasks were not strongly related to their performance
on oral tests. In general, the students did better on written tasks requiring
planned language use. For example, their scores on a written discourse task
closely matched those achieved by native speakers (in contrast to their grammar
scores in the oral interview). Swain suggests that such results reflect the
students’ familiarity with similar tasks in school. Pavesi (1986) found significant
differences in the level of acquisition of relative pronoun functions in two
groups of Italian learners of L2 English.12 The formal, classroom learners
outstripped the informal, naturalistic learners. Whereas the former provided
evidence of having acquired the more marked pronoun functions (e.g. genitive
‘whose’), the latter were generally restricted to the unmarked functions (e.g.
subject and object). Also the informal learners displayed more pronoun copies
(e.g. * ‘The girl who John kissed her ...’). Pavesi draws on the discourse
hypothesis to explain the difference. She suggests that difference discourse
modes (planned vs. unplanned) have different ‘indexes of markedness’ which
affect the degree of elaboration achieved by different learners. Thus the
unplanned mode provides less exposure to and fewer opportunities for using
the more marked relative pronoun functions.
The discourse hypothesis should be considered nothing more than an
interesting idea until further evidence is forthcoming. Long (1988a) argues
that the hypothesis is unlikely on the grounds that although planned discourse
may provide greater exposure to complex, marked linguistic features, these
will not be perceptually salient to the learner. This argument, however, does not
refute the hypothesis, which claims only that learners will master that type of
discourse (and related linguistic features) in which they participate. It does not
claim that participation is the sole causative factor of acquisition. The discourse
hypothesis needs to be viewed as part of a complete theory, which will also
need to account for how internal factors enable the learners to attend to
certain linguistic features when they are ready and not before.
122 Instructed Second Language Acquisition

The Collaborative Discourse Hypothesis

A number of researchers have documented how the process of constructing
discourse in two-way interaction enables learners to produce new grammatical
structures. Wagner-Gough (1975) studied conversational data between native
speakers and Homer, a six-year-old Iranian boy learning L2 English in a
naturalistic setting. She gives instances of what she calls the use of an
incorporation strategy by Homer. This involves ‘borrowing’ a chunk from
the preceding discourse and then extending it by affixing some element to the
beginning or end, for example (1975:165):

Mark: Come here:

Homer: No come here.

Jufy: Where’s Mark?

Homer: Where’s Mark is school. [= Mark is at school.]

Ellis (1985a) points out that the incorporation strategy provides an explanation
for the ubiquitous ‘no + V’ construction in L2 acquisition. In this sense then
‘conversation is the crucible for language acquisition’ (Long and Sato, 1984).
Another form of assistance provided by discourse is that found in vertical
constructions. Here the learner produces an utterance over several turns, as
in this example from naturalistic acquisition in Hatch (1978b:407):

P: Oh-oh!
J: What?
P: this (points at an ant)
J: It’s an ant.
P: ant

According to Scollon (1976) these vertical constructions function as the

precursors of horizontal structures (e.g. ‘Oh-oh, this ant’) in LI acquisition.
That is, what the child can at first only do across turns and across speakers
she is later able to do within a single utterance. Researchers such as Hatch
have suggested that the same process occurs in L2 acquisition.
Collaborative discourse of the kind described above is fairly common in
naturalistic conversations involving L2 learners, but is less well documented in
classroom contexts. Ellis (1985c) gives examples of such discourse in conver¬
sations between a teacher and two pupils and demonstrates how it helped
them to produce ‘new’ structures. However, Ellis’s study involved withdrawing
the learners from the normal classroom setting. Faerch (1985:186) provides a
number of examples from teacher-controlled classroom discourse:

T: what does it mean when she says she wore thick bull’s eye glasses
S: her glasses were thick
T: like
 Classroom Interaction and Language Learning 123
S: the glasses
T: the eyes of a
S: bull

Both teacher and student contribute to the building of this syntagm. There
seems every reason to believe that such collaboration is not uncommon in
classrooms.13
The question that arises is what role collaborative discourse plays in ac¬
quisition. Long and Sato (1984) point out that conversational assistance could
be seen as facilitating communication rather than acquisition. Faerch (1985)
takes up this point, suggesting that where the classroom is concerned scaffolded
structures serve hie function of helping the learner to take part in the
discourse and nothing else. Faerch and Kasper (1986) suggest that it is
unlikely that learners preserve a mental representation of a vertical structure,
as it evolves over several turns. This objection, however, does not apply to
utterances resulting from the use of an incorporation strategy. Both Faerch
and Sato (1986) make the point that even if collaborative discourse is credited
with making a positive contribution to syntactic structures, it is doubtful
whether it plays any part in the acquisition of morphological features (such as
regular past). It is possible that collaborative discourse plays a significant part
in early acquisition, but it is doubtful whether all interlanguage rules can
emerge in this way. Collaborative discourse may help the learner to develop
the resources required in the pragmatic mode but not those needed in the
syntactic mode.

The Topicalization Hypothesis

The role played by the learner in nominating and controlling the topic of the
discourse in language acquisition has been commented on by both first and
second-language researchers. Wells (1985) contrasts the interactional styles of
mothers who are ‘supportive’ and those who are ‘tutorial’. The former allow
their children the opportunity to initiate topics which they help along by
means of continuing moves. The latter cast their children in the role of
respondents while the adult controls the discourse through the use of closed
questions and evaluations of the child’s contributions. Wells argues that the
supportive style results in speedier acquisition. Hatch (1978b) finds that in
naturalistic interactions between native speakers and child L2 learners the
latter are generally allowed to nominate topics and that this provides a basis
for building conversations. Long (1983c) includes at the top of his list of
potentially facilitative foreigner-talk strategies ‘Relinquish Topic Control’. He
notes that native speakers often pass control of current and subsequent
conversational topics to the native speaker and that ‘implicit willingness to talk
about whatever the non-native speaker feels comfortable with is pervasive’
(1983c:132). Ellis (1984a) argues that topic control by the learner facilitates
classroom acquisition. He also claims that fluency-work in the classroom (i.e.
124 Instructed Second Language Acquisition
activities designed to allow the learners to use their existing resources in
natural communication) is enhanced by handing topic control over to the
learner (Ellis, 1986).
The topicalization hypothesis, then, claims that acquisition is facilitated if
the learner is able to nominate and control the topic of a conversation.
Evidence in support of the hypothesis is available in Slimani’s (1987) study
referred to earlier. She found that ‘whatever is topicalized by the learners,
rather than the teacher, has a better chance of being claimed to have been
learned’ (p. 321). A learner would benefit from a topic initiated by another
student to a much greater extent than from a topic initiated by the teacher. A
learner also gained greater benefit if the topic was initiated by another student
rather than by himself. Slimani suggests that when a student topicalized he
offered learning opportunities to his classmates. She also suggests that learners
may benefit more from listening to exchanges in which other students are
involved than in participating in exchanges themselves. Eler study was carried
out in a classroom characterized principally by formal language teaching.
There is a need for similar research to be carried out in classrooms providing
more meaning-focused communication.
Most classrooms provide very little opportunity for student topic control. In
Slimani’s study, for instance, the teacher tended to monopolize the discourse.
Other studies reporting the directorial role exerted by the teacher and its
effect on discourse have already been reviewed in chapter 4. The implications
of the topicalization hypothesis for pedagogy are clear. Opportunity has to
found for the learner to initiate discourse by topic selection.
The topicalization hypothesis emphasizes the importance of a particular
kind of learner output — that which occurs when the learner is given the
opportunity to nominate and control the topic of conversation. Slimani’s
research suggests, however, that learner output may be of less benefit to the
learner who produces it than to other learners.14 In this respect, the topicalization
hypothesis is as much a reception-based as a production-based theory. Learners
advance by attending to input and they are more likely to attend closely if the
input derives from another learner. It has been classified as a production-
based theory here, because it claims that learner production in the classroom
is important for acquisition — in contrast to reception-based theories.

Review
In this section, a number of theories which emphasize the role of learner
production in classroom interaction have been described and evaluated. The
output hypotheses emphasizes the gain to acquisition that accrues from the
learner being ‘pushed’ to produce utterances that are precise and appropriate.
Such pushing may encourage the learner to select those variable forms in her
interlanguage that are closest to target-language norms. The discourse hypo¬
thesis claims that the type of language use the learner participates in functions
as a determinant of acquisition. Participation in the planned mode is required
 Classroom Interaction and Language Learning 125
to achieve higher levels of grammatical competence because only in this type
ol language use is there full grammaticalization. According to the collaborative
discourse hypothesis, learners are able to produce syntactical structures beyond
their competence with the help of ‘scaffolding’ provided by an interlocutor.
This hypothesis helps to explain the occurrence of certain transitional con¬
structions in the initial stages of language acquisition. Finally, the topicalization
hypothesis suggests that learner production is important when learners have
the chance to nominate topics and to control their development. This kind of
output can serve as valuable input for other learners.
To conclude this section, two points need to be emphasized. First, although
the theoretical positions which have been oudined emphasize the importance
ol learner output*, they do so in the context of interaction. That is, production
is seen as valuable not only in itself, but because it contributes to discourse.
Interaction is a joint venture and ultimately it is not possible to isolate the
separate contributions of the speakers or, perhaps, of production and compre¬
hension. Second, the lour hypotheses are not in competition with each other
and no one hypothesis can claim to provide a complete explanation of classroom
L2 learning.

Conclusion

The purpose of this chapter has been to examine a number of theoretical

positions concerning the relationship between instruction and L2 learning. In
doing so, ‘instruction’ has been defined as ‘the interaction that takes place
inside the classroom’. It is important to emphasize that this perspective has
not excluded formal language teaching. The kinds of discourse that result
from form-focused language work have been treated as constituting a particular
kind of interaction. All classroom discourse, irrespective of whether it derives
from form or meaning-focused instruction can be considered as interaction of
one kind or another. This perspective has been informed by Allwright’s
(1984a: 158) claim that ‘it is through the joint management of interaction in
the classroom that language learning itself is jointly managed.’
The theories described in this chapter have outstripped empirical research.
There has been very little research which has attempted to establish a direct
relationship between interaction and L2 acquisition. Indeed there are major
difficulties in designing such research. More promising is research designed
to investigate whether certain kinds of interaction promote comprehension.
However, this still leaves the problem that comprehension and acquisition are
not the same thing (see discussion on p. 105). Production-based hypotheses
such as the output and discourse hypotheses also lack a firm empirical base.
In addition, it should be noted that not all the research reported in this
chapter has taken place inside the classroom. Much of it has been experimental in
nature, although it has typically involved classroom learners.
126 Instructed Second Language Acquisition
What conclusions are possible? In answering this question it is useful to
distinguish three possible roles for classroom interaction: a strong role, a weak
role and a zero role. According to the strong role, interaction in some way
determines the way learners acquire a L2. Ellis (1984a; 1985c), for instance,
claims that syntactic structures develop out of learning how to do conversations
and draws on the evidence provided by studies of collaborative discourse.
Swain suggests that pushing learners to be precise and appropriate in some
way induces them to acquire complex structures which they would otherwise
ignore. According to the weak role, interaction facilitates but does not determine
interlanguage development. This is the position taken up by proponents of the
input and interaction hypotheses. The learner is responsible for selecting
features from the input as intake (e.g. by means of the language-acquisition
device). However, certain types of communication (i.e. those that ensure
comprehensible input) make the learner’s task easier by providing her with
sufficient samples of just those features she is ready to acquire. According to
the zero role, interaction only triggers learning, while some learning can take
place without any interaction (cf. White, 1987). The position that I lean
towards is that all three roles are possible. Some structures probably are learnt
through the process of constructing discourse and some structures are con¬
solidated by being ‘pushed’. Certain kinds of interaction are especially helpful
in enabling the learner to obtain intake. Many linguistic rules are acquired
with little or no help from input or output. This position is motivated by my
recognition of the many-faceted nature of language acquisition and my convic¬
tion that different grammatical properties can be acquired in different ways.
From the point of view of language pedagogy it is more important to
consider how interaction can be organized in the classroom in order to
promote acquisition than to argue about the precise role of communication.
What kind of communication is ‘optimal’? The answer to this question will
need to draw on the insights provided by both reception and production-
based theories.
An optimal communicative environment for classroom language learning
can be considered one where:

1 There is clear separation of the student’s LI and the target language

(Fillmore, 1985). This is required so that the students feel the need to
communicate in the L2.
2 The students are involved and interested in what is being talked about.
In order to achieve this it will be necessary to ensure that the students
are given opportunities to control the topic of conversation. Students
need to be allowed more initiative than is common in most language
classrooms.
3 Both the teacher and the students make efforts to be understood. This
calls for the use of effective interactional strategies on the part of all the
participants to ensure that comprehension is achieved. Some strategies
which have been found to work are the repeated use of patterns and
 Classroom Interaction and Language Learning 127
routines by the teacher and repetitive lesson formats (Fillmore, 1976).
The interactional strategies found in the negotiation of meaning may
also help, as may input rich in directives in the early stages (Ellis,
1984a).
4 The students are encouraged to produce utterances which tax their
linguistic resources — at least some of the time. In order to achieve this
it will be necessary for the teacher to vary the type of question asked to
suit the students’ level of proficiency (i.e. not to ask only simple display
questions). It will also be necessary for the students to be able to
initiate discourse so that they have the opportunity to perform a range
of speech acts, requiring varied linguistic resources. Teachers should
also encourage learners to reformulate their own utterances that cause
comprehension problems. This can be effectively achieved by means of
clarification requests.
5 The students have the opportunity to participate in the kind(s) of
discourse (planned and/or unplanned) which correspond to their com¬
municative needs outside the classroom.
6 In the initial stages the teacher helps to provide scaffolding for the
production of structures which are too complex for learners to produce
by themselves.
7 In the later stages, learners have adequate access to planned discourse
rich in marked linguistic features.
8 Learner production is not forced. Learners need to be free to choose
when to produce.

Optimal interaction of this kind is most likely to occur in meaning-focused

instruction, although some features of it can also be found in form-focused
instruction. It certainly does not follow that formal instruction is entirely
without value, even if this is treated as ‘interaction’.

NOTES

1 Krashen recognizes that comprehending input is not the same as acquisition. If

the learner is not ‘open’ to input because of negative attitudes, it will not reach
those parts of the brain responsible for acquisition. In this sense, then, compre¬
hensible input is not ‘sufficient’.
2 I am indebted to White (1987) for the reasoning behind this example.
3 Krashen provides no evidence to show that the methods he believes facilitative of
acquisition provide more comprehensible input than those he does not consider
effective. His argument is circular, in fact.
4 An account of Krashen’s theory of how the skill of writing is developed can be
found in Krashen (1984).
5 Two kinds of models have been used to explain the process of L2 acquisition;
sequence models and growth models. Early L2 acquisition research (such as the
morpheme studies) was based on a sequence model; learning progressed one
128 Instructed Second Language Acguisition
step at a time. Later research, in particular that associated with the study of
learner variability (e.g. Huebner, 1979) has emphasized the holistic nature of
interlanguage systems and the dynamic nature of interlanguage change.
6 There are relatively few studies of the L2 comprehension process. Exceptions are
Carrell (1983) and Wolff (1987). These researchers disagree, however, over
something as basic as whether L2 learners, like LI users, are adept at top-down
processing.
7 To claim that learners acquire past tense interrogatives from such exchanges is an
example of the kind of ‘leaps’ that Hatch (1983b) warns about.
8 Clarification requests and confirmation checks are used to deal with communication
problems. Teachers who dominate the classroom discourse may leave little room
for learners to negotiate in this way. However, comprehension checks can be
executed by the teacher without any loss of discourse control.
9 Nunan (1987) reports on a similar study to that of Long et al. (1984). A teacher
was encouraged to focus on the use of referential questions in a communicative
language lesson in order to relate the content of the lesson to the students’ own
lives. Nunan reports that ‘the effect was immediately apparent’. Whereas an
earlier lesson had resulted in little genuine communication, the lesson following
the ‘training’ was characterized by such features as content-based topic nominations
from the learners, student/student interactions, an increase in the length and
complexity of student turns, the negotiation of meaning between teacher and
learners and even an example of a student disagreeing with the teacher.
10 Further evidence in support of the output hypothesis can be found in a study
carried out by Mitchell, Parkinson and Johnstone (1981). They correlated various
teaching behaviours in 16 French as a foreign language classrooms in Scotland
with pupils’ scores on comprehension, narration and structure tests. Significant
relationships between the test scores and the amount of pupil talk (as opposed to
listening or reading) were found. Seliger (1977) also found a positive correlation
between learners’ participation in classroom interaction and learning outcomes.
However, not all studies have produced similar results; amount of production
does not always correlate significantly or even positively with acquisition. It should
be noted that these studies have not examined ‘pushed output’ and, therefore, do
not speak directly to the output hypothesis as formulated by Swain.
11 The distinction between informal/unplanned and formal/planned discourse has
been made by a number of commentators. Olson (1977) distinguishes the notions
of utterance and text. The former occurs in oral communication where language
is typically inexplicit and unformalized because the participants are able to make
use of nonverbal cues. Text refers to written language; it is decontextualized and
formally explicit. Olson argues that language development proceeds from the
ability to produce utterances to the ability to produce texts and that one of the
main functions of schooling is to facilitate this development. Callfee and Freedman
(1980) distinguish natural and formal language, which correspond closely to utter¬
ance and text, except that they consider the distinction to be independent of
speech and writing. Cummins (1981:11) talks of ‘a continuum of contextual
support available for expressing or receiving meaning’. For Cummins, therefore,
the distinction between context-embedded communication and context-
reduced communication represents a continuum rather than a dichotomy. The
general distinction represented by these various terms is clearly widely supported
in the theoretical literature. This literature tends to assume that language users
 Classroom Interaction and Language Learning 129
progress from informal/unplanned language use to formal/planned language use.
This is certainly true for LI acquisition and perhaps also for the kinds of second
language learners that Cummins describes (i.e. bilingual learners in Canada).
However, it need not to be the case. It is possible for learners who have already
acquired formal/planned language use in their LI to start with this in L2
acquisition. Such learners will find context-reduced communication easier to
handle than context-embedded communication. This is the position advanced by
the discourse hypothesis.
12 Pavesi’s study was based on the Accessibility Hierarchy for relative pronouns
(Comrie and Keenan, 1978). This was derived from a study of relative clauses in
a large number of different languages. An implicational scale for relative pronoun
functions was established:
4

Subject > Direct object > Indirect object > Object of preposition >
Genitive > Object of comparative

According to this scale the most likely function of a relative pronoun is that of
Subject and the least likely Object of comparative. Furthermore, the scale predicts
that if a language permits relative pronouns with a given relative function (e.g.
Direct object) then it will also permit relative clauses with pronoun functions
higher in the hierarchy (i.e. Subject) but not with pronoun functions lower down
(i.e. those to the right of Direct object).
13 It should be noted, however, that collaborative discourse can only occur if the
learner has the opportunity to participate in an exchange involving several turns.
Discourse that is rigidly controlled by the teacher so that learners are allocated
only one move (i.e. responding) may prohibit collaboration. This may be one of
the effects of a teaching style that makes extensive use of display questions.
14 Slimani’s conclusion flies in the face of the output hypothesis, as it suggests that
acquisition occurs when the learner is involved in listening rather than in production.
It should be noted, however, that Slimani’s study focused chiefly on vocabulary
acquisition (her learners’ reported uptake consisted almost entirely of lexical
items). It is possible that different aspects of communication are responsible for
different aspects of acquisition. The interactions which Slimani investigated oc¬
curred in the context of form-focused instruction. The results Slimani obtained,
therefore, may not generalize to interaction in meaning-focused instruction.
6 Formal Instruction
and Language
Learning

Introduction

Teaching can be viewed in two different ways: (1) as interaction and (2) as
formal instruction. In the last chapter we investigated the teaching—learning
relationship by considering teaching as interaction. In this chapter, we will
look at the relationship between formal instruction and learning. We will be
concerned with direct pedagogic intervention — attempts to influence the way
interlanguage develops through formal instruction which (1) focuses on some
specific property of the target language and (2) tries to make the learner aware
of what the correct grammatical use of the form is. The central question we
will consider is whether formal instruction facilitates acquisition and, if it
does, in what way. We will review research that has examined the effect of in¬
struction on both the rate/success of acquisition and on the process/sequence
of acquisition.

The Effect of Formal Instruction on the Rate and Level

of L2 Acquisition

Long (1983b) reviewed a total of eleven studies that examined the effect of
formal instruction on the rate/success of L2 acquisition. Six of these studies
showed that instruction helped, two produced ambiguous results,10 and three
indicated that instruction did not help. All the studies used designs involving
comparisons between learners receiving instruction and those experiencing
exposure with or without, instruction. In other words, they attempted to find
an answer to the question ‘Does L2 instruction make a difference?’ by asking
whether instruction or exposure produced the more rapid or higher levels of
learning.
On the basis of this review Long claimed that ‘there is considerable evidence to
indicate that SL instruction does make a difference’ (1983b:374). He also
claimed that instruction was beneficial (1) for children as well as adults, (2) for
intermediate and advanced students, (3) irrespective of whether acquisition
 Formal Instruction and Language Learning 131
was measured by means of an integrative or discrete-point test, and (4) in
acquisition-rich as well as acquisition-poor environments.2 Instruction was
more effective than exposure in promoting L2 acquisition.
As an example of the studies that Long reviewed we will consider one
(Krashen, Jones, Zelinski and Usprich, 1978) in some detail. The subjects
were 116 ESL students who had experienced different amounts of instruction
and exposure. The students were given three tests: the Michigan Test of
English Proficiency, a free composition (which was graded by dividing the
total number of words each learner wrote by the number of errors she made),
and a cloze test. The test scores were correlated with both the number of years
each student reported having spent in an English-speaking country and the
number of years' of formal English study. The correlations between the test
scores and the measures of amounts of exposure and instruction were all
statistically significant. However, the effect of formal language study was
much stronger than that of informal contact. For example, 25 per cent of the
total variance on the Michigan test was accounted for by instruction but only
3.2 per cent by exposure. The authors concluded that ‘formal instruction is a
more efficient way of learning English for adults than trying to learn it in the
streets’ (1978:260).
Since Long’s review article there have been a number of other studies. We
will consider three. Weslander and Stephany (1983) examined the effects of
instruction on 577 children with limited English proficiency in Grades 2 to 10
in public schools in Iowa. Students who received more instruction did better
on the Bilingual Syntax Measure, a test devised by Burt, Dulay and Hernandez
(1973) to elicit natural speech. The effects were strongest at lower levels of
proficiency in the first year of schooling and diminished in subsequent years.
These results contrast with those obtained by Hale and Budar (1970). This
study (which Long included in his review3) found that learners who spent two
to three periods each day in special Teaching English as a Second or Other
Language (TESOL) classes ‘were being more harmed than helped’.
Similar results were obtained by Ellis and Rathbone (1987). They examined
the effects of class attendance on the levels of learning achieved by 39 adult
learners of L2 German. The setting was a foreign-language classroom in this
case, so there was limited opportunity for naturalistic exposure. The instruction
was very formal, consisting largely of controlled practice of grammatical items.
Overall attendance during a six-month period was found to correlate significantly
with a number of measures of grammatical learning,4 including those based
on elicited natural speech. This suggests a positive effect for instruction.
However, no relationship was found between attendance and achievement in
the first three months of the study. Ellis and Rathbone speculate that this may
be because the instruction had a delayed rather than immediate effect, a point
that will be taken up later.
The third study, Spada (1986), provides new insights which are of con¬
siderable importance. Whereas the studies that Long examined in his review
sought to compare the effects of instruction with the effects of exposure,
132 Instructed Second Language Acquisition
Spada set out to establish whether there was any interaction between type of
contact and type of instruction. She examined the effects of instruction and
exposure in 48 adult learners enrolled in an intensive six-week ESL course in
Canada. Spada found that although type and amount of contact appeared to
account for variation in some aspects of the learners’ proficiency before the
effects of instruction were considered, it did not account for differences in the
learners’ improvement during the course. She concludes that contact is a less
powerful predictor of differences in learners’ L2 abilities than instruction.
However, Spada also found that the type of instruction interacted with the
amount of contact individual learners experienced: ‘contact positively accounted
for differences in learners’ improvement on the grammar and writing tests
when the instruction was more form-focused, and negatively accounted for
differences on these measures when the instruction was less form-focused.’
(1986:97). In other words, instruction based on direct intervention worked
better than instruction based on indirect intervention where grammar and
literacy were concerned when the learners had opportunity for plentiful informal
contact outside the classroom. The implication of this study is that learners
require both formal instruction and informal exposure and that the two
together work better than either on its own.
Long’s (1983b) review has been widely cited as demonstrating that formal
instruction has a positive effect on L2 acquisition. There are, however, a
number of reasons for exercising caution. We will examine some of the major
ones.

1 As Long admits, many of the studies failed to control for overall amount
of combined contact and instruction. That is, the results obtained by some
studies showing that instruction was more important for learning than contact
might reflect greater overall opportunity for acquisition inside and outside of
the classroom. To counter this objection, Long points to studies (e.g. Krashen
and Seliger, 1976) which showed that more exposure did not result in higher
proficiency in learners who had been matched for the amount of instruction
they had received. The results of these studies contrasted with those that
showed that differences in the amount of instruction were positively related to
differences in proficiency in learners who had received the same overall
amount of exposure. Long argues that these contrasting results enable us to
be more confident in finding a positive effect for instruction in studies which
failed to control for overall acquisition opportunities. However, if such studies
are excluded on the grounds of this design flaw, the number of studies in
Long’s original review which unambiguously show that instruction helps is
reduced to two.
2 An intervening variable — which Long does not consider, but which
Krashen, Jones, Zelinksi and Usprich (1978) do - is that of the learners’
motivation. Students who are highly motivated to learn are likely to enrol in
classes. Those who are less strongly motivated will keep away. Therefore, the
 Formal Instruction and Language Learning 133
positive effect found for instruction may simply reflect stronger motivation on
the part of the classroom learners.
3 Another serious problem concerns the assumption made by the authors
of all the studies Long reviewed that the amount of formal instruction ex¬
perienced by learners can be equated with the number of years spent in the
classroom. None of the studies obtained any data about the nature of the
classroom processes which took place in the name of instruction. We do not
know for certain, therefore, that the instruction was form-focused. But even if
we accept that this was likely for at least a large proportion of the total
teaching time, there is still a problem. As we have seen, formal instruction can
be viewed as ‘interaction’ rather than as an attempt to intervene directly in
interlanguage development. It is possible, therefore, that the positive effects of
instruction derived not from the fact that learners were focusing on form but
from the communicative properties of the interactions which occurred. Long
is aware of this problem and seeks to immunize against it by arguing that there
is evidence to suggest that instruction is beneficial even in settings where
the learners have plenty of opportunity for negotiation of meaning outside
the classroom (i.e. ‘acquisition-rich environments’). According to Long,
this indicates that the instruction worked, not because it provided comprehensible
input but because it required learners to focus on form. However, it is possible
to find counter-arguments. Long admits one possibility — namely, that the so-
called acquisition-rich environments were not so rich after all. Another is that
learners were able to ‘let in’ more input in a classroom context because they
felt more secure and more relaxed than in face-to-face interactions with native
speakers in naturalistic settings. The main problem is that, given the design of
the studies Long examined, there is no evidence to enable us to make an
informed decision between these various arguments.

Despite these misgivings, it seems reasonable to assume that formal in¬

struction is of value in promoting rapid and higher levels of acquisition.
Long’s review and the subsequent studies can be seen as providing tentative
support to this claim. There is also the experience of countless successful
language learners, which testifies to the value of form-focused study (e.g.
Pickett, 1978). Spada’s study is salutary in that it encourages us to recognize
that comparing the differential effects of informal contact and instruction is, in
fact, simplistic and potentially misleading. Clearly learners need both. Further¬
more, formal instruction may work best when there are also opportunities for
informal language use. In other words, the effects of instruction may not be
absolute - as Long’s review suggests - but contingent upon certain conditions
being met.
The research which has been considered in this section is summarized in
table 6.1. It provides no information about the process of classroom language
learning. In the next section, we examine research which has investigated this
issue.
 dj

-TO
2<
S-^o
^
-4-J
X
a
a
•*->
a
x 4-J.
no
a
<2 3
4—1
C/2
O
■4—>
a
X a
-a o
a +_ a
03 p C
C/2

a
1I a o 2
: |
dj
X
O
a
04 a a c/T o wT 2
3 s
_D C/5 c/T o S3 X) c -c
C X c JC &
a a J2< a
o - C/5 C/2
a G
^ C/2 w ^ D,
o C/2 o --9" ^ u 3 CD
o
<u '•O .&.& *D
X D x «
•a c£’y X D
•a o. w !8£ o 'x
a; "O rj no
y "3 “ O o D O y 3 ^
g ^ XI a no C 'a a
j_, ^3 (3 a no ^ c"§ j= c P a
To
c
d
H
d
J-1
o
•a
o
D
H
•5 K
a
C/2 u o
#a t- X3
o D
•a H
D a
|
.S3 S
p .2 §!•§
o
•a
O
.5
X
_
a
C/2
o
a
C/2
O
a y a
C/2
O -s §E
3 y
o
<5 a
5 2 t:
o e- <2
I i! § 2 |
a

o &• c &
i S-& &
CQ d .a d PQ dj .3 w.s a w.S2

c/2 C/2
D a;
C/2
0>
rn a <^
C D .3 « c S
on ‘3 ^ o '3 *"
D .3 #G
Cl, D a, <d o. «
bi O *3 > > o
c Q
a; D \q a* CU 0/ D *£3 u *a <D a
o > >
X CL) •a w 2 « 2 •a
ctf U ^ D d S3 fo O &0 oS
'</) 8 a J-h ^ <D « 4J o
'd kb o kb feb O
cr d 53 D 5 -S OJ
o 04 04 r—I 04 Q
03
eg
-o
a
to to to a
u O o 5^
cn 5> a C a a a
</)
0 .5 .2h 3 .3 .3 .3 .3
o *3 *3 ’3 ‘3 *3 •3 no
o cl§, X <42 <42 D <J2 <42 (D D
13 S’ O O O O O a u
C/5
CO a. 0,22
__ D
kx
CD ’Sb __ (D
Cl.- 3 |
0 23 ^ 22 > <D 22 > > •*-*
d d^ a no
f3 5 s < 2$ CQ 3P <D ►a c«
1—
0
.c
C/2 o
c <42 X X T3 X
X <D
o C/2 D no no no
bo .a
c D D D
o to *bb <2 D .a .>3 0
bo bo
■4—'
o -S’
cC w So as 3 3
a 5/5
3
1—
a I D ‘D
O <42
a Ii D
<0) bo a bo
CO
c j§ 3 <D T3 T3 —1
a
X a
T3o .aS 53
3 5b *a 5b
nO C C 13 g no a
o < J2 < e U <a < J2
CO
o
Q) nO
3= a/ d
0) c> bo.ti _
d
as g £3 T3
<D
bo t: -a -a -a
£ CTJ no
0> dJ <u CD <D iD D
| pi
X -5 &
§
&
*3
a a
D D C3 D D D D D
be -a _ o a
CO c c^ T3 CD c a a a
g> ,|.S on "O C/2 C/2 C/2 c/2 c^
i dj i D 1 <D
X3 S; a 6 a S X hJ "I o 'S
2 O <D X X X ti X -*-? X a
3
*—*
O cS 3 3 X CD 2 co S CD 2 CO S CO 2 co S
</) X X CO x W CD U2 CO X CD dJ CO X CO W co
76
o 0
*k_ o bo
CL 0)
<D
E CD cl x
m T3 00
X oo no 13 a ON x
T3
a ON a a 4-J CQ ON
CTj ON 03 D
Cl T3 a
co a a <2 S-H
D D
CD o 2 °o D
l_, X X X 00 3 a
rS D C/2 C/2 x
2 c 03 o3 'S C/2
OS -a ^ *C H ON H ON «c3 Os <2 a
U U C. QQ yi Cl Nd Ic D
 0 13
”0 "5 Vh
-C -G > 0
+-•
0
4-H
0
0
_
a
0
G G
§ 'g
cj |1
C/3 C/3
0^ a> -a 0
0 O s-a “
-a no 43 0 g
tH O
G G « CV-O
0 O O ^/3
*G *G •c g S
0 0
g o-
H3 4—1 w

<L>
Ih
,-03(73
C 03
a £,43
CQ co 2

C/3
o0 *-H

si
ii
<U ‘5i
<L>
£ <u
o J4
U <£

<u
t: C-4/3
S
o
Jp O
o
O
o
.£
c/3
o Gw
no
<

no no no ■J nO
<L>
.+->
<L> .£
£
<v 03
.tn
C/3
d
o3
*£ '£ ‘G *£
D D P3 p u
G
G a G
•'—1
G
««-H o3
•’”' C/3 C/3 C/3
<U <L> fj
hJ ■4-1 J +-*
co a CO 03 c/3 C/3 •M <U
r Hh co
W CO w CO W W CO u W

<u
G
O
LO 12 Sc
o
ON
I03
r-* O §2 VO
00
ON ^4- P4 ON
oo <V nO
G
03 ON nO G* G ^
G
£
£
o3
o3
«»
H 03 p 03
O 155 Qh oc nO
<L> 0/ 03
3 3
W2
03 Oh

(J-. W w CO
136 Instructed Second Language Acguisitlon

The Effect of Instruction on the Process of L2

Acquisition

There have been a number of studies that have investigated whether instruction
influences the process of L2 acquisition. These studies have sought to establish
die effects of instruction in two ways: (1) by comparing classroom and natural¬
istic acquisition and (2) by means of classroom experiments designed to
ascertain whether teaching specific items results in their acquisition.

Comparisons between Classroom and Naturalistic L2

Acquisition
Research that falls into this category has studied a number of‘process’ features;
L2 errors, the sequence of acquisition of grammatical morphemes and the
sequence of development of syntactical structures such as relative clauses and
word order.

Error Analysis
Errors provide evidence of the processes involved in interlanguage development
(cf. chapter 3). Comparisons of the errors found in naturalistic and classroom
L2 acquisition, therefore, help to show whether the processes involved in the
two kinds of acquisition are the same or different. There are three possibilities:

1 the errors are the same;

2 instruction enables learners to avoid errors commonly found in natural¬
istic acquisition;
3 instruction results in different kinds of errors.

If (1) is true it can be concluded that the process of interlanguage development is

immune to instruction; (2) provides evidence that instruction facilitates acqui¬
sition; (3) is more difficult to interpret; it suggests that instruction does affect
the process of acquisition, but not necessarily in a positive way.
The available evidence indicates that instruction is powerless to prevent
developmental errors occurring. Felix (1981) reports on a study involving
thirty-four children studying L2 English in their first year in a German high
school. He examined errors relating to negation, interrogation, sentence types
and pronouns. Here we will consider the results he obtained for negation.
Naturalistic learners typically go through an early stage in which negative
utterances have the form of ‘no’ + verb - for example:

No you playing there.

Mariana no coming today.
 Formal Instruction and Language Learning 137
Felix found a few examples of this kind of negative in the children’s spontaneous
classroom speech:

It’s no my comb.

but they were rare, perhaps because there was little opportunity for spontaneous
language use in the classroom. However, Felix found that even in drill
situations, the students failed to produce correct negative sentences 45 per
cent of the time. On these occasions they produced sentences of the kind:

Doesn’t she eat apples.

4

instead of the correct negative:

She doesn’t eat apples.

The use of ‘doesn’t’ in this way is not found in naturalistic acquisition.

However, Felix suggests it provides evidence of similar processes at work, as
‘doesn’t’ was monomorphemic for the learners and appeared in sentence-
initial position. In other words, these classroom learners were using ‘doesn’t’
in the same way as naturalistic learners use ‘no’. On the basis of the results
obtained for negation (and similar results for the other structures), Felix
concluded that the parallels between tutored and naturalistic learners were
striking.
In a similar study, Felix and Simmet (1981) concluded that the processes of
acquisition of the English pronoun system were also unaffected by instruction.
The subjects were seventy students in a classical language high school and a
modern language high school in Germany. During the course of a school year,
they received instruction in the use of English pronouns. This instruction took
the form of explanations and practice of each separate pronoun form in an
incremental fashion. However, learner data obtained from classroom recordings
indicated that the students did not master pronouns as whole forms, but as
bundles of semantic features, with one feature being acquired at a time. Thus,
gender was acquired first, followed in turn by person, number and case. Felix
and Simmet provided evidence to show that this feature-oriented approach
was typical of both LI and naturalistic L2 acquisition. They argued that the
students made use of ‘instruction-independent learning strategies’.
Whereas these two studies indicate that the process of acquisition cannot be
manipulated by instruction, the third error-analysis study we will examine
suggests that instruction can have an effect. However, the effect is not the
desired one; instruction does not prevent ‘natural’ errors but can interfere
with natural processing. Lightbown (1983), in a study referred to in the
previous chapter, reports on the acquisition of a number of grammatical
features by seventy-six ESL students in a Canadian high school. In this study,
138 Instructed Second Language Acquisition
learner data were obtained outside the classroom by means of a communication
game. We will consider the results for V-ing. Lightbown found that the
frequency and accuracy of V-ing declined from Grade 6 to Grade 7. At the
earlier time the learners produced V-ing correctly as in:

He’s taking a cake

while at the later time they substituted the uninflected verb form:

He take a cake.

V-ing is acquired early in naturalistic L2 acquisition without any evidence of

the kind of subsequent decline observed in these classroom learners. Lightbown
suggests that the explanation for the unusual reversion is to be found in the
formal instruction the learners experienced. ‘Over-learning’ of V-ing occurred
as a result of intensive instruction in this feature towards the end of Grade 5.
This led to frequent and accurate use of the form during Grade 6. However,
instruction in V-ing was discontinued in this year, while instruction in the use
of the present simple tense was stepped up. This resulted in the learners
dropping V-ing in favour of simple forms in Grade 7. Lightbown contrasts the
‘rote learning’ of V-ing with the natural process by which V-ing is learnt in
meaning-focused language use. She comments: ‘By forcing learners to repeat
and overlearn forms which have no associated meaning to contrast them with
any other form(s), we may be setting up barriers which have to be broken
down before the learners can begin to build up their own interlanguage
systems,’ (1983:239). In other words, formal instruction can temporarily inter¬
fere with the natural process of language learning.
To sum up, the evidence provided by these error analysis studies suggests:

1 In some structures such as English negation and the pronominal system,

instruction is unable to circumvent the natural processes of acquisition.
2 Instruction can result in some structures such as V-ing being ‘over-
leamt’.
3 ‘Over-learning’ has to be overcome before real interlanguage develop¬
ment can take place.

The evidence, then, suggests that formal instruction either has no effect or a
deleterious effect on L2 acquisition.

Morpheme Acquisition Studies

The 1970s saw a number of cross-sectional studies of English morpheme
acquisition by naturalistic or mixed L2 learners (e.g. Dulay and Burt, 1974b;
Bailey, Madden and Krashen, 1974; Larsen-Freeman, 1976b). The procedure
followed by these studies has been described in chapter 3 (see section on
 Formal Instruction and Language Learning 139
performance analysis). The results obtained were used to make claims that
there was a more or less standard of ‘natural’ order of accuracy/acquisition;5
this was not influenced by the learners’ LI, age or whether the medium of
production was speech or writing.
Performance analysis has provided a convenient basis for investigating a
number of interrelated questions:

1 Is there any difference between the order of instruction and the order
of acquisition?
2 Is it possible to alter the ‘natural’ order of acquisition by means of
instruction?
3 Do instrutted learners follow the same order of acquisition as untutored
learners or a different order?

We will examine some of the research which has sought answers to these
questions.
The available research indicates that the order of acquisition of English
grammatical morphemes does not follow the order in which these items are
taught. Turner (1979) investigated three eighteen-year-old Spanish learners
of L2 English who were enrolled on an intensive English programme at the
University of Texas. The order of acquisition they manifested in both natural
speech and elicited data was not strongly related to the order of instruction as
stated in their teachers’ lesson plans. Turner notes that the instruction might
have been expected to have influenced the acquisition of simple structures
such as plural -s and past tense -ed. However, the learners failed to show any
acquisition of these forms. Turner also administered the Michigan Diagnostic
Grammar Test and found that the learners scored correctly on items testing
knowledge of the past tense. Drawing on Krashen’s distinction, Turner con¬
cluded that instruction has little effect on ‘acquisition’ but has a considerable
effect on ‘learning’. Makino (1980) obtained very similar results. He used
written data based on picture stimuli from a total of 777 learners of L2
English in Grades 8 and 9 in rural and urban high schools in Japan. No
relationship between the teaching order of morphemes in the two textbooks
that were used and the observed acquisition order was found.
A number of other studies have investigated whether instruction in the
accurate production of specific morphemes results in a disturbance of the
‘natural’ acquisition order. For example, Perkins and Larsen-Freeman (1975)
examined the effects of one month’s intensive instruction in articles, past
irregular and possessive -s on twelve Venezuelan students recently arrived in
the United States. They found that the instruction did not result in any
change in the accuracy order of five morphemes in the spontanenous speech
of the learners from the beginning to end of the period of study. They
concluded that the instruction did not influence the route of acquisition.6
Comparative studies of the acquisition orders found in tutored and untutored
learners have produced mixed results. Krashen, Sferlazza, Feldman and
140 Instructed Second Language Acquisition
Fathman (1976) found similar orders for the two kinds of learners. So too did
Makino in the study referred to above. However, Sajavaara (1981) found a
disturbed ‘natural’ order among Finnish students, while Fathman (1978)
found both similarities and differences between adolescents acquiring English
informally in the United States and those studying English in formal classrooms
in Germany. It is difficult to interpret these findings as the studies used
different types of data (e.g. written vs. spoken) and also some of the so-called
untutored learners may have received formal instruction.
Lightbown’s study (1983; 1987), however, does suggest that instruction can
affect the order of acquisition of grammatical morphemes. She found that the
order evident in school students’ spontaneous speech differed from Krashen’s
(1977) ‘natural’ order. The auxiliary was performed more accurately and V-
ing less accurately by the classroom learners. Lightbown argues that these
differences were the result of the ‘distorted’ input derived from instruction.
However, she found that the effect was only temporary. Natural processing
took over once the structures were no longer the object of instructional
attention.
The most important comparative study which has been carried out to date
is that by Pica (1983; 1985a). Pica studied the acquisition orders of three
groups of six learners: (1) instruction only, (2) purely naturalistic and (3)
mixed (i.e. learners who received formal instruction and had opportunity for
informal contacts). The data consisted of hour-long audiotaped conversations
between each subject and the researcher. Pica calculated the rank orders for
eight morphemes using the standard method (i.e. based on suppliance of
morphemes in their obligatory contexts). She also obtained accuracy scores
that took into account over- as well as under-suppliance of each morpheme in
order to build up a picture of the learners’ abilities to perform in a completely
target-like manner.
Pica found no evidence that the accuracy order of the instructed group was
disturbed. The morpheme rank order of this group was significandy correlated
with the orders obtained for both the naturalistic and mixed groups. It also
correlated strongly with Krashen’s (1977) ‘natural’ order. Pica comments: ‘all
groups of subjects, across all language contexts, exhibited a highly similar
overall rank order or morpheme suppliance in obligatory contexts’ (1983:479).
However, when Pica investigated over-suppliance of grammatical morphemes
(i.e. overgeneralizations such as ‘teached’ and the use of morphemes in non-
obligatory occasions as in ‘He lived in London now’), she found considerably
more instances in the instructed than in naturalistic group. For example, the
instructed group overused V-ing to a much greater extent, a result that
replicates Lightbown’s finding. Also when Pica investigated the accuracy
levels of individual morphemes, group differences were apparent. In the case
of the plural -s inflection, suppliance was greater in the instructed group than
in the mixed or naturalistic groups. The naturalistic learners tended to omit
plural -s on nouns premodified by quantifiers (e.g. ‘three boy’) - a pattern of
language use frequently noted in pidgins. Differences were also evident on
 Formal Instruction and Language Learning 141
third person -s. The percentage accuracy of this morpheme in the instructed
group was 63 per cent, but only 22 per cent and 25 per cent respectively in
the mixed and naturalistic groups.
Pica is cautious in making claims for instruction. She points out diat the
sample (only eighteen subjects, six per group) was very small and that the
study was restricted to adult native speakers of Spanish. She also emphasizes
that any claim must be restricted to interlanguage production and that no
conclusions can be drawn about acquisition or ultimate attainment.1 But Pica
suggests that the results support the prevailing view that learners contribute a
great deal of‘natural ability’ to their acquisition of a second language. Instruction
does have some effects, however; it can inhibit the use of non-standard forms
and it can, in certain circumstances, have a positive effect on levels of
accuracy. Pica (1983) suggests that instruction can help learners to outgrow
the use ol pidgin-like constructions that are communicatively effective but
ungrammatical. Pica (1985a) argues that the impact of instruction is related to
the linguistic complexity of different features:

1 Production of simple morphemes like plural and third person -s is

improved by instruction.
2 Instruction can result in over-production of forms like V-ing (i.e.
learners attach -ing to base verbs where its use is not required). Such
forms are less simple than -s forms because they perform a range of
functions.
3 Instruction has no impact on complex structures such as article a. This
is because the rules for form—function relationships in such structures
are not readily transparent.

Pica’s work, therefore, introduces the interesting possibility that instruction

has a direct effect on acquisition, but that this impact is selective — a point to
be returned to later.
It is difficult to come to any firm conclusions regarding the effect instruction
has on the process of acquisition on the basis of the morpheme studies
reviewed above. The bulk of the studies found that the rank accuracy order of
tutored learners did not differ from that of naturalistic learners. Morpheme
studies are methodologically flawed, however (see chapter 3). The assumption
that accuracy order is the same as acquisition order is almost certainly unjusti¬
fied. Also, the ‘accumulated entities’ view of L2 acquisition (Rutherford,
1987) is misplaced. This view is inadequate for a number of reasons. The
learner does not work on one structure at a time but on the whole system or
sub-system. Morpheme studies had their heyday in the 1970s. Pica’s classroom
study was one of the last to be published. Their value to our understanding of
L2 acquisition is extremely doubtful.
The detailed study of individual morphemes — of the kind illustrated in
Pica’s research — is informative, however, as it throws light on those linguistic
factors that mediate between instruction and learning. Such research needs to
be complemented by studies of syntactical features.
142 Instructed Second Language Acquisition

The Acquisition of Syntactical Features

Most modem linguists would consider the syntax of a language of greater

theoretical interest than its morphology. Arguably, the same holds for L2
acquisition studies. There has been a general shift in focus from morphemes
to syntax. This shift is reflected in studies of naturalistic and classroom
acquisition.
There are two kinds of syntactical structures which have attracted interest.
The first are transitional structures. These are structures like negatives and
interrogatives that are leamt in a series of stages with the target-language
structure as the final stage. Each stage constitutes a necessary stepping-stone
to the next stage. The second type are the syntactical structures that make up
a structural group. In this case, the structures conform to target-language
norms. Examples are relative-clause structures in English and word-order
rules in German.
There has been a substantial amount of research which has investigated the
L2 acquisition of English negatives and interrogatives in naturalistic settings.
The available research (cf. Ellis, 1984a; eh. 2) suggests that there is predictable
series of stages through which learners pass. Such structures present an
interesting test for the efficacy of instruction, as their interim stages involve
the production of deviant syntactical patterns such as:

No play baseball
Britta no have this
This is flag?
What they are picking?

The key question is whether instruction can help the learner to avoid such
transitional forms and lead her directly to the target structures. We have
already considered evidence from error-analysis studies that suggests this
unlikely (see p. 136). We will now consider a number of longitudinal studies.
Ellis (1982; 1984a) carried out a one-year study of three children aged
between ten and thirteen in a London language unit. Ellis collected samples
of speech that the learners produced in meaning-focused classroom interactions.
He found that the developmental profile for both negatives and interrogatives
was more or less identical to that reported for naturalistic L2 acquisition. For
example, early negatives consisted of anaphoric negation (i.e. ‘no’ by itself or
‘no’ + a separate statement), followed by external negation (e.g. ‘no sir finish’)
and then internal negation (e.g. ‘somebody not driving with this’). The children
failed to reach the next stage, where the negative particle is realized with a full
range of auxiliary verbs. Ellis claims that the children appeared to rely on
exactly the same processes that characterize the acquisition of negatives in a
natural setting, despite the fact that standard negative structures were formally
taught at various points during the year.
 Formal Instruction and Language Learning 143
Ellis’s learners were not pure classroom learners, so caution needs to be
exercised in claiming that instruction has no effect on the acquisition of
transitional structures on the basis of his study. It is possible that exposure to
natural language use overrode the instruction they received.
Weinert’s (1987) study, however, is not subject to the same caveat. She
carried out a four-year study of the acquisition of German negation by
Scottish secondary school children. Data for the study consisted of both the
utterances the learners produced in class and speech elicited by means of a
meaning-focused task performed outside the classroom. Weinert found that
the negative utterances produced by first-year students consisted mainly of
routines and patterns of the kind:
4

Das ist kein_

Ich spiele nicht gern_

This gave the appearance of relatively well-formed negative sentences. How¬

ever, the first year data also included examples of non-standard negative
structures of the kind found in naturalistic acquisition. In the second year the
learners used external negation more frequently:

Nein, ich leme Karate.

The learners gradually switched from external to internal negation between

the second and fourth years, first producing utterances with pre-verbal negation:

Zola Budd nicht spiele Fussball.

and later the standard post-verbal negation:

Ich kann nicht spiele Gitarre.

Weinert argues that the learners constructed their interlanguage systems in

the same way as naturalistic learners. She speculates that instruction might
actually hold learners back by interfering with the organic processes of acqui¬
sition. For example, external negation might have appeared productively in the
first year of acquisition but for the formulaic learning that resulted from
extensive drilling of negative patterns.
One study (Ewbank, 1987), however, does suggest that instructed learners
follow a different route to that of naturalistic learners. Ewbank investigated
the acquisition of L2 German negatives by six university students. She found
that all the learners placed the negator sentence final at one stage of development
(e.g. Es ist Wanduhr nicht). This has not been reported in naturalistic L2
acquisition. Ewbank speculates that it was the result of students’ attempts to
speak in ‘complete sentences’ — a requirement of the instructional method. If
this explanation is accepted, it constitutes further evidence that instruction can
144 Instructed Second Language Acguisition
interfere with the natural processes of acquisition, resulting in deviations from
both target language and interlanguage norms.
Comparative studies of group structures also indicate that instruction has
little effect on acquisition. We will consider the results of two studies — one
of English relative clauses and the other of German word-order rules.
A number of researchers have claimed that the acquisition of relative
clauses conforms to general principles of linguistic markedness. Studies of
naturalistic or mixed learners (e.g. Gass, 1979) have shown that Comrie and
Keenan’s Accessibility Hierarchy (see n 12 to ch. 5) is able to predict the
order in which the grammatical functions of relative pronouns are acquired.
Pavesi (1984 and 1986) reports on a cross-sectional study designed to establish
whether markedness principles applied equally to the formal and informal
acquisition of relative pronouns. Her formal learners were forty-eight EFL
students in an Italian high school; her informal learners were thirty-eight
Italian workers in Edinburgh. She elicited sentences requiring relativization in
different syntactical positions by means of picture cues. Using implicational
scaling, she was able to show that (1) the acquisition of relative pronouns
progressed from unmarked to marked functions and (2) both groups of
learners manifested the same implicational ordering. In other words, there
was no difference in the acquisitional sequence. There were other differences,
however. The formal learners demonstrated greater quantitative development,
i.e. they produced a larger number of the more marked features. Also, the
informal learners produced errors consisting of noun copies, while the formal
learners produced pronoun copies:

The man who my father was talking to the man .... (= noun copy)

The man who my father was talking to him .... (= pronoun copy)

However, Pavesi does not consider formal instruction per se the source of
these differences. Instead she argues that the higher levels of acquisition
shown by the formal learners was the result of exposure to the planned
discourse found in classroom communication (see discussion on p. 121).
The second study investigated the classroom acquisition of German word-
order rules. Ellis (1989b) based his research on the findings of studies of
naturalistic L2 acquisition (e.g. Meisel, 1983; Clahsen, 1985) which show that
learners acquire German word-order rules in a predictable sequence:

1: Adverb preposing This rule states that, optionally, adverbs can be

moved into sentence initial position, e.g.

Da spielen Kinder.
(There play children.)

The rule requires inversion (see below), but naturalistic learners typically
acquire the rule in the first place without inversion,8 e.g.
 Formal Instruction and Language Learning 145
Da Kinder spielen.

2: Particle This rule states that non-finite verbal elements are moved to
clause final position, e.g.

Ich mochte heute Abend ins Kino gehen.

(I would this evening into the cinema go.)

3: Inversion This rule states that a finite verb form precedes the subject of
its clause in certain linguistic contexts, e.g.

Wann gehen wir ins Kino?

(When go we into the cinema?)

4: Verb-end This rule states that the finite verb is placed in final position
in subordinate clauses, e.g.

Er fragte mich, warum ich traurig war.

(He asked me why I sad was.)

In order to explain this acquisitional sequence, Meisel (1983) and Clahsen

(1984) have suggested that development proceeds in accordance with the
cognitive complexity of the different operations involved in the production of
each structure. Psycholinguistic restrictions govern acquisition, such that one
operation serves as the prerequisite for another, more complex operation.
These processing operations are discussed in some detail later (see table 6.3).
Ellis set out to investigate whether a group of thirty-nine learners of
German as a foreign language in beginner courses manifested the same route
of acquisition for die obligatory word-order rules (i.e. particle, inversion and
verb-end). He used data collected by means of a communication task which
the learners performed in pairs on two separate occasions. The results revealed
the same order of acquisition for both the learners individually and as a group,
i.e.

Verb-end > Inversion > Particle

This order did not correspond to either the order in which the diree rules
were first introduced or to the order of emphasis placed on the three rules in
the classroom instruction the learners received in the six months’ duration of
the study. Ellis concluded that the learners followed their own internal ‘syllabus’
in the same way as untutored learners. This syllabus proved immune to
instruction, despite the fact that the learners (college undergraduates) were
well equipped in terms of experience and cognitive skills to benefit from direct
intervention. Ellis noted, however, that the learners were in general successful
in acquiring some knowledge of verb-end, the most advanced of the word-
order rules — in contrast to naturalistic learners who in many cases failed to
146 Instructed Second Language Acquisition
acquire inversion or verb-end in the studies carried out by Meisel and
Clahsen.9 German word-order rules have attracted considerable interest from
classroom researchers. We will consider a number of other studies in a later
section of this review.
These comparative studies of L2 syntax lend further support to the claim
that the effects of instruction are extremely limited. The pervasive finding is
that the overall sequence of acquisition is the same in classroom and naturalistic
settings. There is some evidence to suggest that instruction may help to push
the learner further along the sequence, but there is also some evidence to
suggest that it may inhibit progress by encouraging the use of alternative
strategies of production.

Review

We have now looked at a number of studies which have investigated the effect
of instruction on interlanguage development by means of comparisons between
classroom and naturalistic acquisition. Table 6.2 provides a summary of these
studies.
Comparative studies of the effects of instruction afford a number of insights,
but are difficult to interpret. This is because there is no way of deciding
whether any difference between classroom and naturalistic L2 acquisition is
the result of formal instruction per se or of the special communicative properties
of classroom interaction. Comparative studies provide no information about
the impact of instruction on the acquisition of specific linguistic features. In
Long’s (1983b) terms they shed light on the relative utility of instruction but
not on its absolute value. Experimental studies, however, are not subject to this
limitation, although they do have other problems.

Experimental Studies of the Effect of Instruction

Experimental studies enable us to examine whether explicit instruction directed

at feature x results in x being acquired. The design of these studies follows a
similar pattern. The learners’ existing L2 knowledge is measured by means of
a pre-test. The ‘treatment’ follows in the form of instruction focusing on one
or more specific linguistic features. This usually takes place in the learners’
normal classroom setting. A post-test is then administered in order to determine
what gains in knowledge have taken place. In some studies, a further post-test
is administered some time later in order to establish whether any gains
identified by the first post-test are durable.
Experimental studies fall into three groups: (1) accuracy studies, (2)
acquisition-sequence studies and (3) ‘projection’ studies (i.e. studies that seek
 different from order of

instruction evident in
instruction. Effect of

grammar test.
a?

-G
CJ
X
D
D
IDD H->C/5
CJ

CJ *
C"HH
D
CJ
D a£
C/5
D D
3 o
Oh o G, g 2 C/5
<3 C/5 C/5 C/5 C/5 G
?3
b£ bo o o £
#G 6 G a d D
s
c3 C3
• »H
2 o
o
T3
o
o
G
cti 2cj G
u S- u i-i -4—>
6o
O CO
o C/5
o
cj
C/5
C/5
G
O
*a G
o -a
D
2
D
2 Oh c Gh rj
P^ 'o P^ 2 co o co c3

D d
Comparative studies of instructed and naturalistic L2 acquisition

<§, +->D i~iC/5

JJ D
D
4H
D
CO
I-h
D
—ii|
T3
S' "Bh s
^ D
'Bn 5
a.
o *&> 14
fS 1“ •a
D
ux S.s PP

G G
D <D
*-i N OJ
2 2 -a
os c/5
2cj 20) kb 2
2i
-Sy1 2 2 •S 00
a ’oi *oi G -a i
u i-in 2 S 00
d D -a ™
>-> X
So
U \o co

, T3
—1 2
^ S
as O co
-3 w
.S D a
X c/s > ^
<U .2 o * C/5
g £
"Hb 2 •4D-» §3
S c -5
w « g
t—t
R
D

ON
P CO ON
■h*
Table 6.2

00
CO ON ns
S
£D
£D
00
ON G
Pi pH H
 <D TD
C/5 4-1 CD
re O bo ■*-> *p
bn <u c p O -a <D
i_ *T3 G
ctf <d
■4-* cfa ,| .
45
Uh D
TD 3 3 G -G ■ —* '-4-H i- co G
u X ’2
■G S ^
<u G d
^ © p ^ > -O _, cr

-C g
6 i|
D 4 QJ O t-, ^ O
OJ s c
o ^
(D j-h
0
/
p G
*ap D
fli P
^ « G 5
<D
3
x>
|
_
G G
a OJ
“ QJ
, *-
^
• -J O P! P
•0 43 P .5 "n -a I 2 “_ D
a & 6 -a
o S o G -f L- T3
O >-
«
26 J-h
O
•s ^0-0 J-H 03 G 2
o *■* D 1 o
2 3 czs 4> 2U D 0 L C D
= G
« c 3 * w 4>
303
3 "G
C/3 S G
<D t) D P3
Oh
C/3
£ 6 3
°:ll 3 .g
s
3 S—
u o
^ *P P <u
<D
*3 o3 4> a! 2 G
D
g -a 1
Vh aJ <D
-G G

"2
<d *3
§"2
fcj
4>
l"H
.<D
l—l
CD
T3 r 1=
- § 3 I S
"x
g
<D
£
-5

o
J2

c
C
03
4—'
<D
& G T3
O 5 -B
"S ^

Is
o
2
8 .2
2 a 8
Cl l-
G

6 8 2 -a o
♦ •—I J-H
§ & g ■g.G -fa -G § 8
X> go _
o w c/^
S' 2
<P <73
2
s s .2 o 3
c« cl

c
o
T3
d
C/3
S
^
8
<D
-s
Ctf 3
JO ft O P '%
D C/3 (X
•G
CJ C/3
S 03 D
a
03 G
C/3
P
CJ
G D _. G
^3 O .
Q nb P O _
o P3 Gh
C/3 *P <D O
W3
g.-e
g cl Sr.
3 G T3
s a c/3 ■4—1 o 03 D
2
O
S£
S
y
I;
G G 'S.
•C 2 03 'TD ^ 4>
> .a
^ Ok H
J-H
8.g w
G
-<*
5
G
V3
(U

4<d
- -*
D
>
p QJ
<D
« G "8
4)
0 3
_) .G P
C/5
u bp ■S>
-a
re 2 ^2

6h c G < c G G TD J-h
re § CO *T3 c3 D
C/3 J ^ G
« 82)
c G
E
* >sC ^ hC3
a N ^ G tjJ \ T3 <D
(D C/3 41 D
3^ G £H O ^ oc G
4> -n ai *-4 re 8 42 C/3 _
^ <L) D <D 0 D O P S ^ u
•a 3 ^ ■g .3 "o 0
*3 2 "3 £
CO3 00^ (si 3 E
*-h x oS
•G -£=
(/5 < C/3
* ft
2w
voiiW
C/3 §
«
G

s D
X & \
D Z G
03 P 0
I
w G J-H G 0
o> 03 <D 03
Qh -G <D *a
D <
-§ ‘2 [g cn
Ap 8 G O fan D
'S G C/3
0 G < _c
cyj #G .2 _c
EFL

S: j d -J
high

hJ b
- b P-H b x
w .2 W w w w

G
c <D
c/3
o Wh ,^-v
03 LO
O o 00
J r- t^- 00
00
-a1 ON P r—<
ON
Cvj G ^ cp
CO 55
o * 03
c c 03 00
ON
G
c/3 B) J-H

.0)
13 .§ 2 cO
>
■Q OJ -3 <D
L-
4)
4)
u.
•S
Bj
OJ
O
|2 Cl, b
03
 « D
C/3
V 0)
D c/5 c
nO
G
nO
dJ S
G
be y -a 3 <D o cO o
,_, no CD •G nO C/3 *G *
G G nO CO 03 « m 3 <d O no I °
QJ bo dJ 4—' Oh dJ I -G a
S «• a c ?P « CJ C/3 D t-i
CD i cd ■S GD
6 •*-' G
“ o <D 6 §h c3 C/3
tH QJ 11 d
C/3 -w <G o cj G <d T3 r3 <4H
<u S
l*i o cj U. Ii
.w O £ ^-h ^ u D <D 03 O G _ *G In
cv-0 -S G ^ d. O O C
- E
o3
4-*
O O 3<t| O
o t« p;§ c
D 0 CL)
cd 4-> 4O-J ft
CJ O
S
P c3 _* *G
dJ 03 . "y -a
CO bo 4- G ' jD G -*-1 PE
| 5?2 u. O »H c -a tn
■s^ C S L* Oh C
| C-B
g g <d PI t° -a £'{3 rt >->
cd .G CJ 3 CJ I o c d d
^ J5 03 2 o
2 *G c/3 __: S e3 T3 W
^3 "c« O X> o 3 c
E w 'H >
•G -a
^ C/3 <D O’ E 2 3=
^ o.&2 (D bO •3 '3 U D "> O *G g « »i El
rt Sh O
fl :s si o .2 *
£ o3 ■d1 ®
ff-S V § ^ 1/
o3 O
fc O s
2 d t
|U £ S .3 a 3 « '3 £>
«
E «t ^ g«3
cj V) P
rt ”y E
HE u
§ £-S
t» o 2 « ? g o <*, 3 a o- cr r- pH D dJ
0£.S ctf si 05 "O GO ^0 cS rt -2 !/l O J "3 V3
QJ >
ca'S-S O O I Si
c/3 GcO

D -a .2 dj
aj G dj
-G

y
'be
o C
d *g
•a
>. 03

i03
4#CJ CJ
<D
-G
CJ
dJ
O
Cj
E
03
bo
G
c
o
I.&
<D
u no
CJ
OJ
OJ
a *5
*G CJ C/3 O •d ^ c«
G G- C/3 *G G 4—< 1/3 C/3 #CJ
C/3 a
G O ‘o u G *G
111 S CJ *c
G G
cj O
dJ
s
»s o3
4—> 'S CJ
G O
<L> G
cj
’2 il) c/> G
O O G C3 G
1 i § CJ O s
4-J
T3 O -a C/3 dj
dj (h s 2
i
3
§
fe
3
2
3
G C/3
O 03
C/3 G
O s
CJ
C/3
G 3
u jo3 CJ
G
o
o
CJ
*-G
Cj O Cj cj
U 13 tS 2 "cj C/3 CJ o lj *Sh cz> .2 o3

o <u
4-J 4-»
03 C/3
dj C/3
4-* U 4-< ^
<U dJ <U dJ dj <U dj
G
"S- s 1 § s G
I-I | -5i) «J ‘5b
dj
U i Ui .3 CQ CQ

o C/3
4—'
r—t OJ C/3
C/3 .2 G
TD cd 00 <u O
dJ ^2 G
no dJ 2 co C/3J- G *G
bo G G JUi <U Jh o3
03 Cj O G dj CJ
CJ OJ CJ o dj G
<u C/3 o o
G
<u
l—(
t/3
4—1
C/3
dj

4—> 8.
£ -2 CJ 4—< no
dJ
2 G #g 3 C/3 3 5-h
C/3 dj o 3 dJ no G no <u
-a Tt“ o no > Oj 03 -G
2 &
2 o -G
03
CJ
ro CJ *2 oc bo 3 4—i ON bo
ro CO C/3 NO G nt" 3 CO 3

C hJ .2
.2
C/5
hJ UJ
.2 Ph (Jh fe
3 03
4—1
‘G C/3 c^ ■rt -a C/3
03 e 03
PQ M tb
G G
_g g| 03 •H o
O .2
Gem

o
USA

2 "O G *c3
hJ H C/2 J-H 4—*
fa dj -r
caj 8
w O co uj .S 0 m

00
ON

CO
Jp
ON O'
oo
O'

C/3
dj

%
Ph w
150 Instructed Second Language Acquisition
to establish whether instruction in feature x not only results in the acquisition
of x but also triggers the acquisition of features y ... n).

Accuracy Studies
In accuracy studies the effects of instruction are measured by investigating
whether there are any gains in the accuracy with which specific structures are
performed after the ‘treatment’. We will consider four such studies. First,
Schumann (1978) attempted to improve the accuracy with which Alberto, a
Spanish-speaking learner of English in the United States, produced negatives.
His reason for giving instruction was to discover whether the apparent
‘pidginization’ of Alberto’s English could be overcome. Prior to the instructional
experiment, Alberto’s negatives were mainly of the ‘no + V’ type. The
instruction covered a seven-month period, during which both elicited and
spontaneous negative utterances were collected. There was a marked improve¬
ment in the accuracy of elicited negatives (64 per cent correct as opposed to
24 per cent before the instruction). However, the accuracy of the spontaneous
utterances showed no significant change (20 per cent correct as opposed to
22 per cent before instruction). Schumann concluded that the instruction
influenced production only in test-like situations, while normal communication
remained unaffected.
Second, Lightbown, Spada and Wallace (1980) studied the effects of in¬
struction on the accuracy with which three different structures were produced:
(1) the -s morpheme, used to perform five functions (plural, possessive, third
person singular, copula and auxiliary); (2) ‘be’ in sentences like ‘He is sixteen
years old’; and (3) locative prepositions indicating motion toward a goal (e.g.
‘to’). The subjects were French-speaking children and adolescents learning
English principally in a classroom setting in Canada. Data were collected by
means of a grammaticality-judgement test that required the learners to identify
which sentences were correct and which ones were incorrect and then to
correct the incorrect ones. This test was administered on three occasions —
immediately before the instruction, immediately after and five months later.
The instruction took place over a two-day period and consisted of a review of
the target structures using a list of twenty-five correct and incorrect sentences.
Overall scores immediately after the instruction showed an average of almost
11 per cent improvement for the secondary school students. A control group
showed only a 3 per cent improvement. The Grade 6 students improved
about 7 per cent. The researchers comment: ‘the improvement was clearly
attributable to the period of review instruction’ (1980:164). However, the
scores on the second post-test fell back to a level between those of the first
and second administration. The suggested reason for this is that ‘improve¬
ments .. . were based on the application of knowledge temporarily retained at
a conscious level, but not fully acquired’ (p. 166). This study, then, suggests
that instruction can result in increased accuracy in production but that the
gains may not be long-lasting.
 Formal Instruction and Language Learning 151
Third, Ellis (1984b) investigated the effects of instruction on the pro¬
duction of four semantically appropriate WH pronouns (‘who’, ‘what’, ‘where’
and ‘when’) and of inversion in interrogatives. The subjects in this study were
thirteen children aged between ten and thirteen learning English full-time in a
London language unit. The instruction consisted of three one-hour lessons
involving contextualized practice in both teacher-led and group work. Inter¬
rogative data were collected by means of an elicitation game that required the
subjects to ask questions about a picture. This resulted in relatively spon¬
taneous speech. The game was played by the learners in pairs immediately
before and after the instruction. There was no significant improvement in the
accuracy with which either semantically appropriate WH pronouns or inter¬
rogatives with inversion were produced for the group as a whole. However, a
number of children showed a marked improvement. Ellis speculates that the
within-group differences may have been the result of individual learner factors
or of the way in which opportunities for practice were distributed. To test the
latter possibility he calculated the number of opportunities for practising
‘when’ interrogatives experienced by each child in one of the lessons. Some¬
what surprisingly, children with the fewest opportunities were the ones who
showed the largest gains in accuracy in ‘when’ interrogatives.10
Fourth, Kadia (1988) studied whether formal instruction was successful in
enabling a Chinese student at the University of Toronto to avoid errors in the
placement of pronominal direct objects, as in:

Last time I show Beth it.

He told me that he will call up me this evening.

The pre-test consisted of a substitution task and a grammaticality-judgement

task. Forty minutes of formal instruction consisting of formal explanation and
drill was provided. The subject was observed in informal contexts for nine
weeks following the instruction and a post-test, similar to the pre-test, was
administered two months after the instruction. Kadia found that the instruction
had no real effect on the subject’s spontaneous language production, but there
was some evidence that it aided her controlled production (i.e. in the substitution
task). This study, then, produced similar results to that of Schumann.
Taken together these studies suggest that there are constraints on the
effects that instruction can have on acquisition. The studies by Schumann,
Ellis and Kadia indicate that spontaneous speech production may be impervious
to instruction. Three of the studies (Schumann, Lightbown et al, and Kadia)
lend support to the claim that instruction can improve accuracy in careful,
planned speech production. However, this improvement may disappear over
time, as more ‘natural’ processes take over. Ellis’s study also raises an important
question - how does formal instruction actually contribute to acquisition?
The results he obtained do not support the common-sense assumption that it
is practice that counts. This is a salutary finding as it reminds us of the need
to find out how instruction works as well as whether it works.
152 Instructed Second Language Acguisition

Sequence of Acquisition Studies

The sequence of acquisition studies which we will consider have been based
on the substantial body of research into the naturalistic acquisition of L2
German word-order rules by adult migrant workers and their children and,
more recently, of the L2 acquisition of English by members of ethnic minority
communities in Australia (cf. Pienemann and Johnston, 1987). This research
indicates that there is a natural sequence for the acquisition of these rules (see
p. 144 above), which holds for adults and children and which applies irrespective
of the learners’ LI. The classroom research carried out by Pienemann and his
associates (Pienemann, 1984, 1985 and 1986; Johnston 1987a and 1987b;
Pienemann and Johnston, 1987; Nicholas, 1985; Jansen, 1987) has been
designed to discover whether formal instruction is sufficiently powerful to
disrupt the sequence of acquisition. The results provide us with the most
powerful evidence collected so far that under certain conditions formal instruc¬
tion can influence L2 acquisition. These conditions govern both the nature of
the linguistic feature which is being taught and the timing of the instruction.
The major strength of Pienemann’s research is that it is based on a well-
defined theory of L2 acquisition which has enabled him and fellow researchers to
test specific hypotheses. Pienemann emphasizes that the theory is predictive,
not merely explanatory. We will begin by outlining the theory.
The Multidimensional Model (Meisel, Clahsen and Pienemann, 1981)
distinguishes two sets of linguistic features; developmental and variational.
Developmental features are those that are constrained by developing speech¬
processing mechanisms. Variational features are those that are not so con¬
strained. In other words, the model predicts that certain features will be
acquired in sequence because of the mental operations involved in processing
them, while other features are ‘free’. Variational features, however, are not
entirely random, as their acquisition is dependent upon certain psychosocio-
logical factors such as the learner’s mental make-up and her position and
prospects in the host society (Johnston, 1987b). German and English word-
order rules are examples of developmental features, while copula ‘be’ is an
example of a variational feature.
The processing operations responsible for the acquisition of developmental
features involve the learner’s ability to manipulate syntactic elements in
grammatical strings. Pienemann (1986) and Johnston (1987b) characterize the
developmental stages through which a learner must pass in terms of the
different processing operations involved. These are summarized in table 6.3,
with examples from L2 English. It is important to recognize that each operation
makes possible the production of not just one but a number of structures. The
operations are sequential: the learner’s capacity to perform the operations
involved in one stage entails her already being able to perform those of the
immediately preceding stage. Acquisition, however, does not proceed by means
of sudden jumps from one stage to another; rather it is a continuous process.
That is, the learner continues to work on the processing operations involved
 Formal Instruction and Language Learning 153
Table 6.3 Processing operations involved in the acquisition of grammatical rules

Stage Processing operation Linguistic realization

1 Production relies on non-linguistic Undifferentiated lexical items;

processing devices. Learner has no formulas such as ‘I don’t know’
knowledge of syntactic categories. and ‘I can’t’.
2 Production of simple strings of Canonical word order; intonation
elements based on meaning or questions (e.g. ‘You playing
information focus. Learner still has football?’)
no knowledge of syntactic
categories.
3 Learner is able to identify the Adverb-preposing (e.g. ‘Today I
beginning and end of a string and to play football’); do-fronting (e.g.
perform operations on an element ‘Do you play football?’); neg. + V
in these positions e.g. learner can (e.g. ‘No play football’).
shift an element from beginning to
end of string and vice versa. These
operations are still pre-syntactic.
4 Learner is able to identify an Yes/No questions (e.g. Can you
element within a string and to move swim?); pseudo-inversion (e.g.
this element from the middle of the ‘Where is my purse?’).
string to either the beginning or the
end. This operation is again
characterized as pre-syntactic.
5 Learner is now able to identify WH inversion (e.g. ‘Where are
elements in a string as belonging to you playing football?’); Internal
different syntactic categories. She is neg. (e.g. ‘He did not
able to shift elements around inside understand?’).
the string.
6 Learner is now able to move Q-tags (e.g. ‘You’re playing
elements out of one sub-string and football today, aren’t you?); V-
attach it to another element. This complements (e.g. ‘He asked me
stage is characterized by the ability to play football?’)
of the learner to process across as
well as inside strings.

Source: Based on Johnston (1976); Pienemann, Johnston and Brindley (1988)

in one stage while beginning to acquire those involved in the next. For this
reason it is possible for gaps to appear in the grammatical rules associated
with the processing operations of a particular stage. Although a learner has
mastered a processing operation, she may not apply this to all the grammatical
structures that it is now possible to produce. Pienemann suggests that the
frequency of gaps correlates with the learner’s overall orientation to the
learning task. That is, if a learner is concerned with accuracy, few gaps will
appear, but if she is more concerned with communicative effectiveness some
gaps are likely to occur (see discussion of this distinction with reference to
variational features below). However, the model does claim that it is impossible
for the learner to skip over a stage. Both Pienemann and Johnston make a very
strong claim for this sequence — namely, that it is ‘strictly implicationa!’ and
154 Instructed Second Language Acquisition
not just ‘statistical’ (Johnston, 1987b), In other words, they maintain that the
sequence does not allow any exceptions. All learners follow it.
The developmental features identified by the initial research into naturalistic
L2 acquisition were all word-order rules. However, in subsequent research it
has been shown that the general processing operations can also be applied to
English and German morphology. Pienemann (1987) distinguishes local and
non-local morphemes. Local morphemes require Stage 4 processing operations
i.e. the ability to transfer information to the perceptually salient beginning and
final positions of a string. An example in German is the ge- prefix in develop¬
mental past participles (e.g. ge-denkt; ge-kommt). Non-local morphemes are
those that require the sentence-internal transfer of grammatical information (i.e.
Stage 5 processing operations). An example is English 3rd person -s, where infor¬
mation has to be transferred from the grammatical subject to the verb in order to
determine whether -s should be added. Extending developmental features to
include morphemes has greatly increased the power of the model.
The factors that determine whether a variational feature is or is not acquired
are of a very different kind. Learners differ in the extent to which they orient
towards communicative effectiveness or towards correctness and standardness.
Those learners who give primacy to the instrumental function of language are
likely to make use of simplified and restricted forms. In this way they can
reduce the amount of processing time spent on linguistic features and so
allocate more to the conveyance of meaning-content. Other learners give
greater importance to the integrative function. This involves using language in
such a way as to identify with native speakers of the host community. These
learners are more likely to pay attention to target-language norms. Thus,
whereas the first kind of learner might say ‘I Spanish’, the second kind is
more likely to say ‘I am Spanish’ (Johnston, 1987b: 13). The learner is credited
with limited processing resources and needs to make a decision about how to
put these to best use. Variation between learners is a result of the different
decisions they make.
The acquisition of an L2, then, proceeds on two dimensions, as shown in
figure 6.1. Development on one dimension is independent of development on
the other. Thus a learner like C can progress a long way along the developmental
dimension while remaining oriented towards the instrumental function of
language and, as a result, continuing to manifest simplification and restriction
of variational features. A learner tike B, on the other hand, displays development
on both dimensions, i.e. progresses along both the developmental and variational
axes. Learner A is the opposite, showing little development on either axis.
A number of general hypotheses relating to the effect of instruction on L2
acquisition have been based on this model. These are:

1 Instruction will not enable learners to acquire any developmental feature

out of sequence.
2 Instruction will enable learners to acquire a developmental feature
providing that the processing operations required to produce those
 Formal Instruction and Language Learning 155

Complex processing

Developmental
dimension

\
Simple processing
A

4
Simplified -► Elaborated
Variational dimension

Figure 6.1 The Multidimensional Model of L2 acquisition

Source: Based on Johnston (1987b)

features that precede it in the acquisitional sequence have already been

mastered.
3 Instruction directed at developmental features for which the learner is
not ready may interfere with the natural process of acquisition.
4 Instruction will help learners to acquire variational features.

Hypotheses relating to the acquisition of specific grammatical features have

been tested in a number of studies, which we will now examine.
An experimental study set up to investigate these hypotheses was carried
out by Pienemann (1984). The subjects were ten children aged between seven
and nine years who had been learning L2 German naturalistically. Data were
collected on two occasions — before and after the period of instruction — by
means of interviews between pairs of subjects and between each learner and a
student interviewer. Additional data were also collected by means of hidden
recordings in the children’s playing environment. The instructional targets
were inversion, one of the German word-order rules linked to Stage 5 processing
operations and copula, a variational feature. Pienemann does not provide a
detailed account of the instruction used, but it appeared to involve the use of
dialogues and fairly intensive drilling.
For inversion, the results for only two of the children are given. Teresa
failed to acquire inversion. Pienemann provides evidence to show that prior to
the instruction she had not acquired particle, the word-order rule associated
with the processing operations of the previous developmental stage (Stage 4).
Following the instruction, some of Teresa’s utterances did manifest inversion,
but Pienemann claims that these were all rote-memorized patterns. Whenever
Teresa deviated from the classroom dialogues, she failed to apply inversion.
She provided no evidence of being able to use the rule productively. In
contrast, Giovanni did acquire inversion. He showed that he had already
156 Instructed Second Language Acquisition
acquired particle (but not inversion) before the instruction took place. Following
the instruction, Giovanni was able to apply inversion in a wide range of
linguistic contexts. Pienemann notes that in natural acquisition a new rule is
applied initially only in a small range of linguistic contexts. Giovanni’s advanced
use of inversion suggests more than normal progress. These results, then,
provide support for both hypotheses 1 and 2.
The second part of this study set out to examine the effects of instruction
on the use of the copula. This is a variable feature and one that Clahsen,
Meisel and Pienemann (1983) found to be a particularly reliable index of the
learner’s orientation. The frequency of copula omission in the five subjects
for whom results are provided decreased considerably after the instruction.
Furthermore, acquisition of this feature appeared to proceed independently of
acquisition of the developmental feature (inversion). Hypothesis 4 was sup¬
ported, therefore. However, data obtained from one of these subjects some
nine months later indicated that the rate of copula omission had risen markedly.
It is possible, therefore, that the effects of instruction on the acquisition of
variable features is short-lived, at least for some learners (cf. the results
reported for Lightbown, Spada and Wallace, 1980).
Hypothesis 3 claims that premature learning of a developmental feature
may be counterproductive. Pienemann’s study lends support to this hypothesis.
The results show that for two of the learners the use of the adverb-preposing
rule actually fell away as a result of the instruction on the use of inversion.
Adverb preposing is an example of a rule with Stage 3 processing operations.
One of the consequences of this rule in German is that inversion is also
required:

Jetzt bin ich miide.

(Now am I tired.)

When learners first acquire adverb-preposing, however, they do so without

inversion:

* Jetzt ich bin miide.

(Now I am tired.)

Pienemann suggests that as a result of the instruction these two learners

discovered that when they used adverb preposing they should also apply
inversion. However, they were not able to do this, because they were not
developmentally ready. They knew, therefore, that then- utterances containing
a preposed adverb were incorrect. This led them to withdraw the use of this
rule in order to avoid producing incorrect sentences. Pienemann notes that
this resulted in the loss of an effective means of attention-focusing.
In addition to this important study - which gives support to all four of the
general hypotheses listed above — a number of cross-sectional and longitudinal
studies of the classroom acquisition of L2 German give support to the
 Formal Instruction and Language Learning 157

Multidimensional Model. Daniel (1983) carried out a cross-sectional study of

adult beginners of German as a L2 at a university in Australia. Her results
revealed the same implicational order of acquisition of word-order rules.
Westmoreland’s (1983) study was also cross-sectional and involved similar
kinds of learners. The implicational order followed the same developmental
sequence as in the other studies.
We have already referred to the longitudinal study carried out by Ellis
(1989b). Pienemann (1987) carried out a similar study. The subjects were
three adult classroom learners of L2 German. Once again these were complete
beginners at the start of the study. He documents when specific word-order
rules were introduced in the instruction and was able to show that in the case
of developmental features the instruction had no effect unless a learner was
ready to acquire a new rule. For example, the particle rule was explicitly
taught in week 7 of the course but was not acquired until week 17 by Steven,
until week 15 by Guy and not at all by Vivien (who was studied for only nine
weeks). A similar contrast between the instructional input and the learners’
output was evident for inversion and verb-end. Pienemann draws two major
conclusions from this study: (1) formal learners progress step by step irrespective
of the teaching schedule and (2) the sequence of acquisition is the same in
classroom as in naturalistic acquisition.
This longitudinal study also provides further evidence of avoidance behaviour
induced by premature instruction. Pienemann found that the three classroom
learners differed from naturalistic learners in the way they handled the German
perfect tense. He argues that this structure belongs to Stage 5 (i.e. co-occurs
with the acquisition of inversion). Naturalistic learners tend to make use of
lexical markers of pastness (e.g. temporal adverbials) and a simplified version
of the standard perfect verb structure, e.g.:

* Er hat sprechen.
(He has speak.)

instead of

Er hat gesprochen.
(He has spoken.)

The classroom learners were taught the complexities of the perfect verb
structure at an early stage. However, because they were unable to handle the
necessary processing operations, they tended to avoid using the form altogether,
substituting modal + verb structures. Their chances of conforming to the
norms of correctness were greater with these because the main verb consisted
of the unmodified infinitive.
As a result of this research, Pienemann has put forward the teachability
hypothesis. This states (1985:37): ‘instruction can only promote language
acquisition if the interlanguage is close to the point when the structure to be
158 Instructed Second Language Acquisition
taught is acquired in the natural setting (so that sufficient processing requisites
are developed).’ The hypothesis rules out the possibility that instruction can
help the learner to beat the natural order of developmental features. However,
it does allow a clear role for instruction. First, instruction can facilitate natural
language-acquisition processes if it coincides with when the learner is ready.
This facilitation is evident in three different ways: (1) increased speed of
acquisition, (2) increased frequency in rule-application and (3) application of
the rule in a wider range of linguistic contexts. Also the teachability hypothesis
allows for the positive effect that instruction can have on the acquisition of
variational features. Instruction may serve a particularly important function
here as it may help communicative-oriented learners to avoid early fossilization.
A corollary of the teachability hypothesis is that premature instruction can
actually be harmful. Johnston (1987a) distinguishes three kinds of response to
premature instruction. First, there are learners who attempt to perform struc¬
tures that are beyond them by adapting them to the processing operations they
have acquired. This results in erroneous versions of these structures. Second,
there are learners who substitute developmentally less complex substitutes for
rules that they know they cannot perform. Johnston suggests that this strategy
is common among university students. Thirdly, learners who are very norm-
oriented (i.e. obsessed by correctness) may avoid difficult structures entirely
when instruction draws attention to the errors they make. It is this last
response that is likely to result in interference with the normal progress of
acquisition.
The teachability hypothesis is the most powerful account we have of how
formal instruction relates to learning. As we shall see later, it can explain the
conflicting results found in other studies of classroom L2 acquisition. It is
important, however, to recognize the limitations of the model and the research
that supports it. The theoretical basis of the model has been criticized on a
number of counts. Hulstijn (1987), for instance, has pointed out that Pienemann
does not set quantitative or qualitative criteria by which to judge whether a
specific processing operation has been acquired, nor is it entirely clear whether
the notion of processing operation applies to acquisition or production or
both. The empirical research upon which the hypothesis has been based is
still very limited. Given all the variables involved in teaching—learning, it
cannot yet be concluded that instruction has no impact on developmentally
advanced features. Pienemann and his associates provide very little information
about the kind of instruction their subjects experienced. It is possible that
their failure to find any effect was not because instruction per se was incompatible
with learning such features but because the particular type of instruction they
provided was wrong. It must be said, however, that the strength of the
theoretical framework that underpins the hypothesis - the implicationally
ordered set of processing operations - gives grounds for claiming that in¬
struction is powerless to affect at least some aspects of L2 acquisition.
We now turn to research based on a totally different theoretical framework.
This has produced results that cannot be easily be explained by the teachability
hypothesis.
 Formal Instruction and Language Learning 159

‘Projection’ Studies
The final group of studies in this review of the experimental research is based
on predictions derived from the study of linguistic universals. Two types of
universals can be distinguished: typological universals and universals based on
Chomsky’s theory of Universal Grammar (cf. Ellis, 1985a: ch. 8). The former
have been identified by examining a representative sample of natural languages.
The latter have been identified by studying individual languages in depth in
order to establish the set of principles or parameters which govern the way any
grammar is constructed. Both approaches allow for the existence of impli-
cational universals. These relate the presence of one linguistic property to
the presence of ohe or more other properties. For example the Accessibility
Hierarchy for relative clauses (see ch. 5, n. 12) predicts that if a language
allows for relativization in the object of comparative position, it will also allow
relativization in all positions higher up the hierarchy. Similarly, Universal
Grammar predicts that the pro-drop parameter determines whether a whole
cluster of features (including optional omission of subjects and free inversion
in simple sentences) is or is not present in a language.
Implicational universals allow predictions to be made about the acquisition
of sets of linguistic features. The learner is credited with a projection device
(Zobl, 1983) that enables the acquisition of one rule to trigger the acquisition
of all the other rules that cluster with it. This device explains why learners are
able to acquire a language quickly despite the immense complexity of the task
and the relative poverty of the input they experience.11 One of the interesting
questions that can be asked is whether instruction can activate the projection
device. Specifically, researchers have asked whether instruction directed at
feature x is powerful enough to trigger acquisition of features y, z ... n, which
are implicationally linked to x. They have investigated whether instruction
aimed at marked linguistic features (i.e. features that are difficult to acquire
because they are not universal) can facilitate the automatic acquisition of
unmarked features (i.e. features which are universal and, therefore, easy to
learn).
Gass (1979) based her study on the Accessibility Hierarchy. She investigated
the effects of instruction on adult ESL learners’ acquisition of relativization.
The experimental group were given instruction in recognizing and producing
sentences in which the object of preposition was relativized (i.e. a position low
down in the Accessibility Hierarchy). A control group was given similar
instruction involving sentences in which the subject and object were relativized
(i.e. positions high up in the hierarchy). The results of a sentence-combining
task (administered before and after the period of instruction) showed that the
learners in the experimental group not only succeeded in improving their
scores on object of preposition but also on all the positions higher in the
hierarchy. In contrast, the control group improved their scores on the subject
and object positions but failed to demonstrate any improvement on the lower
positions. Gass’s study, therefore, lends support to the projection hypothesis.
It has been replicated with similar results by Eckman et al. (1988).
160 Instructed Second Language Acguisltion
Zobl’s (1985) study also supports the hypothesis. This investigated the
effects of fifteen minutes of instruction on the acquisition of English possessive
adjectives by approximately forty French-speaking university students in Canada
who were assigned randomly to two groups. Both groups received intensive
oral practice consisting of question and answer and teacher-correction. The
first group received practice directed at the use of possessive adjectives with
non-human entities (e.g. ‘his/her car’); the second group’s instruction involved
examples of human-possessed entities (e.g. ‘his/her sister’). Zobl claimed that
these two features are implicationally ordered, such that acquisition of the
latter (the marked form) would result automatically in the acquisition of the
former (the unmarked form). The experiment bore this out. The second
group of learners showed gains in both features, while the first group showed
gains in neither. Zobl also reports a replication of this study one year later.
The results were the same except that the first group did this time show some
gains in the use of possessive adjectives with non-human entities. However,
these gains (the target of instruction) were still smaller than those demonstrated
by the second group, which received no instruction at all in this feature. Zobl
also examined the errors made by both groups. The first group tended to
overgeneralize ‘his’ and to substitute the developmentally simpler ‘the’. The
second group overgeneralized ‘her’ but were less likely to substitute ‘the’.
The third study (Henry, 1986), investigated whether fifteen adult English
learners of L2 Chinese were able to predict the positioning of relative clauses
with regard to the head noun on the basis of general exposure to word-order
phenomena but without any specific instruction in the use of relative clauses.
The word order of Chinese is basically head-final (i.e. modification generally
precedes the head noun). English is basically head-initial (i.e. modification in
general follows the head noun). The students were asked to translate ten
sentences, some of which contained relative clauses, into Chinese. Those
students who attempted to translate the relative clauses invariably positioned
them before the head-noun. When asked why they thought Chinese had pre-
nominal relative clauses they gave answers like the following:

Well, if you want to say ‘the door of the house’ you have to say ‘frangzi de
menkour’, ‘house of door’, and you put adjectives before the noun, so it just
sounded right that way.

Henry concludes that these students were able to access parameters of word
order and that this enabled them to ‘know’ features that they had not actually
been taught.12
These three studies indicate that the relationship between instruction and
acquisition is much more complex than is generally assumed. We may draw
the following tentative conclusions:

1 Instruction can result in learners acquiring not only those features that
have been taught but also other features that are implicationally associated
with them.
 Formal Instruction and Language Learning 161
2 Instruction in marked features can facilitate the acquisition of umnarked
features, but not vice versa.
3 Instruction in unmarked features may result in learners’ simplifying
their interlanguages, whereas instruction in marked features aids the
process of complexification.

The notion of‘projection’ is an extremely powerful one. Of course, instruction is

not necessary for projection to take place; unfocused input will suffice. However,
instruction may be particularly helpful for the simple reason that it can supply
the learner with plentiful examples of those marked features that will serve to
trigger the acquisition of unmarked features. Also, as Zobl (1985:342) suggests,
marked data may* be necessary if interlanguage construction is to progress
beyond a certain stage of complexity, and the best way to ensure the learner
receives such data in sufficient quantity and attends to it is through formal
instruction.

Review
Table 6.4 provides a summary of the experimental studies which have been
discussed in this section. In contrast to many of the comparative studies, this
research provides convincing evidence that instruction can have a direct effect
on the acquisition of specific linguistic features. It also suggests both the
conditions that have to be met in order for the instruction to work and the
conditions under which instruction will prove most effective. The research is
not entirely comforting to supporters of instruction, however. There is evidence
to indicate that the effects can wear off over time. This suggests that although
instruction may bypass natural processing mechanisms in the short term, these
will eventually reassert themselves. Also, there is an apparent contradiction in
the results obtained from the sequence of acquisition studies carried out by
Pienemann and his associates and those obtained from the projection studies
based on the study of linguistic universals. Whereas the former support the
claim that instruction will only work if it is directed just one step ahead of the
learner’s current level of acquisition, the latter support the claim that instruction
will be more effective if it is directed several steps beyond the learner’s
current level. These claims derive from very different underlying theories. If
we are to make overall sense of the relationship between formal instruction
and learning we will need to develop a comprehensive theory that will explain
and reconcile the two sets of results.
This review of the research has covered (1) studies which have examined
the effect of instruction on the rate and level of success of acquisition and (2)
studies which have investigated whether instruction influences the process of
acquisition. The available research is now quite extensive, although it is also
limited in a number of ways (e.g. the size of the sample, the nature of the
instruction provided and the choice of target structures). Also it is not easy to
interpret. We turn now to consider what answers can be given to the central
question with which we began the review.
 T3 C/5
% P oS
3 * 5 G -G
G
G 0.-3 "2
.S £ - -a
S2 <d
c § 3-2
a-S HO M
§ g a-g s
T3
1)
G •£ m o
§'V o 2
'C 0^3 G -<-* T3
•G .2 <d 9 o G ^
gO <D1 Oo flj -*—'W
c«t t- T3
«L> Uh
C/5
CD
6 S .52 °* <u ^ o
> ?
rtt Oh In TJ *5
c C ^ O *G x> -fl
t:
oj g ^ 2 o e ^ 03
•3 rt O1
oS M > rt t3
C/5 H u o y b£) <D ^
<D C/5 3^2
^ --2 G
So Hcfl H c
•§s:
(>5 c/5 rn o
a >
2 G .2P y « •=
o I B i—i c/5 as .s cct a &

-G
O
a>
<u 2 T3
^ <U
c/5

tf
V2 qj
C/5 ^
■S 2
_ G
Q
a
2O oG a o
G"
«5 rj •a
2 *-o U B 03
G crj -to c
03 •*-* b£
•*-» *a O
O 3 —
G G
O -G O £ C bD P .3 —
ai
G< . a-S 2 ~2 § .y 4-*
CO + CO c/i ci) .2. O -
c

<d G ai C
#P4 a> .j u
c
o
~ G <D * 6 v
—' & =b R< -2
<l> J2 a e-y
V) .5 n u a
u o

3 ^ S bfl
^ ^ a
cr € COJ
o $ -a g «*
C/5 “ £
CO
£ §3
O
s
as
cm

J-H <D
a> a; T3
E E 2
as o3 «2
<D <D T3 ■—- 03 CZl-
i G <75
a <L> -£
Co T3
G G c § -a •—
-a
G
«
as
T3
OS
<u 7G CD
G c/5 C/5
U V
<u fll
CD -G
G G 5o°° O
o w «-> n "O IT)
O
CD
fc
U ^
as vO »“< Ctf 1—l
CD
0>
CD
G G G
© O .2
T3 .2
1
C-O C->
as *3 .s
<«0 as
G *c3
•j—»
C/) J G <D ^ §
<d <sj
D ■c
-2 u 03
"a § u.^
3 .S W 3
o C/5
G C/5
Uh
f. J ^ £j
& c/) 2 75 J J hJ
CO co g co
W co .S W G w
c W
CD
E
l_
CD
Q. 00
X
ON
LU
"a1 00
•st-
a 00 Xi
ON
CO CO 03 00
O'
.CD 6
G
-Q -C "O
.CD cj c3
CO
W
 as those obtained by Gass).
than vice versa (i.e. results
S Ji
fcS
3 u «
rt .2 "P •S.s-S a

HD
G c/2
c3 bo bo bo
C /2 .2 t "P #g G
£ TD c *s
Ctf
*S
.2 o O i2 o ;s
r w *i
C 2
4- > ‘5 '« Q-> o
U G
.S G v <L>
s
§ p S
bp * 4—1
73
5- 4 a
"O u *0 G
G
<L>
oa o .a ^ C/}

<D
4->

a;
g<L>

-a
QJ
bo

G
*-t ,—v
TU c/2

2 ^
° ON G
-a -p
< <

w <u
*c/3 * 02 y
c
4J
p
o
G G
<D O
<l> -a u o
•5 -S &
G I 'S •a
OJ
> hJ < ,
c/jg’a ^*2
W P P w D g

00
00
ON
oc
ON
ON
ON
G
o3
<U
G O
2 o
2
 "O 73 CJ on

« QJ C/2
cj
J.| X
S-H

“5 <3 pH <D

3 g^ -7
5S \0 n
C/5
C/2
^ g -a
Q Crt <U O
<12 t-H G
cl)
C/2 C/2 .2c /2
g 2 ■G
C/2
o >
CJ G <U
CT< t> CJ CVG
Cj
— G CJ
cj

■& 13 -G cj G G
A c 12 ;=* cj
G 2
H <lc > rl
J-H
O o o 4-j
> 15
!> 03 /2

§ 0)
i> CJ
JJ >
CJ
S' £
O £
o
hJ ■J

-S'
a
02
G G
*T3 ^3
< <

X <S\
o
G
CJ
-3
VS
X> h n»
CJ -Q
CJ
§• -i -“J2 Cti
>—' y
7? sj
G G
co 2
w &u U

c
o
O
X
a lO 00
02 00 ON
<d O'
■SJ
-Q X
.(0 o OJ
N I
 Formal Instruction and Language Learning 165

Does Formal Instruction Work?

The answer suggested by this review of the research to the question ‘Does
formal instruction work?’ is a tentative ‘yes’. The evidence is of three kinds.
First, instructed learners appear to outperform naturalistic learners. Second,
there is evidence that instruction aids the acquisition of useful formulas.
Third, instruction can result in the acquisition of some new linguistic rules
and can improve control over existing knowledge.

Rate of Acquisition
In general, classroom learners learn more rapidly and progress further than
naturalistic learners. This provides weak evidence in favour of the claim that
instruction affects acquisition. The evidence is weak because we cannot be
sure that it is the focus on the linguistic code that is responsible for the
advantage. Other factors might be the cause. Classroom learners may be more
motivated to learn. It may be classroom input rather than instruction that is
responsible. Classroom learners may obtain more comprehensible input or
they may benefit from access to marked linguistic forms in planned discourse.
There is, however, a line of argument that supports the view that it is the
focus on form that is responsible. Studies of naturalistic acquisition (e.g.
Schmidt, 1983) have shown that learners sometimes do not develop high
levels of linguistic accuracy even though they do become communicatively
effective. Grammatical competence appears to be partially independent of
other components of communicative competence and to develop it successfully
may require more than successful interaction. Schmidt points out that there
are no well-documented studies of adults who have successfully learnt the
grammar of a L2 solely through interaction. If Schmidt is right and some
degree of conscious attention to form is needed for adults to acquire high
levels of linguistic competence, this would suggest that instructed learners do
better than naturalistic learners precisely because they are encouraged to
focus on form.

Instruction and the Acquisition of Formulas

Instruction seems to facilitate the acquisition of formulas. A number of the
studies in the review bear testimony to this. Weinert (1987) and Pienemann
(1984), for example, both draw attention to the routines and patterns that
learners memorize during the course of instruction and then reproduce later
in their free speech. It is possible that much of what commonly passes for
formal instruction (e.g. pattern drills and dialogues) actually results in the
acquisition of unanalysed units. In other words, although the aim is to help
the learners acquire a rule, the product the learner takes away is a useful
formula or two. In this case, the effect of instruction is not what was intended.
166 Instructed Second Language Acguisltion
However, there are also cases when teachers have set out to teach a formula.
Ellis (1984a) documents how one teacher successfully taught a group of ESL
beginners ‘I don’t know’. Instruction based on notional/functional materials
might also be seen as an attempt to teach ready-made phrases — what Hatch
(1983b) calls ‘canned speech’. Formulas can be taught in much the same way
as vocabulary. They require little processing effort and, therefore, can be
learnt easily.13 However, if the efficacy of instruction goes no further than
formulas, it must indeed be considered limited. A strong case for instruction
must show that it is capable of influencing the acquisition of productive rules.

Instruction and the Acquisition of Rules

A number of studies in the review show that instruction has no effect on the
acquisition of linguistic rules. Felix (1981:109) represents the conclusion that
many researchers have reached when he says: ‘the possibility of manipulating
and controlling the students’ behaviour in the classroom is in fact quite
limited.’ Other research (e.g. Lightbown’s) has shown that instruction may
appear to work initially, only for its effects to wear off after time. This is how
Lightbown (1985c: 102) summarizes her findings:

the learners heard and practised certain language items ... In class, and
for a period of time outside of class, they appeared to ‘know’ these forms
in the sense that they used them correctly in appropriate contexts. Later,
however, some of these ‘correct’ forms disappeared from the learners’
language and were replaced by simpler or developmentally ‘earlier’
forms.

These studies have been used to claim that L2 acquisition involves certain
natural processes that cannot be bypassed. As a result learners follow a
sequence of development which instruction is powerless to change.
On the basis of much of the research, therefore, it would be possible to
argue that instruction has no major impact on the acquisition of linguistic
rules. There is also evidence, however, for two alternative positions. First, a
number of studies indicate that instruction can have an immediate effect.
Second, there is some evidence to suggest that instruction can have a delayed
effect.
Instruction can have an immediate effect providing that certain conditions
are met. These conditions concern the nature of the linguistic structure itself
and the timing of the instruction.

Linguistic Conditions
Linguistic structures need to conform to two criteria to be amenable to
instruction:

1 They must be formally simple, i.e. they must not involve any psycho-
linguistically complex processing operation.
 Formal Instruction and Language Learning 167

2 Form—function relationships must be transparent. A truly simple feature

will be one that performs a single function. A complex feature will be
one that is linked to a number of different functions.

Figure 6.2 gives examples of hnguistic features from English that conform to
these criteria. Only features that meet both criteria are teachable (i.e. those
features listed in cell A). Features that meet one criterion but not the other
will be resistent to instruction (i.e. features in cells B, C and D). It is
interesting to speculate that the most difficult structures will be those that are
both formally complex and. that display opaque form—function relationships
(i.e. features in cell D), but there is no clear evidence for this as yet.
Attempts to teach structures which are not leamable can produce a number
of different results. The instruction may be simply ignored with the result that
the learner falls back on developmentally easier rules or substitutes randomly
from her available repertoire. This is likely if the structures require complex
processing operations. It is also possible that premature instruction aimed at
unleamable structures can interfere with acquisition — by encouraging
the learner to avoid what she finds psycholinguistically difficult. Premature
instruction can also result in the learner abandoning a transitional construction
which serves as a necessary stepping-stone to the acquisition of the target
structure because she is made aware of its incorrectness. In this way, instruction
may actually impede acquisition. If the structure can be easily processed, but
the form—function relationships are opaque (as with V-ing) the learner is
likely to acquire the form but to make unique use of it by establishing her own
system of form—function relationships. The overt manifestation of this is
overuse and overgeneralization.

Form-function relationships

Transparent -*-*- Opaque

A 8

plural -s V-ing
Simple

copula ‘be’ articles

O) (/)
c C
CO o
CO P
CD 03
C D
9
cl o
2.

inversion verb infin.

Complex 3rd person -s14 complements

Figure 6.2 The teachability of different linguistic structures

168 Instructed Second Language Acquisition
Instruction directed at structures that require complex processing operations
can prove successful if it is timed to occur when the learner is developmentally
ready to acquire the new feature. The evidence for this comes from the
research into the acquisition of German word-order rules (e.g. Pienemann,
1984). Each word-order rule involves certain processing operations. These
operations are psycholinguistically ordered such that one serves as a prerequisite
for another. Instruction cannot beat the order, but it can help the learner to
progress more rapidly and improve control over each operation, providing that
the teaching proceeds in conformity with the natural sequence of development.
The capacity of instruction to produce an immediate effect on the acquisition
of new linguistic material must be considered very limited, however. Only
some features — very few, in fact — are teachable at any one time and the
effects of instruction on the acquisition of these features may only be temporary.
Otherwise instruction can only work if the learner is at the right stage of
development. One way of enhancing the effects of instruction would be to
make efforts to ensure that the instruction conforms with what is teachable.
Pienemann (1985:163) has put forward arguments in support of what he calls
‘natural grading’. This requires that the instruction proceed in line with what
the learner is able to process. Pienemann’s proposal has met with the obvious
criticism that teachers are generally not in a position to know when individual
learners are ready to acquire the next rule (Long, 1985b; Lightbown, 1985c).
Pienemann, Johnston and Brindley (1988) have responded to this criticism by
developing ‘an acquisition based procedure for second language assess¬
ment’ — an instrument which teachers can use to profile individual learners’
stage of acquisition. However, by their own admission this instrument is not
yet sufficiently accurate. Also, practical problems would remain regarding how
to ensure that learners at disparate levels of development received the right
instruction at the right time. This would call for individualized teaching of a
highly sophisticated kind.

The Delayed Effect of Instruction

Given the limited immediate effect of instruction, the second position —
namely that instruction has a delayed effect - becomes an attractive one. The
empirical evidence in support of it, however, is very limited - as might be
expected, given the difficulty of designing a study to investigate how instruction at
one time might result in subsequent learning. Ellis and Rathbone (1987)
reported that class attendance correlated significantly with measures of sub¬
sequent acquisition but not with current measures. Lightbown (1983) noted
that instruction in V-ing in Grade 5 affected performance on this structure in
Grade 6. Lightbown (1987) reported a similar delay in instructional effect on
the acquisition of the progressive auxiliary.
Some studies have produced results that do not support a delayed hypothesis
- the study by Lightbown, Spada and Wallace (1980) showed that the effects
of instruction were evident immediately after the instruction but wore off
 Formal Instruction and Language Learning 169
later. Harley (1987) found that the gains which an instructed group of Grade
6 immersion learners experienced in the use of French imparfait and passe
compose in comparison to a control group were not evident three months later,
although in this case it was because the uninstructed group eventually caught
up with the instructed group.
There is a strong logical argument to support the delayed-effect position.
We have seen that the review of the research indicates that instruction
accelerates acquisition and that it is associated with higher levels of ultimate
proficiency. We have also seen that instruction does not seem to result in the
immediate acquisition of new features, except in rather exceptional circum¬
stances. The question that arises is how instruction facilitates learning when it
is, in general, powerless to intervene in the natural processes of interlanguage
development. One explanation is that instruction in some way primes the
learner so that acquisition becomes easier when she is finally ready to assimilate
the new material. As Lightbown (1985c: 108) puts it, ‘formal instruction may
provide ‘hooks’, points of access for the learner.’ According to this view, then,
we should see instruction as an ‘acquisition facilitator’ (Seliger, 1979). It
speeds up learning in the long term and helps to prevent the kind of grammatical
fossilization found in adult naturalistic learners.
How does this happen? We can only speculate. One possibility is that
instruction raises the learner’s consciousness about the existence of linguistic
features which she would otherwise ignore. For example, instruction can help
the learner to ‘know’ — in some declarative sense — that the third person
singular verb requires -s. Initially, however, she is unable to process this form.
But because she ‘knows’ about -s, she is better equipped to perceive it in the
input. When the prerequisite processing operations for -s have been developed,
it is acquired. In this way, declarative knowledge serves as a platform for the
acquisition of subsequent procedural knowledge. Conscious awareness of
forms that contribute little to communicative effectiveness may be necessary to
ensure that they are eventually acquired — at least where adults are concerned.
This explanation is also compatible with one set of results that are otherwise
difficult to explain. We saw that ‘projection’ studies such as that carried out by
Zobl (1985) indicate that instruction directed at marked features can trigger
the acquisition of implicated unmarked features. These findings, it was noted,
appeared to contradict other findings that developmental structures cannot be
learnt out of sequence. Now, in the projection studies the instruments used to
measure the effects of the instruction (typically grammaticality-judgement tasks)
have encouraged the use of conscious grammatical knowledge. If this is so,
the results they have produced may not be so incompatible with the results of
other studies which have measured acquisition using instruments that tap
more spontaneous language use. These latter studies provide information
about the effects of instruction on procedural knowledge. We arrive at the
intriguing possibility that whereas instruction is limited in the effect it can
have on procedural knowledge, it can have a considerable effect on the
acquisition of declarative knowledge, as this is not subject to processing
170 Instructed Second Language Acquisition
constraints. The projection studies provide us with an important clue as to
how teaching directed at conscious knowledge should be organized: it should
focus on marked forms.
This conclusion is partly compatible with the results obtained by Schumann
(1978) and Kadia (1988), which showed that instruction did have an immediate
effect on monitored language use, but none on spontaneous use. However,
Schumann’s and Kadia’s learners failed to show any improvement in spon¬
taneous language use later as well. It may be that these learners needed more
time for the benefits of instruction to show or it is possible that certain social
and psychological conditions have to be met for instruction to have the
delayed effect that is being proposed.
One objection to the delayed-effect hypothesis is Pienemann’s finding that
instruction can have a deleterious effect on the learner’s current interlanguage
(see above). This objection arises, however, only if the instruction entails
practice aimed at developing the procedural knowledge needed in production,
as the processing operations which determine the sequence of acquisition of
developmental features relate only to production. The objection disappears if
the goal of the instruction becomes declarative knowledge and if the instructional
activities are directed at consciousness-raising. This argument, then, constitutes
additional support for an instructional approach involving consciousness-raising
rather than drill.15

Summary

To sum up, there are grounds for believing that form-focused instruction
does help the acquisition of linguistic competence. Instruction can work
directly, that is, it can have an immediate effect on the learner’s ability to
perform the target structures in natural communication. However, not all
structures are teachable. Also teachable structures have to be taught at the
right time. It is likely, therefore, that instruction works indirectly in the main;
that is, it has a delayed effect. According to this interpretation, instruction
contributes to declarative rather than procedural knowledge. Declarative
knowledge serves as a facilitator of ultimate procedural knowledge by helping
to make forms salient that would otherwise be ignored by the learner. Conscious
knowledge of marked forms may help to accelerate learning and may also be
necessary to prevent fossilization.

Conclusion

In this chapter we have considered whether and under what conditions

attempts to intervene directly in the process of L2 acquisition by means of
 Formal Instruction and Language Learning 171
form-focused instruction are successful. We have reviewed a number of
studies which have addressed this issue and examined different theoretical
positions. The conclusion reached was that formal instruction does contribute
to L2 acquisition. In some cases, depending on the nature of the target
structure and when the intervention takes place, instruction can have an
immediate effect. In other cases formal instruction may have a delayed effect,
by providing the learner with more or less explicit grammatical concepts,
which will later help her to attend to these features in the input and so acquire
them procedurally. Learners who receive formal instruction outperform those
who do not; that is, they learn more rapidly and they reach higher levels of
ultimate achievement.
The research “which we have studied has been psycholinguistic in nature. It
has been quantitative and, in many instances, experimental. The focus has
been on measuring the learning that results from instruction and, on the basis
of this, inferring to what extent and in what ways the instruction affected L2
acquisition. Such research contrasts with the research which we considered in
chapter 4, which focused on the classroom behaviours which arise as part of
the teaching—learning process, usually without reference to whether any
learning took place. Studies of the effects of formal instruction have been
product- rather than process-oriented.
One of the difficulties of this product-orientation is that ‘formal instruction’
is treated as an undifferentiated phenomenon. As defined in the introduction
to this chapter, it is instruction that (1) focuses on some specific property of
the target language and (2) tries to make the learner aware of what the correct
form or use of the form is. But such a definition raises a number of problems:

1 It is not clear what unit the definition should be applied to (e.g. ‘course’,
‘lesson’ or ‘sequence’). Most of the research has been based on the assumption
that a whole course or a whole lesson can be classified as ‘formal instruction’.
However, as Bialystok (1981:65) has pointed out, ‘a formal learning situation
encompasses many more features than those which are explicitly designated as
the goal of the lesson.’ Any lesson will be made up of both form and meaning-
focused exchanges. It is difficult to imagine a lesson which is entirely form-
focused, let alone a complete course. It should also be noted that instruction
that is primarily meaning-focused can involve form-focused exchanges, as
when the teacher side-tracks to explain the meaning of a lexical item or deal
with a grammatical problem.
2 The definition also ignores the fact that what is form-focused instruction
to the researcher or the teacher may not be so for the learner. Teachers and
learners do not always share the same goals. There is evidence to show that
when teachers operate in a form-focused way, learners sometimes respond
meaningfully (McTear, 1975).
3 A ‘focus’ on form can be achieved in different ways. There are different
kinds of practice drill. The teacher may or may not provide an explanation of
a grammatical point. It is possible that some kinds of form-focused instruction
172 Instructed Second Language Acquisition
work while others do not. If a researcher finds that the learners have failed to
acquire a particular linguistic feature, this may reflect a failure in the kind of
instruction offered rather than in formal instruction per se.
4 Learners may vary individually in the extent to which they are able to
benefit from formal instruction or from different kinds of formal instruction.
The studies reported in this chapter give no or little recognition to the role of
individual differences in learning style among learners.

These problems arise out of the asocial nature of the research. They can only
be overcome by incorporating a process element into research designs in
order to examine how ‘formal instruction, is negotiated by the classroom
participants. Very few studies have done so to date. It is a characteristic of
much of the research that ‘formal instruction’ is either treated as a given or as
something that can be defined without reference to actual classroom events
(e.g. in terms of a syllabus, textbook content or the aim of a lesson plan).
For this reason, it is wise to treat the results which have been obtained with
some caution. It is not possible to ignore them, however. In order to build a
theory of L2 classroom learning it is essential to consider what learning (if
any) results from formal instruction. The study of learning as a product of
instruction is still in its infancy but already it is an important component of
second-language classroom research. Until we have better studies, which are
truly process—product in design, we will have to proceed with the material to
hand.

NOTES

1 The two studies that produced ‘ambiguous results’ showed, in fact, that instruction
did not promote acquisition i.e. that naturalistic exposure was more effective.
However, Long subjected both studies to careful scrutiny and demonstrated that
the results could be interpreted as lending support to instruction.
2 An ‘acquisition-rich’ environment is one where the learner receives plenty of
comprehensible input outside the classroom. An ‘acquisition-poor’ environment is
one where the learner receives little ‘comprehensible input’ in natural settings.
3 Hale and Budar’s (1970) study was one of the two studies that Long claimed
produced ‘ambiguous results’.
4 There are considerable difficulties in interpreting such a correlation. It cannot be
claimed that attendance caused acquisition to take place. One possible interpretation
is that more regular attendance resulted in greater contact with instruction which
in turn enhanced acquisition. Another interpretation is that learners who were
more strongly motivated attended more. It is not possible to decide whether
attendance is functioning as a measure of motivation or of the direct effects of
instruction.
5 Not all researchers were happy to equate ‘order of accuracy’ with ‘order of
acquisition’. Larsen-Freeman, for instance, preferred to talk only about accuracy.
6 Perkins and Larsen-Freeman also collected data from a translation task. The
results obtained from these data did show some effect for instruction, as the
 Formal Instruction and Language Learning 173
accuracy of some of taught morphemes increased considerably and was reflected
in a disturbed order. This study together with that carried out by Schumann
(1978) — see p. 150 — suggests that instruction can have an effect on careful,
planned language production but not on informal, unplanned language use.
7 The distinction between production and acquisition is an important one. However,
it is not clear whether it is possible to investigate anything other than production,
as the principal means of determining what has been acquired is by studying what
a learner can produce in speech (usually) or writing. If all researchers followed
Pica’s lead they would never have anything to say about acquisition. It is more
important, perhaps, to indicate clearly what kind of production has been studied.
Results are likely to differ according to whether the production is unplanned or
planned. Comparisons between instructed and naturalistic L2 acquisition should
be based on equivalent data (i.e. spontaneous speech). Pica’s study does, in fact,
appear to conform to this requirement.
8 Adverb preposing is, therefore, an example of a transitional structure, as the
utterances that result are non-standard in form. The other word-order rules,
however, reflect target-language norms. Adverb preposing is an optional rule,
whereas the other three rules are obligatory.
9 The learners studied by Meisel and Clahsen were migrant workers in Germany.
Obviously, it is difficult to compare rates of acquisition in such disparate groups
as migrant workers and educationally successful classroom learners. Differences
may have as much to do with their social and educational backgrounds as with the
presence or absence of formal instruction.
10 Ellis’s (1984b) finding that learners who participated the least frequently in
practice showed the biggest gains runs contrary to received opinion that ‘practice
makes perfect’. It suggests that formal instruction consisting of formal practice
may not be the best way to promote L2 acquisition. It does not show that
instruction per se has no effect; instruction that involves consciousness-raising
without practice may still work.
11 The claim that input is impoverished is central to the arguments put forward by
researchers who work within the Universal Grammar paradigm. It should be
noted, however, that researchers operating in different paradigms dispute the
claim.
12 Of course, it is not necessary to invoke a theory of Universal Grammar to explain
Henry’s results. Learners may simply be generalizing existing ‘rules’. They have
discovered that possessed nouns go before the noun (unlike in English), so they
extend this rule to cover relative clauses.
13 It is also possible that teaching formulas (intentionally or unintentionally) can
inhibit the acquisition of linguistic rules — as Weinert (1987) suggests.
14 Not all researchers would agree that 3rd person -s involves complex processing
operations. However, Pienemann argues that it is an example of non-local mor¬
phemes, as the learner has to transfer information from the grammatical subject
to the verb. This would account for its late acquisition. So would other expla¬
nations, however. 3rd person -s is not perceptually salient and has little communi¬
cative value.
15 It should be noted that Pienemann (1985) himself makes it clear that the
teachability hypothesis applies only to L2 production. He does not object to
learners being exposed to a rule beyond their immediate processing capacity in
the input. One might expect, therefore, that he would not object to consciousness-
raising of the same rule.
7 An Integrated Theory
of Instructed Second
Language Learning

Introduction

The aim of this chapter is to develop a theory of instructed second language

acquisition. The intention is not to develop a radically new account of how
language learning is influenced by teaching but rather to provide an economical
explanation which is consistent with the theoretical positions and research
discussed in previous chapters.
All theories are purposeful — that is, they are constructed with a particular
goal in mind. The theory presented here has two primary purposes. The first
is to provide a set of statements or hypotheses about classroom L2 learning
which are testable and, therefore, falsifiable. In this sense, the theory aims to
be scientific, as defined by Popper (1976). The second purpose is to provide
an account of classroom L2 learning that is relevant and accessible to teachers. In
this sense, the theory aims to be appropriate to the needs of a particular set
of users.1
There are plenty of theories of L2 acquisition already in existence.
McLaughlin (1987) provides a detailed examination of five of the most promi¬
nent. Ellis (1985a) considers seven. These theories seek to account for the
interaction of those sets of factors which affect the process and success of L2
learning. In other words, they describe the conditions under which efficient
L2 learning can take place. Spolsky (1988) suggests that a theory should take
account of four sets of factors: (1) the knowledge and skills which the learner
possesses at any given moment in the learning process, (2) various components of
ability, (3) various affective factors such as personality and motivation and (4)
the opportunities for learning. A theory of instructed language learning must
ultimately account for the same set of factors. The difference between such a
theory and a theory of naturalistic L2 learning will rest not on its basic
components but rather on the contents of each component and the way they
interact.
A theory of classroom L2 learning should be clear about what it is seeking
to explain. In particular, it should specify whether it intends to account for
competence or for proficiency. Taylor (1988) argues that there has been
considerable terminological confusion in the use of ‘competence’. He draws
attention to the fact that Chomsky in his initial and more recent uses of the
 Integrated Theory of Instructed Learning 175
term explicitly distinguishes the idea of ‘knowledge’, which is static, from the
idea ol ‘ability to use knowledge’, which is dynamic. Taylor suggests that the
term competence be restricted to the former and proficiency be used to
label the latter.2 This is an important and useful distinction as it contributes to
conceptual clarity and helps us to decide precisely what a theory of classroom
L2 learning which seeks to be relevant to teachers should address. Teachers
are concerned with both how knowledge is acquired and, crucially, with the
learner’s ability to make use of this knowledge. Any theory that seeks to be
relevant to the needs of teachers, therefore, must address both competence
and proficiency. Arguably teachers have little to gain from a theory of com¬
petence by itself (a point that Chomsky has himself acknowledged).3 They
need a theory that helps them to understand and plan for how new knowledge
is formed on the one hand and how learners learn how to use this knowledge
correctly and appropriately on the other. The theory presented in the following
pages is an attempt to build an integrated theory of this kind.
This chapter will begin with an account of cognitive learning theory.
Currently, considerable attention is being paid to cognitive accounts of how
learning takes place and the extent to which they can be applied to L2
learning (McLaughlin, 1987; O’Malley, Chamot and Walker, 1987; Bialystok,
1988). A cognitive theory views language learning as a complex skill which,
like other such skills, involves the use of various information-processing
techniques to overcome limitations in mental capacity which inhibit perform¬
ance. Learning takes place when the learner is able to carry out operations
automatically as a result of practice. In chapter 2 we saw the dangers of
extrapolating from a general theory of learning; language learning cannot be
assumed to be a ‘skill’ of the same kind as playing a piano or playing tennis.
However, cognitive theory, as formulated in the work of Shiffrin and Schneider
(1977) and Anderson (1980; 1983; 1985) provides a much more power¬
ful account of many of the phenomena of instructed L2 learning than
did structuralist-behaviourist theories.
Cognitive learning theory is a powerful explanation of how learners develop
the ability to use their L2 knowledge (i.e. proficiency) and, therefore, is
particularly relevant to the aim of this chapter, but it is unable to provide an
adequate explanation of how this knowledge is acquired in the first place.
This is because it fails to recognize that L2 acquisition is determined, in part,
by linguistic as well as cognitive factors. However, the theory serves as a
platform for developing an integrated theory of instructed L2 learning, which
attempts to account for both proficiency and competence.

Cognitive Theory and Instructed Language Learning

Cognitive theory is the result of extensive research into the role that mental
processing plays in learning. It has benefited in particular from research into
176 Instructed Second Language Acquisition
information processing i.e. how information is stored and retrieved. Cognitive
theory seeks to explain three principal aspects of learning: (1) how knowledge
is initially represented, (2) how the ability to use this knowledge develops and
(3) how new knowledge is integrated into the learner’s existing cognitive
system.

The Representation of New Knowledge

New knowledge is acquired in two stages (Weinstein and Mayer, 1986). First,
the learner selects which features in the environment to pay attention to and
transfers this information into short-term memory. Second, the learner acquires
some or all of these features by transferring the information into long-term
memory for permanent storage. New information can be comprehended by
means of inferencing, i.e. using existing knowledge and the context in which
the new information occurs. However, the learner will tend to ignore new
information that does not accord with her existing knowledge-system. Ausabel
(1971) refers to this as obliterative subsumption.4 It can be overcome if the
learner receives corrective feedback which enables her to compare her own
performance with that of some external model and if she notices the difference.

Developing the Ability to Use Knowledge

Information which has been transferred into long-term memory is not initially
available for use in all performance conditions. Two accounts of how learners
achieve control over new information have attracted the attention of L2
researchers. The first concerns the distinction between controlled and auto¬
matic processing and the second that between declarative and procedural
knowledge.
A number of researchers (e.g. Posner and Snyder, 1975; Schneider and
Shiffrin, 1977) distinguish the mental operations that a person is able to
perform effortlessly and automatically from those that can only be per¬
formed with considerable difficulty and relatively slowly. Learning involves the
automatization of memory nodes, so that information which was initially only
available for use through controlled processing can be handled spontaneously.
Automatization is necessary because the learner has limited processing capability
and needs to free capacity to deal with new information and to form higher-
level plans. For example, a learner who is struggling with the effort of
pronouncing words correctly will have tittle time to pay attention to grammar.
Once pronunciation becomes automatic, however, she is free to attend to
higher-level skills.
This processing model of learning also accounts for why learners vary in
the way they perform different tasks. Tasks vary in the amount of attention
they require and also in where attention needs to be focused. Thus, a tennis
player may be able to perform a copy-book backhand if allowed to focus
 Integrated Theory of Instructed Learning 177
exclusively on how to execute this particular stroke, but may fail to achieve
backhand returns during the course of an actual game. A practice-session and
a competitive match constitute different kinds of tasks.
The distinction between declarative and procedural knowledge comes from
the work of J. Anderson (1980; 1983; 1985). Declarative knowledge involves
‘knowing that’. It consists of such information as the definitions of words,
facts, rules and memory for images and sequences of events. Procedural
knowledge is ‘knowing how’. It is represented in memory in terms of‘production
systems’ consisting of a condition and an action as in this example from
O’Malley, Chamot and Walker (1987): ‘IF the goal is to generate a plural
of a noun and the noun ends in a hard consonant, THEN generate the
noun + /s/.’ 4
Anderson identifies three stages in the learning process:

1 In the cognitive stage the learner makes use of conscious activity. The
knowledge acquired is typically declarative in nature and can often be
described verbally by the learner.
2 In the associative stage, errors in the original declarative knowledge
are detected and corrected and the knowledge is also proceduralized.
During this stage condition—action pairs which are initially represented
in declarative form are gradually converted into production sets. The
initial declarative representation is never lost, however.
3 Finally, in the autonomous stage performance becomes more or
less totally automatic and errors disappear. The learner relies less on
working memory and performance takes place below the threshold of
consciousness.
These two accounts of how new knowledge becomes automatic have much
in common. It is clear that there is a loose correspondence between controlled
processing and declarative knowledge and between automatic processing and
procedural knowledge. Both models are based on the notion that human
beings have limited information-processing abilities and, therefore, need to
activate certain kinds of knowledge rapidly and easily.
There is, however, one important difference, which can be encapsulated in
the distinction between ‘attention’ and ‘consciousness’. It is possible for a
learner to pay attention without doing so consciously. Controlled processing —
in Shiffrin and Schneider’s terms — can take place with or without any
conscious awareness on the part of the learner. Also the learner may or
may not be able to articulate what she has done. In contrast, declarative
knowledge — in Anderson’s terms — involves the learner’s conscious attention
and can be explicitly formulated. Only when this knowledge is subsequendy
proceduralized does consciousness disappear. This distinction is of considerable
importance when it comes to applying the theory to language learning, as it
concerns the central question of whether L2 knowledge starts out as conscious
and explicit or whether it is subconscious and implicit.
178 Instructed Second Language Acguisition

Integrating Knowledge
The third aspect of learning involves the way in which the acquisition of new
information leads to a restructuring of the learner’s knowledge system. As
learning takes place, the existing system is modified in order to take account
of the new information. Karmiloff-Smith (1986) (cited in Mclaughlin, 1987)
describes how the process of restructuring takes place. In the first place adults
and children master components of a new task but do not impose any overall
organization on the information they obtain. Later, however, organization is
created as learners attempt to simplify, unify and gain control over the internal
representation of their knowledge. Whereas in the first stage learners are
unduly influenced by input data and are responsive to corrective feedback, in
the second stage they are more strongly influenced by their own mental
schemas and under-utilize external feedback. There is a third stage, however,
when the learner is able to balance environmental and mentalistic influences
without threatening the stability of the knowledge system. It should be noted
that according to this view of learning the manner in which knowledge is
represented varies at different stages of development.

Cognitive Theory and L2 Acquisition

There have been a number of attempts to apply cognitive learning theory to
L2 acquisition. These have been based on the assumption that L2 learning is
no different in kind from any other kind of complex skill learning. O’Malley,
Chamot and Walker (1987:288), for instance, comment: ‘The theory suggests
that language-related codes and structures are stored and retrieved from
memory much like other information, and that language acquisition follows
the same principles of learning as do other complex cognitive skills.’J. Anderson
(1983:398) writes: ‘little direct evidence exists to support the view that language
is a unique system.’ Language is, therefore, deprived of the unique status
attributed to it by generative linguists in the Chomskyan tradition and by those
researchers who seek to explain L2 learning with reference to the principles of
Universal Grammar (cf. Cook, 1985; 1988). The extent to which cognitive
theory can account for L2 learning depends crucially on the extent to which
language constitutes a general skill or a specific faculty of the human mind.

Strategic Competence

One aspect of cognitive learning theory which has received considerable

attention in L2 acquisition research is procedural knowledge/ Faerch and
Kasper (1983) identify a number of sub-components of procedural knowledge.
Reception procedures involve such strategies as bottom-up and top-down
processing and inferencing. Production procedures consist of the various
 Integrated Theory of Instructed Learning 179
strategies which learners use both to plan and to monitor their output.
Conversational procedures concern the devices needed to communicate ap¬
propriately and to construct coherent discourse. Communication strategies are
used to solve problems which the learner experiences in the planning and
execution of speech as a result of inadequate resources. It is by these means
that the learner is able to automatize L2 knowledge for both comprehension
and production. Knowledge can only become ‘procedural’ when it can be
manipulated in the various ways described by Faerch and Kasper. Learners
need to develop strategic competence (Canale and Swain, 1980).

Bialystok’s Mbdel

The most fully worked out cognitive theory of L2 learning is that developed
by Bialystok (Bialystok, 1979; 1981; 1982; 1983; 1988; Bialystok and Sharwood-
Smith, 1985). The theory has subtly changed over the years (with concomitant
changes in terminology),6 but the central premises have remained intact,
reflecting the general distinctions discussed above. The following account of
Bialystok’s theory is based on her later publications.
Like other cognitive psychologists who have addressed L2 learning, Bialystok
explicitly affirms the principle that language is processed by the human mind
in the same way as other kinds of information (Bialystok, 1988:32). Language
proficiency is described with reference to two dimensions: an analysed factor
and an automatic factor. The analysed factor concerns the extent to which the
language learner is aware of the structure of her linguistic knowledge. In
unanalysed knowledge, which is characteristic of the early stages of L2 learning,
the learner is not aware of the structure and organization of knowledge. As
learning takes place awareness increases, enabling the learner to identify the
formal structure. Awareness takes the form of ‘a propositional mental rep¬
resentation’ of linguistic knowledge which may or may not be conscious to the
learner. Bialystok is emphatic that the degree of analyticity is not linked to
consciousness and is not explicitly represented in the mind of the learner: ‘it is
erroneous to equate analyzed knowledge with articulated knowledge, or knowl¬
edge of rules’ (Bialystok, 1988:40).
However, analysed knowledge does make ‘articulated knowledge’ and
metalingual knowledge possible. Its real significance ties in the fact that it can
be operated on by the learner and so is available for language uses of the kind
required in formal education (e.g. academic essay writing). Learners who only
have access to unanalysed L2 knowledge will be restricted to the kinds of
language use for which this is appropriate (e.g. everyday conversation). The
process by which knowledge gradually becomes analysed during the course of
L2 acquisition corresponds to the general process of cognitive restructuring
described above.
The automatic factor concerns the relative access which the learner has to
180 Instructed Second Language Acquisition
Automatic
fluent speakers highly skilled and
literate users

native speakers specialized uses of

in ordinary language, e.g.
conversation rhetorical

Non- __ -► Analysed
analysed

L2 learners at L2 formal learners

early stages

children learning
LI

Non-automatic

Figure 7.1 Bialystok’s two dimensions of language proficiency

Source: Bialystok (1988:37)

knowledge. Knowledge gains in automaticity as learning takes place. It is

reflected in easier access and more fluent performance. Automaticity is achieved
through practice.
These two factors are independent of each other. Thus development on the
analysed dimension occurs independently of development on the automatic
dimension. Learners begin with unmarked knowledge, i.e. knowledge that is
non-automatic and unanalysed. The simultaneous development of both dimen¬
sions is difficult, so learners are likely to concentrate on one factor or the
other. Some learners will choose to go for automaticity, others will concentrate
on analysing their knowledge by breaking down formulaic expressions acquired
as unanalysed chunks. The learning context will also influence what kind of
acquisition takes place. Informal L2 learners are more likely to emphasize
automaticity, formal learners will stress analyticity. Figure 7.1 illustrates the
different kinds of learners and language uses predicted by the model.

Cognitive Theory and Instructed L2 Learning

There has been no attempt to develop a specific cognitive theory of instructed

L2 learning. However, both general cognitive theory and Bialystok’s tri¬
dimensional model have been applied to the classroom. Key issues which have
been addressed are the relationships between (1) practice and automaticity,
(2) classroom interaction and procedural knowledge and (3) formal instruction
and analysed knowledge.
One view holds that practice is the key to the development of automaticity.
 Integrated Theory of Instructed Learning 181
McLaughlin (1987) adopts this position, claiming that considerable training is
necessary before automatic processing can take place. Errors occur because
the learner has not yet automatized performance and does not have the time
to make use of controlled processing. Errors can also occur as a result of
obliterative subsumption. Practice can help to activate mental nodes, strengthen
connections, increase speed of access and thereby reduce errors. Corrective
feedback — an essential component of practice — can help to overcome the
effects of obliterative subsumption by helping the learner to preserve the
essential attributes of the new linguistic forms (Seliger, 1977).
Johnson (1988) suggests that a number of criteria have to be met before
correction can be effective in eliminating errors (or ‘mistakes’, as he prefers to
call them):7 *

1 The learner must have a desire or need to eradicate the mistake.

2 She must be able to form an internal representation of what the correct
behaviour looks like.
3 She must realize that her performance is flawed.
4 She must have the opportunity to perform in real conditions.

The latter point is particularly significant, as it constitutes the principal

difference between a cognitive and a behaviourist view of practice. Whereas
audiolingual learning theory emphasized the controlled practice of specific
forms in specially contrived exercises, cognitive learning theory stresses the
need for practice in real operating conditions. Only in this way will the learner
succeed in achieving automatic control of her knowledge for use in authentic
communication.
Cognitive theory also emphasizes the importance of ensuring that the
classroom provides adequate opportunities for the proceduralization of knowl¬
edge. We saw in chapter 4, however, that many classrooms fail in this respect.
Classroom process research has shown that interaction is qualitatively different
in classroom and natural discourse and that this is reflected in the learning
and communication strategies learners employ. It can be argued that in many
cases language classes do not really prepare students for participation in
ordinary face-to-face interaction because they do not provide the conditions
needed to ensure that full automatization of L2 knowledge takes place.
According to cognitive theory, formal instruction can help to increase the
learner’s analysed knowledge. Focusing on specific linguistic forms and en¬
couraging the learner to manipulate these with varying degrees of awareness
facilitates restructuring and also builds propositional representations of L2
knowledge. This enables the learner to participate in the kinds of decontex-
tualized language use which require analysed knowledge. The aim, then, is
awareness-raising rather than automaticity via proceduralization. This is not
so very different from the aim of cognitive code learning theory (see chapter
3), one of the main assumptions of which was that perception and awareness
of second language rules are separate from their actual use. One of the
182 Instructed Second Language Acquisition
legacies of audiolingualism is that the two functions of instruction — the
development of automatic and analysed knowledge — have been conflated.

Problems with Cognitive Accounts of Classroom L2

Learning
Cognitive learning theory offers a powerful account of how L2 learning takes
place. Nevertheless, it is inadequate in two important respects where instructed
L2 learning is concerned. First, the theory cannot easily account for the
presence of acquisitional sequences. Second, it is rather vague regarding the
role of explicit knowledge.
The existence of acquisitional sequences that are immune to instruction is
well attested (see chapter 6). The fact that instruction appears powerless to
affect the order in which developmental structures are acquired cannot be
accounted for by cognitive learning theory. Neither automatization nor re¬
structuring can explain why a word-order rule such as verb-end emerges at a
particular stage of development in the interlanguage of L2 German learners.
These processes are also inadequate to account for why instruction in a
marked linguistic feature such as relative pronoun as object of preposition can
trigger the acquisition of associated unmarked features such as subject and
object relative pronouns (Gass, 1979; Eckman et al., 1988). According to
Bialystok’s theory learners begin with knowledge that is non-automatic and
non-analysed. But this theory provides no means of specifying which non¬
automatic and non-analysed items of knowledge are developed first and which
later. McLaughlin (1987) attempts to protect the theory against this criticism
(which he fully acknowledges) by referring to Sajavaara’s (1978) proposal that
predictable acquisitional sequences reflect routines that are fully automatic at
the very moment they emerge and, therefore, are not subject to the process of
automatization that governs other knowledge. Such a suggestion is not very
convincing, however, as it is not clear why this happens to some knowledge
forms and not others and, more crucially, because the available empirical
evidence suggests that developmental structures such as German word-order
rules are not fully automatized at the point of emergence but, in fact, take
time to become so. The basic assumption of cognitive learning theory — that
language learning and use are not different in kind from other kinds of skill
learning and use — is not fully justified.
This criticism does not constitute a total rejection of cognitive learning
theory, however. It demonstrates the need for a comprehensive theory that
incorporates both a cognitive explanation of how learners develop the ability to
use their knowledge in different kinds of tasks (i.e. a proficiency theory) and
a linguistic explanation of how learners acquire a knowledge of L2 rules (i.e. a
competence theory).
Bialystok and Sharwood-Smith (1985) offer the outlines of such a theory.
They distinguish two components of L2 acquisition: knowledge and control.
Knowledge has both a quantitative and qualitative dimension. The former
 Integrated Theory of Instructed Learning 183
concerns how much the learner knows. The latter concerns the way in which
the language system has been organized and represented in the learner’s
mind, which changes over time. Control refers to the processing system
which is used during actual performance. An analogy with a library is used to
make the distinction clear. The number of books corresponds to ‘quantitative
knowledge’ and the system for arranging them on the shelves to ‘qualitative
knowledge’. The procedures involved in retrieving a given book from the
library are analogous to ‘control’. Bialystok and Sharwood-Smith emphasize
that control involves both knowing the retrieval procedures and being able to
operate them with speed and efficiency. A cognitive theory of language learning,
then, will inform principally about control procedures and their application
(i.e. procedural knowledge or what Bialystok calls the automatic factor). It also
explains how learners develop the ability to use their knowledge in different
kinds of language use (i.e. by means of analysing knowledge). A linguistic
theory — which is accommodated in the model — is needed to explain how
knowledge of L2 properties enters interlanguage. Cognitive and linguistic
theories, as currently formulated, are not as complementary as Bialystok and
Sharwood-Smith’s model suggests, however. In particular, it is not clear what
qualitative aspects of an L2 grammar are governed by linguistic or cognitive
factors.
The second major problem of a cognitive account of L2 acquisition
concerns the role of explicit knowledge. Explicit knowledge refers to knowledge
which is consciously represented in the mind of the learner and which can be
articulated on request. Any theory of instructed language learning, however,
will need to provide a clear account of the relationship between explicit and
implicit knowledge. Anderson does offer an account, but this is unsatisfactory,
because it does not conform with the known facts of L2 acquisition. He
claims that using declarative knowledge makes it automatic so that it can be
used below the level of consciousness. Learning begins with conscious attention
on the part of the learner, resulting in declarative knowledge. O’Malley,
Chamot and Walker (1987) argue that this is true in both naturalistic and
classroom learning. In the case of the former conscious attention will be
focused on the functional uses of language, while in the latter attention is
focused on formal aspects. In naturalistic learning learners will be conscious
of routines and patterns, while in classroom learning they will conscious of
‘rules’. The fact that learners — particularly adults — are conscious of some
aspects of what and how they are learning is not in dispute, neither is the fact
that there are differences both among individual learners and in acquisitional
settings regarding what aspects of the L2 are initially dealt with consciously.
The problem ties in conflating automaticity with consciousness. Anderson is,
in fact, claiming that implicit knowledge of linguistic forms is dependent upon
prior explicit knowledge, but the available evidence suggests otherwise and
supports Krashen’s claim that learners (including adults) can acquire an L2
subconsciously. However, it does not follow from this that explicit knowledge
has no role to play in the development of this grammar.
184 Instructed Second Language Acquisition

Summary
This section has examined in some detail various cognitive learning theories
in order to determine to what extent they can account for instructed L2
learning. The following conclusions have been reached:

1 In developing a theory of L2 learning it is necessary to distinguish

‘knowledge’ and ‘control’ and to take account of both linguistic and
cognitive factors.
2 Cognitive learning theory provides an adequate account of how learner’s
obtain control over L2 knowledge; free practice enables controlled pro¬
cessing to become automatic. Learners need appropriate interactional
opportunities to activate strategies for proceduralizing knowledge.
3 Cognitive theory can also provide some insights into how learners
restructure their L2 knowledge in order to make it available for use in
an increasing range of tasks, including those requiring decontextualized
language use.
4 A cognitive theory is unable to explain the sequence of acquisition of
grammatical structures.
5 A cognitive theory is unable to account for the part played by explicit
knowledge and cannot, therefore, fully explain how formal instruction
contributes to L2 acquisition.

A comprehensive theory of instructed L2 learning will need give recognition

to the fact that some aspects of L2 acquisition (those that are associated with
the ability to use L2 knowledge) are governed by general cognitive factors,
while other aspects (those associated with how an L2 grammar is constructed)
are governed by linguistic factors.

An Integrated Theory

The theory described in this section seeks to explain how instructional input
(in its various forms) provides the learner with the data needed to construct
her interlanguage. It recognizes that different aspects of L2 learning require
different kinds of explanation and that neither a purely linguistic nor a purely
cognitive framework will provide a complete explanation.

L2 Knowledge is Differentiated
An initial premise is that L2 knowledge is differentiated. Explicit and implicit
knowledge are held to be different in kind and to be stored separately in the
brain. Explicit knowledge is conscious and declarative. Implicit knowledge is
subconscious and procedural, although not necessarily fully automatic. Neither
 Integrated Theory of Instructed Learning 185
kind of knowledge is developmentally primary — that is, the learner may
acquire a particular ‘rule’ explicidy in the first instance and then proceed to
acquire the same rule implicidy at a later point, or vice versa. Only im¬
plicit knowledge is developmental, however; it is manifested as acquisitional
sequences.
The claim that implicit and explicit knowledge constitute discrete types of
knowledge runs contrary to the views of other theorists (e.g. Faerch, Haastrup
and Phillipson, 1984). According to the models proposed by these theorists
knowledge is continuous rather than dichotomous. Explicit and implicit forms
represent the two poles of a continuum. Explicit knowledge is defined as
metalingual knowledge and implicit knowledge as knowledge that can be used
without any reference to a formal rule. Intermediate levels of knowledge are
reflected in the learner’s ability to decide that speech is or is not in accordance
with an L2 rule or her ability to describe a rule in her own (but not in
metalingual) terms. Faerch et al’s position, however, conflates the two processes
of consciousness-raising and automatization. They state (1984:203), for
instance:

The process of developing more consciousness about implicit knowledge

is part of the general process of consciousness-raising. A process in the
opposite direction — the learner gradually developing an ability to use a
certain rule for productive and receptive purposes without being aware
of this — is usually referred to as an automatization process.

Such a model does not allow for the possibility that both explicit and implicit
knowledge can be accessed with varying degrees of automaticity.
The claim that explicit and implicit knowledge are distinct is in accord with
Krashen’s Monitor Model; ‘acquired’ knowledge is implicit and ‘learnt’ knowl¬
edge is explicit. Krashen (1985) advances a number of arguments in favour of
this position. Some learners - like ‘P’ (Krashen and Pon, 1975) - demonstrate
explicit knowledge of some very simple rules without being able to perform
these in everyday communication. Conversely, they demonstrate implicit
knowledge of some very complex rules of which they have no conscious
knowledge. A study by Hulstijn and Hulstijn (1984) showed that learners with
no explicit knowledge of two Dutch word-order rules nevertheless demonstrated
competence in the use of the rules. Seliger (1979) asked twenty-nine mono¬
lingual children, eleven bilingual children and fifteen adult ESL learners to
state the rule for using ‘a’ and ‘an’ and then elicited samples of speech from
the same subjects, allowing them time to focus on form if they wanted to. He
found no relationship between the subjects’ performance and conscious knowl¬
edge of the rule even though many subjects believed that the rule had guided
their speech. Kadia (1988) reports that her subject (an adult Chinese learner
of English in Canada) had ‘a rich metalinguistic knowledge’ but that the
relationship between her ability to state the rules for transitive phrasal verbs
and ditransitive verbs and her spontaneous production of these rules was very
186 Instructed Second Language Acguisitlon
poor. There was a better relationship where written test performance was
concerned. These studies lend support to the claim that the two types of
knowledge are separate.
It is, however, not easy to decide between these two positions. As McLaughlin
(1978b) has pointed out, it is very difficult to tell what kind of knowledge a
learner is using in any single performance. We cannot distinguish very easily
editing by ‘feel’ (which involves implicit knowledge) from editing with the
‘monitor’ (which involves explicit knowledge).8 Whether the two knowledge
types are perceived as dichotomous or continuous depends to a large extent
on whether the theorist posits any interface between them (i.e. whether
explicit knowledge can be converted into implicit and vice versa). The theory
presented here, like Krashen’s and unlike that of Faerch et al. adopts a non¬
interface position.
It is important to recognize (contrary to the position taken up by Faerch et
al.) that explicit knowledge does not require metalingual expertise. It can be
articulated with varying degrees of precision and varying degrees of technicality.
It can also vary in the extent to which it is descriptively adequate, i.e. accounts
for implicit knowledge. A linguist’s description will be more adequate than a
teacher’s, which in turn is likely to be nearer the mark than the representation
formed by the learner (Seliger, 1979). Explicit knowledge, then, can be
differentiated according to how it is articulated and how elaborate it is
(Sharwood-Smith, 1981).
Implicit knowledge can also be differentiated. It can consist of the rules by
which utterances are constructed or it can consist of ready-made chunks of
language (cf. Ellis, 1985a:167—75). Pawley and Syder (1983) argue that
an important component of the native speaker’s linguistic knowledge is the
hundreds of thousands of ‘lexicalized sentence stems’ which every speaker
has stored in her memory. These are defined as ‘a unit of clause length or
longer whose grammatical form and lexical content is wholly or largely fixed’
(1983:190). Examples are such expressions as ‘Need any help?’ and ‘I see
what you mean.’ Pawley and Syder suggest that there is a continuum between
fully productive rules and lexicalized sentence stems. Because speakers are
able to retrieve the latter as wholes they economize on processing effort and
thereby gain in fluency. A number of researchers have noted the existence of
formulaic expressions in the speech of L2 learners (e.g. Fillmore, 1976;
Krashen and Scarcella, 1978; Ellis, 1984). This distinction between stereo¬
typical expressions and productive rules also seems to correspond in part to
Bialystok’s notion of unanalysed and analysed knowledge.
Productive implicit rules can also be differentiated. Research based on
typlogical universal or Universal Grammar distinguishes marked and unmarked
linguistic knowledge. For example, the Accessibility Hierarchy (see chapter 6)
posits that a relative pronoun functioning as subject as in:

The footballer who scored the goal has been mobbed.

 Integrated Theory of Instructed Learning 187
is less marked than a relative pronoun functioning as object of preposition
as in:

The house that we moved into was built in the 1920s.

Similarly, a verb which takes two noun phrases as in:

They offered the hospital a donation.

can be considered more marked than one that takes a noun phrase followed
by a preposition phrase as in:
4

They explained the problem to me.

(Mazurkewich, 1985). Functional models of language such as that developed

by Givon (1979) also differentiate linguistic knowledge, in this case according
to whether it is typical of unplanned/informal or planned/formal language
use. The discourse hypothesis discussed in chapter 5 is based on this distinction.
The recognition that knowledge is differentiated is central for understanding
the role played by instructional input in L2 learning, as different kinds of
input result in different kinds of knowledge and, more significantly, different
kinds of input are needed to achieve acquisition of different kinds of knowledge.

Control
As in the framework proposed by Bialystok and Sharwood-Smith, control is
considered to be separate from knowledge. Control applies to both explicit
and implicit knowledge. Thus the learner is able to access both conscious and
subconscious linguistic rules with varying degrees of automaticity. Newly
acquired knowledge is typically accessed only by means of controlled processing.
With opportunities for free practice, however, it can be processed more
rapidly and efficiently. The development of control is reflected in both increased
accuracy in the performance of specific linguistic features and in increased
fluency (e.g. fewer pauses and longer runs between pauses).

Instructional Input
A basic distinction has been drawn between meaning and form-focused
instruction. In the case of the former the learner is engaged in communication
where the primary effort involves the exchange of meaning and where there is
no conscious effort to achieve grammatical correctness. In the case of the
latter the learner is engaged in activities that have been specially designed to
teach specific grammatical features. The input that derives from these two
kinds of instruction differs with regard to its communicative properties (e.g.
188 Instructed Second Language Acguisition
meaning-focused instruction is likely to afford the learner an opportunity to
listen to and to perform a greater range of language functions than will form-
focused instruction) and also with regard to the kind of response it typically
evokes in the learner (e.g. form-focused instruction encourages the learner to
reflect on the formal features of the language while meaning-focused instruction
encourages semantic processing).
As we saw in chapter 6, the instructional input in many lessons will be
mixed, affording the learner the opportunity to attend to both meaning and
form — to experience language or to study it. Teachers shift the focus as the
lesson unravels — at one moment engaging the learners in meaningful com¬
munication and at another directing their attention to the linguistic code. It
may also be possible, however, to distinguish whole lessons and even whole
courses of instruction according to whether the emphasis is primarily on
meaning or form (Spada, 1987). A form-focused lesson will always afford the
learner input that goes far beyond the specific property which is the instructional
target. Ultimately, of course, it is the learner and not the textbook or the
teacher that determines in what way the input is attended to. An activity
intended by the teacher to focus the learner’s attention on form may be
treated by her as an opportunity for meaningful communication, or vice versa.
However, given that the raison d'etre of the classroom is to learn the language,
the learner may be predisposed to attend to form. The evidence obtained
from many classrooms (even so-called communicative ones) is that the inter¬
actions that take place are far removed from those found in natural settings
(see chapter 4).

Learning Style
Learning style — the learner’s affective and cognitive orientation to
learning — also plays an important part in instructed language learning,
although there has not been space in this book to deal with this aspect. Not all
learners are receptive to instructional input. Only learners with a positive
affective orientation to learning the L2 are likely to benefit from the instruction.
Tolerance of ambiguity and motivation govern the extent to which individual
learners are open to input and also how frequently they participate in classroom
activities, either actively or as ‘silent speakers’ (Reiss, 1983).
Even if a learner is receptive to input, she may not be cognitively suited to
the type of instructional input provided. An experiential learner, whose natural
learning style predisposes her to focus on communication, may be held back if
she is faced with an input that is predominantly form-focused. Likewise, a
studial learner who is required to participate extensively in meaning-focused
instruction may be inhibited. In addition, learners who find that their learning
style is incompatible with the instructional style may develop anxiety and so
lose confidence and motivation, perhaps even to the point of abandoning
instruction.
 Integrated Theory of Instructed Learning 189

Instruction and the Acquisition of Explicit Knowledge

The theory claims that different kinds of instruction typically result in different
kinds of L2 knowledge. In order to examine how this happens we will
consider first the relationship between instruction and explicit knowledge and
then go on to consider how instruction influences the acquisition of implicit
knowledge. It needs to be stressed straight away, however, that there is no
simple correlation between form-focused instruction and explicit knowledge
or between meaning-focused input and implicit knowledge.
Classroom learners typically make use of their cognitive problem-solving
skills in order to derive conscious rules about the form of the L2 grammar.
They will do this whether the method of instruction invites them to or not.
Arguably, however, form-focused instruction facilitates the acquisition of ex¬
plicit knowledge, although very little is known about how effective instruction
is in teaching explicit knowledge. Krashen (1982) argues that ‘learning’ is
limited to linguistic forms that are simple in structure and transparent in
function (see chapter 3). Seliger (1979) argues that the rules learners construct
are typically anomolous — blurred reflections of the teacher’s pedagogic rules,
which in turn are imprecise versions of the linguist’s rules. Seliger describes a
number of characteristics of both conscious learner-formed rules and pedagogic
rules. They go beyond the information given, they do not provide accurate
descriptions of the speaker’s internal knowledge and they do not correspond
to actual language behaviour. The picture painted by Krashen and Seliger is
overly pessimistic, however. Diary studies (e.g. Ellis and Rathbone, 1987)
reveal that learners are able to acquire conscious representations of quite
complex rules and that they actively seek to do so. Doubtless many of these
representations lack precision and accuracy, at least in their original formu¬
lation, but they can be worked on, gradually refined and extended to account
for new features which the learner observes in the input or new information
obtained subsequently (cf. Kadia, 1988).
In chapter 6 the possibility was floated that linguistic theory explains how
explicit rather than implicit knowledge is acquired. One of the hypotheses
based on this theory that is receiving considerable attention at the moment is
that unmarked properties of the L2 emerge in interlanguage before marked
ones. Current models of generative grammar (e.g. Chomsky, 1981) provide a
rich source of information from which to derive operational hypotheses regard¬
ing the role of markedness. One of the advantages of this approach in the
study of L2 acquisition is that ‘very specific claims’ can be advanced and
tested (Rutherford, 1987). Another hypothesis that is attracting attention is
that instruction directed at marked linguistic features serves to trigger acquisition
of associated unmarked features, i.e. learners are able ta project their explicit
knowledge beyond that which they have actually been taught. This is clearly
speculative at the moment, but should it prove to be correct, it would indicate
that there is a preferred order of presentation of explicit knowledge.
190 Instructed Second Language Acquisition

Instruction and the Acquisition of Implicit Knowledge

Instruction contributes to the acquisition of implicit knowledge in a number
of ways. Both meaning-focused and form-focused instruction can facilitate
learning.

The Role of Meaning-focused Instruction

Meaning-focused instruction supplies the learner with input for processing.

One well-documented way in which this occurs is through ‘scaffolding’ (cf.
chapter 5). Learners build structures vertically with the support provided by
their own and others previous utterances. They then learn the structures they
build in this way and subsequently produce them horizontally. This route to
linguistic knowledge is of limited value, however, except perhaps for beginners.
Of greater importance is access to comprehensible input. Despite reservations
regarding the role which this plays (see chapter 5), it is hard not to conclude
that comprehensible input functions as the primary source of data in L2
acquisition. One way in which this occurs is by supplying the learner with
ready-made chunks — lexicalized sentence stems — which can be memorized
as wholes and used to perform communicative functions that are important to
the learner. In this case, there may be a fairly direct relationship between
input and acquisition, with the learner’s innate language-learning mechanisms
not involved. Research has shown that even in a classroom context certain
chunks or formulas occur with considerable frequency (Ellis, 1984a). Input
frequency together with communicative need probably act as the determinants
of acquisition of this kind of implicit knowledge.
Input frequency is probably much less important where productive implicit
rules are concerned. Two conditions govern whether input becomes intake:

1 the learner must attend subconsciously to the presence of a specific

feature in the input (i.e. notice the feature without reflecting on it);
2 the learner must have reached a stage of development which makes it
possible to incorporate this feature into her interlanguage system.

Both conditions are necessary for acquisition to take place. Neither attention
nor readiness is sufficient by itself. Cognitive mechanisms govern which
features learners attend to. For example, there is ample evidence in both LI
and L2 acquisition to show that elements that occur at the beginnings and
ends of sentences are more salient than elements that occur sentence-medially
(Slobin, 1973; Clahsen, 1984). Linguistic factors govern whether the feature
attended to is learnable. For example, universal principles of grammar (see
above) may guide the process of acquisition or there may be a hierarchy of
 Integrated Theory of Instructed Learning 191
linguistic processing operations (see p. 153) which determine when a specific
feature is acquirable.
In emphasizing the role played by linguistic universals or linguistic processing
operations, however, it is not necessary to discount the contribution of compre¬
hensible input (chapter 5). From the point of view of a theory of instructed L2
learning, comprehensible input is of special importance because it enables us
to address the question about how ‘intervention’ should take place. We can
speculate that neither input which is entirely unproblematic nor input which is
so problematic that it is entirely incomprehensible is likely to encourage the
learner to attend to specific linguistic features. Long’s (1983a) claim that
negotiated input is best for learning may not have been empirically demonstrated
but it is supported by the common-sense assumption that attention to ‘new’
features will be encouraged if a degree of problematicity occurs. It is worth
noting that it is the problematicity of such exchanges not their successful
resolution which is the crucial factor in determining whether or not the
learner attends to the new features.
Meaning-focused instruction also provides opportunities for learner output
and this, too, contributes to the acquisition of implicit knowledge. The output
hypothesis (chapter 5) states that learners need to be ‘pushed’ into producing
output that taxes their existing linguistic resources in order to develop a full
grammatical competence. The discourse hypothesis claims that learners need
the opportunity to participate in a variety of discourse modes (ranging from
the planned to the unplanned). Linguistic knowledge is tied to the particular
type of discourse in which it is learnt. Access to planned/formal discourse
may be particularly important for acquiring implicit knowledge of the marked
properties of the L2.

The Role of Form-focused Instruction

Implicit knowledge can also be acquired via form-focused instruction, although

there are a number of constraints governing the teachability of linguistic
structures (chapter 6). Successful instructional intervention in interlanguage
development occurs in the case of:

1: Lexicalized sentence stems Routines and patterns can be memorized

as wholes. Some instructional methods (e.g. the notional/functional approach)
are particularly well suited to ensuring that the learner is equipped with a
supply of ready-made phrases. Beginner learners, in particular, need access to
such phrases before they can begin interacting with native speakers.

2: Developmental structures that are ‘learnable’ These are structures

which are acquired in a fixed acquisitional sequence as a result of the
processing operations they require of the learner (chapter 6). Direct instruction
192 Instructed Second Language Acguisition
will only result in the acquisition of a structure that the learner is ‘ready’ for,
i.e. the next one in the sequence from that which she has already acquired.
Direct instruction aimed at structures which are more than one step ahead of
the learner’s current interlanguage will founder and may even interfere with
the acquisitional process.

3 Variational Structures Not all linguistic structures are subject to devel¬

opmental processing. Some — like copula ‘be’ — are variational in the sense
that they can be acquired at different stages of development and in the sense
that they are acquired by some learners but not by others, in accordance with
their sociopsychological attitudes.
The role of form-focused instruction is restricted where implicit knowledge
is concerned. In practical terms, only instruction directed at formulaic phrases
is likely to prove profitable. The principal source of implicit knowledge is
meaning-focused instruction.

Instruction and the Acquisition of Control

A widely held view is that learners gain control over their L2 knowledge
through practice. For example, Sharwood-Smith (1981:166) writes:

it is quite clear and uncontroversial to say that most spontaneous perform¬

ance is attained by dint of practice. In the course of actually performing
in the target language, the learner gains the necessary control over its
subjects such that he or she can use them quickly without reflection.

Sharwood-Smith’s subsequent comments make it clear that by practice he

means ‘controlled practice’.9 That is, he believes that by consciously rehearsing
structures, the learner proceduralizes them. Sharwood-Smith cites no empirical
evidence to support this position, nor in fact is there much. The available
research suggests that controlled practice may result in spontaneous productive
use only of those linguistic forms which the learner is ‘ready’ to handle (Ellis,
1988b). In the case of certain structures (variational features, for example)
practice may lead to increased accuracy, but perhaps only temporarily (see
chapter 6). Thus controlled practice appears to have little long-term effect on
the accuracy with which new structures are performed. It also has little effect
on fluency as measured by temporal variables. Ellis and Rathbone (1987), for
instance, found no relationship between the amount of controlled practice ab
initio learners of German took part in and speech rate.
How then do learners achieve control? The answer lies in meaning-focused
instruction. This provides the conditions the learner needs to activate those
procedures that are responsible for both automatizing knowledge and for
compensating for lack of it. In order to develop control the learner needs to
practise in ‘real operating conditions’. Only in this way will the learner
develop the strategic abilities needed to perform her competence accurately
and fluently.
 Integrated Theory of Instructed Learning 193

The Role of Explicit Knowledge in L2 Acquisition

The theory maintains that explicit knowledge does not turn into implicit
knowledge (see above). Given that implicit knowledge is the primary goal of
most language teaching, this raises the question ‘What is the role of explicit
knowledge?’ Krashen claims that its role is extremely limited. In the integrated
theory proposed here, however, explicit knowledge is of major importance.
The way explicit knowledge is hypothesized to work has already been
described in chapter 6. Briefly, it is claimed that conscious knowledge functions
as an acquisition facilitor, enabling the learner to ‘notice’ L2 features in
meaning-focused input which would otherwise be ignored. The features
themselves, however, are acquired from this input in accordance with the
learner’s internal cognitive and linguistic processing faculties. Figure 7.2
provides a schematic representation of this process.
A diary study by Schmidt and Frota (1986) lends support to this view. They
claim that the learner they studied (called ‘R’) benefited from formal instruction
because he obtained information about the structure of the target language
(Portuguese) which could have been derived only with great difficulty from
interaction alone. However, R did not learn everything he was taught in the
sense that he was able to use it spontaneously in subsequent interaction.
Rather, ‘R heard and used what he was taught if he subsequently heard it and
noticed it’ (1986:279). The diary study kept by R shows that he tended to
notice specific features in the input immediately after they had been taught
and that he felt that conscious awareness of what was present in the input was
a definite aid to learning. Schmidt and Frota argue that R was able to notice
the gap between his own output and the input as a result of the explicit
knowledge he had acquired.
The ‘notice the gap’ hypothesis helps to explain a number of aspects of
instructed language learning. For example, it provides a way of accounting for
how corrective feedback might aid learning. Corrective feedback juxtaposes
the interlanguage form with the target-language form, putting the learner in
an ideal position to notice the difference. Such feedback serves as a way of
raising the learner’s consciousness about the lack of correlation between the
target language and her own interlanguage.10 It should be noted, however,
that the corrective feedback which Schmidt and Frota refer to occurs in the

explicit
1 knowledge
1
1
1
1
1
meaning- learner notices
focused and attends to linguistic implicit
input linguistic processing knowledge
features

Figure 7.2 The role of explicit knowledge in the acquisition of implicit knowledge
194 Instructed Second Language Acquisition
overall context of natural (and, therefore, meaning-focused) interaction and
not in the context of specially contrived language practice. The hypothesis also
suggests why instructed learners do better than naturalistic learners — they
know more ‘facts’ about the language. Schmidt and Frota speculate that those
learners whose explicit knowledge is the most developed will learn the most.
The theory, therefore, allocates a significant role to explicit knowledge,
while maintaining that there is no direct interface between the two types of
knowledge. Two general points need to be emphasized. First, it follows from
the theory that formal instruction directed at explicit knowledge can facilitate
the process of acquiring implicit knowledge. The learner is not just a ‘language-
acquisition device’ but can take conscious decisions, which will aid learning.
Second, the effects of explicit knowledge will be indirect and delayed. One
reason for this is that the explicit knowledge may be ‘fuzzy’ (see discussion on
p. 189). The main reason is that noticing and attending to developmental
linguistic features is no guarantee of their acquisition. The learner also has to
be ‘ready’ to assimilate them into her interlanguage.

Output
Output can be generated in a number of ways. The learner can rely entirely
on implicit knowledge. In doing so, the learner is likely to make extensive use
of formulaic expressions and vertical constructions in the early stages so as to
offset the processing difficulty that arises from trying to use non-automatic
implicit knowledge productively. The learner can also ‘borrow’ from her LI
(Corder, 1978a) by translating using whatever implicit and explicit L2 knowledge
comes to hand. Learners can also edit their output using either kind of
knowledge (i.e. editing by ‘feel’ vs. editing by ‘monitor’). Which means of
generating output a learner chooses will depend, in part, on whether she is
oriented towards correctness or towards fluency. If the former, she is more
likely to resort to explicit knowledge to compensate for the lack of implicit
knowledge. It will also depend on the nature of the communication task, in
particular on whether there is a felt need to focus on form (Hulstijn and
Hulstijn, 1984; Hulstijn, 1988).
The learner’s output also contributes to acquisition. One way is by being
pushed to produce sentences which stretch the limits of her competence (the
output hypothesis). It also contributes in another way. The learner’s output is
part of the totality of input available for processing. Thus utterances generated
in the ways described above feed into the learner’s own acquisition device.
This means that ‘utterances initiated by means of explicit knowledge can
provide feedback into implicit knowledge’ (Sharwood-Smith, 1981:166), so
providing a further way in which explicit knowledge can facilitate acquisition.
As Schmidt and Frota observe, however, such ‘auto-input’ is not always
beneficial, as, often enough, learner output is erroneous. Thus learners may
learn their own, incorrect solution to a problem. This may account for why
some errors are so persistent and why learners often feel that that their errors
 Integrated Theory of Instructed Learning 195
are not really errors. However, learners also learn their correct productions.
Thus a linguistic form that has just entered the learner’s interlanguage can be
strengthened by means of her regular attempts to use it.

Summary
The integrated theory is summarized in diagrammatic form in Figure 7.3.
The principal hypotheses of the model are:

1 Instructional input is ‘filtered’ by the learner in accordance with her

learning style. The learner will be predisposed to attend to linguistic
features in the input if she responds positively (cognitively and affectively)
to the type of instruction.
2 Implicit and explicit knowledge constitute different kinds of knowledge
and are stored separately.
3 Instructional intervention can take the form of meaning-focused and
form-focused activities. Learners can respond to classroom input as
‘interaction’ or as ‘formal instruction’.
4 Explicit knowledge is derived largely from form-focused instruction.
5 Implicit knowledge is derived largely from meaning-focused instruction.
For acquisition to take place the learner must attend to specific linguis¬
tic features in the input and be ready to incorporate these into her
interlanguage.
6 Implicit knowledge can also be taught directly if (a) the learner is devel-
opmentally ‘ready’ or (b) the target forms are not subject to developmental
constraints.
7 Explicit knowledge serves to sensitize the learner to the existence of
non-standard forms in her interlanguage and thus facilitates the acqui¬
sition of target-language forms in accordance with (5) above.
8 Control over linguistic knowledge is achieved by means of performing
under real operating conditions in meaning-focused language activities.

Form-
focussed
input

Meaning-
focused
input

Figure 7.3 A model of instructed L2 acquisition

196 Instructed Second Language Acguisition
9 The learner’s own output serves as input to the language-processing
mechanisms. Deviant output that is subject to corrective feedback
raises consciousness and facilitates acquisition of the correct forms.

Conclusion

This chapter began with an outline of cognitive learning theory, as this has
been applied to language teaching. This theory is able to explain some of the
known facts about the relationship between instruction and learning. In par¬
ticular, it provides a convincing account of how learners obtain control over
their L2 knowledge. However, cognitive learning theory is unable to explain
the sequence of L2 acquisition nor does it convincingly elucidate the relationship
between explicit and implicit knowledge and the role of instruction in this.
The integrated theory of instructed L2 acquisition builds on cognitive
learning theory but also tries to incorporate those aspects of language learning
that depend on linguistic and psycholinguistic factors and which, it is claimed,
are responsible for the way in which linguistic knowledge enters the learner’s
interlanguage. Descriptions of some of the linguistic factors are to be found in
current accounts of Universal Grammar (e.g. Cook, 1988). Some of the
psycholinguistic factors are dealt with in discussions of how developmental
features are acquired (e.g. Pienemann, 1986).
The theory proposed in this chapter rests, in part, on the hypothesis that
explicit and implicit knowledge are distinct types of knowledge which are
stored separately. This hypothesis is controversial; it is probably true to say
that theorists are most divided on this issue. The main argument for claiming
separation is that the available evidence indicates that instruction is often
powerless to convert explicit into implicit knowledge. In other words, the
relationship between the two types of knowledge is not one of automaticity.
Although the integrated theory adopts a non-interface position, it differs
from other such theories (such as the Monitor Model) in allocating a major
role to explicit knowledge. The theory hypothesizes that explicit knowledge
functions as a facilitator of implicit knowledge by making the learner conscious of
linguistic features in the input which otherwise might be ignored. Explicit
knowledge helps the learner to notice marked forms.
The integrated theory is able to resolve the central paradox of instructed
language learning. Instruction frequently fails to result in the direct acqui¬
sition of new linguistic structures, yet instruction results in faster learning and
higher levels of achievement. In the main, it is the learner who is in charge of
both what can be learnt and when it can be learnt, not the teacher.11 But the
teacher has a definite role to play both by ensuring that there are adequate
opportunities for meaning-focused communication to foster the acquisition
of implicit knowledge and also by helping the learner to develop explicit
knowledge.
 Integrated Theory of Instructed Learning 197

NOTES

1 The two goals — building a theory that is both scientific and appropriate — are
in potential conflict. A scientific theory which formulates statements with the
precision needed to ensure testability may not be accessible to many teachers.
Equally, a theory which is appealing to teachers (e.g. the Monitor Model) may not
be very scientific in Popper’s sense. The dangers of trying to be both scientific
and appropriate are not always recognized.
2 The distinction between ‘knowledge’ and ‘ability’ should not be confused with the
distinction between ‘linguistic’ and ‘sociolinguistic/pragmatic’ aspects of language.
The idea of ‘knowledge’ and ‘ability to use’ applies equally to both aspects. It
follows, therefore, that ‘proficiency’ is concerned with both the ability to use
linguistic knowledge (correctness) and the ability to use pragmatic rules (appropri¬
ateness). The focus of this book is on the linguistic rather than pragmatic aspects
of language.
3 It is worth noting that whereas linguists have emphasized competence as knowledge,
educationalists have emphasized competence as behaviour (cf. Taylor, 1988:160).
4 Overgeneralization errors can be seen as examples of obliterative subsumption.
For example, the learner who says:

* My father made me to go to bed.

may do so because she has internalized a rule for infinitival complements which
states that the infinitive occurs with ‘to’, as in:

My father asked me to go to bed.

and has ignored the fact that the verb ‘make’ does not conform to this rule in
sentences in the active voice.
5 The term ‘procedural knowledge’ is used to refer to two different aspects of
learning: (1) to knowledge which has been proceduralized and is, therefore,
automatic; (2) to the strategies or procedures that are involved in the processing
of knowledge for actual use. The two are interrelated, as knowledge becomes
proceduralized as a result of the learner mastering the procedures required to
process it.
6 In early accounts of the theory Bialystok referred to explicit and implicit knowledge.
She abandoned these terms because she did not wish to suggest that consciousness
was a factor and also because she wished to distinguish the automaticity and
analysability of L2 knowledge. Her initial distinction tended to conflate these.
7 Errors are the result of a lack of competence; mistakes are the result of the
learner’s failure to perform her competence (cf. ch. 2). Arguably, mistake-correction
has a better chance of succeeding because the learner already knows the linguistic
structure. Error-correction can only work if new knowledge is created.
8 McLaughlin argues that the distinction between controlled and automatic process¬
ing is more tenable because it can be empirically tested by means of controlling
the amount of time learners have to perform a particular task. In contrast, the
explicit/implicit knowledge is not amenable to empirical study and, therefore, is
not falsifiable. One response to this might be to argue that if a theory aims to
provide a ‘reasonable interpretation’ rather than a ‘scientific explanation’ this
198 Instructed Second Language Acgulsltlon
objection is irrelevant. However, it is possible to conceive of methods of testing
the separation of implicit/explicit knowledge by using introspective research tech¬
niques which invite learners to comment on what kinds of knowledge they used in
performing different tasks.
9 The term ‘practice’ is used with widely different meanings in language teaching.
On the one hand it can refer to the kind of controlled practice associated with
audiolingual pattern drills. On the other it can refer to the regular use of the L2
in real communication. Such usage of the term does not contribute to clear
thinking and it might be better to restrict it to ‘controlled practice’.
10 In contrast to the positive contribution of corrective feedback supplied by the
learner’s interlocutors (i.e. ‘other-correction’), the learner’s attempts to correct
her own output appear to be of little value for acquisition. Schmidt and Frota
(1986) claim that R’s self-corrections had no discemable effect on his interlanguage
development.
11 Corder (1980:11) concludes a perceptive review of what L2 acquisition research
has to say about the teaching of grammar by arguing that ‘teaching should be
coordinated with learning and not learning with teaching’. However, this does not
mean that no grammar teaching should take place. It means that the learner must
be left to make use of grammar teaching in her own way. This is precisely the
position adopted in this chapter.
8 Final Comments

Introduction

In the preceding chapters we have explored different ways of investigating

how a second language (L2) is learnt through instruction. On the basis of the
insights obtained from this survey I have proposed a theory of instructed
language learning. It is now time to stand back and to identify what the main
questions are for which answers are still needed and what kinds of research
might help us to find these answers. First, though, I want to try to pinpoint
some of the problems inherent in much of the research undertaken to date.

Problems of L2 Classroom Research

The L2 classroom researcher seeks to show how instructional events cause or

impede the acquisition of a second language. In order to achieve this, it is
necessary to (1) identify which instructional events are significant, (2) find
valid and reliable measures of the L2 learning that takes place and (3) be able
to demonstrate that the relationship between instructional events and learning
is in some way causal. Researchers have faced a number of problems in
carrying out each of these tasks.
There are different ways of identifying ‘significant’ instructional events.
One way is to base selection on some explicit theory. This is the approach
recommended by Long (1988b). He argues that classroom researchers should
choose ‘psycholinguistically relevant design features’ as a basis for enquiry.
The distinction between form-focused and meaning-focused learning activi¬
ties — which Long proposes — is derived from Krashen’s Monitor Model.
Another way is to examine the methodological recommendations of language
educationalists. Teacher training handbooks (e.g. Gower and Walters, 1983;
Harmer, 1983) represent a rich source of information regarding those instruc¬
tional events which experienced professionals in the field consider significant.
Research which looks at the role of practice (e.g. Seliger, 1977) or of small-
group work is a good example of this approach.
200 Instructed Second Language Acquisition
Both approaches are problematic, however. Constructs such as ‘form vs.
meaning-focused instruction’ and ‘practice’ are notoriously difficult to oper¬
ationalize in terms of actual social or cognitive behaviour in the classroom.
The same problem faces researchers who have tried to investigate the ‘method’
construct (see chapter 1), though on a much larger scale. The problem is that
a priori definitions of classroom events often turn out to be imprecise, incom¬
plete and inaccurate. Allwright (1988b) argues: ‘any a priori characterization
of learning and teaching environments is bound to be suspect for fundamental
research purposes.’ Allwright believes that the characterization of instructional
events should ‘emerge from research’ rather than be imposed upon it externally.
The difficulty of identifying potentially significant instructional events, then,
confronts the researcher with a dilemma. Should she opt for theory-led
research or should she concentrate on collecting data in order to arrive at
accurate accounts of what actually happens in the classroom? Or is it possible
to operate deductively (as Long wants), while making certain that the instruc¬
tional events studied really do take place (as Allwright wants)?
The problems associated with the measurement of learning are no less
tractable. There is no direct window through which the researcher can peer to
discover how the learner is learning or what she has learnt. How, then, can
the researcher discover what the effects of the learner’s participation in
instruction are? One solution, adopted by some researchers, is to try to tap the
learners’ intuitions about the L2. This is usually undertaken by means of
some kind of grammaticality-judgement task (i.e. a task that requires the
learner to state whether individual sentences are well-formed or not). Another
solution is to collect samples of naturally occurring language use and analyse
these to see if there is any evidence to show that the learner is able to employ
target features correctly. Clearly the concept of ‘learning’ that underlies these
two approaches to data collection is very different. Not surprisingly, the
results obtained diverge considerably.
These two approaches to the measurement of learning in L2 acquisition
research are supported by different theories of language and of language
learning. The measurement of intuitions is undertaken by researchers who
work within a theory based on the competence/performance distinction. Their
aim is to establish what the learner ‘knows’ rather than what she ‘does’. The
analysis of actual language use is undertaken by researchers who base their
study of L2 acquisition on cognitive theories of language processing. As long
as there are competing theories of language and language learning, there will
be differences in opinion regarding how learning should be measured.
The question remains, though, as to what constitutes valid data for the
study of classroom L2 acquisition. In order to answer this question, it is surely
important to consider the purpose of the research. In the case of classroom
research, the aim is to investigate how learners develop the ability to use the
L2 accurately, appropriately and fluently, as this is the goal of instruction. It
is, therefore, surely more important to pay attention to research that addresses
what learners can do in real-life situations than what they know but may be
 Final Comments 201

unable actually to use. In terms of the distinction developed in chapter 7,

classroom research is directed at understanding how instruction contributes to
proficiency rather than to competence. The measures of learning that are
chosen should reflect this aim. Thus we ought to place greater confidence in
measures of learning derived from natural language behaviour than in those
obtained from artificial grammaticality-judgement tasks.
The most serious problem, however, is how to investigate the relationship
between instructional events and learning outcomes. How can the researcher
show that learners acquire new linguistic knowledge (or greater control over
existing knowledge) as a result of taking part in specific instructional events?
One possible way is to aim at explanation-through-description — that is, to
build up a detailed understanding of how instructional events are enacted in
the classroom and, by so doing, illuminate the factors that appear to influence
whether learning takes place or not. This is the approach adopted by many
classroom process researchers (see chapter 4). The problem, however, is that
usually no evidence is collected to demonstrate whether any learning actually
takes place. A second way is to try to establish a relationship between the
occurrence of specific types of classroom behaviour (such as participation by
the learner or the teacher’s use of referential questions) and measures of
learning. The problem here is that it is only possible to demonstrate that
instruction and learning correlate; it is not possible to show that instruction
causes learning. A third way is to carry out some kind of carefully designed
experiment in order to examine whether a specific kind of ‘treatment’ results
in predicted learning outcomes. Experiments do allow the researcher to make
firmer statements about causation, but they have their own problems. For one
thing, experiments can often result in an artificial teaching—learning environ¬
ment. However, the experimental approach has led to some interesting results
in the case of form-focused instruction (see chapter 6). The main problem lies
in trying to set up experiments to show whether and to what extent different
kinds of meaning-focused instruction facilitate learning. As we have seen,
there is an absence of such studies.
It is important to recognize these problems. The research which this book
has reported must be treated circumspectively because of them. We should
not expect the research to produce clear-cut answers, if only because it begs
the kinds of questions outlined above. We would do better to see the research -
even the experimental studies - as contributing to understanding, rather than
providing explanations.

Questions for Future Research

The questions which future research will need to address will be very similar
to those which have informed research to date. They key questions are likely
to be:
202 Instructed Second Language Acquisition
1 What kinds of instructional events are ‘significant’ for the study of L2
acquisition? How can these events be reliably identified?
2 To what extent is the distinction between form-focused and meaning-
focused instruction a real one? Is it possible to identify specific classroom
behaviours that are characteristic of each type of instruction?
3 How do learners acquire implicit knowledge of the L2 as a result of
comprehending meaning-focused input in the classroom?
4 To what extent and what ways can form-focused instruction contribute
to the acquisition of implicit knowledge of the L2?
5 What forms does explicit knowledge of the L2 take?
6 How can instruction contribute most effectively to the development of
explicit knowledge?
7 To what extent and in what ways does explicit knowledge facilitate the
acquisition of implicit knowledge?
8 How can learners increase their control over existing L2 knowledge
(i.e. become more fluent) in a classroom setting?
9 What individual learner factors influence the way a learner responds to
instruction?

These questions derive from the theory of instructed second language acqui¬
sition presented in chapter 7. Individual studies will need to focus on more
narrowly formulated questions, based on these general issues. For example,
research based on question 5 might try to investigate the nature of the
learners’ metalingual competence in specific domains of linguistic knowledge
(e,g. verb tenses; word-order rules; overt discourse markers).

The Case for Hybrid Research

What kind of research is needed to seek answers to these questions? It is
unlikely that any one type of research will succeed. Progress in understanding
classroom language acquisition is most likely if researchers are prepared to
adopt a variety of qualitative and quantitative methods. There is a continued
need for ethnographic research directed at identifying significant classroom
processes and also for experimental research to try to establish causative links
between instructional variables and learning. There is also a case for the use
of introspective methods to investigate how individual learners respond to
instructional events. Above all, there is a need for hybrid research, i.e.
research that combines the procedures of both hypothesis-forming and
hypothesis-testing methods.
There has, in fact, been relatively little mixed research. Researchers have
tended to fall into camps and to argue the case for their preferred approach.
The old controversy between those who favour an experimental approach and
those who favour ethnographic enquiry or the use of introspective methods is
very much alive in L2 classroom research. But it is a sterile and unnecessary
debate.
 Final Comments 203
Let us look at an example of how hybrid research might help us to probe a
difficult research question. We have seen that research which has sought to
investigate how learners acquire new linguistic knowledge by comprehending
input has made little progress. There are a number of reasons for this:

1 Qualitative research (e.g. discourse analysis) has been unable to demon¬

strate a causative relationship between input and acquisition.
2 Experimental studies have gone some way towards demonstrating how
specific kinds of input promote comprehension, but have been unable
to show that comprehension promotes acquisition.
3 It is difficult to manipulate the input in order to investigate how specific
linguistic features affect acquisition without drawing the learner’s con¬
scious attention to the features, i.e. encouraging the learner to process
the L2 as form rather than for meaning.
4 It is not clear how we can measure the acquisition that results from
exposure to meaning-focused input.

Ellis (1989c) outlines a research proposal that might begin to overcome these
problems. He suggests the use of partially artificial reading texts, which have
had non-standard rules inserted into them. To encourage the students to
attend to meaning rather than to form they are asked to do a faster reading
exercise. To investigate what the subjects actually do as they read, a number
of them are interrupted and asked to state what they have just been thinking
about (cf. Cavalcanti, 1987). Or selected subjects are asked to read the text
aloud and a careful study is made of how they ‘read’ the doctored sentences in
the text. In this way, qualitative information about how learners process the
text, including the artificial elements, can be undertaken. In addition, compre¬
hension can be measured in order to obtain some summative measure of how
much has been understood. Finally, without forewarning, the subjects can be
asked to state in what ways the grammar of the artificial text differs from that
of the standard language or, alternatively, can be asked to translate standard
sentences into the artificial language, in order to measure what information
about the ‘new’ linguistic features they have been able to internalize.
Hybrid research, such as this proposed study, incorporates both experimental
and qualitative procedures. It seeks to establish cause and effect, but also to
uncover the processes involved in language use and language learning. It aims
to provide explanations, while at the same time increasing understanding.
The arguments in favour of hybrid research are well rehearsed. In chapter
1 we saw that the comparative method studies foundered because of the
absence of any qualitative information regarding what actually took place in
the classroom in the name of the different methods. In chapter 4 we saw that
Long (1980) stressed the importance of a process element in classroom
experiments in order to provide detailed information about the instructional
‘treatment’ under investigation. There is also research that shows that different
learners respond differently to the same instructional events (e.g. Slimani,
204 Instructed Second Language Acguisition
1987) and which demonstrates the importance of eliciting information from
learners about how they learn (e.g. Faerch and Kasper, 1987). But the lesson
has not really been learnt. There are few L2 classroom studies that combine
experimental design with ethnographic and/or introspective techniques de¬
signed to shed light on the underlying processes involved in both instruction
and learning.

Conclusion

It is easy to be pessimistic about the results achieved to date. Chaudron

(1988:180) concludes his own review of L2 classroom research as follows:

Despite the obvious increase in the amount of classroom-oriented research

in recent years, few of the suggestions offered here can be made with
great confidence, for the existing research is difficult to synthesize.

It is not only a problem of synthesis. There are, as we have seen, problems

with the research itself. One way forward is better research. Chaudron is
confident that, as we learn how to do better research, we will be able to make
more confident statements about how instruction contributes to acquisition.
It might be wise, though, not to place too much faith in better research. For
a start, old problems tend to be ignored. This might not matter too much,
however, if we do not look to research to provide answers and instead accept
that it is only likely to provide insights or clues about what happens when
teachers try to intervene in the process of language learning. We will always
need to interpret the clues with the help of common sense based on our
practical experience of what works and does not work in the classroom. In this
way we can avoid becoming consumers of research and, instead, build our
own theories of how learning takes place through instruction.
Bibliography

Adams, S. 1982, Scripts and the recognition of unfamiliar vocabulary: Enhancing

second language reading skills. Modem Language Journal, 66:155—9.
Adjemian, C. 1976, On the nature of interlanguage systems, Language Learning,
26:297-320.
Allen, P., Frohlich, M. and Spada, N. 1984, The communicative orientation of
language teaching: An observation scheme. In Handscombe, J., Orem, R. and
Taylor, B. (eds), On TESOL ’83: The Question of Control, Washington, DC,
TESOL.
Allwright, R. 1975, Problems in the study of the language teacher’s treatment of
learner error. In Burt, M. and Dulay, H. (eds), On TESOL 75, Washington, DC,
TESOL.
Allwright, R. 1980, Topics, turns and tasks: Patterns of participation in language
learning. In Larsen-Freeman (1980).
Allwright, R. 1983, Classroom-centred research on language teaching and learning: a
brief historical overview. TESOL Quarterly, 17:191—204.
Allwright, R. 1984a, The importance of interaction in classroom language learning,
Applied Linguistics, 5:156—71.
Allwright, R. 1984b, Why don’t learners learn what teachers teach? — The interaction
hypothesis. In Singleton, D. and Little, D. (eds) Language Learning in Formal and
Informal Contexts, Dublin, IRAAL.
Allwright, R. 1988a, Observation in the Language Classroom, London, Longman.
Allwright, R. 1988b, The characterization of teaching and learning environments:
Problems and perspectives. Paper given at Conference on Empirical Research
into Foreign Language Teaching Methods, Bellagio, Italy.
Anderson, J. 1980, Cognitive Psychology and its Implications, San Francisco, Freeman.
Anderson, J. 1983, The Architecture of Cognition, Cambridge, Mass.; Flarvard University
Press.
Anderson, J. 1985, Cognitive Psychology and its Implications (2nd ed.), New York,
Freeman.
Andersen, R. (ed.) 1981, New Dimensions in Second Language Acquisition Research,
Rowley; Mass., Newbury House.
Andersen, R. (ed.) 1983, Pidginization and Creolization in Second Language Acquisition,
Rowley, Mass., Newbury House.
d’Anglejan, A. 1978, Language learning in and out of classrooms. In Richards (1978).
Anthony, M. 1963, Approach, method and technique, English Language Teaching,
17:63-7.
206 Instructed Second Language Acquisition
Asher, J. 1977, Learning Another Language through Actions: The Complete Teacher's
Guidebook, Los Gatos, Calif., Sky Oaks Productions.
Aston, G. 1986, Trouble-shooting in interaction with learners: The more the merrier,
Applied Linguistics, 7, 128—43.
Ausabel, D. 1971, Some psychological aspects of the structure of knowledge. In
Johnson, P. (ed.), Learning: Theory and Practice, New York, Thomas Y. Crowell.
Bailey, K. 1983, Competitiveness and anxiety in adult second language learning:
Looking at and through the diary studies. In Seliger and Long (1983).
Bailey, N., Madden, C. and Krashen, S. 1974, Is there a ‘natural sequence’ in adult
second language learning?, Language Learning, 24:235—44.
Barasch, R. and James, V. (forthcoming), Answers to Krashen, Rowley, Mass., Newbury
House.
Barnes, D. 1969, Language in the secondary classroom. In Barnes, D. (ed.) Language,
the Learner and the School, Harmondsworth, Penguin.
Barnes, D. 1976, Prom Communication to Curriculum, Harmondsworth, Penguin.
Bellack, A., Kliebard, H., Hyman, R. and Smith, F. 1966, The Language of the
Classroom, New York, Teachers College Press.
Belyayev, B. 1963, Thy Psychology of Teaching Foreign Languages, Oxford, Pergamon
Press.
Bialystok, E. 1979, Explicit and implicit judgements of L2 grammaticality, Language
Learning, 29:81 — 104.
Bialystok, E. 1981, Some evidence for the intergrity and interaction of two knowledge
sources. In Andersen, R. (1981).
Bialystok, E. 1982, On the relationship between knowing and using linguistic forms,
Applied Linguistics, 3:181—206.
Bialystok, E. 1983, Inferencing: Testing the ‘hypothesis-testing’ hypothesis. In Seliger
and Long (1983).
Bialystok, E. 1988, Psycholinguistic dimensions of second language proficiency. In
Rutherford, W. and Sharwood-Smith, M. (eds), Grammar and Second Language
Teaching, Rowley, Mass., Newbury House.
Bialystock, E. and Sharwood-Smith, M. 1985, Interlanguage is not a state of mind: .An
evaluation of the construct for second-language acquisition, Applied Linguistics, 6,
101-17.
Bley-Vroman, R. 1983, The comparative fallacy in interlanguage studies: The case of
systematicity, Language Learning, 33:1 — 17.
Bley-Vroman, R. 1988, The fundamental character of foreign language learning. In
Rutherford, W. and Sharwood-Smith, M. (eds), Grammar and Second Language
Teaching, Rowley, Mass., Newbury House.
Bloom, L. 1970, Language Development: Form and Function in Alerging Grammars,
Cambridge, Mass., MIT Press.
Bloomfield, L. 1942, Outline Guide for the Practical Study of Foreign Languages, Baltimore,
Linguistic Society of America.
Bowers, R. 1980, The individual learner in the general class. In .Altman, H. and James,
C. (eds), Foreign Language Teaching: Meeting Individual Needs, Oxford, Pergamon.
Breen, M. 1985a, The social context for language learning — neglected situation?
Studies in Second Language Acquisition, 1, 135—58.
Breen, M. 1985b, Authenticity in the language classroom, Applied Linguistics, 6,
60-70.
Briere, E. 1978, Variables affecting native Mexican children’s learning Spanish as a
 Bibliography 207
second language, Language Learning, 28:159—74.
Brock, C. 1986, The effects of referential questions on ESL classroom discourse,
TESOL Quarterly, 20:47—59.
Brooks, N. 1960, Language and Language Learning. New York, Harcourt Brace and
World.
Brown, R. 1970, Psycholinguistics, New York, Free Press.
Brown, R. 1973, A First Language: The Early Stages, Cambridge, Mass., Harvard
University Press.
Brumfit, C. 1979, Accuracy and fluency as polarities on foreign language teaching,
Bulletin CILA, 29:89—99.
Brumfit, C. 1984, Communicative Methodology in Language Teaching, Cambridge,
Cambridge University Press.
Bruton, A. and Samuda, V. 1980, Learner and teacher roles in the treatment of oral
error in group work, RELC Journal, 11, 49-63.
Burt, M., Dulay, H. and Hernandez, E. 1973, Bilingual Syntax Measure, New York,
Harcourt Brace Jovanovich.
Burt, M. and Kiparsky, C. 1974, The Gooficon: A Repair Manual for English, Rowley,
Mass., Newbury House.
Callfee, G. and Freedman, S. 1980, Understanding and comprehending. Paper pre¬
sented at the Center for the Study of Reading, Urbana, Illinois.
Canale, M. and Swain, M. 1980, Theoretical bases of communicative approaches to
second language teaching and testing, Applied Linguistics, 1:1—47.
Cancino, H., Rosansky, E. and Schumann, J. 1974, Testing hypotheses about second
language acquisition, Working Papers on Bilingualism, 3:8—96.
Cancino, H., Rosansky, E. and Schumann, J. 1978, The acquisition of English
negatives and interrogatives by native Spanish speakers. In Hatch (1978a).
Carrell, P. 1983, Three components of background knowledge in reading comprehen¬
sion, Language Learning, 33:183—207.
Carroll, J. 1963, Research on teaching foreign languages. In Gage, N. (ed.), Handbook
of Research on Teaching, Chicago, Rand McNally.
Carroll, J. 1966, The contributions of psychological theory and educational research to
the teaching of foreign languages. In Valdman (1966b).
Carroll, J. 1967, Foreign language proficiency levels attained by language majors near
graduation from college, Foreign Language Annals, 1:131—51.
Catford, J. 1959, The teaching of English as a foreign language. In Quirk, R. et al.,
The Teaching of English, London, Seeker and Warburg.
Cathcart, R. 1986, Situational differences and the sampling of young children’s school
language. In Day (1986).
Cathcart, R. and Olsen, J. 1976, Teachers’ and students’ preferences for correction of
classroom conversation errors. In Fanselow, J. and Crymes, R. (eds), On TESOL
76, Washington, DC, TESOL.
Cavalcanti, M. 1987, Investigating FL reading performance through pause protocols.
In Faerch and Kasper (1987).
Cazden, C., Cancino, H., Rosansky, E. and Schumann, J. 1975, Second Language
Acquisition Sequences in Children, Adolescents and Adults, Final Report, US Depart¬
ment of Health, Education and Welfare.
Chastain K. 1969, The audiolingual habit theory versus cognitive code-
learning theory: Some theoretical considerations, International Review of Applied
Linguistics, 7:97—106.
208 Instructed Second Language Acquisition
Chastain, K. 1971, The Development of Modem Language Skills: Theory to Practice,
Philadelphia, Center for Curriculum Development.
Chaudron, C. 1977, A descriptive model of discourse in the corrective treatment of
learners’ errors, Language Learning, 27:29—46.
Chaudron, C. 1983, Foreigner-talk in the classroom — An aid to learning? In Seliger
and Long (1983).
Chaudron, C. 1985, Comprehension, comprehensibility and learning in the second
language classroom, Studies in Second Language Acquisition, 7:216—32.
Chaudron, C. 1986a, Teachers’ priorities in correcting learners’ errors in French
immersion classes. In Day (1986).
Chaudron, C. 1986b, The role of simplified input in classroom language. In Kasper
(1986).
Chaudron, C. 1988, Second Language Classrooms, Cambridge, Cambridge University
Press.
Chaudron, C. and Richards, J. 1986, The effect of discourse markers on the compre¬
hension of lectures Applied Linguistics, 7:113 — 27.
Chihara, T. and Oiler, J. 1978, Attitudes and attained proficiency in EFL: A socio-
linguistic study of adult Japanese speakers, Language Learning, 28:55 — 68.
Chomsky, N. 1957, Syntactic Structures, The Hague, Mouton.
Chomsky, N. 1959, Review of Verbal Behavior by B.F. Skinner, Language, 35:26—58.
Chomsky, N. 1965, Aspects of the Theory of Syntax Cambridge, Mass., MIT Press.
Chomsky, N. 1966, Linguistic theory. In Mead, R. (ed.), Northeast Conference on the
Teaching of Foreign Languages, Menasha, Wis., George Banta.
Chomsky, N. 1981, Lectures on Government and Binding, Dordrecht, Foris.
Chomsky, N. 1986, Knowledge of Language: Its Nature, Origin and Use, New York,
Praeger.
Clahsen, H. 1984, The acquisition of German word order: A test case for cognitive
approaches to L2 development. In Andersen R. (ed.), Second Languages: A Cross-
linguistic Perspective, Rowley, Mass., Newbury House.
Clahsen, H. 1985, Profiling second language development: A procedure for assessing
L2 proficiency. In Hyltenstam and Pienemann (1985).
Clahsen, H., Meisel, J. and Pienemann, M. 1983, Deutsch als Zweitsprache, Der Spacher-
werb ausldndischer Arbeiter, Tubingen, Gunter Narr.
Clark, J. 1969, The Pennsylvania Project and the ‘Audiolingual vs Traditional’ question,
Modem Language Journal, 53:388—96.
Cohen, A. 1987, Using verbal reports in research on language learning. In Faerch and
Kasper (1987).
Comrie, B. and Keenan, E. 1978, Noun phrase accessibility revisited, Language,
55:649-64.
Cook, V. 1969, The analogy between first and second language learning, IRAL.
Reprinted in Lugton and Heinle (1971).
Cook, V. 1985, Universal grammar and second language learning, Applied Linguistics,
6:2-18.
Cook, V. 1988, Chomsky's Universal Grammar: An Introduction, Oxford, Basil Blackwell.
Corder, P. 1967, The significance of learners’ errors. International Review of Applied
Linguistics, 5:161—9.
Corder, P. 1973, Introducing Applied Linguistics, Harmondsworth, Penguin.
Corder, P. 1974, Error analysis and remedial teaching. Paper presented at the first
overseas conference of the International Association of Teachers as a Foreign
 Bibliography 209
Language, Budapest, Also in Corder (1981).
Corder, P. 1975, The language of second language learners: The broader issues,
Modern Language Journal, 59:409—13.
Corder, P. 1976, The study of interlanguage. In Proceedings of the Fourth International
Congress of Applied Linguistics, Munich, Hochschulverlag. Also in Corder (1981).
Corder, P. 1977, Language teaching and learning: A social encounter. In Brown, D.,
Yorio, C. and Crymes, R. (eds), ON TESOL 77, Washington, DC, TESOL.
Corder, P. 1978a, Language distance and the magnitude of the language learning task,
Studies in Second Language Acquisition, 2/1.
Corder, P. 1978b, Language-learner language. In Richards (1978).
Corder, P. 1980, Second language acquisition research and the teaching of grammar,
BAAL Newsletter, 10.
Corder, P. 1981, Error Analysis and Interlanguage, Oxford, Oxford University Press.
Coulthard, M. 1977, An Introduction to Discourse Analysis, London, Longman.
Crookes, G. 1986, Task classification: A cross-disciplinary review, Technical Report No.
4, Honolulu, Center for Second Language Classroom Research, University of
Hawaii at Manoa.
Cross, T. 1977, Mothers’ speech adjustments: the contribution of selected child
listener variables. In Snow and Ferguson (1977).
Cummins, J. 1981, Bilingualism and Minority-language Children, Toronto, Ontario
Institute for Studies in Education.
Dakin, J. 1973, The Language Laboratory and Language Learning, London, Longman.
Daniel, I. 1983, On first-year German foreign language learning: A comparison of
language behaviour in response to two instructional methods. Unpublished Ph.D.
dissertation, University of Southern California.
Davies, A., Criper, C. and Howatt, A. (eds) 1984, Interlanguage, Edinburgh, Edinburgh
University Press.
Day, R. (ed.) 1986, Talking to Learn: Conversation in Second Language Acquisition,
Rowley, Mass., Newbury House.
Delamont, S. 1976, Interaction in the Classroom, London, Methuen.
Dickerson, L. 1975, Interlanguage as a system of variable rules, TESOL Quarterly,
9:401-7.
Donaldson, W. 1971, Code-cognition approaches to language learning. In Lugton and
Heinle (1971).
Doughty, C. and Pica, T. 1986, Information gap tasks: Do they facilitate second
language acquisition?, TESOL Quarterly, 20:305—25.
Duff, P. 1986, Another look at interlanguage talk: Taking task to task. In Day (1986).
Dulay, H. and Burt, M. 1973, Should we teach children syntax?, Language Learning,
23:37-53.
Dulay, H. and Burt, M. 1974a, You can’t learn without goofing. In Richards (1974).
Dulay, H. and Burt, M. 1974b, Natural sequences in child second language acquisition,
Language Learning, 24:37—53.
Dulay, H. and Burt, M. 1975, Creative construction in second language learning. In
Burt, M. and Dulay, H. (eds), New Directions in Second Language Learning, Teaching
and Bilingual Education, Washington, DC, TESOL.
Eckman, F., Bell, L. and Nelson, D. 1988, On the generalization of relative clause
instruction in the acquisition of English as a second language, Applied Linguistics,
9:1-20.
Edmondson, W. 1985, Discourse worlds in the classroom and in foreign language
210 Instructed Second Language Acquisition
learning, Studies in Second Language Acquisition, 7:159—68.
Edmondson, W. 1986, Some ways in which the teacher brings errors into being. In
Kasper (1986).
Ellis, R. 1981, The role of input in language acquisition; some implications for second
language teaching, Applied Linguistics, 2:70—82.
Ellis, R. 1982, Discourse processes in classroom second language development. Un¬
published Ph.D. thesis, University of London.
Ellis, R. 1984a, Classroom Second Language Development. Oxford, Pergamon.
Ellis, R. 1984b, Can syntax be taught? A study of the effects of formal instruction on
the acquisition of WH questions by children, Applied Linguistics, 5:138—55.
Ellis, B. 1984c, The role of instruction in second language acquisition. In Singleton,
D. and Little, D. (eds), Language Learning in Formal and Informal Contexts, Dublin,
IRAAL.
Ellis, R. 1985a, Understanding Second Language Acquisition, Oxford, Oxford University
Press.
Ellis, R. 1985b, The LI = L2 hypothesis: A reconsideration, System, 13:9—24.
Ellis, R. 1985c, Teacher—pupil interaction in second language development. In Gass
and Madden (1985).
Ellis, R. (1985d), Sources of variability in interlanguage, Applied Linguistics,
6:118-131.
Ellis, R. 1986, Developing interlanguage through fluency, Focus, 4:23—39.
Ellis, R. 1987a, Interlanguage variability in narrative discourse: Style shifting in the use
of the past tense, Studies in Second Language Acquisition, 9:1—20.
Ellis, R. 1987b, Contextual variability in second language acquisition and the relevancy
of language teaching. In Ellis (1987c).
Ellis, R. (ed.) 1987c, Second Language Acquisition in Context, Englewood Cliffs, NJ,
Prentice-Hall.
Ellis, R. 1988a, Investigating language teaching: The case for an educational perspective,
System, 16:1 — 11.
Ellis, R. 1988b, The role of practice in classroom language learning, Feanga, 8:1—25.
Ellis, R. 1989a, Sources of intra-learner variability in language use and their relationship
to second language acquisition. In Gass, S., Madden, C., Preston, D. and Selinker,
L. (eds), Variation in Second Language Acquisition: Psycholinguistic Issues, Clevedon,
Avon, Multilingual Matters.
Ellis, R. 1989b, Are classroom and naturalistic acquisition the same?: A study of the
classroom acquisition of German word order rules, Studies in Second Language
Acquisition, 11:305—28.
Ellis, R. 1989c, Comprehension and the acquisition of grammatical competence. Paper
given at the Conference on Comprehension and Language Learning, Ottawa,
Canada.
Ellis, R. and Rathbone, M. 1987, Fhe Acquisition of German in a Classroom Context,
London, Ealing College of Higher Education.
Els, van T., Bongaerts, T., Extra, G., Os, van C. and Janssen-van-Dieten, A. 1984,
Applied Linguistics and the Learning and Teaching of Foreign Languages, London,
Edward Arnold.
Erickson, F. and Mohatt, G. 1982, Cultural organization of participation structures in
two classrooms of Indian students. In Spindler, G. (ed.), Doing the Ethnography of
Schooling: Educational Ethnography in Action, New York, CBS College Publishing.
Ervin-Tripp, S. 1974, Is second language learning like the first?, TESOL Quarterly,
8:111-27.
 Bibliography 211
Ewbank, L. 1989, The acquisition of German negation by formal language learners. In
Van Patten, B., Dvorak, T. and Less, J. (eds), Foreign Language Learning: A
Research Perspective, Rowley, Mass., Newbury House.
Faerch, C. 1985, Meta talk in FL classroom discourse, Studies in Second Language
Acquisition, 7:184—99.
Faerch, C., Haastrup, K. and Phillipson, R. 1984, Learner Language and Language
Learning, Clevedon, Avon, Multilingual Matters.
Faerch, C. and Kasper, G. 1980, Processes in foreign language learning and communi¬
cation, Interlanguage Studies Bulletin, 5:47—118.
Faerch, C. and Kasper, G. 1983, Procedural knowledge as a component of foreign
language learners communicative competence. In Boete, H. and Flerrlitz, W.
(eds), Kommunikation im Sprach-Unterricht, Utrecht. Rijksuniversiteik.
Faerch, C. and Kasper, G. 1986, The role of comprehension in second-language
learning, Applied Linguistics, 7:257—74.
Faerch, C. and Kasper, G. (eds), 1987, Introspection in Second Language Research,
Clevedon, Avon, Multilingual Matters.
Fanselow, J. 1977, Beyond Rashomon — Conceptualising and describing the teaching
act, TESOL Quarterly, 11:17—39.
Fathman, A. 1975, The relationship between age and second language productive
ability, Language Learning, 25:245—53.
Fathman, A. 1976, Variables affecting the successful learning of English as a second
language, TESOL Quarterly, 10:433—41.
Fathman, A. 1978, ESL and EFL learning: similar or dissimilar? In Blatchford, C. and
Schacter, J. (eds), On TESOL ’78: EFL Policies, Programs, Practices, Washington,
DC, TESOL.
Felix, S. 1981, The effect of formal instruction on second language acquisition,
Language Learning, 31:87—112.
Felix, S. and Simmet, A. 1981, Natural processes in classroom L2 learning. Revised
version of a paper presented at the 3eme Colloque, Groupe de Recherche sur
PAcquisition des Languages, Paris.
Fillmore, W. 1976, The second time around: Cognitive and social strategies in second
language acquisition. Unpublished Ph.D. thesis, Stanford University.
Fillmore, W. 1979, Individual differences in second language acquisition. In Fillmore,
C., Kempler, D. and Wang, W. (eds), Individual Differences in Language Ability and
Language Behavior, New York, Academic Press.
Fillmore, W. 1982, Instructional language as linguistic input: Second language learning in
classrooms. In Wilkinson, L. (ed.), Communicating in the Classroom, New York,
Academic Press.
Fillmore, W. 1985, When does teacher talk work as input? In Gass and Madden
(1985).
Fisiak, J. (ed.) 1981, Contrastive Linguistics and the Language Teacher, Oxford, Pergamon.
Fishman, J. 1964, Language maintenance and language shift as fields of inquiry,
Linguistics, 9:32—70.
Flanders, N. 1970, Analyzing Teacher Behavior, Reading, Mass., Addison-Wesley.
French, F. 1949, Common Errors in English, London, Oxford University Press.
Freudenstein, R. 1977, Interaction in the foreign language classroom. In Burt, M.,
Dulay, H. and Finocchiaro, M. (eds), Viewpoints on English as a Second Language,
New York, Regents.
Fries, C. 1948, As we see it, Language Learning, 1:12-16.
Gaies, S. 1977, The nature of linguistic input in formal second language learning:
212 Instructed Second Language Acguisitlon
linguistic and communicative strategies. In Brown, H., Yorio, C. and Crymes, R.
(eds), On TESOL 77, Washington, DC TESOL.
Gaies, S. 1983a, The investigation of language classroom processes, TESOL Quarterly,
17:205-18.
Gaies, S. 1983b, Learner feedback: An exploratory study of its role in the second
language classroom. In Seliger and Long (1983).
Gass, S. 1979, Language transfer and universal grammatical relations, Language Learning,
27:327-44.
Gass, S. 1982, From theory to practice. In Rutherford, W. and Hines, M. (eds), On
TESOL ’81, Washington, DC, TESOL.
Gass, S. and Selinker, L. (eds), 1983, Language Transfer in Language Learning, Rowley,
Mass., Newbury House.
Gass, S. and Madden, C. (eds) 1985, Input in Second Language Acquisition, Rowley,
Mass., Newbury House.
Gass, S. and Varonis, E. 1985, Task variation and nonnative/nonnative negotiation of
meaning. In Gass and Madden (1985).
George, H. 1972, Common Errors in Language Learning: Insights from English, Rowley,
Mass., Newbury House.
Givon, R. 1979, On Understanding Grammar, New York, Academic Press.
Glahn, E. and Holmen, A. (eds), 1985, Learner Discourse, Anglica et Americana,
Copenhagen, University of Copenhagen.
Gleitman, L., Newport, E. and Gleitman, H. 1984, The current state of the motherese
hypothesis, Journal of Child Language, 11:43—79.
Gower, R. and Walters, S. 1983, Teaching Practice Handbook, London, Heinemann
Educational.
Gregg, K. 1984, Krashen’s Monitor and Occam’s Razor, Applied Linguistics,
5:79-100.
Gregg, K. (forthcoming). The variable competence model of second language acquisi¬
tion, and why it isn’t, Applied Linguistics.
Gremmo, M., Holec, H., Riley, P. 1977, Interactional structure: the role of role,
Melanges Pedagogiques, University of Nancy, CRAPEL.
Gremmo, M., Holec, H. and Riley, P. 1978, Taking the initiative: some pedagogical
applications of discourse analysis, Melanges Pedagogiques, CRAPEL.
Grittner, F. 1968, Letter to the Editor, Newsletter of the National Association of Laboratory
Directors, 3:7.
Grotjahn, R. 1987, On the methodological basis of introspective methods. In Faerch
and Kasper (1987).
Hakansson, G. 1986, Quantitative studies of teacher talk. In Kasper (1986).
Hakuta, K. 1976, A case study of a Japanese child learning English as a second
language, Language Learning, 26, 321—51.
Hale, T. and Budar, E. 1970, Are TESOL classes the only answer? Alodern Language
Journal, 54:487—92.
Hamayan, E. and Tucker, R. 1980, Language input in the bilingual classroom and its
relationship to second language achievement, TESOL Quarterly, 14:453—68.
Harley, B. 1987, Functional grammar in French immersion: a classroom experiment.
Paper given at 1987 AILA Conference in Sydney, Australia.
Harmer, J. 1983, The Practice of English Language Teaching, London, Longman.
Hatch, E. 1974, Second language learning — Universal, Working Papers on Bilingualism,
3:1-18.
 Bibliography 213
Hatch, E. (ed.), 1978a, Second Language Acquisition, Rowley, Mass., Newbury House.
Hatch, E. 1978b, Discourse analysis and second language acquisition. In Hatch
(1978a)
Hatch, E. 1978c, Discourse analysis, speech acts and second language acquisition. In
Ritchie, W. (ed.), Second Language Acquisition Research, New York, Academic Press.
Hatch, E. 1978d, Acquisition of syntax in a second language. In Richards, J. (ed.),
Understanding Second and Foreign Language Learning, Rowley, Mass., Newbury
House.
Hatch, E. 1983a, Psycholinguistics: A Second Language Perspective, Rowley, Mass., Newbury
House.
Hatch, E. 1983b, Simplified input and second language acquisition. In Andersen R.
(1983).
Hatch, E., Shapira, R. and Gough, J. 1978, ‘Foreigner-talk’ discourse, ITL Review of
Applied Linguistics, 39/40, 39—59.
Hatch, E. and Wagner-Gough, J. 1976, Explaining sequence and variation in second
language acquisition, Language Learning, Special Issue No. 4:39—57.
Hawkins, B. 1985, Is an ‘appropriate response’ always so appropriate? In Gass and
Madden (1985).
Hendrickson, J. 1978, Error correction in foreign language teaching: Recent research,
theory and practice, Modem Language Journal, 67:41—5.
Henry, A. 1986, Linguistic theory and second language teaching. Paper given at GILT
Workshop on Acquiring Language and Learning Languages, Cumberland Lodge,
Windsor Great Park.
Henzl, Y. 1973, Linguistic register of foreign language instruction, Language Learning,
23, 207-22.
Henzl, V. 1979, Foreign talk in the classroom. International Review of Applied Linguistics,
17, 159-67.
Holley, F. and King, J. 1975, Imitation and correction in foreign language learning. In
Schumann, J. and Stenson, N. (eds), New Frontiers in Second Language Learning,
Rowley, Mass., Newbury House.
Holmen, A. 1985a, Analysis of some discourse areas in the PIF data and in classroom
interaction. In Glahn and Holmen (1985).
Holmen, A. 1985b, Distribution of roles in learner—native speaker interaction. In
Glahn and Holmen (1985).
Holmes, J. 1978, Sociolinguistic competence in the classroom. In Richards (1978).
House, J. 1986, Learning to talk: talking to learn. An investigation of learner performance
in two types of discourse. In Kasper (1986).
Howatt, A. 1984, A History of English Language Teaching. Oxford, Oxford University
Press.
Huang, J. and Hatch, E. 1978, A Chinese child’s acquisition of English. In Hatch
(1978a).
Huebner, T. 1979, Order-of-acquisition vs. dynamic paradigm: A comparison of
method in interlanguage research, TESOL Quarterly, 13:21—8.
Huebner, T. 1983, Longitudinal Analysis of the Acquisition of English Ann Arbor, Mich.
Karoma Publishers.
Hughes, A. 1983, Second language learning and communicative language teaching. In
Johnson, K. and Porter, D. (eds), Perspectives in Communicative Language Teaching,
New York, Academic Press.
Hughes, A. and Lascaratou, C. 1982, Competing criteria for error gravity, ELT
214 Instructed Second Language Acguisition
Journal, 36, 175—82.
Hulstijn, J. 1987, Onset and development of grammatical features: Two approaches to
acquisition orders, Paper given at Interlanguage Conference, Trobe University,
Melbourne.
Hulstijn, J. 1988, Implicit and incidental second language learning: Experiments in
processing of natural and partly artificial input. Mimeograph to be published in
Dechert, H. (ed.), Interlingual Processing, Tubingen, Gunter Narr.
Hulstijn, J. and Hulstijn, W. 1984, Grammatical errors as a function of processing
constraints and explicit knowledge, Language Learning, 34:23—43.
Hyltenstam, K. and Pienemann, M. (eds) 1985, Modelling and Assessing Second Language
Acquisition, Clevedon, Avon, Multilingual Matters.
Jakobovits, L. 1968, Implications of recent psycholinguistic developments for the
teaching of a second language, Language Learning, 18:89—109.
Jakobovits, L. 1970, Foreign Language Learning: A Psycholinguistic Analysis of the Issue,
Rowley, Mass., Newbury House.
James, C. 1974, Linguistic measures for error gravity, AVLA Journal, 12:3—9.
Jansen, L. 1987, The development of word order in formal German second language
acquisition. Unpublished paper, Australian National University.
Johnson, K. 1988, Mistake correction, ELT Journal, 2:89—96.
Johnston, M. 1987a, Second language acquisition research: A classroom perspective.
In Johnston and Pienemann (1987).
Johnston, M. 1987b, Second language acquisition research in the Adult Migrant
Education Program. In Johnston and Pienemann, (1987).
Johnston, M. and Pienemann, M. 1987, Second Language Acquisition: A Classroom
Perspective, NSW Adult Migrant Education Service.
Kadia, K. 1988, The effect of formal instruction on monitored and spontaneous
naturalistic interlanguage performance, TESOL Quarterly, 22:509 — 15.
Karmiloff-Smith, A. 1986, Stage/structure versus phase/process in modelling linguistic
and cognitive development. In Levin, J. (ed.), Stage and Structure: Reopening the
Debate, Norwood, Ablex.
Kasper, G. 1985, Repair in foreign language teaching, Studies in Second Language
Acquisition, 7:200 —15.
Kasper, G. 1986, (ed.), Learning, Teaching and Communication in the Foreign Language
Classroom, Aarhus, Aarhus University Press.
Kelch, K. 1985, Modified input as an aid to comprehension, Studies in Second Language
Acquisition, 7:81 — 90.
Kellerman, E. 1984, The empirical evidence for the influence of the LI in interlanguage.
In Davies, A, Criper, C. and Howatt, T. (eds), Interlanguage, Edinburgh, Edinburgh
University Press.
Kellerman, E. and Sharwood-Smith, M. (eds), 1986, Crosslinguistic Influence on Second
Language Acquisition, Oxford, Pergamon.
Kennedy, G. 1973, Conditions for language learning. In Oiler and Richards (1973).
Kleifgen, J. 1985, Skilled variation in a kindergarten teacher’s use of foreigner talk In
Gass and Madden (1985).
Klein, W. 1986, Second Language Acquisition, Cambridge, Cambridge University Press.
Kramsch, C. 1985, Classroom interaction and discourse options, Studies in Second
Language Acquisition, 7:169—83.
Krashen, S. 1973, Lateralization, language learning and the critical period, Language
Learning, 23:63—74.
 Bibliography 215
Krashen, S. 1977, The Monitor Model for adult second language performance. In
Burt, M., Dulay, H. and Finocchiaro, M. (eds), Viewpoints on English as a Second
Language, New York, Regents.
Krashen, S. 1980, The theoretical and practical relevance of simple codes in second
language acquisition. In Scarcella and Krashen (1980).
Krashen, S. 1981, Second Language Acquisition and Second Language Learning, Oxford,
Pergamon.
Krashen, S. 1982, Principles and Practice in Second Language Acquisition, Oxford,
Pergamon.
Krashen, S. 1984, Writing: Research, Theory and Application, Oxford, Pergamon.
Krashen, S. 1985, The Input Hypothesis, London, Longman.
Krashen, S., Jones, C., Zelinksi, C. and Usprich, C. 1978, How important is instruction?,
ELTJournal, 32:257—61.
Krashen, S. and Pon, P. 1975, An error analysis of an advanced ESL learner: The
importance of the Monitor, Working Papers on Bilingualism, 7:125—9.
Krashen, S. and Scarcella, R. 1978, On routines and patterns in language acquisition
and performance, Language Learning, 28:283—300.
Krashen, S. and Seliger, H. 1976, The role of formal and informal linguistic environ¬
ments in adult second language learning International Journal of Psycholinguistics,
3:15-21.
Krashen, S., Seliger, H. and Hartnett, D. 1974, Two studies in adult second language
learning, Kritikon Litterarum, 3:220—8.
Krashen, S., Sferlazza, V., Feldman, L. and Fathman, A. 1976, Adult performance on
the SLOPE test: More evidence for a natural sequence in adult second language
acquisition, Language Learning, 26:145—51.
Krashen, S. and Terrell, T. 1984, The Natural Method: Language Acquisition in the
Classroom, Oxford, Pergamon.
Labov, W. 1969, The study of language in its social context, Studium Generale,
23:3-87.
Lado, R. 1957, Linguistics across Cultures: Applied Linguistics for Language Teachers, Ann
Arbor, University of Michigan.
Lado, R. 1964, Language Teaching: A Scientific Approach. New York, McGraw Hill.
Larsen-Freeman, D. 1976a, The acquisition of grammatical morphemes by adult
learners of English as a second language, Unpublished Ph.D. thesis, University of
Michigan.
Larsen-Freeman, D. 1976b, An explanation for the morpheme accuracy order of
learners of English as a second language, Language Learning, 26/1, 125—35.
Larsen-Freeman, D. (ed.) 1980, Discourse Analysis in Second Language Acquisition
Research, Rowley, Mass., Newbury House.
Lenneberg, E. 1967, Biological Poundations of Language, New York, Wiley.
Lester, M. (ed.) 1971, Readings in Applied Transformational Grammar, New York,
Holt, Rinehart and Winston.
Lightbown, P. 1983, Exploring relationships between developmental and instructional
sequences in L2 acquisition. In Seliger and Long (1983).
Lightbown, P. 1984, The relationship between theory and method in second language
acquisition research. In Davies et al.(1984).
Lightbown, P. 1985a, Great expectations: second-language acquisition research and
classroom teaching, Applied Linguistics, 6:173 —89.
Lightbown, P. 1985b, Input and acquisition for second language learners in and out of
216 Instructed Second Language Acquisition
classrooms, Applied Linguistics, 6:263—73.
Lightbown, P. 1985c, Can language acquisition be altered by instruction? In Hyltenstam
and Pienemann (1985).
Lightbown, P. 1987, Classroom language as input to second language acquisition. In
Pfaff, C. (ed.), First and Second Language Acquisition Processes, Rowley, Mass.,
Newbury House.
Lightbown, P., Spada, N. and Wallace, R. 1980, Some effects of instruction on child
and adolescent ESL learners. In Scarcella and Krashen (1980).
Long, M. 1977, Teacher feedback on learner error: Mapping cognitions. In Brown,
H., Yorio, C. and Crymes, R. (eds), On TESOL 77, Washington, DC, TESOL.
Long, M. 1980, Inside the ‘black box’: methodological issues in classroom research on
language learning, Language Learning, 30:1—42. Also in Seliger and Long (1983).
Long, M. 1981, Input, interaction and second language acquisition. In Winitz, H.
(ed.), Native Language and Foreign Language Acquisition, Annals of the New York
Academy of Sciences, 379.
Long, M. 1983a, Native speaker/non-native speaker conversation in the second language
classroom. In Clarke, M. and Handscombe, J. (eds), On TESOL ’82: Pacific
Perspectives on Language Learning and Teaching, Washington, DC, TESOL.
Long, M. 1983b, Does second language instruction make a difference? A review of the
research, TESOL Quarterly, 17:359—82.
Long, M. 1983c, Native speaker/non-native speaker conversation and the negotiation
of meaning, Applied Linguistics, 4:126—41.
Long, M. 1985a, Input and second language acquisition theory. In Gass and Madden
(1985).
Long, M. 1985b, A role for instruction in second language acquisition: task-based
language training. In Hyltenstam and Pienemann (1985).
Long, M. 1988a, Instructed interlanguage development. In Beebe, L. (ed.), Issues in
Second Language Acquisition: Multiple Perspectives, Rowley, Mass., Newbury House.
Long, M. 1988b, Focus on form: A design feature in language teaching methodology.
Paper presented at the Conference on Empirical Research on Second Language
Learning in Institutional Settings, Bellagio, Italy.
Long, M. 1988c, Maturational constraints on language development. Mimeograph,
University of Hawaii.
Long, M., Brock, C., Crookes, G., Deicke, C., Potter, L. and Zhang, S. 1984, The
effect of teachers’ questioning patterns and wait-time on pupil participation in
public high school in Hawaii for students of limited English proficiency, Technical
Report No. 1, Honolulu, Center for Second Language Classroom Research, Uni¬
versity of Hawaii at Manoa.
Long, M. and Larsen-Freeman, D. (forthcoming), Introduction to Second Language
Acquisition Research, London, Longman.
Long, M. and Porter, R. 1985, Group work, interlanguage talk and second language
acquisition, TESOL Quarterly, 19:207-28.
Long, M. and Sato, C. 1983, Classroom foreigner talk discourse: Forms and functions
of teachers’ questions. In Seliger and Long (1983).
Long, M. and Sato, C. 1984, Methodological issues in interlanguage studies: An
interactionist perspective. In Davies et al. (1984).
Lorscher, W. 1986, Conversational structures in the foreign language classroom. In
Kasper (1986).
Lugton, R. and Heinle, C. (eds) 1971, Toward a Cognitive Approach to Second-Language
 Bibliography 217
Acquisition, Philadelphia, Center for Curriculum Development.
Lynch, A. 1988, Speaking up or talking down: foreign learners’ reactions to teacher
talk, ELTJournal, 42:109—16.
McHoul, A. 1978, The organization of turns at formal talk in the classroom, Language
and Society, 7:183 — 213.
Mackey, W. 1965, Language Teaching Analysis, London, Longman.
McLaughlin, B. 1978a, Second Language Acquisition in Childhood, Hillsdale, NJ, Lawrence
Erlbaum.
McLaughlin, B. 1978b, The Monitor Model: some methodological considerations,
Language Learning, 23:309—32.
McLaughlin, B. 1985, Second Language Acquisition in Childhood, vol. 2: School-Age
Children, 2nd edn, Hillsdale, NJ, Lawrence Erlbaum.
McLaughlin, B. 1987, Theories of Second Language Acquisition, London, Edward Arnold.
Macnamara, J. 1973, The cognitive strategies of language learning. In Oiler and
Richards (1973).
McNeill, D. 1965, Some thoughts on first and second language acquisition. Mimeograph,
Center for Cognitive Studies, Harvard University.
McNeill, D. 1966, Developmental psycholinguistics. In Smith, F. and Miller, G. (eds),
The Genesis of Language: A Psycholinguistic Approach, Cambridge, Mass., MIT
Press.
McNeill, D. 1970, The Acquisition of Language, New York, Harper and Row.
McTear, M. 1975, Structure and categories of foreign language teaching sequences.
In Allwright, R. (ed.), Working Papers: Language Teaching Classroom Research,
University of Essex, Department of Language and Linguistics.
Makino, T. 1980, Acquisition order of English morphemes by Japanese secondary
school students, Journal of Hokkaido University of Education, Section IV (Humanities),
30/2; 101-48.
Marton, W. 1981, Contrastive analysis in the classroom. In Fisiak (1981).
Mason, C. 1971, The relevance of intensive training in English as a foreign language
for university students, Language Learning, 21:197—204.
Mazurkewich, I. 1985, Syntactic markedness and language acquisition, Studies in
Second Language Acquisition, 7:15—36.
Mehan, H. 1974, Accomplishing classroom lessons, In Cicourel, A. et ah, Language
Use and School Performance, New York, Academic Press.
Mehan, H. 1979, Learning Lessons: Social Organization in the Classroom, Cambridge,
Mass., Harvard University Press.
Meisel, J. 1983, Strategies of second language acquisition: More than one kind of
simplification. In Andersen, R. (ed.), Pidginization and Creolization as Language
Acquisition, Rowley, Mass., Newbury House.
Meisel, J., Clahsen, H. and Pienemann, M. 1981, On determining developmental
stages in natural second language acquisition, Studies in Second Language Acquisition,
3:109-35.
Mitchell, R. 1985, Process research in second-language classrooms, Language Teaching,
18:330-51.
Mitchell, R. 1988a, Evaluation of foreign language teaching projects and programmes.
Paper given at Conference on Empirical Research into Foreign Language Teaching
Methods, Bellagio, Italy.
Mitchell, R. 1988b, Communicative Language Teaching in Practice, London, CILT.
Mitchell, R., Parkinson, B. and Johnstone, R. 1981, The Foreign Language Classroom:
218 Instructed Second Language Acguisition
An Observational Study, Stirling Education Monographs, No. 9, University of
Stirling.
Modem Language Materials Development Center 1961, Teacher’s Manual, New York,
Modem Language Materials Development Center.
Morton, F. 1960, The language laboratory as a teaching machine. In Oinas, F. (ed.),
Language Teaching Today, International Journal of American Linguistics, 26, 1.
Morton, F. 1964, Audio-lingual language programming project. Unpublished report,
Contract No. OE-3-14-012, US Office of Education, Language Development
Branch.
Moscowitz, G. 1970, The Foreign Language Teacher Interacts, Minneapolis; Association
for Productive Teaching.
Mowrer, O. 1960, Learning Theory and Behavior, New York, Wiley.
Naiman, N. 1974, The use of elicited imitation in second language acquisition research,
Working Papers on Bilingualism, 2:137—53.
Newmark, L. 1963, Grammatical theory and the teaching of English as a foreign
language. In Harris, D. (ed.), The 1963 Conference Papers of the English Language
Section of the National Association for Foreign Affairs. Also in Lester (1971).
Newmark, L. 1966, How not to interfere in language learning, International Journal of
American Linguistics, 32:77—83. Also in Lester (1971).
Newmark, L. and Reibel, D. 1968, Necessity and sufficiency in language learning,
International Review of Applied Linguistics in Language Teaching, 6:145—64. Also in
Lester (1971).
Nicholas, H. 1985, Learner variations and the teachability hypothesis. In Hyltenstam
and Pienemann (1985).
Nickel, G. 1973, Aspects of error analysis and grading. In Svartvik, J. (ed.), Errata:
Papers in Error Analysis, Lund, Gleerup.
Norton, R. 1975, Measurement of ambiguity tolerance, Journal of Personality Assessment,
39:607-18.
Nunan, D. 1987, Communicative language teaching: Making it work, English Language
Teaching Journal, 41:136—45.
Nystrom, N. 1983, Teacher-student interaction in bilingual classrooms: Four
approaches to error feedback. In Seliger and Long (1983).
Oiler, J. & Richards, J. 1983, (eds), Focus on the Learner, Rowley, Mass. Newbury
House.
Olson, D. 1977, From utterance to text: The bias of language in speech and writing,
Harvard Educational Review, 47:257—281.
O’Malley, J., Chamot, A. and Walker, C. 1987, Some applications of cognitive theory
to second language acquisition, Studies in Second Language Acquisition, 9:287-306.
Osgood, C. 1953, Method and Theory in Experimental Psychology, New York, Oxford
University Press.
Palmer, H. 1921, The Principles of Language Study, London, Harrap.
Pavesi, M. 1984, The acquisition of relative clauses in a formal and in an informal
setting: Further evidence in support of the markedness hypothesis. In Singleton,
D. and Little, D. (eds), Language Learning in Formal and Informal Contexts, Dublin,
IRAAL.
Pavesi, M. 1986, Markedness, discourse modes and relative clause formation in a
formal and informal context, Studies in Second Language Acquisition, 8:38—55.
Pawley, A. and Syder, F. 1983, Two puzzles for linguistic theory; nativelike selection
 Bibliography 219
and nativelike fluency. In Richards, J. and Schmidt, R. (eds), Language and
Communication, London, Longman.
Penfield, W. and Roberts, L. 1959, Speech and Brain Mechanisms, New York, Atheneum
Press.
Perkins, K. and Larsen-Freeman, D. 1975, The effect of formal language instruction
on the order of morpheme acquisition, Language Learning, 25:237—43.
Philips, S. 1972, Participant structures and communicative competence, Warm Springs
children in community and classroom. In Cazden, C., John, V. and Hymes, D.
(eds), Functions of Language in the Classroom, New York, Teachers College Press.
Pica, T. 1983, Adult acquisition of English as a second language under different
conditions of exposure, Language Learning, 33:465—97.
Pica, T. 1985a, Instruction on second-language acquisition, Applied Linguistics,
6:214-22.
Pica, T. 1985b, Linguistic simplicity and leamability: Implications for syllabus design.
In Hyltenstam and Pienemann (1985).
Pica, T. 1988, Interlanguage adjustments as an outcome of NS—NNS negotiated
interaction, Language Learning, 38, 45—73.
Pica, T. and Doughty, C. 1985a, The role of group work in classroom second
language acquisition, Studies in Second Language Acquisition, 7:233—49.
Pica, T. and Doughty, C. 1985b, Input and interaction in the communicative classroom: a
comparison of teacher-fronted and group activities. In Gass and Madden (1985).
Pica, T., Holliday, L., Lewis, N. and Morganthaler, L. 1989, Comprehensible output
as an outcome of linguistic demands on the learner, Studies in Second Language
Acquisition, 11:63—90.
Pica, T. and Long, M. 1986, The linguistic and conversational performance of experi¬
enced and inexperienced teachers. In Day (1986).
Pica, T., Young, R. and Doughty, C. 1987, The impact of interaction on comprehension,
TESOL Quarterly, 21:737-58.
Pickett, G. 1978, The Foreign Language Learning Process, London, British Council.
Pienemann, M. 1984, Psychological constraints on the teachability of languages, Studies
in Second Language Acquisition, 6:186—214.
Pienemann, M. 1985, Leamability and syllabus construction. In Hyltenstam and
Pienemann (1985).
Pienemann, M. 1986, Is language teachable? Psycholinguistic experiments and hypoth¬
eses, Australian Working Papers in Language Development, 1/3.
Pienemann, M. 1987, Determining the influence of instruction on L2 speech processing.
Unpublished paper, University of Sydney.
Pienemann, M. and Johnston, M. 1987, Factors affecting the development of language
proficiency. In Nunan, D. (ed.) Applying Second Language Acquisition Research,
National Curriculum Resource Centre, Adult Migrant Education Program,
Australia.
Pienemann, M., Johnston, M. and Brindley, G. 1988, Constructing an acquisition-
based procedure for second language assessment, Studies in Second Language
Acquisition, 10, 2:217—44.
Pimsleur, P. 1960, Report of the conference on psychological experiments related to
second language learning. Mimeograph report of conference held in Los Angeles.
Politzer, R. 1961, Teaching French: An Introduction to Applied Linguistics, Boston, Ginn.
Politzer, R. 1965, Foreign Language Learning: A Linguistic Introduction, Englewood
Cliffs, NJ, Prentice-Hall.
220 Instructed Second Language Acguisition
Politzer, R. 1970, Some reflections on ‘good’ and ‘bad’ language teaching behaviors,
Language Learning, 20:31—43.
Politzer, R., Ramirez, A. and Lewis, S. 1981, Teaching standard English in the third
grade: Classroom functions of language, Language Learning, 31:171—93.
Popper, K. 1976, Unended Quest, London, Fontana-Collins.
Porter, P. 1986, How learners talk to each other: Input and interaction in task-centred
discussions. In Day (1986).
Posner, M. and Synder, C, 1975, Attention and cognitive control. In Solso, R. (ed.),
Information Processing and Cognition: The Loyola Symposium, Hillsdale, NJ, Lawrence
Erlbaum.
Prator, C. 1969, Adding a second language, TESOL Quarterly, 3/2.
Ramirez, A. and Stromquist, N. 1979, ESL methodology and student language
learning in bilingual elementary schools, TESOL Quarterly, 13:145—58.
Ravem, R. 1968, Language acquisition in a second language environment, International
Review of Applied Linguistics, 6:175 — 85.
Reiss, M. 1983, Helping the unsuccessful language learner, Canadian Modem Language
Review, 39:257—66.
Reynolds, P. 1971, .4 Primer in Theory Construction, Indianapolis, Bobbs-Merrill.
Richards, J. (ed.) 1974, Error Analysis, London, Longman.
Richards, J. (ed.) 1978, Understanding Second and Foreign Language Learning, Rowley,
Mass., Newbury House.
Richards, J. and Rogers, T. 1986, Approaches and Methods in Language Teaching: A
Description and Analysis, Cambridge, Cambridge University Press.
Riley, P. 1977, Discourse networks in classroom interaction: some problems in com¬
municative language teaching. Melangues Pedagogiques, University of Nancy,
CRAPEL.
Riley, P. 1985, Discourse and Learning, London, Longman.
Rivers, W. 1964, The Psychologist and the Foreign Language Teacher, Chicago, University7
of Chicago, Press.
Rivers, W. 1968 Teaching Foreign Language Skills, Chicago, University Press of Chicago
Press.
Rosansky, E. 1976, Methods and morphemes in second language acquisition research,
Language Learning, 26:409—25.
Rosing-Schow, D. and Haastrup, F. 1982, The use of communication strategies in
classroom and spontaneous discourse. Unpublished MS, Department of English,
University of Copenhagen.
Rulon, K. and Creary, J. 1986, Negotiation of content: teacher-fronted and small-
group interaction. In Day (1986).
Rutherford, W. 1987, Second Language Grammar: Learning and Teaching, London,
Longman.
Rutherford, W. and Sharwood-Smith, M. 1985, Consciousness-raising and universal
grammar, Applied Linguistics, 6:274—81.
Rutherford, W. and Sharwood Smith, M. (eds) 1988, Grammar and Second Language
Teaching: A Book of Readings, Rowley, Mass. Newbury House.
Sacks, H., Schegloff, E. and Jefferson, G. 1974, A simplest systematics for the
organization of turn taking in conversation, Language, 50:696-735.
Sajavaara, K. 1978, The Monitor Model and monitoring in second language speech
communication. In Gingras, R. (ed.), Second Language Acquisition and Foreign
Language Learning, Washington, DC, Center for Applied Linguistics.
 Bibliography 221
Sajavaara, K. 1981, The nature of first language transfer: English as L2 in a foreign
language setting. Paper presented at the first European-North American
Workshop in Second Language Acquisition Research, Lake Arrowhead, Calif.
Sato, C. 1981, Ethnic styles in classroom discourse. In Hines, M. and Rutherford, W.
(eds), On TESOL ’81, Washington, DC, TESOL.
Sato, C. 1986, Conversation and interlanguage development: Rethinking the connection.
In Day (1986).
Scarcella, R. and Higa, C. 1981, Input, negotiation and age differences in second
language acquisition, Language Learning, 31:409—37.
Scarcella, R. and Krashen, S. (eds) 1980, Research in Second Language Acquisition,
Rowley, Mass., Newbury House.
Schachter, J. 197^4, An error in error analysis, Language Learning, 24:205 — 14.
Schachter, J. 1986, In search of systematicity in interlanguage production, Studies in
Second Language Acquisition, 8:119—34.
Schegloff, E., Jefferson, G. and Sacks, H. 1977, The preference for self-correction in
the organization of repair in conversation, Language, 53:361-82.
Scherer, A. and Wertheimer, M. 1964, A Psycholinguistic Experiment in Loreign Language
Teaching. New York, McGraw Hill
Schinke-Llano, L. 1983, Foreigner-talk in the content classroom. In Seliger and Long
(1983).
Schmidt, R. 1983, Interaction, acculturation and the acquisition of communicative
competence. In Wolfson, N. and Judd, E. (eds), Sociolinguistics and Second Latiguage
Acquisition, Rowley, Mass., Newbury House.
Schmidt, R. and Frota, S. 1986, Developing basic conversational ability in a second
language: A case study of an adult learner of Portuguese. In Day (1986).
Schneider, W. and Shiffrin, R. 1977, Controlled and automatic human information
processing in detection, search and attention, Psychological Review, 84:1—66.
Schumann, J. 1978, The Pidginization Process: A Model for Second Language Acquisition,
Rowley, Mass., Newbury House.
Scollon, R. 1976, Conversations with a One Year Old, Honolulu, University of Hawaii.
Seliger, H. 1977, Does practice make perfect? A study of interaction patterns and L2
competence. Language Learning, 27:263—75.
Seliger, H. 1978, Implications of a multiple critical periods hypothesis for second
language learning. In Ritchie, W. (ed.), Second Language Acquisition Research, New
York, Academic Press.
Seliger, H. 1979, On the nature and function of language rules in language teaching,
TESOL Quarterly, 13, 359—69.
Seliger, H. and Long, M. (eds) 1983, Classroom Oriented Research in Second Language
Acquisition, Rowley, Mass., Newbury House.
Selinker, L. 1972, Interlanguage, International Review of Applied Linguistics, 10:209—30.
Selinker, L. and Lamendella, J. 1978, Two perspectives on fossilization in interlanguage,
Interlanguage Studies Bulletin, 3:143—91.
Sharwood-Smith, M. 1981, Consciousness-raising and the second language learner,
Applied Linguistics, 2:159—69.
Shiffrin, R. and Schneider, W. 1977, Controlled and automatic information processing,
II: Perceptual learning, automatic attending and a general theory, Psychological
Review, 84:127—90.
Simon, A. and Boyer, E. (eds) 1967, Mirrors for Behavior, Philadelphia, Researcher for
Better Schools, 71.
222 Instructed Second Language Acquisition
Sinclair, J. and Brazil, D. 1982, Teacher Talk, Oxford, Oxford University Press.
Sinclair, J. and Coulthard, M. 1975, Towards an Analysis of Discourse, Oxford, Oxford
University Press.
Skinner, B. 1957, Verbal Behavior, New York, Appleton Century Crofts.
Slimani, A. 1987, The teaching-learning relationship: Learning opportunities and
learning outcomes, and Algerian case study. Unpublished Ph.D. thesis, University
of Lancaster.
Slobin, D. 1973, Cognitive prerequisites for the development of grammar. In Ferguson,
C. and Slobin, D. (eds), Studies of Child Language Development, New York, Holt,
Rinehart and Winston.
Smith, P. (Jr.) 1970, A Comparison of the Cognitive and Audiolingual Approaches to Foreign
Language Instruction: The Pennsylvania Foreign Language Project. Philadelphia, Center
for Curriculum Development.
Smith, P. (Jr.) 1971, Toward a cognitive approach. In Lugton and Heinle (1971).
Snow, C. and Ferguson, C. (eds) 1977, Talking to Children, Cambridge, Cambridge
University Press.
Snow, C. and Hoefnagel-Hohle, M. 1978 Age differences in second language acqui¬
sition. In Hatch (1978a).
Snow, C. and Hoefnagel-Hohle, M. 1982, School age second language learners’
access to simplified linguistic input, Language Learning, 32:411 — 30.
Spada, N. 1986, The interaction between type of contact and type of instruction: Some
effects on the L2 proficiency of adult learners, Studies in Second Language Acquisition,
8:181-200.
Spada, N. 1987, Relationships between instructional differences and learning outcomes:
A process—product study of communicative language teaching. Applied Linguistics,
8:137-55.
Spolsky, B. 1988, Bridging the gap: A general theory of second language learning,
TESOL Quarterly, 22:377-96.
Stenson, N. 1975, Induced errors. In Schumann, J. and Stenson, N. (eds), New
Frontiers in Second Language Learning, Rowley, Mass., Newbury House.
Stem, H. 1981, Communicative language teaching and learning: toward a synthesis. In
Alatis, J. and Altman, H. and Alatis, P. (eds) 1981, The Second Language Classroom,
New York, Oxford University Press.
Stem, H. 1983, Fundamental Concepts of Language Teaching. Oxford, Oxford University
Press.
Stevick, E. 1976, Memory, Meaning and Method, Rowley, Mass., Newbury House.
Stevick, E. 1980, Teaching Languages: A Way and Ways, Rowley, Mass., Newbury
House.
Stockwell, R. and Bowen, J. 1965, The Sounds of English and Spanish, Chicago,
University of Chicago, Press.
Stockwell, R., Bowen, J. and Martin, J. 1965, The Grammatical Structures of English and
Spanish, Chicago, University of Chicago Press.
Stubbs, M. 1976, Language, Schools and Classrooms, London, Methuen.
Swain, M. 1985, Communicative competence: some roles of comprehensible input and
comprehensible output in its development. In Gass and Madden (1985).
Swain, M., Allen, P., Harley, B. and Cummins, J. 1989, The Development of Bilingual
Proficiency Project, Ontario Institute of Educational Studies, University of Toronto.
Swain, M. and Lapkin, S. 1982, Evaluating Bilingual Education: A Canadian Case Study,
Clevedon, Avon, Multilingual Matters.
 Bibliography 223
Tarone, E. 1983, On the variability of interlanguage systems, Applied Linguistics,
4:143-63.
Tarone, E. 1988, Variation in Interlanguage, London, Edward Arnold.
Tarone, E., Cohen, A. and Dumas, G. 1976, A closer look at some interlanguage
terminology: A framework for communication strategies, Working Papers on
Bilingualism, 9:76—90.
Taylor, D. 1988, The meaning and use of the term ‘competence’ in linguistics and
applied linguistics, Applied Linguistics, 9:148—68.
Terrell, T., Gomez, E. and Mariscal, J. 1980, Can acquisition take place in the
classroom? In Scarcella and Krashen (1980).
Tong-Fredericks, C. 1984, Types of oral communication activities and the language
they generate: A comparison, System, 12:133—46.
Turner, D. 1979, The effect of instruction on second language learning and second
language acquisition. In Andersen, R. (ed.), The Acquisition and Use of Spanish as
First and Second Languages, Washington, DC, TESOL.
Upshur, J. 1968, Four experiments on the relation between foreign language teaching
and learning, Language Learning, 18:111—24.
Valdman, A. 1966a, Introduction. In Valdman (1966b).
Valdman, A. (ed.), 1966b, Trends in Language Teaching, New York, McGraw Hill.
Valdman, A. 1978, On the relevance of the pidginization—creolization model for
second language learning, Studies in Second Language Acquisition, 1:55—77.
Van Lier, L. 1982, Analysing interaction in second language classrooms. Ph.D.
dissertation, University of Lancaster, Lancaster, England.
Van Lier, L. 1988, The Classroom and the Learner, London, Longman.
Varonis, E. and Gass, S. 1985, Non-native/non-native conversations: A model for the
negotiation of meaning, Applied Linguistics, 6:71—90.
de Villiers, J. and de Villiers, P. 1973, A cross-sectional study of the acquisition of
grammatical morphemes in child speech, Journal of Psycholinguistic Research,
2:267-78.
Wagner-Gough, J. 1975, Comparative Studies in Second Language Learning, CAL-
ERIC/CLL Series on Language and Linguistics, 26.
Walker, R. and Adelman, C. 1972, Towards a Sociography of the Classroom. Mimeo,
Report to Social Science Research Council.
Wardhaugh, R. 1970, The contrastive analysis hypothesis, TESOL Quarterly,
4:123-30.
Wardhaugh, R. 1971, Teaching English to speakers of other languages: The state of
the art. In Lugton and Heinle (1971).
Watson, J. 1924, Behaviorism, New York, Norton.
Weinert, R. 1987, Processes in classroom second language development: The acquisition
of negation in German. In Ellis (1987c).
Weinstein, C. and Mayer, R. 1986, The teaching of learning strategies. In Wittrock,
M. (ed.), Handbook of Research on Teaching (3rd edn), New York, Macmillan.
Wells, G. 1985, Language Development in the Pre-School Years, Cambridge, Cambridge
University Press.
Wesche, M. 1981, Language aptitude measures in streaming, matching students with
methods and diagnosis of learning problems. In Diller, K. (ed.), Individual Differ¬
ences and Universals in Language Learning Aptitude, Rowley, Mass., Newbury House.
Wesche, M. and Ready, D. 1985, Foreigner talk in the university classroom. In Gass
and Madden (1985).
224 Instructed Second Language Acquisition
Weslander, D. and Stephany G. 1983, Evaluation of an English as a second language
program for southeast Asian students, TESOL Quarterly, 17:473 — 80.
Westmoreland, R. 1983, L2 German acquisition by instructed adults. Unpublished,
University of Hawaii at Manoa.
White, L. 1985, Is there a logical problem of second language acquisition, TESL
Canada, 2:29—41.
White, L. 1987, Against comprehensible input: The Input Hypothesis and the develop¬
ment of second-language competence, Applied Linguistics, 8:95 — 110.
Widdowson, H. 1975, The significance of simplification, Studies in Second Language
Acquisition, 1:11 — 20.
Widdowson, H. 1978, Teaching Language as Communication, Oxford, Oxford University'
Press.
Wilkins, D. 1974, Second-language Learning and Teaching, London, Edward Arnold.
Wilkins, D. 1976, Notional Syllabuses, Oxford, Oxford University Press.
Willing, K. 1987, Learning Styles in Adult Migrant Education. National Curriculum
Resource Centre, Adelaide, Australia.
Winitz, H. 1978, A reconsideration of comprehension and production in language
training, Allied Health and Behavioral Sciences, 1:272—315.
Wode, H. 1976, Developmental sequences in naturalistic L2 acquisition, Working
Papers on Bilingualism, 11:1 — 13.
Wolff, D. 1987, Some assumptions about second language text comprehension, Studies
in Second Language Acquisition, 9:307 — 26.
Zobl, H. 1983, Markedness and the projection problem, Language Learning,
33:293-313.
Zobl, H. 1985, Grammars in search of input and intake. In Gass and Madden (1985).
Index

accessibility hierarchy 129nl2, 144, 159, Bailey, Madden and Krashen 55, 138
186 Barnes 78, 81
accuracy vs fluency 28 behaviourist psycholog}' 5, 9, 19, 21 — 2
acquisition rich vs acquisition poor Belyayev 28
environments 131, 133, 172n7 Bialystok 170, 175, 179-80, 197n6
acquisition vs development 63n2 Bialystok and Sharwood-Smith 182-3,
acquisition vs learning 41, 57, 59, 63n3, 187
139, 185 Bley-Vroman 4, 49
acquisition vs production 173n7 Bloom 8, 34
affective factors 57, 58—9, 116, 133 (see Bloomfield 7, 19, 23
also motivation) Brock 82
age factor 31n2, 36, 42, 43-4, Brooks 6, 19, 20, 22, 23, 24, 25, 31
110-11, 139 Brown 8, 34, 37
Allen, Frohlich and Spada 91—2 Brumfit 2, 28, 30
Allwright 11, 12, 45, 64, 66, 67, 68, 69, Bruton and Samuda 71
72, 73, 84, 89, 91, 100, 200 Burt and Kiparsky 54
analysed vs unanalysed L2 knowledge
179, 180 Canale and Swain 117, 179
analysis in language learning (see also Cancino et al. 49, 52
consciousness-raising) capability continuum 120
vs analogy 24, 28—9 caretaker input 102, 105
and rule perception 39 Carroll 18, 26, 38
Anderson 7, 175, 177, 178 Cathcart 81
d’Anglejan 2 Cathcart and Olson 70
anthropological observation see Cazden et al. 48
observation Chastain 9, 26, 35, 38, 39, 40, 43
articles 50, 141 Chaudron 70, 71, 72, 73, 75, 76, 79, 102,
Asher 103, 119 104, 204
Aston 115, 116 Chomsky 3, 8, 17, 25, 26, 27, 31, 33—4,
audiolingualism 6—8, 9, 10, 15, 18, 39, 62-3, 175, 189
19-31 Clahsen 144, 145, 190
Ausabel 178 Clahsen, Meisel and Pienemann 156
automaticity of L2 knowledge 179—80, Clark 10
185 classical conditioning 22, 31n6
avoidance behaviour 156, 157, 158, 167 classroom interaction
226 Index
and L2 acquisition 11, 12—13, 18n9, corrective feedback, see error treatment
73-4, 79-80, 84-5, 90, ch.5 creative construction 47
vs naturalistic interaction 16, 75, 81, critical period hypothesis 43
90, 111, 187 Crookes 112
structure of 86 crosslinguistic influence 29
types of 85—6, 88—9, 187
classroom language learning Dakin 27, 28, 31
definition 1—3, 17nl declarative knowledge 7, 8, 28, 169, 176,
vs naturalistic language learning 1—2, 177
9, 13, 17n2, 130, 135, 136-48, developmental errors 9
173n7 diary studies 84, 189, 193—4
classroom management discourse 78 discourse hypothesis 95, 96, 119—21,
classroom process research 4, 10, 11 — 12, 187, 191
ch.4 domains of language acquisition 1
cognitive anti-method 32, 35 — 8, 40, Donaldson 35, 38
60-1 Duff 112, 113
cognitive code learning theory/method 9, Dulay and Burt 9, 27, 46, 47, 138
10, 15, 28, 32, 35, 38-40, 60-1,
181 Eckman et al. 13, 159, 162, 182
cognitive learning theories 7—8, 18n6, Edmondson 16, 70, 74, 85, 90
28, 31n9, 175-84, 196, 200 Ellis 13, 15, 16, 27, 49, 58, 63, 76, 82,
collaborative discourse hypothesis 95, 96, 83, 88-9, 92, 95, 102, 113, 120-1,
122-3, 129nl3 122, 123-4, 126, 127, 135, 142-3,
communication strategies 144, 151, 159, 162, 165, 173, 174,
teachers’ 78, 80—1 186, 190, 192, 203
learners’ 52, 82—3, 179 Ellis and Rathbone 75, 113, 131, 135,
communicative competence 53, 56, 117 168, 189, 192
communicative language teaching 2, 30 Els et al. 45
comparative method studies 10, 64, 202 Erickson and Mohatt 69
competence vs performance 34, 35, 43, error analysis 45—6, 136—8
200 error evaluation 72
competence vs proficiency 174—5, 182, errors 7, 9, 24—5, 32-3, 36—7, 45—6,
197n2, 201 63n8, 74, 194, 197n4
comprehensible input 16, 57, 58, 100, error treatment 54—5, 59, 66, 70—4,
103, 133, 165 92n3, 181, 193, 198nl0
comprehensible output hypothesis 17, 95, Ervin-Tripp 27
96, 117-19, 128nl0, 190, 191, Ewbank 143—4
194 explicit L2 knowledge 7, 24, 39, 183,
comprehension process 105, 128n6, 191 184-7, 189, 193-4, 195, 196
Comrie and Keenan 129 extrapolation
consciousness-raising 15, 169—70, from general learning theories, 3, 6-8,
173nl0, 185, 193, 196 21, ch. 2, 175
constitutive enthography 68 from LI acquisition research 3, 8—10,
contrastive anlaysis hypothesis 25, 29, 45 33-44, 109-110
controlled vs automatic processing 176, from naturalistic L2 acquisition
183, 197n8 research 44—60
Cook 9, 27, 41, 42, 52, 178, 196 legitimacy of 61 — 2
copula- ‘be’ 94, 152, 156
Corder 9, 10, 26, 29, 35, 37, 41, 45, 51, Faerch 122—3
53, 54, 56, 62, 96, 194, 198nll Faerch and Kasper 82, 104, 105, 123,
 Index 22 7

178-9, 204 Harley 169

Faerch, Haastrup and Phillipson 185 Hatch 48, 79, 95, 106, 122, 123, 128,
Fanselow 65, 67 166
Felix 136—7, 166 Hatch, Shapira and Gough 76
Felix and Simmet 137 Hatch and Wagner-Gough 97
Fillmore, 113, 115, 126, 127 Henry 160, 164, 173
foreign language in the elementary school Henzl 66, 75, 79
(FLES) programmes 20, 22 Holley and King 72
formal instruction Holmen 82, 84
and L2 acquisition 11, 12—13, 125; see House 81—2
especially chap. 6, 189, 191—2 Howatt 7, 20
as interaction 93—4 Huebner 49, 50
delayed effect o‘f 97, 131, 147, 149, Hughes 62
168-70 Hughes and Lascaratou 72
effect on acquisition of rules 166-8 Hulstijn and Hulstijn 185, 194
effect on acquisition of formulas 1,
143, 155, 165-6, 173nl3, 186, ignorance hypothesis 44
190, 191, 194 immersion programmes 103, 104,
effect on rate of acquisition 15, 130—5, 117-18, 121
165 implicit L2 knowledge 7, 183—7, 190-2,
effect on route of acquisition 15, 195, 196
136-64 incorporation strategy 122
and explicit knowledge 189 information-processing 175—7
and implicit knowledge 190—2 input
options in 14, 187—8 input features 70, 74—6
projection and 159—61, 169 input features vs interaction features
formal vs informal language learning 2, 70, 107
36, 130-48, 183 input hypothesis 57, 95, 96, 99—107
form-function analysis 49—50, 141, simplified input 101, 105, 107
167-8 vs intake 96
form vs meaning-focused instruction interaction; see also classroom interaction
93-5, 187, 190-4 interactional features 70, 76—81
fossilization 52, 158, 170 interactional modifications
frequency hypothesis 95, 96—9 and teacher training 114
Fries 7, 19, 20 changes over time 113
in group work 113
Gaies 66, 69, 75, 79, 83, 89 in nautrlistic vs classroom settings 111
Gass 144, 159, 162, 182 nature of 103, 107—11
Givon 119-20, 187 interactional styles 123
Gower and Walters 28, 199 interaction analysis 65, 67—8, 91—2n2
grammar-translation 10, 18, 20, 38 interaction hypothesis 91, 95, 96,
grammaticality judgement tasks 200 107-16, 118
Gregg 51, 59, 104, 105 instrumental conditioning 22
Gremmo, Holec and Riley 76—7 interlanguage 33
group-work 13, 113, 198 applications of 52—6
theory of 51—2
habit-formation 6, 22—4, 27—8, 30, interlanguage talk 100
31n4
Hakansson 75, 76 Jakobovits 35, 37, 39, 40, 43
Hale and Budar 131, 135, 172 James 46, 54
228 Index
Johnson 8, 28, 187 Makino 139
markedness 121, 144, 159, 160—1, 165,
Kadia 151, 162, 170, 185, 189 169, 186-7, 189
Kasper 71 McLaughlin 7, 15, 20, 44—5, 50, 58, 59,
Kelch 102 63, 68, 69, 104, 174, 175, 178, 182,
Kennedy 35 — 6, 41 186, 197n8
Kleifgen 75, 79 McNeill 34, 37, 40
knowledge vs control 183, 184, 187 McTear 72, 85, 171
Krasmch 86 meaning-focused instruction 16, 94
Krashen 10, 15, 16, 18, 33, 41, 57-62, mean length of utterance (MLU) 34, 102
63, 73, 85, 95, 99-107, 118, 127, Mehan 65, 68, 88
140, 185 mentalist theories of language acquisition
Krashen, Jones, Zelinksi and Usprich 8-9, 26, 32
131, 132-3, 135 Meisel 144, 145
Krashen, Sferlazza, Feldman and Meisel, Clahsen and Pienemann 48, 152,
Fathman 140—1. 156
Krashen and Seliger 2, 63 micro-ethnography 69
Krashen and Terrell 57 Mitchell 4, 69, 78, 79-81, 83, 90
Mitchell, Parkinson and Johnstone 128
LI acquisition research 33 — 5, 37, 97, models of L2 acquisition 127 —8n5
105, 109, 123 Monitor Model, the 10, 18, 33, 56—60,
LI interference 24—5, 29, 37, 42—3, 61, 100, 185, 196, 197
194 criticisms of 59—60, 199
L2 = LI hypothesis 4, 22, 27, 32, morpheme studies 47, 138—42, 147—8
41-4, 137 Moscowitz 67
Lado 6, 19, 20, 23, 25 motivation 42, 132—3, 153, 165
language acquisition device 34, 36, 42, Multidimensional Model 152—3, 155
43, 52, 100
Larsen-Freeman 97, 98, 138, 172 Natural Approach, the 59
lateralization 43, 63n4 naturalistic language learning 2, 3,
laws of learning 23 44-50
leamability 27, 152—3, 166—8 natural order of L2 acquisition see
learner’s language, in the classroom sequence of L2 acquisition
81-5 negatives 48, 122, 136-7, 142-3, 149,
learning ‘set’ 31n5 150, 162
learning style 188 negotiation of meaning 12, 16, 107—9,
Lenneberg 26 115-16, 133
Lightbown 13, 61-2, 97, 99, 137-8, Newmark 26, 36, 44
140, 166, 169 Newmark and Reibel 9, 35, 37, 39, 40,
Lightbown, Spada and Wallace 150, 156, 52
162, 168-9 non-participant observation 68
linguistic universals 159, 187, 191 Nunan 128
listening methods 119 Nylstrom 72, 73
Long 4, 5, 13, 16, 18, 43, 68, 70, 72, 73,
95, 99, 102, 104, 107-16, 121, 123, obligatory occasion analysis 46
130, 132, 133, 146, 168, 191, 199, obliterative subsumption 176, 177, 181,
203 197n4
Long and Sato 46, 48, 77, 78, 79, 97, 98, observation, in the classroom 64-9,
102, 122, 123 92n7
Lorscher 87 observational schedules 67—8
 Index 229
O’Malley, Chamot and Walker 175, 177, 146, 147, 159-61, 169
178, 183 pronouns 137, 160, 164
operant conditioning 23, 31n6
optimal communicative environment questions, in classroom discourse, 78, 82,
126-7 111, 128n9
output hypothesis, see comprehensible
output hypothesis Reibel 36
relative clauses 142, 144, 159, 160, 162
participant ethnography 68 remedial language teaching 53, 63n7
Pavesi 121, 144 repair 71, 72, 73—4
Pawley and Syder 186 research methods 4—5, 11, 15, 45—50,
Penheld and Roberts 43 64, 66-8, 92n7, ch.8
performance analysis 46—9, 139 research questions 201 — 2
Perkins and Larsen-Freeman 139 Riley 16, 67, 90
phrase structure rules 34 Rivers 19, 26, 27, 28, 35
Pica 13, 118, 140-1, 173 Rutherford 15, 141, 189
Pica and Doughty 109, 113
Pica and Long 75, 76, 111, 114 Sajavaara 140
Pica, Holliday, Lewis and Morganthaler Sato 84, 115, 123
118 scaffolding 122, 123, 190
Pica, Young and Doughty 103, 111 Scarcella and Higa 79, 110
Pickett 133 Schachter 46, 50
Pienemann 13, 152—8, 162, 165, 168, Scherer and Wertheimer 10, 64
173, 196 Schinke-Llano 75
Pienemann and Johnston 152, 153 Schmidt 118, 165
Pienemann, Johnston and Brindley 153, Schmidt and Frota 193—4, 198nl0
168 Schumann 150, 162, 170, 173
Pimsleur 18 Scollon 122
planned vs unplanned discourse 119—21, Seliger 43, 66, 128, 169, 181, 185, 186,
128—9nll, 144 189, 199
plural-s 140, 141, 150 Selinker 50
Politzer 22, 24, 25, 65 Selinker and Lamendella 52
Politzer, Ramirez and Lewis 77, 79, 81 sequencing and grading of linguistic
Popper 174, 197nl items 39—40, 41
Porter 113 sequence of language acquisition 9, 13,
possessive-s 150 15, 18, 34, 39, 47-9, 57, 138-42,
poverty of the stimulus 62—3nl, 173nll 146-9, 152-8, 168, 172n5, 182,
practice 184, 185, 191
controlled/pattem practice, 23, 28, 29, Sharwood Smith 15, 51, 59, 186, 192,
30, 66, 75, 151, 160, 165, 171-2, 194
173nl0/ll, 192, 198n9, 198 Sinclair and Coulthard 67, 72, 88
free/creative practice 39, 175, 177, Skinner 7, 19, 20, 21, 23, 31
181-2, 184, 187 Slimani 115, 124, 129, 203-4
pragmatic vs. syntactic mode 120, 123 Smith 10, 38, 64
Prator 38, 41, 43 Snow and Hoefnagel-Hohle 98, 110
procedural knowledge 7, 8, 28, 176, 183, Spada 131, 135, 188
197n5 Spolsky 174
production, role in L2 acquisition see Stenson 74
comprehensible output hypothesis Stem 2, 17, 62
projection of grammatical knowledge 13, strategic competence 178—9
230 Index
structural approach 20 uptake 115, 129n4
structuralistic linguistics 6, 19 Universal Grammar 34, 63n6, 157,
Swain 17, 95, 117-18, 121 173nll, 178, 186, 196
Swain, Allen, Harley and Cummins 117, U-shaped pattern of development 34
118
Swain and Lapkin 103 Valdman 24, 25, 56
syllabus organization 55 — 6 Van Lier 4, 11, 64, 71, 72, 73-4, 87, 89,
92
Tarone, 95, 119—20 variation in L2 acquisition 49—50, 51,
task-based learning 112 55, 119-20, 152, 154, 156, 192
Taylor 174-5, 197n3 Varonis and Gass 115
teachability hypothesis 157—8, 173nl5 verb-ing 138, 141
teacher roles 77 vertical constructions 122
teachers’ attitudes 92n7
teacher’s paradox 85 Wagner-Gough 49, 122
teacher talk 66, 70, 76—7 Wardhaugh 25, 29, 35
Terrell et al. 101 Watson 19, 20, 21
third person -s 141, 150, 173nl4 Weinert 143, 165, 173
Tong-Fredericks 112 Wells 97, 105, 109, 123
topicalization hypothesis 95, 96, 123—4 Wesche and Ready 75
Total Physical Response 103 Weslander and Stephany 131
transformational rules 24 White 43, 104, 105, 106, 126, 127
transitional competence see interlanguage Wilkins 31
transitional structures 142 (see also Wode 48
interrogatives, negatives and word Wolff 128
order rules) word order rules 142, 144—6, 152,
triangulation 69 155-7, 160, 162, 173n8
Turner 139
turn-taking 86—7 Zobl 43, 106, 159, 160, 164, 169
 Instructed Second Language Acquisition
Hod Ellis

How does classroom language learning take place? How does an under¬
standing of second language acquisition contribute to language teach¬
ing? In answering these questions, Rod Ellis reviews a wide range of
research on classroom learning, developing a theory of instructed
second language acquisition which has significant implications for
language teaching.

The early chapters of this book trace the attempts to explain classroom
language learning in terms of general theory of learning (behaviourism)
and the study of naturalistic language learning. The middle chapters
document the attempts of researchers to enter the ‘black box’ of the class¬
room in order to describe the teaching-learning behaviours that take
place there and to investigate to what extent and in what ways instruction
results in acquisition.

The book concludes with a theory of classroom language learning. This

theory advances an explanation of the relationship between explicit and
implicit linguistic knowledge and in so doing accounts for how both form-
focussed and meaning-focussed instruction contribute to second
language acquisition in the classroom.

Rod Ellis is Professor of Applied Linguistics at Temple University Japan.

Previously he was Head of the English Language Teaching Department
in Ealing College of Higher Education. He has published extensively in
the field of second language acquisition. Understanding Second Lan¬
guage Acquisition (1985) won the BAAL prize for the best book published
in applied linguistics that year. Second Language Acquisition in Context
(1987), a collection of papers he edited, won the Kenneth W. Mildenberger
prize, awarded by the Modern Language Association.

ISBN D-b31-lbSDS-X

BLACKWELL 9 780631 162 25

Oxford UK & Cambridge USA