LANGUAGE,

SCHOOLS AND
CLASSROOMS
Michael Stubbs

ROUTLEDGE LIBRARY EDITIONS:

EDUCATION
ROUTLEDGE LIBRARY EDITIONS:
EDUCATION

LANGUAGE, SCHOOLS AND

CLASSROOMS
LANGUAGE, SCHOOLS AND
CLASSROOMS

MICHAEL STUBBS

Volume 200
ROUTLEDGE

Routledge
Taylor &. Francis Group

LONDON AND NEW YORK

First published in 1976
Second edition published in 1983
This edition first published in 2012
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Simultaneously published in the USA and Canada
by Routledge
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 1976 and 1983 Michael Stubbs
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in writing from the
publishers.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without intent
to infringe.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library

ISBN 13: 978-0-415-61517-4 (Set)

eISBN 13: 978-0-203-81617-2 (Set)
ISBN 13: 978-0-415-50104-0 (Volume 200)
eISBN 13: 978-0-203-12856-5 (Volume 200)

Publisher’s Note
The publisher has gone to great lengths to ensure the quality of this reprint but
points out that some imperfections in the original copies may be apparent.

Disclaimer
The publisher has made every effort to trace copyright holders and would
welcome correspondence from those they have been unable to trace.
MICHAEL STUBBS

Language,
schools
and
classrooms
Second edition

METHUEN
London and New York
First published in 1976 by British Library
Methuen& Co. Ltd Cataloguing in Publication Data
11 New Fetter Lane, London EC4P 4EE Stubbs, Michael
Reprinted 1978 and 1979 Language, schools and
Second edition 1983 classrooms.—2nd ed.—
Reprinted 1985 (Contemporary sociology of
the school)
1. Schoolchildren—Language
Published in the USA by 2. Teachers—Language
Methuen&Co. I. Title II. Series
in association with Methuen, Inc. 371.1'02 LB1027
733 Third Avenue, New York, NY 10017 ISBN 0-416-35640-0

Library of Congress
© 1976 and 1983 Michael Stubbs Cataloging in Publication Data
Stubbs, Michael, 1947-
Language, schools and
classrooms.
Printed in Great Britain (Contemporary sociology of
by Richard Clay & Co Ltd the school)
The Chaucer Press Bibliography: p.
Bungay, Suffolk Includes index.
1. Language arts—Great
Britain.
All rights reserved. No part of this book may be 2. Children—Great Britain
reprinted or reproduced or utilized in any form —Language.
or by any electronic, mechanical or other 3. Sociolinguistics.
means, now known or hereafter invented, in- I. Title. II. Series.
cluding photocopying and recording, or in any LB1576.S871983 372.6
information storage or retrieval system, with- 83-7972
out permission in writing from the publishers. ISBN 0-416-35640-0 (pbk.)
 Contents
Editor's introduction 7

Foreword to the second edition 11

1 Why is language important in education? 15

The language of education? 16
Language, learning and classrooms 16
Teacher training and language study 21

2 Some basic sociolinguistic concepts 24

Language and attitudes to language 25
The primitive language myth 30
Standard and nonstandard English 32
Language structure and language use 37
The implication of such distinctions 43

3 Bernstein's theory of restricted and elaborated codes 46

The work of Basil Bernstein 47
Bernstein's early work 49
Bernstein's later work (1973) 56
Do the codes exist? 61
Some possible confusions 63
Conclusions 65

4 Labov and the myth of linguistic deprivation 67

Languages, logic, explicitness and grammar 68
Nonstandard languages as media of education 72
The myth of linguistic deprivation 74
Labov and Bernstein 79
West Indian children in British schools 80
A pseudo-problem? 84

5 The need for studies of classroom language 88

Reasons for studying classroom language 88
Our ignorance of classroom language 90
The rationale for naturalistic studies 94
Teachers as researchers 97
 6 Studies of classroom language 99
Commentaries on classroom dialogue 100
The structure of classroom dialogue 107
The classroom as a sociolinguistic setting 111
Studying social processes in classrooms 114

7 Teaching and talking: the hidden

curriculum of classroom talk 118
The hidden curriculum 119
The framing of educational knowledge 121
Discourse structure and assumptions about teaching 124
The social construction of children's ability 128
Teaching as talking: some cross-cultural data 129

8 Towards a sociolinguistic analysis

of language in education 133
Language as evidence for educational statements 134
Language is organized 136
Criteria for studies of language in education 140

9 Some topics for investigation 142

Further reading 149

References and name index 151
Subject index 158
 Editor's
introduction

Less than a decade ago the 'Contemporary Sociology of the

School' series was conceived. Its purpose was to bring together
the new and often complex sociological explorations of events in
and around the school and its classrooms in a way in which they
could be understood and made use of by teachers and other
professional workers. An important part of the purpose was also
to bring together, with similar clarity, the relevant range of
theoretical orientations and research strategies, for without
these any new understanding could only be incomplete. The
enterprise has been an outstanding success. With the help of an
able and enthusiastic team of authors, a group of books has been
produced which has been used by tens of thousands of students.
The distinctive red volumes have become key texts in their own
right in universities and colleges throughout the world.
There is little doubt that the series has made an important
contribution to sweeping away many of the misleadingly easy
and often unexamined assumptions of the 1960s - such as those
about the achievements of working-class children, girls and
members of ethnic minorities. The books have illustrated the
7
ways in which individual teachers' and students' definitions of
situations can influence events, how perceptions of achievement
can not only define achievement itself, but also identify those
who achieve; how expectations about schooling can help to
determine the nature and evaluations of schools.
The books explore the main areas of the sociology of the school
in which new understandings of events are available. Each
introduces the reader to the new interpretations, juxtaposes
them against the longer-standing perspectives and reappraises
the contemporary practices of education and its consequences.
Each author in the series has worked extensively in his or her
areas of specialism and has been encouraged not only to intro-
duce the reader to the subject but also to develop, where
appropriate, his or her own analyses of the issues. Yet though
each volume has its distinctive critical approach, the themes and
treatments of all of them are closely interrelated. The series as a
whole is offered to students who seek understanding of the
practice of education in present-day societies, and to those who
wish to know how contemporary sociological theory may be
applied to the educational issues of these societies.
A new development in the series is the introduction of
'readers' to accompany several of the volumes. These contain a
range of papers, many not previously published, which have
been selected by the authors of the original volumes to augment
and develop their analyses and to help readers to extend their
understanding of the fields.
Since the publication of the earlier volumes the pace of
research and theoretical development in many of the areas has
been rapid - development in which the authors themselves have
been actively involved. Nowhere has such development been
more rapid than in the study of the language of the school. In
this volume Michael Stubbs has revised and rewritten his text
which was first published in 1976. He has taken account of
remarkable advances in the study of dialect and minority lan-
guage, the debate on language deficit and the new emphases on
the use of language following the publication of major reports in
many countries, notably the Bullock Report in Britain. But he
8
also re-emphasizes and sharpens his original concern with the
practice of the classroom, showing how the individual teacher
can investigate the language of schools and thereby better
understand and utilize the regional, social and ethnic diversity
they offer. As before the book is characterized by a clarity of
presentation that itself constitutes a valuable example of lan-
guage use.
John Eggleston

9
 Foreword to
the second
edition

In the seven years since the appearance of the first edition of this
book in 1976, a great deal of important work has been published
on language in education, and there is probably even greater
recognition than seven years ago that a systematic understand-
ing of language is very important to teachers. The major
findings and recommendations of the Bullock Report, A Lan-
guage for Life (HMSO, 1975), are still widely cited and influen-
tial: a considerable achievement for an official report (see Ch.
1.3).
All the topics which I discussed in the first edition therefore
still seem relevant, and I have not omitted anything major in this
revised edition. The main changes have been to revise thorough-
ly and bring up to date all the references and to expand and make
more explicit several sections.
Since 1976 many valuable books have made accessible to
teachers facts and ideas which are of central importance to
education. This work falls into the following main categories.
(1) A great deal of work has been published on the sociology of
language in Britain. This includes:
11
 (a) Descriptions of dialects and accents of British English,
using modern methods of urban dialectology. For ex-
ample, Macaulay (1978), Milroy (1980) and Cheshire
(1982) provide details of nonstandard dialects in Glasgow,
Belfast and Reading respectively, and J. C. Wells (1982), a
comprehensive study of English accents. Hughes and
Trudgill (1979) summarize some of the main facts and
principles.
(b) Descriptions of other dialects of English, in particular
Caribbean English: see V. K. Edwards (1979), Sutcliffe
(1981), Le Page (1981).
(c) Facts about the distribution of ethnic minority lan-
guages in Britain: see Rosen and Burgess (1980). The work
of the Linguistic Minorities Project based at the London
University Institute of Education is very important and will
soon be available in book form.
All of this work provides badly needed information about
dialects and languages in Britain.
(2) There has been increased interest in reading and writing as
social activities, especially in the ways in which printed and
written material are actually used in school classrooms,
as opposed to mainly psychological, experimental studies
of reading ability: for example, Lunzer and Gardner
(1978).
(3) Language disability, in its clinical and pathological aspects,
is also an important topic for at least some teachers,
although too specialized to be discussed in this book. It is
now, for the first time, well provided with introductory and
more advanced textbooks which are accessible to teachers:
see Crystal (1980) and the associated series edited by
Crystal, Studies in Language Disability and Remediation
(Edward Arnold).
(4) A large amount of work on classroom language and interac-
tion has continued to be published. Although the descrip-
tions are in some ways more sophisticated, work in this area
does not seem essentially different from what was available
in 1976. There are, however, excellent studies which relate
12
 theory, description and educational practice: a model in this
respect is by Willes (in press).
I am particularly aware that this book does not provide any
substantial discussion of such major topics as the place of ethnic
minority languages in British schools, and the concept of liter-
acy. However, not everything can be covered in a short book,
and books which attempt to put across basic concepts in a small
space have their own merits. Much relevant material on both
these topics is provided in the companion reader to this book,
edited by myself and Hilary Hillier (1983), and a discussion
of literacy is provided by Language and Literacy (Stubbs,
1980).
One very striking development over the past few years is that
an increasing number of professional academic linguists have
become interested in language in education. For example, two
associations of professional linguists, the British Association for
Applied Linguistics and the Linguistics Association of Great
Britain, contribute to the Committee for Linguistics in Educa-
tion, which was set up in 1978 and has active programmes of
co-operation with teachers. These two organizations are also
affiliated to the National Congress on Languages in Education,
which was set up in 1976 to study all aspects of foreign language,
second language and mother-tongue teaching in Britain, and has
already produced many useful reports. It is encouraging that
many academic linguists have taken on the responsibility of
selecting, interpreting and presenting up-to-date knowledge
about language in a way which makes clear its great social and
educational importance.
However, it is very disappointing that this flood of excellent
academic work has been hindered by the way in which much
teacher training has been severely restricted by the closure of
courses and whole colleges. In many cases it has been courses in
English language and linguistics in education that, as recently
established, have been among the first to vanish.
In the preparation of this edition I have received helpful
suggestions from Margaret Berry and Patrick J. Finn. I am also
most grateful to the teachers and pupils who allowed me to
13
observe and tape-record them and therefore provided many of
the ideas discussed here.

Michael Stubbs
Nottingham, 1983

14
 1

Why is language
important in
education?

There is probably general agreement among educationalists that

language is somehow a crucial factor in a child's education. But
there is certainly no general agreement on precisely how
'language' and 'education' are related. How, for example, is
language related to learning? How is a child's language related,
if at all, to his success or failure at school? Does it make sense to
call some children's language 'restricted'? What kind of lan-
guage do teachers and pupils use in the classroom? Does a
child's dialect bear any relation to his or her educational ability?
What is the significance of the fact that over a hundred languages
are spoken in Britain? Should special educational provision be
made for the very high concentrations of speakers of immigrant
languages in several areas of the country? These are socially very
important questions, because what is at stake is people's beliefs
about the place of language in schools and classrooms, and
people's tolerance - or lack of it - of regional, social and ethnic
variation in language.
This book has two main aims. The relations between lan-
guage and education are complex, and it is important not to
15
oversimplify them. The first aim is therefore to give readers the
necessary concepts for disentangling some of the complexities
and for understanding the basic issues in the debate. The second
is to suggest some ways in which students and teachers them-
selves can observe and study how language is used in schools and
classrooms.

1.1 The language of education?

(1) See Janet, mother. See Janet play. See John and father.
(2) A: Now, who can tell me what a discjockey is? Brian.
B: It's on the radio, the man who says what records are
going to be played.
A: Yes, good. On the radio. You get someone who
announces the records and says, 'Now for Mrs Smith of
22 High Street, Easthampton'. . .
Even if you did not already know that this book is about
language in schools, you would doubtless recognize that these
fragments of language were taken from teaching contexts. They
are genuine, but stereotypical, examples. (1) is the artificial
language unfortunately still found in some reading books for
young children: language in a style that no young children (or
adults!) would normally use in speech. And (2) is a fragment of
tape-recorded classroom dialogue with a teacher asking a ques-
tion, not because he or she wants to find out something, but to
find out if the pupil knows something; and then reformulating
the pupil's answer in his or her own words. Schools and
classrooms sometimes make strange linguistic demands on
pupils.

1.2 Language, learning and classrooms

Imagine going into a dozen different classrooms in different
kinds of school. In some classrooms pupils might be doing
practical activities, like painting or woodwork, or nonverbal
activities, like mathematical calculations. But in most class-
rooms including maths and woodwork classes, of course, you
16
would find teachers and pupils talking, reading and writing:
that is, engaged in linguistic activities. Language is a central fact
in schools and classrooms, and there are therefore several simple
but important reasons why it deserves careful study by anyone
concerned with education. Some of these reasons are as follows.
Schools and classrooms are pervasive language environments.
Pupils are dealing with language for most of the day: with the
spoken language of the teacher or of other pupils, and with the
written language of books. There is a sense in which, in our
culture, teaching is talking. Research on traditional, relatively
formal, chalk-and-talk classrooms shows that, on average,
teachers tend to talk for about 70 per cent of classroom time
(Flanders, 1970). If pupils remain in school between the ages of
4 and 16 years they may have over 8000 hours of teacher-talk to
listen to! Many cultures have quite different concepts of
teaching and learning through practical demonstration, super-
vised participation, observation and trial and error (see Ch. 7.5).
But teaching as we know it is almost inconceivable without
language. For us, teaching and learning typically comprise
linguistic activities such as: lecturing, explaining, discussing,
telling, questioning, answering, listening, repeating, para-
phrasing and summarizing. One main aim of this book is
therefore to suggest some ways in which such activities can be
studied in everyday classroom dialogue.
It is sometimes said that 'every teacher is an English teacher'
and that 'every lesson is an English lesson'. By this is meant that
a teacher of any subject (not just English, but also maths,
geography or chemistry) has to teach the language of his subject.
A teacher cannot easily separate teaching facts or concepts in,
say, chemistry, from teaching his pupils how to use appropriate
terminology, how to construct a coherent argument and how to
understand books about chemistry. We do not expect a pupil's
essay on chemistry to read like an essay on English literature:
not only the subject matter but also the style is quite different.
For example, we do not expect a chemistry essay to begin:

Few things can be more beautiful than to watch crystals grow

17
 before your eyes in various shapes and hues. Yesterday, my
friend and I dropped some warm, strong copper-sulphate
solution onto a microscope slide, and watched with delight as
the liquid cooled and the tiny crystals took shape.

More appropriate (that is, stylistically conventional) would be:

Two or three drops of warm, concentrated copper-sulphate

solution were placed on a microscope slide. As the solution
cooled, crystals were deposited. Solutions were selected so
that varieties of crystal forms and colours were investigated.

The subject-specific language of an academic subject may have

an intellectual function. We are all familiar with the kind of
academic specialist (possibly a teacher) who is only happy
thinking about his subject in its specialist terminology, and is
unable to explain things in everyday language to non-specialists.
(Linguists are perhaps particularly guilty of this!) It is often
easier to use one technical term to explain another, using terms
like counters to be shuffled around, rather than thinking about
what they actually mean and relate to in real experience. Con-
versely, such a teacher may be unable to recognize a valid idea
from a pupil if it is not expressed in the style and terminology he
is used to. It is not only a matter of specialist jargon - this is
usually obvious and everyone is aware of it - it is also a question
of the whole style of language considered appropriate for
academic discussion (see Ch. 6.1).
Keddie (1971) gives a striking example of how a teacher may
pay more attention to a pupil's style of language than to the idea
being expressed. In a science lesson, pupils were shown pictures
of a foetus in a womb. One boy asked: 'How does it go to the
toilet?' A more conventional formulation might be: How does it
dispose of waste products? This is a sensible question, and
shows that the pupil is thinking for himself, but the boy could
not express himself in this more conventional style, and a
teacher later commented that 'he must have been joking'. It is
important not to confuse what a pupil says with how he says it.
I was once castigated by the editor of a book on classroom
18
research for using too many TV in an article I had sent him. He
changed several of my sentences from, for example, 'I will argue
that. . .'to I t will be argued that. . .'At one level, this did not
change my meaning, but it changed the overall style.
Some people would argue further that language is somehow
related to thinking, learning and cognitive development. The
precise relationship between language and thinking is complex
and little understood, and has provided philosophical con-
troversy for centuries: Does thought depend on language or vice
versa? This formulation of the question is probably too vague to
be answerable. But it is at least plausible that higher levels of
abstract thinking are supported by language. It is difficult to
think about abstract concepts (say, 'exponence' or 'feedback')
unless we have convenient labels for them. We are on safer
ground, and able to say something more specific by rephrasing
the question as: Are there describable linguistic routines
through which pupils acquire information and understanding?
Phrased in this way, it is possible to study actual classroom
dialogue in actual classrooms, and to study how a teacher
controls a lesson by opening or closing various learning possi-
bilities to pupils (see Chs 6 and 7). (I will discuss the language-
thought problem below only from this relatively restricted point
of view, but readers should be aware that many different
theories have been proposed by psychologists and philosophers
to try and explain how language and thinking are related. A
discussion of this large literature is beyond the scope of the
present book, but interested readers should consult Greene
(1975), who provides a short introduction to the area.)
An often-heard catchphrase in recent years has been that
'educational failure is linguistic failure'. This superficially sim-
ple statement covers a highly complex problem, and debate over
it has often generated more heat than light. In some sense, it is
clear that if a school considers a pupil's language to be inad-
equate, then he or she will probably fail in the formal education-
al system. But this is a tautology which follows automatically
from the pervasive language environment on which schools
depend. It merely raises the central question of what linguistic
19
 demands schools make on pupils. What does 'linguistically
inadequate' mean? Does it mean that the child's language
somehow constrains his or her thinking? Or that it causes
communication problems with a teacher who may speak a
different regional or social dialect? Or that the child cannot read
or cannot spell? Or that the child speaks a dialect with low social
prestige to which teachers react by calling it 'ugly' or 'slovenly',
and in rejecting the child's language reject the child? Clearly,
these are all very different kinds of inadequacy. We must
beware of oversimplifying the complex relationships between
language, social class, educational success and the linguistic
demands of everyday life in the classroom.
Changes in academic fashion have variously attributed educa-
tional failure to IQ (in the period 1920-40), to home environ-
ment (in the 1950s and 1960s), and now (since the 1960s) to
language. No single factor, however, can explain why some
pupils fail where others succeed. Language, social class, home
environment and intelligence are all interdependent, and any
simple model of language in education will be oversimple.
Certainly, any theory which claims to relate a child's language
directly to his or her educational success can easily be shown to be
inadequate (see Chs 3 and 4). It is the responsibility of teachers and
other educationalists, as well as sociologists, to understand the
complexities of the debate in this area.
Finally, language is important in education because it is
socially very important. No dialect is inherently superior or
inferior to any other. What we know as standard English (SE) is
largely based on the social-class dialect of upper-income groups
in south-east England, and is only 'standard' due to historical,
geographical and social accidents. (SE is descended from the
dialect of educated Londoners which, in the Middle Ages,
began to acquire social prestige and to turn from a regional
dialect into a social-class dialect. See Ch. 2.3.) But speakers may
be at a severe social disadvantage if they use nonstandard dialect
forms or have nonstandard accents; However groundless such
judgements may be, people are nevertheless judged intellec-
tually on the basis of minor differences in pronunciation and

20
superficial features of their language. There is no linguistic
justification for such judgements, but it is an important social
fact that A may be thought less intelligent than B because of his
regional accent. (It is also reported in the psychological litera-
ture that people who wear spectacles are typically assumed to be
more intelligent than those who do not!) Such sociolinguistic
phenomena may be crucial in classrooms where teachers and
pupils speak different dialects. So it is important that educa-
tionalists can distinguish between the characteristics of lan-
guage itself and the power of people's stereotyped attitudes to
language (see Ch. 2.1).
All this might suggest a reformulation of the previous point.
Rather than speaking of a pupil's linguistic failure', it might be
more accurate to speak of sociolinguistic barriers between
pupils and the educational system. This point is discussed at
length in Chapters 3 and 4.
For these reasons, then - the pervasive language environment
of schools; the difficulty of separating conventional styles of
language from the content of academic subjects; the complex
relationships between language, thinking and educational suc-
cess; and the power of social attitudes to language - it is
important for everyone concerned with schools and classrooms
to give language careful study. It is important for very practical
reasons that teachers should understand, for example, the
nature of people's reactions to different styles and regional
varieties of language, and the nature of possible differences
between their own language and the language of their pupils.
And it is important, too, that students of sociology should
understand the complex and often indirect nature of the rela-
tionship between language, social class, social groups and
education.

1.3 Teacher training and language study

It is because language concerns all teachers that the influential
Bullock Report makes a strong recommendation that: 'A sub-
stantial course of language in education (including reading)
21
should be part of every primary and secondary-school teacher's
initial training, whatever the teacher's subject or the age of the
children with whom he or she will be working' (HMSO, 1975, p.
515). I strongly endorse this view and intend this book to cover
some of what I believe should go into such a course.
There is one interesting, but disturbing, finding in the Bul-
lock Report about the qualifications of teachers. In a survey of
2000 schools in England, the committee found that almost a third
(32.8 per cent) of teachers ofEnglish had no formal qualifications in
English, that is, no degree, drama qualification, B.Ed, with an
English component and so on. Now, of course, formal qualifica-
tions may bear little relation to a teacher's actual classroom skill:
a teacher with no formal qualifications whatsoever may do
excellent work. And it may be that English is often taught by
teachers who have had linguistic training through the classics or
modern languages. Further, an English degree may not, in any
case, be an appropriate qualification for teaching English in a
school. Many English degrees contain little language work, and
there is no obvious relationship between, say, studying Beowulf
and teaching essay writing to 14 year olds! The point is,
however, that English is frequently regarded as a subject which
can be taught, often part time, by non-specialists who have had
little or no linguistic training. No one would regard physics
teaching in this amateur way.
Reasons for this neglect of language training for teachers are
not hard to find. Different authorities often disagree sharply
about what an English teacher should be trying to do. In recent
years, justified dissatisfaction with unimaginative teaching of
traditional grammar has given way in many places to a 'creative
writing' approach. But this approach often leaves teachers
wondering if they are not trying to develop the aesthetic sensi-
bility or even moral character of their pupils, rather than teach
them about language. And more recently, there has been a
further reaction against an exclusive reliance on creative writ-
ing, and a feeling that a more structured approach is required.
But it is often not clear what this approach should be. One
obvious reason for the neglect of language work in teacher

22
training is the lack of appropriate material. There are many
introductions to linguistics on the market, but these are not
usually designed for educationalists' interests. (An exception is
Crystal, 1976.) Neither are there many suitable courses on
language for educationalists, although appropriate syllabuses
have now been proposed for student teachers and teacher
trainers (HMSO, 1975), and the Open University offers two
courses on language in education (PE232 and E263).
The approach taken in this book is that educationalists and
sociologists of education should have a clear understanding of
some of the central concepts of recent work in sociolinguistics,
in so far as these are relevant to language in education. It is
reasonable to believe that certain educational problems could be
handled more successfully if teachers, and also educational
researchers, had a clearer understanding of the sociolinguistic
forces at work in schools and classrooms, of the ways in which
language varies within a speech community, and of the attitudes
which such language variation inevitably provokes. Such an
understanding requires (a) attention to the observed details of
language use in schools and classrooms. How do teachers and
pupils actually talk to each other? But it also requires (b) a
coherent framework to make sense of such observations.
Observations are of no interest in themselves, unless we can
relate them to general principles of language use in social
contexts. These are the two points which the rest of the book
tries to illustrate in detail.

23
2

Some basic
sociolinguistic
concepts

There has been much debate in educational and sociological

circles since the mid-1960s about the precise relationship be-
tween education and language. One question often asked is:
Does a child's language affect his success or failure at school?
And if so, how? Many people believe that a child's language is a
crucial cause of his educational success or failure. Another
question often asked is: How does a teacher's language affect his
pupils' learning? The aim of this chapter is to provide the reader
with some of the basic sociolinguistic concepts necessary to
understand the kinds of relationships which exist between
language and educational processes. (By sociolinguistics I simply
mean studies of language and how it is used in different social
contexts, such as homes, factories, schools and classrooms.)
Disentangling basic concepts is not merely an academic exer-
cise, but essential to anyone who wants to understand the issues
involved in the debate. What precisely is meant, for example, by
the often quoted phrase: 'Educational failure is linguistic fail-
ure'? Does this phrase, in fact, make sense? Is there any precise
meaning which can be attached to assertions that some pupils
24
are 'linguistically inadequate'? And what would constitute evi-
dence for such assertions?

2.1 Language and attitudes to language

The first distinction it is crucial to be clear about is the distinc-
tion between language itself and the deeply entrenched attitudes
and stereotypes which most people hold about language.
It is difficult to overestimate the importance of people's
attitudes and beliefs about language. It is almost impossible, for
example, to hear someone speak, without immediately drawing
conclusions, possibly very accurate, about their social-class
background, level of education and what part of the country
they come from. We hear language through a powerful filter of
social values and stereotypes. As a precise example of what I
mean by linguistic stereotypes, consider this fragment from a
recorded classroom lesson which was based on a discussion of
examples of dialect speech. The pupils have a transcript in front
of them.
Teacher: You can see on the bottom of your sheet, 'We ain't
got no money.' That is typically a London accent - the
tendency to drop the aitch off the front of words, d'you see?
It's a lazy way of speaking.
Just these few comments embody several pieces of confused and
dangerous linguistic folklore. The first is the moral censure
('lazy') which is attributed to a regional or social dialect feature.
The second is the way this moral disapproval is backed up with
pseudo-linguistic arguments. In the example the teacher
quotes, there is nowhere to 'drop an aitch' from! The teacher
may mean that 'ain't' is related to the standard form 'haven't'.
But we cannot make the form standard by saying 'hain't'! 'Ain't'
is now simply a dialect form of the negative. But 'dropping
aitches' is a linguistic stereotype which is widely believed to
characterize London speech, and it is thought to be 'lazy' or
'slovenly'. In fact, dropping of word-initial 'aitch' is found in
the casual speech of educated speakers from most parts of the
25
 country. The fragment also reveals other confusions: between
spoken and written language; and between 'accent' or pronun-
ciation and nonstandard grammar (e.g. the double negative).
But we will leave these for the moment (see further Chs 2.3 and
4.1).
The point is that British people are very sensitive to the social
implications of dialect and accent, and the characteristic speech
of our large cities, especially Birmingham, east London, Liver-
pool, Newcastle and Glasgow, is often regarded as 'slovenly'
and 'ugly'. Giles (1971) carried out experiments in which
people listened to standard and regional dialects. In fact, they
heard the same speaker using different language varieties, but
they did not know this! Speakers of SE were perceived as more
ambitious, more intelligent, more self-confident and more reli-
able. Such judgements may be manifestly unfair, but it is
an important social fact that people judge a speaker's intelli-
gence, character and personal worth on the basis of his or her
language. We ought to be aware of the power of such social
stereotyping.
It has been confirmed in many other studies, and is probably
obvious from everyday experience, that a speaker's language is
often a major influence on our impression of his or her
personality. In particular it has been shown (in a Canadian
study) that teachers evaluate pupils academically on the basis of
their voices, and also their physical appearance, even when they
have available relevant academic work on which to form their
judgements, such as written compositions and art work (Selig-
man et al., 1972). That is, a teacher may base serious systematic
judgements about a pupil's intellectual abilities on totally
irrelevant information. It is important for teachers to be aware
that this tendency to linguistic stereotyping can mean that
pupils may 'look and sound intelligent', and therefore to be
aware of the misleading clues often used in evaluating them.
As a more detailed example of the weight people often attach
to superficial features of language, consider this extract from an
interview I recorded with two Edinburgh schoolgirls, aged 14.
We were discussing a tape-recording of some dialect speech.
26
 R: Well, they sound sort of as if they weren't very well
brought up their selves, the way they were talking.
MS: Mmhm - what are you thinking of in particular?
H: Their grammar's pretty awful.
MS: What's pretty awful about it?
H: It only sort of went in a little bit. (Quoting from recording.)
MS: What's wrong with that?
H: Well, you don't sort of say that, do you?
MS: Well, what in particular?
H: It's bad English.
MS: Why?
H: Well, it just sounds bad English.
MS: Which bit of it then, or is it all. . .?
H: It only sort0fwent in.
MS: So, you don't say sort of}
H: I keep saying sort of yeah, but you're not meant to say sort
of.
MS: Well, I mean you said, em,you don't sort of say that, I
think.
H: I know - you're not meant to say that sort of thing - and I
know I shouldn't.
MS: Why not?
H: It just doesn't sound right. It sounds as though you're
Tarzan - Me Tarzan you Jane - Me speak English - sort of-
I'm saying it again, aren't I?
MS: Well, don't you think it's quite a useful expression?
H: You get into the habit of using it, I won't say it again. I'll
persevere and I won't say it. You get used to saying it if you
hear other people saying it - you know you sort of - I'll
never do it! - you associate that sort of thing with people
who haven't really been taught to say it better.

The girls interpret the language of the tape as evidence that the
speakers have been badly brought up, and are not far off the
level of an (intellectually?) primitive Tarzan. But, when they are
pressed, all they seem to be objecting to is the use of sort of. This
is an expression which H herself uses constantly in the extract
27
 (in spite of her efforts not to!) and which most people use in
informal conversation. Again, sort of is a feature of language
which has acquired the status of a stigmatized stereotype. What
are we to think, though, of an educational system which has so
tied this girl in knots over a small and superficial linguistic item?
I conducted this interview as part of a series of discussions
with Edinburgh children. I asked them to listen to recordings of
boys from east London, and asked them to comment on what
they heard. (The children did not know where the speakers
came from or who they were.) One of the most striking things
was the way in which the children singled out isolated features of
speech as particularly reprehensible. These included the use of
you know and we was. In general, the children tended to be
hyperconscious of a very few stigmatized features, which were
therefore made to carry a great weight of social significance. The
recordings often elicited quite unjustified extrapolations like:
'He sounds like a skinhead from his voice.' Such is the power of
linguistic stereotypes!
Some teachers might like to carry out such an experiment for
themselves. It could form the topic of a lesson in English, or
social science. The teacher could record people with different
accents and dialects from the radio or television or from real life,
and discuss with pupils why some speakers sound 'posh' or
'educated' or 'working class', and why such judgements may be
very misleading.
There is evidence that such stereotypes are transmitted at
least partly by schools. Very little work has been done on
institutional attitudes to language, but Milroy and Milroy
(1974, and in prep.) have done work on teachers' attitudes to
language in Northern Ireland. They have evidence that colleges
of education are particularly sensitive to such linguistic atti-
tudes, that they screen applicants for acceptability of speech,
and that they attempt subsequently to 'improve' candidates'
speech. And in Glasgow, Macaulay (1978) interviewed about
fifty teachers; he found that almost a third of them thought that
the school should try to change the way pupils speak. Some
teachers implied that because some children could not use the

28
language of the school, they were therefore less 'able' - thus
basing far-reaching intellectual judgements on children's
speech. (Chs 3 and 4 will explore fully whether such judgements
have any basis.) Such findings about teachers' attitudes to
speech are, for the present, rather impressionistic, but they
could be corroborated or modified by anyone reading this book
from his or her own classroom observations. It would be most
important, for example, to see whether comments made by
teachers in interviews correspond to the way they actually
attempt to modify their pupils' language in the classroom.
It is clear at any rate that schools in our society have always
been very sensitive to the social meaning of different language
varieties. In some extreme social situations, children have
actually been forbidden to speak their own language altogether,
and even punished for using their native language in schools.
This has been true in the past in Britain for Welsh and Scots
Gaelic speakers (Trudgill, 1974, p. 134) and in the USA for
American Indian children (Hymes, 1972). These may only be
extreme and explicit examples of the disapproval of children's
language often found in schools today. Scots Gaelic is now used
in primary schools in north-west Scotland, and Welsh is actively
encouraged in Welsh schools and is probably on the increase as a
second language (Sharp, 1973). But in the very recent past in
Wales, prisoners have been forbidden to speak Welsh to visitors
(The Times, 28 April 1972). It is important to appreciate that
language differences can provoke strong feelings of language
loyalty and group conflict, and are therefore often a critical
factor in education.

What is to be done?
One of the most important tasks in teacher training and in
teaching pupils about language is to undermine such
stereotypes. In itself, language diversity is no news to teachers:
it is often all too evident in many schools. We require a way of
making sense of the different kinds of diversity: geographical,
ethnic, social and stylistic. (A detailed framework for under-
29
standing language diversity is provided by the companion read-
er to this book.) However, we require more than a mere
understanding in principle.
One approach which has been proposed (for example, by
Trudgill, 1975a) is called 'appreciation of dialect differences'.
Trudgill argues that we should try to change people's attitudes,
not their language (p. 69), and that we should try to increase
people's tolerance of different accents and dialects (p. 101). This
is important as far as it goes, and it is true that increased
understanding can often lead to tolerance. But it is probably
necessary also to attack the stereotypes directly, and to show up
the mechanisms which transmit linguistic attitudes.
It is common, for example, for varieties of language and
therefore their speakers to be identified on the basis of the kind
of isolated and superficial features illustrated above. It is when
such superficial features of accent or dialect are lumped together
as 'bad English' and used in order to attach categoric labels to
people that real violence can be done. Readers will be familiar
with many labels for linguistic, social and ethnic groups. One
British example is the term teuchtar, which is used pejoratively
in Scotland to refer to Gaelic speakers or Highlanders.
This book is intended for teachers and student teachers, but it
should provide much material for teaching pupils about lan-
guage. A discussion of linguistic stereotypes can lead naturally
to a more general discussion of how people are labelled as 'dull',
'adolescent', 'delinquent' or 'immigrant', and how such labels
can influence how these people are perceived and treated.

2.2 The primitive language myth

Having now distinguished sharply between attitudes to lan-
guage and language itself, let us look at some features of
languages and dialects.
It is accepted by linguists that no language or dialect is
inherently superior or inferior to any other, and that all lan-
guages and dialects are suited to the needs of the community
they serve. A notion that one dialect is, say, more aesthetically
30
pleasing than another is, as we have already seen, a culturally
learned notion which generally reflects the social prestige of the
dialect speakers, and not inherent properties of the dialect itself.
The social prestige of groups of speakers, as it were, rubs off on
their language.
Linguists long ago dispelled the myth that there are primitive
tribes who speak 'primitive languages' with only 200-300
words, and simple grammar. It is now known that there is no
correspondence at all between simplicity of material culture and
simplicity of language structure, and all the world's languages
have been shown to have vastly complex grammatical systems.
However, the primitive language myth often lives on in a
pernicious form, in the unfounded belief that the language of
low-income groups in rural or urban industrial areas is somehow
structurally 'impoverished' or 'simpler' than SE. There is no
linguistic basis for such a belief. Fieldwork in urban and rural
areas of Britain and the USA has demonstrated in detail that
such dialects are inherently systematic and rule-governed, deep-
ly organized systems of great complexity.
It is true, of course, that some languages arc functionally more
highly developed than others. Thus English is an international
language, with a highly standardized writing system and is used
in a wide range of functions from everyday casual conversation
to writing scientific papers. Many hundreds of the world's
languages have no writing systems and cannot therefore serve
the same range of functions (Stubbs, 1980). It is also clear that
the native language of an Amazonian Indian is unlikely to be
well suited to discuss civil engineering. It is well known that
languages reflect, in their vocabulary, the needs and interests of
their speakers. Thus English does not need many different
words for snow (which Eskimo has) or several dozen words for
camel (which Arabic has).
None of these points, however, affect the central issue that all
languages and dialects are vastly complex structural systems.
(This will be illustrated further in Ch. 4.1.)

31
2.3 Standard and nonstandard English
I have already used terms such as standard language and dialect
without discussing what they mean. Such terms are in common
use, and although people often think their meaning is obvious,
they turn out to be rather elusive. The definition of SE is
complex, but it is important to give a rather careful and detailed
definition, because SE has a special place in the education
system in Britain. There are therefore several topics discussed
elsewhere in this book which depend on a consistent definition.
For example, there is the potential confusion between dialect
and style (Ch. 1.2), the common confusion between restricted
code and nonstandard English (Ch. 3), and the question of the
relationship between Caribbean varieties of English and British
English (Ch. 4.5).

Terminology and a preliminary definition

Several terms are used for SE including 'BBC English', the
'Queen's English' and 'Oxford English'. These terms are not
very precise, and the third in particular is very out of date in its
assumptions about British society. But they do no harm if they
are not taken too literally. Another term, however, does not
refer to the same thing at all. This is 'Received Pronunciation'
(RP). This is a socially prestigious accent, and refers only to
pronunciation, whereas SE refers to a dialect defined by gram-
mar and vocabulary. There is a peculiar relationship between
RP and SE. All users of RP speak SE: this is not logically
necessary, but merely a fact about language in Britain. On the
other hand, only a few speakers of SE use RP. For example, I
speak SE with a regional west of Scotland accent. There is, in
fact, no standard accent of English.
A fairly satisfactory preliminary definition of SE can be
provided simply by listing examples of its main uses. SE is the
variety of English which is normally used in print, and more
generally in the public media (hence BBC English), and used by
most educated speakers most of the time. It is the variety used in
32
the education system and therefore the variety taught to learners
of English as a foreign language. These examples tell any native
speaker roughly what is meant by SE. On the other hand, they
leave unclear whether SE is a predominantly written variety,
and whether it is a prescriptive norm imposed by the education
system or a description of the language which some people
actually use.

Dialect and style

There is one widespread confusion which is easily disposed of.
SE is a dialect, and like any other dialect there is stylistic
variation within it. That is, SE may be either formal or casual
and colloquial. The following sentences are all SE:
I have not seen any of those children.
I haven't seen any of those kids.
I haven't seen any of those bloody kids.
Speakers of SE can be as casual, polite or rude as anyone else,
and can use slang, swear and say things in bad style or bad taste.
This all has to do with stylistic variation or questions of social
etiquette, and not with dialect. The following sentence is not
SE:
I ain't seen none of them kids.
It is not incorrect SE: it is simply not SE at all. The double
negative, the use of them as a demonstrative adjective and the use
of ain't, are perfectly regular grammatical features which
characterize many nonstandard dialects of British English.
Vocabulary can also be regional and nonstandard: for example,
bairns is regionally restricted to Scotland and northern England.

Regional versus social dialects

One important point about SE is that it is not regionally
restricted as nonstandard dialects are. There is some regional
variation between the SE used in England, Scotland, Wales and
33
Ireland, but very much less than in nonstandard varieties. In
fact, there is a remarkably uniform international SE. Again
there are small differences among the standard varieties used in
Britain, North America, South Africa, Australia, New Zealand
and the Caribbean. But the differences (in vocabulary, grammar
and spelling) are minor given its very large number of speakers
over an enormous geographical area.
It follows that SE is not a regional or geographical dialect. It is
a social dialect: that dialect which is used by almost all 'edu-
cated' speakers. It is intuitively obvious that there is much more
variation in the language used by working-class people in
Britain than by middle-class people. Thus two businessmen
from Aberdeen and Exeter would have little difficulty in under-
standing each other; but two farm labourers from Aberdeen and
Exeter would speak very differently. Trudgill (1975a) diagrams
the relation between social and regional diversity as follows:
Social \ariation

Regional variation

As we move up the social-class scale, there is less regional

variation in dialect, though even right at the top there is still a
little.

Standard, written and spoken English

There is stylistic variation within both spoken and written SE.
However, spoken English is more variable due to the conven-
tion that most written (certainly printed) English is fairly
formal. A very simple example of this is the convention against
using forms such as don't, doesn't and isn't in printed books. This
rather crude diagram illustrates several related points:
34
 formal
formal
WRITTEN

SPOKEN

casual

This is intended to indicate that spoken and written SE are (a)

partly separate, but (b) overlapping. There is (c) variation in
both, but (d) more variation in the spoken form. It shows that (e)
the most formal written English is more formal than the most
formal spoken, and (f) conversely with casual spoken English.
And finally (g) as spoken English becomes more formal, it
moves closer to written English.
If we add a nonstandard dialect, the following diagram
represents further complications:

Standard Nonstandard
formal
formal formal
WRITTEN

casual
SPOKEN

casual casual

This is intended to indicate (a) that all written language is on the

standard side of the line, and (b) that as nonstandard English
becomes more formal, it moves towards SE, and may or may not
coincide with SE.
The very strong social convention that all printed English is
standard means that only occasional examples of dialect poetry
and the like get into print. If a language is highly standardized,
this implies in fact that it has a written form. This is because
standardization implies deliberate codification and planning of
the language by dictionary makers, grammar-book writers and
the like. This brings us to the special relationship between SE
and the education system, for it is the education system which is
a powerful instrument for promoting such codified norms of
language.
35
Usage versus prescription
People, including dictionary makers and schoolteachers,
observe what they think is good usage. This may well be a
mixture of local prejudice about what is a 'good accent', some-
times outdated norms of educated usage, and notions of written
or even literary language which may be inappropriate to spoken
English. On this basis they may formulate rules which may
become quite rigid. Once established, such rules can become
self-perpetuating. SE is used by prestigious people for presti-
gious purposes. The prestige of the speakers rubs off on the
language, and the circle continues. SE, the social elite who use
it, and the public functions it serves become inseparable.

Different types of definition

I have now used by implication several different types of
definition of SE. SE is most closely related to regional dialects
spoken in the south-east of England. This looks like a geo-
graphical definition. But, as I have pointed out, SE is no longer
regionally restricted. In fact, this is a disguised historical defini-
tion: SE developed historically out of a dialect used in London,
especially in the court. This shows in turn the need for a social
definition. Historically SE spread because of the prestige of its
users, and is now the social dialect used by educated middle-
class speakers from all over Britain (and with minor variations in
many other countries). I have also used a functional definition.
SE is the variety used in print, in education and as an interna-
tional language. These definitions are not prescriptive: they do
not say who ought to use SE for what purposes. They are
descriptive: they describe certain social facts which govern how
it is, as a matter of fact, used. These conventions are socially and
politically loaded, but I have not passed any judgement on
whether they are desirable or not.
However, it is easy to see how the borderline between a
descriptive and a prescriptive definition breaks down. The
reason for one further confusion should now be clear. SE is
36
prestigious and because of its speakers and its uses it is simply
more visible than other varieties. The very fact that it is the
variety used in print makes it more visible. People therefore
come to think of SE as the language. They confuse one socially
predominant dialect with 'the English language'. This is clearly
an extreme idealization, since the whole point of my discussion
has been that English is a cluster of many different styles and
dialects. This is one way in which a descriptive definition turns
into a prescriptive one. Other varieties may simply be defined
out of existence as 'bad English' or 'not real English'.
The sociolinguistic view of SE is often misrepresented. It is
often said, for example, that sociolinguists argue that SE is just
another dialect, like any other, with no privileged position.
This is not so. It is quite evident that SE holds a special position.
What sociolinguists emphasize is that this special position is not
due to any linguistic superiority. It is due to a complex of
historical, geographical and social factors which I have tried
briefly to summarize.
For further details and many examples of social and regional
variation in standard and nonstandard varieties of English, see
Hughes and Trudgill (1979) and Trudgill and Hannah (1982).
For further discussion of the relation between spoken and
written English, see Stubbs (1980).

2.4 Language structure and language use

A native speaker of a language knows, largely unconsciously, a
vast number of facts about his language. Much of this know-
ledge involves the ability to understand complex grammatical
relationships within sentences. For example, given two sen-
tences such as:
The teacher asked John to go (1)
The teacher told John to go (2)
a speaker of English knows that they mean almost the same
thing, except that asked implies more politeness than told. But
contrast these two sentences:
37
 The teacher asked John what to do (3)
The teacher told John what to do (4)

We know that, although the surface grammar of (3) and (4) is

identical, the sentences express different underlying relations.
In (3) it is the teacher who will do something. In (4) it is John
who will do something. Our knowledge of our language com-
prises awesomely complex knowledge of such surface and
underlying grammatical relationships: that is, knowledge about
language structure. If you are unconvinced that such knowledge
is awesomely complex, consider further such examples as these:

I expected John to be examined by the doctor (5)

I persuaded John to be examined by the doctor (6)
I expected the doctor to examine John (7)
I persuaded the doctor to examine John (8)

How do we know that (7) is a paraphrase of (5)? But that (8) is

not a paraphrase of (6)? We have no trouble understanding such
sentences and the relations between them. But you would find it
very difficult indeed to make explicit exactly why we understand
sentences in this way. Recall, then, that any dialect of any
language is based on many hundreds of such structural rela-
tionships.
There is much more to language, however, than grammatical
structure. A speaker of English knows that both (9) and (10) are
grammatically normal:

Give me that comic (9)

Could I have that comic, please (10)

But he also knows that (9) and (10) are used in quite different
social situations. (9) and (10) have the same referential meaning:
they could refer to the same event. But they clearly have quite
different social meanings. A teacher might say (9) to a pupil.
The pupil might later say (10) to the teacher. A main task for
sociolinguists is to specify such relationships between the use of
language and different social situations and social relationships.
38
Grammatical and communicative competence
The ability to speak a language is not, therefore, only the ability
to produce grammatical sentences. Hymes (1967) sums it up
neatly: 'A child capable of using all grammatical utterances, but
not knowing which to use, not knowing even when to talk and
when to stop, would be a cultural monstrosity.' Knowing a
language thus involves knowing how to say the right thing in the
appropriate style at the right time and place. It involves complex
knowledge of how to say what, to whom, when and where. This
knowledge of how to use language appropriately in social
situations is termed communicative competence. (The term is
Hymes's.) Although children have acquired the bulk of a vast
complex grammatical system by the time they begin school at 4
or 5 years, they continue to acquire the sociolinguistic system
until well into their teens or beyond.

Language varieties
Different language is used in different situations, so we can say
that a language is not a uniform object. It is a basic principle of
sociolinguistics that there are no single-style speakers (Labov,
1972b). That is, everyone is multidialectal or multistylistic, in
the sense that they adapt their style of speaking to suit the social
situation in which they find themselves. It is intuitively clear,
for example, that a boy does not speak in the same way to his
teachers, his parents, his girlfriend or his friends in the play-
ground. Imagine the disastrous consequences all round if he
did! His way of talking to his teacher will also change according
to the topic: answering questions in class or organizing the
school sports. People adapt their speech according to the person
they are talking to and the point behind the talk. These are social
rather than purely linguistic constraints.
As a more general example of what I have in mind by language
varieties, consider the following rather mixed bag of different
varieties or styles of English, spoken and written: BBC English,
Cockney, officialese, journalese, lecture, church sermon. These
39
language varieties differ along several dimensions, notably re-
gional/geographical, social class and functional/contextual. But
their description involves questions of the same order: who says
what? to whom? when? where? why? and how? In addition,
more than one dimension is typically involved in any one of the
varieties. For example, 'BBC English' implies not only that the
speaker is likely to come from a certain region (southern Britain)
and belong to a certain social class (educated middle class), but
also implies a relatively formal social situation (probably not
casual conversation in a pub). Some of what I have listed as
language varieties might be thought of rather as speech situa-
tions. But speech and situation are not entirely separable in this
way. For example, it is not simply that certain social situations
demand that a teacher 'gives a pupil a dressing down'. By 'giving
a pupil a dressing down' the teacher may create a certain social
situation!
Note the importance of the concept of language variation
when discussing children's language. A teacher may tend to
think of a child's language in a stereotyped way, as though the
child was a one-variety speaker. But the teacher typically sees
the child in only a narrow range of social situations in the
classroom, and may forget that the child also controls other
language varieties. In other words, many teachers are unaware
that all speech communities use ranges of different language
varieties in different social contexts; yet this is an elementary
sociolinguistic fact. Conversely, many teachers maintain the
fiction that there is only one 'best' English for all purposes, and
that this is the only English proper to the classroom. Yet a
moment's thought or observation will convince any teachers
that they themselves use many varieties of language throughout
the day, depending on the purpose or context of the com-
munication. This is not reprehensible, implying a chameleon-
like fickleness, but a basic sociolinguistic fact about language
use all over the world.

40
Correctness or appropriateness?
The concept that different language varieties are suited to
different situations can be summed up in the distinction which
is often drawn between correctness and appropriateness of
language. Many of us were taught at school some version of the
doctrine of correctness: that 'good English' means grammati-
cally correct SE; and that the use of colloquialisms, slang or
nonstandard forms is 'bad English'. No linguist would nowa-
days take this prescriptive attitude. Contemporary linguistics is
strictly descriptive: it describes what people do, and does not try
to prescribe what they ought to do. This does not mean, of
course, that 'anything goes'. If a pupil writes a letter to a
prospective employer which is full of colloquialisms or nonstan-
dard forms, he will have to be warned of the conventions of
English usage. It is not that such forms are wrong in any
absolute sense, but that they are considered inappropriate to
this social occasion - applying for a job.
Macaulay (1978) interviewed personnel managers, a careers
officer and the director of an employment agency in Glasgow, to
investigate the importance employers attach to speech in inter-
viewing school leavers for jobs. He discovered that most em-
ployers feel that speech is important and may be crucial at the
interview stage. Only a few of those interviewed said that an
applicant's accent was important, but there were many com-
plaints about 'slovenly speech'. There is no reason why such
views about language should not be discussed openly with
pupils in schools: in the English classroom, for example. As
Macaulay says, social judgements about language, particularly
about accent, are treated as a taboo subject even less mention-
able than sex or money.
To say that a piece of language is 'wrong' is therefore to make
a judgement relative to a social situation. It may be felt just as
inappropriate to use colloquialisms and regional dialect forms in
a job interview, as it is to use very formal language over a drink
with some friends in a pub. In the first instance one is likely to be
thought uncouth, impolite, socially gauche or uneducated.

41
(However unjustified such judgements may be, it is only fair to
warn pupils that people do base harsh social judgements on
surface characteristics of other people's speech. The ultimate
aim here must be to make more people more tolerant of linguis-
tic diversity.) In the second case one risks being thought aloof,
stand-offish, 'lah-de-dah' or a bit of a snob.
It follows, then, that within SE (as I showed in Ch. 2.3) there
is stylistic variation according to social context. Thus SE has
formal and informal styles in both writing and speech. So, the
use of colloquial forms, slang or swear words are all quite normal
within SE. They simply define the style as informal: they do not
define it as nonstandard. As a speaker of SE moves between
different social situations, he or she will style-shift. But precisely
the same functions may be served by other speakers shifting
between dialects. For example, many West Indian children in
Britain are bidialectal, using a form of Creole English in the home
and a more formal language variety, much closer to SE, in school
(see Ch. 4.5).
A teacher may often find that he or she wants both to defend
some controversial form (e.g. split infinitives) because it is
nowadays in widespread usage, but also to warn pupils not to use
it when, say, writing a job application. In other words, the
question 'What is correct English?' is too oversimple to answer.
(Mittins, 1969, gives an entertaining and sensible discussion of
this.) To say, therefore, that someone's English is 'wrong' is to
make not a linguistic, but a sociolinguistic judgement.

Production and comprehension

Suppose then that a teacher observes that a child uses language
inappropriate (in the teacher's terms) for the classroom. Is this
because the child does not know the forms the teacher thinks
appropriate? Or because the child knows the forms but does not
realize that it is appropriate to use them in this situation? If a
teacher observes that a child never produces a particular linguis-
tic item (word, sentence-type, etc.) this may mean several
things: (a) that the child neither knows nor understands the
42
item; (b) that the child understands the item, but never uses it in
his own speech; or (c) that the child both knows and uses the
item, but the teacher has never observed the child in a situation
where the child finds it appropriate and necessary to use it. We
all have a passive knowledge of many aspects of our language,
words and constructions, which we understand but never
actively use. An adult's passive vocabulary, for example, typi-
cally includes several hundred words which he understands but
does not use. And most of us can understand certain styles of
language, say the language of the courtroom, which we could
not, however, competently use ourselves. Similarly, young
children understand many things that their parents say to them
long before they can actively produce the same items. That is,
speakers have asymmetrical linguistic systems: they can per-
ceive and understand linguistic distinctions which they do not
(or cannot) themselves make.
A simple experiment by Labov (1969) illustrates this distinc-
tion sharply. He asked Negro youths in the USA, who were
speakers of nonstandard dialect, simply to repeat the sentence:
'I asked Alvin if he knows how to play basketball.' The boys
were unable to repeat this SE sentence, and instead regularly
produced a nonstandard sentence such as 'I axt Alvin does he
know how to play basketball.' That is, they produced a version
which differed from the original in details of surface grammar.
Clearly, the boys had understood the meaning of the original
sentence, since they could immediately and correctly translate
the sentence into their own dialect. Yet they could not produce
the surface grammar of the target sentence. We must therefore
be very careful before we equate the inability to use a particular
grammatical form with the inability to understand it or the
concept which underlies it.

2.5 The implication of such distinctions

These distinctions between different aspects of linguistic com-
petence are not merely academic. They show at once that any
claim to relate 'language' directly to 'education' is almost certain
43
to be so oversimple as to be meaningless. Is one talking about:
comprehension or production? language structure or language
use? prescriptive norms of correctness or appropriateness to
social context? grammatical or communicative competence? the
child's language itself or the school's attitudes to his language? It
should already be clear how oversimple it is to say that a child's
language directly determines his success or failure at school. The
child uses different varieties of language in different social situa-
tions, say home and school. The teacher may (rightly or wrong-
ly) regard the child's language as inappropriate to the classroom.
The child's language may also provoke negative attitudes in the
teacher, perhaps because the child speaks a low-prestige dialect.
These attitudes may be transmitted to the child. Even if the
teacher expresses no overt disapproval of the child's language,
the teacher's own language may still be different from the child's
in the direction of prestige varieties, and this in itself may be an
implicit condemnation of the child's language. The child will be
aware that people with more prestige and authority than him
speak differently, and may draw his own conclusions. Such a
complex of sociolinguistic factors may lead cumulatively to
educational problems for a child.
Thus a child's language may be a disadvantage in his educa-
tional progress: not because his language is itself'deficient', but
because it is different. These distinctions may seem initially to
recall the Feiffer cartoon which runs:
I used to think I was poor. Then they told me I wasn't poor, I
was needy. Then they told me it was self-defeating to think of
myself as needy, I was deprived. Then they told me deprived
was a bad image, I was underprivileged. Then they told me
underprivileged was overused. I was disadvantaged. I still
haven't got a dime, but I have a great vocabulary.
But it is important to ask just how the disadvantage arises. Is it
'in' children's language? Or does it arise rather from people's
attitudes to language differences? If you believe that children's
language can be 'deficient', then you might be tempted to try
and 'improve' their language in some way. If you believe, on the
44
contrary (as this book argues), that the concept of language
deficit does not make much sense, and that there is nothing
wrong with the language of any normal child, then you will
probably believe that schooling should not interfere with chil-
dren's dialects. This is not to say that teachers should not try to
develop and extend their pupils' competence in different
varieties of language. A view that all dialects are valuable is quite
compatible with attempts to extend the functional range of
children's language. After all, we are all developing our com-
petence in different styles of language throughout our whole
lives. And if you believe (as this book also argues) that linguis-
tic disadvantage arises largely from people's intolerance and
prejudice towards language differences, then you will probably
try to change people's attitudes to language (see Ch. 2.1).
Having begun to disentangle some sociolinguistic concepts,
we are now in a position to discuss some specific studies of
language in education.

45
3

Bernstein's theory
of restricted and
elaborated codes

It is an important social fact that working-class (WC) children

do not do as well at school as middle-class (MC) children. This is
not to say that no WC children are successful in the formal
education system. But a WC child has much less chance than a
MC child of, say, reaching university. This fact is not in dispute:
it poses a problem for educationalists and demands an explana-
tion. Nor is there any dispute over the fact that the language
used by WC children is characteristically different from the
language of MC children. This is an elementary observation
about sociolinguistic diversity in an industrialized society like
Britain or the USA.
Much disagreement centres, however, on the interpretation
of these two facts. Are they, for example, causally related? One
type of theory which has been frequently proposed over the past
ten years or so is that the difficulties of WC children at school are
caused by their language. The WC child's language, according to
such theories, is said to be somehow unsuitable for the type of
intellectual or cognitive activities which are the basis of educa-
tion. The gist of the argument over the next two chapters will be
46
that such a proposition is, as yet, unproven. There is no firm
evidence that differences in language are causally and directly
related to differences in intellectual ability. Certainly, no one
has yet satisfactorily explained precisely what features of a
child's language contribute to his educational success. How-
ever, such theories are widely accepted by teachers and taught in
colleges of education, and it is important to understand the
issues involved. They are widely accepted because they have a
certain immediate plausibility, and this is all the more reason for
carefully examining the evidence said to support them.
Note for the present that alternative theories are still open, for
example: the two facts are not causally related - they are both
caused by something else - or, the two facts are related, but not
directly.

3.1 The work of Basil Bernstein

The best known name in the language and education field is
Basil Bernstein, who is Professor of Sociology of Education at
London University. His views have filtered down, often in
out-of-date or inaccurate formulations, through most colleges of
education, and have entered staff room folklore. Most British
teachers have at least a passing acquaintance with the theories
for which he is primarily known. These theories concern social-
class differences in language, and propose causal relationships
between a child's social class, his language and his success or
failure at school. In particular, Bernstein is known for the
concept of 'restricted' and 'elaborated' code, for the theory that
some WC children may not have access to elaborated code, and
for the theory that this partly explains their problems at school.
(The work of Bernstein and his colleagues is published in
three volumes, Class, Codes and Control. To save space, I will
refer to these as CCC 1, 2 and 3. CCC 1 and 3 are Bernstein's
own collected papers. CCC 2 is mainly experimental work by
his colleagues. The papers in CCC 3 do not deal directly with
language and will be referred to only briefly in later chapters.
CCC 1 is mostly easily available in a Paladin paperback, and my
47
page references are to this (the 1973 edition containing a
postscript by Bernstein which is not in the original 1971 edi-
tion). References to CCC 2 and 3 are to hardback editions
published by Routledge & Kegan Paul.)
In fact this particular question about educability (does a
child's language affect his progress at school?) is, for
Bernstein, only part of a much more general question which he
calls the 'structure and process of cultural transmission' (CCC
7, p. 19). Cultural transmission is Bernstein's term for socializa-
tion. The problem is: How do children come to learn what is
socially appropriate behaviour? How do they acquire the ex-
pectations, assumptions and ways of seeing the world of their
family, friends or teachers? By using the term cultural transmis-
sion, Bernstein emphasizes that what the child acquires are
'symbolic orders': ways of organizing experience. This sounds
very abstract, and it is. Bernstein has justifiably complained that
his critics have frequently not recognized just which level of
problem he is dealing with (CCC / , p. 262). It should be clear
simply from Bernstein's terms and concepts (e.g. 'codes', 'sym-
bolic orders') that he is not just talking about social-class
stratification and saying in a roundabout way that WC children
do badly at school.
It is difficult to assess Bernstein's work for several practical
reasons which it is best to clear out of the way at once. First, his
views have altered in crucial respects since his first paper was
published in 1958. It is reasonable of Bernstein to argue that
where research leads is more important than where it started,
but this provides a severe practical problem for readers.
Bernstein has complicated this problem, for he has republished
a collection of his papers (CCC / ) , which span the years 1958 to
1973 and contain many contradictory statements. He has also
published experimental papers (CCC 2) based on his early and
now outdated theoretical position. He admits that this ex-
perimental work is based on a 'much coarser theoretical posi-
tion' than he now holds (CCC 2, p. 36). There is therefore the
danger that casual readers never get past the early papers which
Bernstein now recognizes as inadequate statements of the com-

48
plex relationships between language, social class and educabil-
ity. Second, because of the social and educational implications
of his work, his views have often been seized upon by others in
support of language policies in education. In the course of this,
Bernstein's views have often been oversimplified, misused or
distorted. At certain points below, I comment on how some
people have understood Bernstein, rather than on what he has
himself said: this is the nature of the debate. Third, Bernstein's
style is highly abstract, heavily loaded with sociological terms,
and contains few examples of actual language in use. The work is
therefore often difficult to understand.

3.2 Bernstein's early work

In the out-of-date version in which Bernstein's theories are most
widely known, the argument runs thus. There are two different
kinds of language, restricted and elaborated code, which are
broadly related to the social class of speakers. MC speakers are
said to use both codes, but some WC speakers are said to have
access only to restricted code, and this is said to affect the way
such speakers can express themselves and form concepts. This
is claimed to be particularly important in education, since
'schools are predicated upon elaborated code' (CCC 7, p. 212).
Few detailed examples of the codes were given, but elaborated
code was said to be characterized by accurate and complex
grammar, frequent use of prepositions, impersonal pronouns,
passive verbs, and unusual adjectives and adverbs. Conversely,
restricted code was said to be characterized by short, grammati-
cally simple, often unfinished sentences; frequent use of short
commands and questions, categoric statements ('Do as I tell
you'); simple and repetitive use of conjunctions; rigid and
limited use of adjectives and adverbs.
The theory is appealing because it appears to provide a
linguistic explanation of why WC children are less successful than
MC children at school. In this early version of the theory - to
which Bernstein no longer subscribes - a direct relation was claimed
between social class and the codes to which a speaker had access:
49
a crude correlation between forms of language and social class
(CCC / , p. 226). Since Bernstein no longer holds this view, and
since it has been criticized in detail by others (see Ch. 3.6 for
some references), I will not discuss it here.
What I will do, however, in order to point out some of the
dangers in an unqualified acceptance of this type of position, is
discuss an experimental study based on this early theoretical
position. This will also allow me to give specific illustrations of
some of the concepts discussed in Chapter 2, and to illustrate the
need to develop sufficiently sophisticated sociolinguistic con-
cepts to handle the relationships between language, social class
and social context.

Hawkins's experiment
Consider, then, a much quoted experiment by Hawkins (1969)
who set out to study the relationship between the language of 5
year old children and their social class (MC or WC). He divided
over 300 children into MC and WC on the basis of the occupa-
tional and educational status of their parents. The children were
then interviewed and given tasks to do. One task was to look at a
series of four pictures showing some boys playing football,
kicking a ball through a window, being shouted at by a man and
running away. They then had to tell the story in the pictures.
Hawkins quotes two 'slightly exaggerated' versions of the same
story as told, respectively, by MC and WC children. That is,
although he must have collected over 300 actual stories, he
invents hypothetical illustrations for his argument - he gives no
reason for this. (I will return later to this crucial lack of real
linguistic data in Bernstein's work.) The two versions are as
follows.
(1) Three boys are playing football and one boy kicks the ball
and it goes through the window - the ball breaks the window
and the boys are looking at it - and a man comes out - and
shouts at them - because they've broken the window - so
they run away - and then that lady looks out of her window -
and she tells the boys off.
50
(2) They're playing football and he kicks it and it goes through
there - it breaks the window and they're looking at it and he
comes out and shouts at them because they've broken it - so
they run away - and then she looks out and she tells them
off.

Hawkins argues that the two stories differ in elaboration and in

the amount of information communicated. (1) is explicit about
what happens, but (2) cannot be understood without the pic-
tures and therefore makes greater demands on the listener. This
is because (1) uses many nouns to convey information, whereas
(2) depends largely on pronouns. He argues further that these
'considerable differences' between WC and MC children's lan-
guage 'may well have important cognitive consequences' (CCC
2, p. 91), since WC children have reduced possibilities of
modification and qualification (you cannot modify a pronoun
with an adjective). The data for these arguments are statistically
significant differences in the numbers of nouns and pronouns
used by the two groups of children.
This argument falls down in several ways. Consider this
startling statement:

The second of the versions makes enormous demands on the

listener. It means that the context (i.e. the pictures) must be
present if the listener is to understand who and what is being
referred to. It assumes the listener can see the pictures. (CCC
2, p. 87.)

But the listener (researcher) can see the pictures! He does not
need to be told explicitly who 'they' are, or where they kick the
ball. Version (2) is perfectly appropriate in a situation in which
there is no reason to use elaborated or explicit language. One can
stand Hawkins's argument on its head and say that the WC
version takes appropriate account of the listener's knowledge;
while the MC version is full of redundant information. This is
completely the opposite of Hawkins's conclusion: perhaps it is
the MC children who are incapable of adapting their language to
51
the context, since they treat the researcher as someone who must
have the most obvious things explained to him!
Where has Hawkins gone wrong? Principally, he crucially
ignores the social context in various ways. He fails to discuss
how language is adapted to its context of use (see Ch.2.4). He
studies language use in an experimental interview setting, not
natural language use. Perhaps the WC children were simply
more overawed by the situation than the MC children: remem-
ber they were 5 years old! Using an experimental situation does
not 'control' the context; it just provides a different context with
different rules (see Ch. 5.3). Perhaps the WC children felt less at
home than the MC ones, alone for half an hour with a strange
adult; and perhaps they did not realize the verbal game they
were being asked to play. As Wight (1975) says about such
adult-child contexts: 'What can you tell a man about his picture
that he can't already see and doesn't already know?' The
experiment might indicate that young MC children were better
at role-playing, that is, at pretending the researcher could not
see the pictures. The fact that Hawkins fails to consider the
context means that he fails to consider the crucial distinction
between knowledge of language and use of language (see Ch.
2.4). It could well be that the WC children knew how to use
elaborated language, and would do so if they thought it
appropriate. But this situation failed to elicit such language.
The experiment shows that the WC children used certain lan-
guage; it does not show that they would not have used different
language in different circumstances (if the listener had not
known what was in the picture, for example).
Hawkins's experiment uncovers potentially interesting dif-
ferences in the language of WC and MC children. But, due to the
inadequacy of his socioiinguistic framework, he has no way of
speculating sensibly on what the differences mean.
Note a most important point. Bernstein's own comments on
this experiment are much more guarded and reasonable than the
conclusions Hawkins draws in the original paper. Bernstein
points out that differing use of explicit and implicit language
does not mean that WC children do not have access to explicit,
52
elaborated language (CCC 1, p. 219). What the experiment
shows is differences in the use of language in a specific context,
and Bernstein maintains that it would not be difficult to imagine
a context in which the WC children would produce speech
rather like the MC ones. All the experiment shows is that this
experimental context did not elicit elaborated language from the
WC children. The generalization which is warranted is: 'the
social classes differ in terms of the contexts which evoke certain
linguistic realizations'. To repeat, however, we can draw no
conclusions whatsoever about which contexts do elicit which
styles of language in such children.
Eight years after his 1969 article, Hawkins (1977) published a
full-length book (in a series edited by Bernstein) which presents
a substantial reanalysis and reinterpretation of his 1969 data. In
this book Hawkins takes, like Bernstein, a much more cautious
line and admits that his earlier work had crucially ignored the
social context and the functions of utterances in conversation.
However, in being more cautious, Hawkins may have so
weakened his claims that there is little left to argue with. For
example, he drops all references to children's cognitive abilities,
claiming simply that WC and MC children speak differently
from each other in experimental situations. He also claims that
the children talk differently because of speech differences in
their home backgrounds (pp. 200 ff.). This would be interesting
if it could be documented. But there is no observation to support
the claim: only hypothetical interview data (pp. 64, 202).
Mothers were asked how they would talk to their children in
different situations.
Despite such limitations, the book is valuable. First, it is a
warning that widely divergent interpretations can be placed on
data about social-class variation in language use. Second, it
demonstrates that everything is vastly more complex than a
crude dichotomy between elaborated and restricted code often
implies. Hawkins's analysis is careful and painstaking. What we
are dealing with are differences in the relative frequency of items
which all the children use: not with an absolute difference
between two groups. However, it is precisely such data which
53
can be taken as evidence against the codes theory. If there are no
absolute differences, if it is all highly context-dependent and all
a matter of relative frequencies in usage, then this provides no
support for the concept of two discrete underlying codes.
Note one interesting suggestion from Bernstein about the
relevance of such experimental situations to the school situation
(CCC 7, p. 278). He argues that there is a broad analogy
between experimental settings and the test situation in school:
similar social assumptions underlie them. They are both situa-
tions in which the child is in an abnormal setting, isolated from
other children, with an adult who provides minimal support,
and where he must do tasks very different from what he is used
to outside school. For reasons which are not yet known, MC
children are more successful (i.e. do what teachers/experimen-
ters want) in both such settings. One obvious conclusion to draw
from this is, however, that it is schools which should change, not
children. (Bernstein does not draw this conclusion.) One danger
is, of course, that a strange or hostile experimental or test
situation may be taken as evidence of a child's total linguistic
ability (see Ch. 4.3). Bernstein does not make this mistake but it
is something to guard against. These differing interpretations of
the same data emphasize the need for extreme caution in
drawing general conclusions about the relationships between
language, social class and cognitive ability.
Note, then, that Bernstein no longer talks of a direct relation
between language and social class. He uses two intervening
concepts: (a) he distinguishes language from use of language;
and (b) he distinguishes different social contexts in which lan-
guage is used.

Language forms and language functions

Turner (1973, CCC 2) gives a precise linguistic example of this
crucial point that it is not possible to make sense of social-class
differences in language use unless one looks at the social con-
texts which control language. For example, one way of threaten-
ing someone is to use a grammatically complex utterance such
54
as: 'If you do that again, I'll hit you.' In an experiment with
young children, Turner found that WC children used more
threats, and therefore more grammatically complex language
than MC children. Note also that this form of threat is gramma-
tically very similar in form to rational appeals of the type: 'If you
eat that now, you won't want your dinner.' These examples
make clear the importance of not confusing grammatical com-
plexity of form with the social function of the language. There is
no simple correspondence between the complexity of language
(e.g. at the grammatical level) and the complexity of functions,
cognitive or social, which it serves.
In a relatively early paper (1965), Bernstein himself appears
critically confused about the relation between linguistic form
and cognitive processes. He quotes two short extracts of res-
tricted code. These are the only examples of restricted code
of more than a couple of sentences in length in the whole of
CCC / , so it is important to look at them closely. Here is one
of them:
It's all according like these youths and that if they get into
these gangs and that they must have a bit of a lark around and
say it goes wrong and that they probably knock someone off I
mean think they just do it to be big getting publicity here and
there. (CCC 7, p. 158)
Bernstein does not comment on the details of the extract, but
says it will illustrate various points about restricted code: name-
ly, that there is a 'rigid range of syntactic possibilities', leading
to 'difficulty in conveying linguistically logical sequence and
stress' (p. 157).
Note first that we have no information about the context of the
language, except that the speaker is 17 years old. Nor are we
given any information about the intonation, or changes in speed
or stress which often structure and organize spoken language.
Let us concentrate, however, on the words on the page. Clearly,
the language is neither particularly memorable nor particularly
effective. But it certainly does not show thoughts 'strung
together like beads on a frame rather than following a planned
55
sequence' (p. 158) or any difficulties in handling a logical
argument. There is no inability to convey an abstract argument
involving hypothetical cases and interdependent propositions.
(If some youths get in a gang and if something goes wrong when
they are having a lark about, then someone might get killed.)
There is no problem in understanding what is meant, as this
translation into more formal (written) style indicates. There-
fore, the language is not tied to its context. There is, therefore,
no support for the contention that such language 'initiates and
sustains . . . radically different . . . conditions of learning' (p.
157).
In his early papers, Bernstein claimed that restricted code has
a 'logically simpler language structure' (CCC 7, p. 47) than
elaborated code. This kind of statement (which confuses com-
plexity of linguistic form with complexity of thought) is quite
misleading and Bernstein no longer makes it. (Gazdar, 1979,
provides a thorough analysis of this confusion.) But in later
papers (e.g. 1973) he still insists that different forms of language
have different cognitive effects, since they 'focus experience' in
different ways (1973, p. 211). If, however, such statements
cannot be related to actual language in use, then it is not clear
what value they have.

3.3 Bernstein's later work (1973)

A major difference between Bernstein's earlier and later work is
an increase in the depth of complexity and abstraction at which
the theory is formulated. I will summarize the position which
Bernstein proposes in 'Social class, language and socialization',
his most recent paper (1973) in CCC 1. It is still the work on
sociolinguistic codes for which Bernstein is best known, yet he
has not published any papers on the codes since 1973, although
he has published papers on other topics: a fact noted by several
commentators. Nor has he published any overview which
attempts to reconcile the contradictions in his work between
1958 and 1973.
In Bernstein's first published paper (1958) he talks of public
56
and formal language (these terms were later changed to re-
stricted and elaborated code respectively) and discusses these as
observable speech varieties. And in the earliest paper on linguis-
tic codes (1962) he implies that restricted and elaborated codes
can be partly defined in terms of features of language structure
(e.g. CCC 7, p. 93). An elaborated-code user has more alterna-
tives to choose from, while a restricted code is more predictable
in its syntactic choices. But by the 1965 paper quoted above,
Bernstein no longer defines the codes as actual language
varieties. They are defined as abstract frameworks at a psycho-
logical level of verbal planning, and 'only at this level could they
be said to exist' (CCC 7, p. 154). Nevertheless, there is the
implication that by inspecting actual extracts of speech, one can
tell whether they are related to elaborated or restricted code
(e.g. CCC / , p. 158, and see the extract quoted and discussed
above, p. 55). In the later work, then, the relation between the
codes and what people actually say is defined as much more
complex and abstract.
Bernstein now distinguishes between sociolinguistic codes
(elaborated and restricted) and speech variants (elaborated and
restricted) which realize the codes (CCC 7, p. 200). The codes
are now defined as entirely abstract, underlying, interpretive
procedures which 'generate' different speech variants in dif-
ferent contexts. Thus Bernstein no longer claims a direct rela-
tionship between a speaker's social class and the codes he uses.
He has introduced two intervening concepts of speech variant
and context.
Speech variants are not actual language either. They are
defined as 'contextual constraints upon grammatical-lexical
choices' (CCC 7, p. 200). No detailed specification is given of
these choices, but it is implied (p. 203) that restricted speech
variants are characterized by a reduced range of syntactic
alternatives and a narrow range of lexis (vocabulary): that is,
they are defined in the same way as codes earlier were. Consis-
tent with this increase in abstraction, speech variants are not,
however, defined in terms of linguistic forms, but in terms of
meanings. Thus, elaborated speech variants are said (p. 202) to
57
realize universalistic meanings, i.e. meanings which are explicit
and not tied to a given context (p. 199). Conversely, restricted
speech variants realize particularistic meanings which are implicit
and take for granted many shared meanings between speaker
and hearer. (Compare Bernstein's comments, quoted above, on
the Hawkins experiment.) It will be clear that this increase in
abstraction has led Bernstein a very long way from observable
language forms.
In order to discover whether someone is an elaborated or
restricted-code user, one has to look at the language he or she
uses in what Bernstein calls the four critical socializing contexts
(CCC 1, p. 206): regulative, e.g. being told off by mother;
instructional, e.g. the classroom; imaginative, e.g. in play; and
interpersonal, e.g. in talk with others where the child is made
aware of emotional states. If the linguistic realizations of these
four contexts are 'predominantly' in terms of restricted speech
variants, then the deep structure of the communication is said to
be a restricted code. Conversely for elaborated code.
Note one most important point which has caused much
confusion. Codes are defined as abstract, underlying principles
which regulate communication and generate speech. People do
not speak codes, just as they do not speak grammar: both
grammar and codes are abstract, underlying systems. It is
therefore incoherent to refer to a child as a 'restricted-code
speaker'. One can only talk of a child who tends to use restricted
speech variants in certain contexts.
Bernstein further adds to the complexity of the theory by
distinguishing two family types (p. 209). In positional families,
there is said to be clear-cut definition of the status of different
members of the family, as 'father', 'grandmother' and so on. In
person-centred families these status distinctions are blurred in
favour of members' unique characteristics. Bernstein claims,
without giving any details, that the communication structure in
these different family types is differently focused, such that we
should expect restricted code in positional families and elab-
orated code in person-centred families (p. 211). He claims that
both family types can be found in both MC and WC, but implies
58
that, at the present time, positional families are more character-
istic of the WC.
What has Bernstein now said? What is the form of his theory?
He is attempting to formulate a theory which relates a child's
social class, family background, language use and cognitive
style. He still claims (CCC 7, p. 209) that access to the codes is
'broadly related to social class' and that there 'may well be
selective access to elaborated code' (p. 208); that is, some WC
speakers do not have access to elaborated code but most (?) MC
speakers have access to both. This is because (so runs his
argument) there is selective access to the role systems which
evoke the use of the codes (p. 208). Within positional families we
should 'expect' restricted code (p. 211). But both positional and
person-centred families are found in both WC and MC (p. 209).
Finally, to know if a given piece of language realizes restricted or
elaborated code, we should have to observe the language use
across the four critical socializing contexts - for an elaborated
speech variant (defined in terms of meanings, not observable
linguistic forms) may realize a restricted code, and vice versa.
From the way I have summarized this position, it will be clear
that I believe the theory is now formulated in such a way that no
testable claim is now being made. The model apparently being
proposed might be summarized thus:

social classes: working class middle class

family type: positional person-centred

codes: restricted elaborated

speech variants: restricted elaborated

59
The arrows represent links between levels. The vertical col-
umns represent 'expected' links. Thus WC families may be
expected to be positional, but they may be person-centred. A
positional family would lead one to expect code restriction, but
if it should show (unspecified) signs of being person-centred,
then the children might be able to switch codes.
When the theory is set out in this way, it becomes clear that it
is not a real theory. No real predictions are made, for example,
about which forms of family are related to which codes. All we
have are 'expectations', unsupported by evidence. More im-
portant, it is not possible to refute such a theory from em-
pirical evidence and observations. Genuine scientific theories
make predictions that certain things (X) will happen under
certain conditions, and that other things (Y) will not happen.
The statement may involve specifying a statistical probability
that X will occur: that is, a definable margin of error may be
specified in the prediction. If Y does happen, the theory must be
altered or abandoned, since it does not fit the facts. But Bern-
stein's model sets no real constraints on what may happen. An
example of an elaborated speech variant may be expected to be a
realization of elaborated code, but it may realize restricted code
too. We can only know which code it realizes by investigating
whether the critical socializing contexts of the speaker were
predominantly characterized by restricted or elaborated speech
variants - but this is uninvestigable.
Gordon (1981, p. 81) has made the following valiant attempt
to give a succinct statement of what Bernstein has actually
claimed in two papers published in 1970 and 1971:

social class family structure roles modes of early

socialization roles . . . modes of perception . . . access to
codes codes speech educational attainment

An arrow indicates a claimed causal link; the dots indicate a link

which is irremediably obscure in Bernstein's writings. Gordon
notes that 'roles' occurs twice, and that the whole chain of
causality is particularly obscure after the second occurrence. He
60
notes also that the first and last entities in the chain are simply
taken for granted.

The lack of linguistic analysis in Bernstein's work

The major limitation in Bernstein's work, from the point of view
of its ability to make precise and testable statements, is the
almost total lack of linguistic exemplification of his theories. Most
of the papers are entirely abstract speculations with no precise
linguistic data. Where Bernstein does provide brief linguistic
examples and data, this is either (a) invented, i.e. imagined,
fictional anecdotes (e.g. CCC 7, p. 201), or (b) hypothetical,
e.g. mothers were asked what they would say if their child was
naughty, or (c) taken from experimental test situations, as in
Hawkins's experiment. There are no extended real-life exam-
ples of language use anywhere in Bernstein's papers: no more
than a couple of conversational fragments in footnotes (e.g.
CCC 7, p. 283), and no examples whatsoever of actual language
use between mothers and children in the home or between
teachers and pupils in the classroom.

3.4 Do the codes exist?

More radically, one has to question the linguistic evidence for
the very existence of the codes. Early on in his work, Bernstein
dropped the notion that the codes exist as observable language
varieties. The only sense in which the codes could now be said to
exist is as hypothetical constructs intended to explain social-class
differences in language use and cognitive orientation. More
precisely, then, are the codes a useful hypothesis? Do they, for
example, explain the experimental evidence on language use
which Bernstein and his colleagues have collected? And the
answer is: No.
The experimental evidence that is said to support the concept
of code is always of the form that one group of speakers (say MC)
tends to use more of a certain kind of language than another
group (WC). The differences are never absolute (cf. Hawkins's
61
experiment). What the data display is one group using more of
the same features than another group. Both groups use the same
linguistic items (e.g. complex noun groups) but one group uses
them more frequently. On the basis of these relative differences
in frequency, Bernstein nevertheless argues for a sharp contrast
between two underlying modes of speech: restricted and elab-
orated. On the basis of a measure of relative frequencies, how-
ever, a notion of a continuum might be more appropriate: a cline
of continuously variable speech behaviour.
There is also a more fundamental worry over such counts of
relative frequency in the use of linguistic items. Suppose in 5
minutes' speech a WC child uses simple pronouns most of the
time (he, she, it, etc.) and just once uses a complex noun group
(e.g. the very old man), then this proves that he or she does know
how to use complex noun phrases, not that he or she doesn't.
Statistical statements about language are therefore always evi-
dence against fundamental differences in language capacity, not
for them. All the empirical work (in CCC 2) shows is rather mi-
nor differences in relative frequency of some grammatical items,
between WC and MC groups. Trudgill puts this succinctly:
If we look at things from a purely linguistic point of view - all
the theorizing of the past sixteen years appears to have
reduced to evidence that, in situations more artificial and
alien to them than to middle-class children, working-class
children use a higher proportion of pronouns. Is this what it
has been all about? (1975b)
Various writers (e.g. Trudgill, 1975b; Labov, 1969; Kochman,
1972) have for this reason suggested that Bernstein is here
dealing with mere differences of style.
Bernstein may, however, have hit on a more important
difference than this. It may be that in choosing characteristic
styles of language, WC and MC speakers are reflecting different
preferred modes of discussion and therefore different value
systems about what it is important to elaborate and make
explicit. Educationally, this would still be a very important
thing to demonstrate. But still, one has to emphasize that value
62
systems cannot be read unproblematically off statistical features
of language use. And what one certainly cannot argue is that
such language use constrains cognitive orientation in any abso-
lute sense. Such statistical data are also evidence against the
notion that language can constrain a speaker's thought, since
one cannot be constrained by mere tendencies in one's language
(cf. Jackson, 1974). What has to be explained is why WC
children do not frequently use linguistic forms they quite clearly
know (and which tend to be valued by teachers).
I am here drawing attention again to the distinction between
knowledge of language and use of language (see Ch. 2.4). It is
important to realize that Bernstein claims always to be talking
about speech and therefore use of language. But one is still left
with the complete lack of explanation, within the theory, for
why different social groups use different forms of language in
different contexts. There is, for example, no attempt to analyse
the different values which different social groups place on
different forms of language. In other words, the experimental
evidence on differences between MC and WC speakers' use of
language is important. But it is unclear how this empirical
evidence is related, if at all, to the codes theory.
I am suggesting that Bernstein's theory has an unclear and
unacceptable status. First, it is now formulated in a way which
makes it untestable. Second, it has a loose, if not actually
contradictory relationship to the experimental evidence said to
support it. The experimental tests have uncovered interesting
social-class differences in language use, but they are not tests of
the theory. And the theory does not explain the experimental
evidence.

3.5 Some possible confusions

Educationalists have often used Bernstein's work in support of
educational policies. (It is clear that Bernstein's theories have
educational implications, but no one seems quite clear just what
these implications are!) Bernstein's work has been used, for
example, to support the over simple statement that 'educational
63
failure is linguistic failure'. This is unfair to Bernstein. It is true
that the most apparent concomitant of educational failure is
often a pupil's language. This is what is immediately visible and
audible to the teacher, and often the only evidence of academic
ability that a teacher has to go on. But beyond a pupil's
language, Bernstein emphasizes a highly complex set of factors:
family background, social class and so on. The child's language,
in so far as this is observable to teachers, is literally the most
superficial aspect of his sociolinguistic competence.
It is important to make explicit some features of language
which must not be confused with the notion of restricted code
(cf. Trudgill, 1975a, Ch. 5). First, it is clear that the mismatch
between school and pupil is not simply due to a pupil's lack of
words. A child may not have a well-developed vocabulary, but
this cannot constrain his thinking in any absolute way, since he
can simply learn more words (as we all do all our lives). Any
vocabulary can be expanded, and a child with a poor vocabulary
may still be a fluent speaker if he has other linguistic strategies to
compensate. Nor are the child's problems due to impoverished,
simple or deficient grammar. There is no evidence that the
language structure used by socially disadvantaged groups is less
complex than that, say, of SE (see Chs. 2.2 and 4.1). Nor are
such educational problems due directly to nonstandard dialect.
One major source of confusion has been to identify nonstandard
dialect with restricted code. Bernstein himself is quite clear that
these are quite distinct concepts, and that one can have elabor-
ated and restricted variants of both standard and nonstandard
English:
There is nothing, but nothing, in the dialect [of nonstandard
speakers] as such which prevents a child from internalizing
and learning to use universalistic meanings. (CCC / , p. 224)
(One might add, however, that Bernstein himself is partly to
blame for this confusion, since his rare examples of restricted
code have often been in nonstandard dialect.)

64
3.6 Conclusions
Bernstein's work is widely taught on courses in education and
sociology, and is accepted as established fact by many people,
but his theories have now attracted a considerable amount of
fundamental criticism. Among the most critical reviews, the
reader is referred to: Labov (1969), who argues that no detailed
specification has been given of the central concept of code, and
that the experimental results are artefacts of the experimental
situations; Rosen (1973), who questions Bernstein's concept of
social class and the lack of linguistic data; Jackson (1974), who
argues that Bernstein's work 'fails by rather obvious intellectual
tests' in that the theory is untestable and unrelated to linguistic
evidence; Trudgill (1975b, quoted above), who questions the
lack of linguistic exemplification and argues that the language
differences Bernstein found are simply differences in style; and
Gordon who calls the work 'pseudo-theory' and 'fundamentally
unscientific' (1981, p. 73). Bernstein replies to some of his
critics in the introduction and postscript to CCC 1 (1973), and
the introduction by Halliday to CCC 2 is a statement by a
linguist more sympathetic to Bernstein's argument. There is
clearly still room for debate (see Ch. 4.4).
For alternative reviews of this area, readers are referred to:
Dittmar (1976), who provides a Marxist critique of Bernstein's
work (and also Labov's, see Ch. 4) and an annotated compre-
hensive bibliography of work up to 1976 on the concept of verbal
deprivation; J. R. Edwards (1979), who provides further refer-
ences up to 1979; Stubbs (1980, Ch. 7), who provides a slightly
broader critique of the debate over verbal deficit than is given
here; and Gordon (1981), who provides an excellent brief
summary of Bernstein's work and makes explicit some assump-
tions underlying verbal deficit theory.
No critic of Bernstein's has ever denied that there are social-
class differences in language, or that these differences are
somehow related to educational problems faced by WC children.
What is in dispute is the nature of the relationship. MC children
tend to score higher than WC ones in tests of academic ability;

65
and MC language is likely to be closer to the standard language.
Both these propositions may be true, but it remains to be
demonstrated that they are causally related.
The main point of my own discussion of Bernstein's work has
been to urge extreme caution in interpreting the findings (which
are extremely interesting in themselves) of social-class differ-
ences in language use. It is not clear that the work provides any
explanation for these differences, and alternative explanations
are still open. Their importance might lie, for example, in
people's attitudes to such differences, rather than in the differ-
ences themselves (see Ch. 2.1).
Often Bernstein's work has been taken simply as further
evidence of inequalities in the education system. This is unfair,
and treats the work at a trivial theoretical level. Atkinson (1981)
provides a detailed argument that Bernstein's work is to be
interpreted as a contribution to the tradition of European
structuralist sociology, and that his work on sociolinguistic
codes is only one part of this. But theories have to explain data,
and if the data do not fit, then it is the theory which has to be
altered or abandoned. It seems, therefore, that Bernstein's
theories will remain uncorroborated until they can be formu-
lated in such a way that they can be closely related to linguistic
data collected by observation and recording in homes and
classrooms. Chapter 8 will suggest how some of Bernstein's
concepts can be related to observations inside classrooms, and
will discuss how his later work seeks also to show how social
class acts on the distribution of knowledge in society and
thereby relates to social power and control.
At present, however, the naturalistic data which are available
from other sources only adds to the difficulty in interpreting
Bernstein's abstract model. This is the topic of the next chapter.

66
 4

Labov and the

myth of
linguistic deprivation

William Labov is Professor of Linguistics at the University of

Pennsylvania, USA. Since the mid-1960s, he has revolutionized
the study of sociolinguistics. He has published important work
on the way a speaker's language reflects his or her social class,
and descriptions of varieties of nonstandard English used by
different ethnic groups in the USA, for example Blacks and
Puerto Ricans in New York. Many of his papers are on technical
aspects of sociolinguistics and language change, but he has also
written many influential articles on language in education,
including articles on the relationship between reading problems
and nonstandard English and peer-group membership, and oral
story-telling styles. Several of his articles on educational topics
are collected in Labov (1972a), and other theoretical papers on
language and social class are collected in Labov (1972b). A great
deal of work has been published in both the USA and Britain to
develop Labov's work on the relation between varieties of
English and educational problems. Outstanding contributions
include Macaulay (1978) on language and education in Glasgow;
Milroy (1980) on nonstandard English in WC areas of Belfast;
67
and V. K. Edwards (1979) and Sutcliffe (1981) on British Black
English.
Labov's work gains its power from two sources. First, it is
based on long-term, intensive fieldwork and participant observa-
tion in the speech communities he has investigated. Characteris-
tically, Labov has tape-recorded speakers, not only in inter-
views, but in situations which were as natural as possible. To
study the language of Black adolescents in the urban ghettoes of
New York, for example, he spent time getting to know his
informants, used a Black colleague as one of the investigators,
and observed and recorded his informants in their usual sur-
roundings in Harlem, in the gangs with which they spent their
time. Second, his arguments are based closely on detailed
analyses of the actual language recorded: not on surface analyses
of a few dialect features, but on analyses of nonstandard dialects
in depth, as self-consistent language systems. Labov's work is
therefore important for its linguistically detailed analyses of
spontaneous language collected as far as possible in its natural
social context of use.
One of Labov's most important papers is 'The logic of
nonstandard English' (1969). If I had to recommend one single
paper to readers of this book it would be this one, and a
summary of the main linguistic concepts in this paper provides a
good introduction to an approach to language in education
which is very different to the approach we saw in Chapter 3. The
central concern of the article is to show that the concept of verbal
deprivation is a myth, unsupported by linguistic and anthropo-
logical evidence. To show this, Labov discusses the misunder-
standings which are possible about the relationships between
language, concept formation, explicitness and logic.

4.1 Languages, logic, explicitness and grammar

Labov points out a first possible confusion between logic and
explicitness. A criticism often raised against pupils' speech by
teachers is that it is 'badly connected' and inexplicit. Teachers
68
often feel this about Black English Vernacular (BEV) which has
sentences like: 'he my brother' (SE: 'he's my brother'). But
there are many languages which do not use the verb to be in such
sentences, for example Russian: 'on moj brat' (literally: 'he my
brother'). Other languages which use no verbal connective in
such sentences include Hungarian and Arabic. Such languages
may be foreign, but they are not inexplicit because the verb to be
is not used in certain types of sentences. Further, it would be
ludicrous to argue that a Russian had a defective concept of
existential relationships, just because of this detail in the gram-
mar of his language. One must therefore be quite clear that if we
teach children to insert connectives and use standard syntax in
such sentences, then we are only teaching them slightly dif-
ferent forms of surface grammar. We are neither developing
their logic or concepts, nor teaching them to be explicit.
A comparable example occurs with BEV forms like: 'He come
yesterday' (SE: 'he came yesterday'). Failure to mark explicitly
the past tense in the verb does not indicate a failure to perceive
past time. It merely means that in BEV come is in the same class
as verbs like put and hit in SE (cf. 'I always put it there, I put it
there yesterday').
It is also easy to confuse logic and grammar. Many non-
standard dialects of British and American English use double or
multiple negatives such as: 'I don't know nothing' (SE: 'I don't
know anything'). It is sometimes said that such sentences are
illogical on the grounds that if I don't know nothing, then I do
know something. One might call this the pseudo-algebraic view
of language: two negatives make a positive. (Proponents of this
view never explain why the two negatives might not add
together to make a stronger negative!) And again, many lan-
guages use double negatives (e.g. French: 'je N'en sais RIEN'.
Spanish: 'yo NO se NADA'). Again, these languages may be
foreign, but they are not illogical just because they often use two
participles to negativize a sentence. And within such languages,
including BEV, the use of such double negatives is regular and
rule-governed.
We must be careful, then, not to confuse logic, grammar and
69
explicitness, or to confuse the conventions of SE grammar with
universal canons of logic and thought.

The complexity of grammatical rules

In general, Labov points out that we have almost no knowledge
of the cognitive correlates of grammar. One reason for this is
that the grammatical system which a normal native speaker of
any dialect or language commands is much more complex than
most people realize, and much more complex than the type of
conceptual operation typically tested in 'intelligence tests'.
Labov gives the following example.
All speakers of any dialect of English can correctly use words
such as any, one and ever, although the rules governing them are
highly complex. One cannot, for example, use anyone in sen-
tences with a simple past or progressive tense. That is, one
cannot say: 'Anyone went to the party' or 'Anyone is going to the
party'. These constraints mean that speakers know uncon-
sciously that anyone has a feature (+ hypothetical), since it is
possible to say: '//anyone went to the party,. . .'
Another complex and abstract property of anyone is that it has
a feature (+ distributive). That is, if we want just one more
player for a game, we can say to a group of people: 'Do any of
you want to play?' We do not, in this case, say: 'Do some of you
want to play?' With any we are considering a group as indi-
viduals, that is, distributively.
The importance of such analyses of the details of language is
that they reveal the cognitive complexity involved in any speak-
er's use of language. In particular, it becomes absurd to accuse
any dialect of any language of being logically deficient, when any
dialect is based on many hundreds of such rules involving such
abstract underlying concepts and principles.

Nonstandard dialects as consistent linguistic systems

Speakers of a standard language often interpret grammatical
deviations from the standard as random 'errors' due to 'ignor-
70
ance', 'carelessness' or 'slovenly' speech. Labov exposes this
faulty equation of linguistic competence with SE. It is all too
easy to assume that if someone's language is different from SE,
then it is deficient. But there is no linguistic evidence for equating
such differences with deficit. The work of Labov and others in
the Black speech communities of the inner city areas in the USA
has documented in detail that BEV is a coherent, systematic,
highly structured, rule-governed linguistic system. The notion
that the language of socially improverished groups is 'deficient'
or structurally underdeveloped rests on a serious misunder-
standing of the nature of human language (cf. Ch. 2.2). In fact,
BEV is very closely related to SE with which it shares the bulk of
its grammatical and lexical systems. And the differences be-
tween the two language varieties are systematic, not random. All
the evidence demonstrates that when any child (unless severely
mentally retarded) comes to school at the age of 5, he has control
of a complex linguistic system, so complex in fact that linguists
are not yet able to describe it fully.
Teachers may often overestimate the differences between
their pupils' nonstandard language and SE. In a study of
Glasgow speech (which replicates some of Labov's work in New
York), Macaulay (1978) found that many teachers believed that
their WC pupils spoke 'two languages': 'English' in the class-
room and 'Glas^yegian' elsewhere. But the children could not
be held to speak two languages under any reasonable definition
of 'language'. There is not even any evidence of a Glasgow
dialect, distinct from other varieties of English. At most chil-
dren shift in style inside and outside the classroom, and these
style-shifts are mainly in pronunciation, with a few superficial
grammatical and lexical differences. Pronunciation or accent is,
of course, literally the most superficial and easily observable
aspect of language, and often attracts more attention than it
merits. (Consider again in this connection the debate over
whether Bernstein has put his finger on fundamental differences
in 'code' or has merely documented stylistic differences between
WC and MC speakers. See Ch. 3.4.)

71
4.2 Nonstandard languages as media of education
When it is realized that nonstandard languages and dialects are
highly complex, coherent language systems, based on many
hundreds of abstract patterns and regularities, it must also be
admitted that there is no linguistic reason why such language
varieties should not be the media of instruction in schools. It is
believed in many speech communities that nonstandard dialects
are fit only for non-serious purposes (such as casual conver-
sation, telling anecdotes, joking) but are not suited to discussion
of intellectually demanding topics. Teachers have been known
to assert that There are things that you can't say' in such-and-
such a dialect, implying that the dialect is deficient in vocabu-
lary. But the vocabulary of English is a resource available to all
its dialects: any word can be used in any nonstandard utterance.
Characteristically, there are stylistic constraints on the co-
occurrence of items, so that sentences such as: T h e m linguists
don't know nothing about morphophonemics' sound odd. But
such stylistic conventions cannot constitute a constraint on the
speaker's thinking. There is no evidence, then, that any non-
standard language variety is, in itself, an unsuitable medium for
education and intellectual discussion.
Strong, independent confirmation of this comes from work
on pidgin and creole languages (see Hall, 1972). Pidgins are
languages which develop as simplified versions of a source
language (such as English, French and Portuguese) in restricted
situations in which speakers of the source language have to
communicate with a native population. Thus, Melanesian Pid-
gin English developed from English, and pidgins in Haiti,
Martinique, Guadeloupe and Mauritius developed from
French. Such pidgins are much simpler in grammar and lexis
than their source language, irregularities being reduced or
eliminated. And functionally a pidgin is much simpler than its
source, being used only as a lingua franca in limited trading
situations, and by definition not being a native language of any
of those who speak it. But although structurally simpler than
'normal' languages, pidgins are still real languages, governed by

72
complex rules which are difficult to learn. (The commonsense
meaning of 'pidgin English' as a crude communication system
with no proper grammar and only a few dozen words is very
misleading.) A Creole is a pidgin which has then become the
native language of a group of people. Structurally (or linguistic-
ally) it may be identical to the pidgin from which it developed.
But functionally it now serves a group in all their everyday
communication, and is learned by children as their native
language.
Despite their relative structural simplicity, pidgins and
Creoles are quite satisfactory media for conveying complex
technological information, and manuals in fields such as medi-
cine have been prepared in them. The belief of policy makers
that such languages are not suitable as media of instruction is not
therefore supported by the linguistic evidence. Creoles are
perfectly satisfactory media of communication in schools. In-
deed, it is only sensible in many cases to make children literate
initially in the only language they know (Stubbs, 1980).
Note, however, that the local population themselves, who
speak a Creole as their native language, may hold the Creole in
such low esteem that they do not wish their children to be taught
in it. It is a finding of sociolinguistic studies in different parts of
the world that speakers of social groups who are low in prestige
relative to another social group tend to downgrade their own
ethnic-linguistic group. Thus both English Canadians and
French Canadians regard Canadian French speakers less
favourably than English speakers, both groups sharing the
community-wide stereotype of French Canadians as relatively
second-rate people (Lambert, 1967). And speakers of BEV
(Mitchell-Kernan, 1972) and of Glaswegian English (Macaulay,
1978) have very ambivalent attitudes towards their own lan-
guage. Although it symbolizes group loyalty for them, they tend
to regard prestige SE as 'better'. As always in the sociolinguis-
tics of language in education, speakers' attitudes are crucial.

73
4.3 The myth of linguistic deprivation
One reason for supposing that Negro children from the ghetto
areas have deficient language has been the assumption that they
do not receive enough 'verbal stimulation' in the home. But
Labov has shown that this too is false, and that Negro children
typically hear more well-formed sentences than MC children and
participate in a highly verbal culture. The high value placed on
verbal skills in Afro-American culture is now well documented
by field studies (e.g. Labov, 1972a; Kochman, 1972). Expertise
in speech is much more highly valued than in MC culture (which
places a high value on written language) and the Black commun-
ity has a rich oral tradition quite alien to Whites.

Experiments on language 'deprivation'

If direct observation of Black speech communities immediately
dispels notions of 'linguistic deprivation', where do such no-
tions come from? One source is artificial experiments with
mothers and children. In a much quoted study, Hess and
Shipman (1965) investigated how 163 Negro mothers from
different social-class backgrounds set about teaching their 4
year old children tasks, such as sorting toys by colour and
function. They found differences in the teaching strategies of
mothers in different social-class groups, and claim that such
data provide evidence of 'cultural deprivation' in lower social-
class groups which 'depresses the resources of the human mind',
due to a 'lack of cognitive meaning in the mother-child com-
munication system'. I have argued above that such concepts are
not supported by linguistic or anthropological evidence, so what
evidence do Hess and Shipman provide, and why do I find it
unconvincing?
Hess and Shipman found that after the teaching sessions the
MC children had learned the tasks better than the WC children,
and they attempt to explain this by reference to the mothers'
language. They found that the MC mothers talked more to their
children, used more abstract words, and more syntactically
74
complex language. They claim that these differences reveal a
'deprivation of meaning' in the WC mother-child relationship.
But as we have seen already (Ch. 3.4), such comparisons cannot
demonstrate that a group of speakers is 'deprived' of anything.
The WC mothers did talk to their children, did use abstract
words, and did use syntactically complex language. (In fact, the
differences quoted between the highest and lowest status groups
are not very large. For example, mean sentence length was 11*4
words for the college educated group and 8*2 words for the WC
group from an unskilled occupational level with fathers absent
and on public assistance. No levels of statistical significance are
quoted, which makes such figures impossible to interpret.)
Therefore, the experiment shows that WC mothers do have
access to the meanings which can be conveyed by such linguistic
resources - they are not 'deprived of meanings' at all. In the
experimental situation, WC mothers apparently chose not to
express such meanings quite as often as MC mothers, but that is
quite a different matter.
In fact, the statistical linguistic analysis appears to bear no
relation to the mother-child dialogues which were the data.
Hess and Shipman quote some extracts from these dialogues.
One mother explained a task to her child as follows:
All right. Susan, this board is the place where we put the little
toys; first of all you're supposed to learn how to place them
according to colour. Can you do that? The things that are all
the same colour you put in one section . . .
Another mother was much less explicit:
M: I've got some chairs and cars, do you want to play the
game? OK, what's this?
C: A wagon?
M: Hm?
C: A wagon?
M: This is not a wagon. What's this?
Clearly, the second teaching style is less adequate than the first.
And I imagine that the difference in educational effect of the two
75
mothers' teaching styles is considerable, if this is indeed how
these children are regularly spoken to. But the inadequacy has
nothing to do with average sentence length or the number of
complex sentences! The second mother simply does not explain
what the child should do. What does require explanation is why
WC mothers, in this situation, are not motivated to explain
things in sufficient detail in their children. The superficial
analysis of language form which Hess and Shipman provide is
largely irrelevant to this question.
As always with experiments in laboratory settings, there is
also the criticism that the highest status group (defined here as
college educated) were simply more at home in the test situa-
tion. All the mothers were brought to the university for testing,
although only the college educated group would ever have been
inside such an institution before. The differences in WC and
MC teaching strategies might be an artefact of the test situation
itself. Certainly, the tests were heavily culture biased in favour of
the MC mothers and children, since they involved playing with
the kind of toys much more likely to be found in MC homes.
Always in work of this kind we come up against the danger of
interpreting speakers' use of language in a strange or threaten-
ing context (a 'test') while doing unfamiliar tasks, as their total
verbal capacity. We must always be particularly critical of the
sources of data in such experiments, and ask whether the
apparent linguistic incompetence of certain social groups is not
in fact generated by the research itself. Such research, in other
words, fails to take into account that the social situation of the
research itself is likely to be a crucial determinant of data (see
Ch. 5.3).

The social context as determinant of language use

A major finding of sociolinguistics is that the social context is the
most powerful determinant of verbal behaviour. Fieldwork
with Black children (e.g. Labov, 1969) has shown that they
produce vivid, complex language in unstructured situations
with friends, but may appear monosyllabic and defensive in
76
asymmetrical classroom or test situations where an adult has
power over them. Philips (1972) has found exactly the same with
American Indian children: that they are expressive outside the
classroom, but silent, reticent and defensive inside it with their
White teachers (see Ch. 7.5).
The very notion of 'testing' someone by asking him questions
to which there are correct verbal answers is a strongly culture-
bound notion not shared by many social groups. Thus the type
of behaviour found by Hess and Shipman may merely represent
the different sociolinguistic constraints which control language
in different social groups. Certainly there is little connection
between the kind of language demanded by such 'tests' and the
kind of language valued by WC Negro culture. So it simply does
not make sense to take language used in tests as a measure of
total language capacity - and far less as a measure of cognitive
capacity. Interview, interrogation, test or classroom situations
simply do not tap the verbal ability of such speakers.
It is important to realize also how different social groups may
react in widely different ways to what are apparently the 'same'
situations. MC Whites take it for granted that teaching and
testing involve teachers asking questions and pupils giving
answers. But different ethnic groups interpret questioning in
radically different ways. Thus WC Blacks may identify ques-
tioning with the prying questions of welfare agencies. Some
American Indian groups regard direct questions as an unforgiv-
able invasion of personal privacy (Dumont, 1972). Hawaiian
children will often talk freely to adults whom they believe to be
interested in hearing them, but will refuse to answer direct
questions (Boggs, 1972). In other words, different groups have
quite different sociolinguistic assumptions about how and when
it is appropriate to talk to different audiences.

The social pathology model

What Labov is arguing against is the social pathology model of
language and intelligence. The social pathology model starts
from the undeniable differences in language and culture between
77
 different social groups, say WC Negroes and MC Whites in the
USA, and from the undeniable fact that, compared with
Whites, Negro children are unsuccessful at school. It then
falsely interprets differences as the cause of failure at school, by
falsely interpreting differences as deficits. The model then charac-
teristically goes on to argue that these deficits are transmitted by
the family environment by, for example, inadequate child-
rearing practices. The door is then open for the argument that
to cure the 'deficits', what is required is interventionist
programmes (known also as compensatory, enrichment or
Heads tart programmes): that is, pre-school programmes
designed to improve or enrich the child's supposedly deficient
language.
But it is now generally admitted that such interventionist
programmes have failed. That is, giving the underprivileged
Negro children intensive linguistic training before they start
school does not help their school achievement. The reason is
simply that such programmes are designed to cure deficits that
are not there. The work of Labov and others has shown in detail
that Negro children do not have deficient language. So, teach-
ing a child SE in pre-school programmes will only teach
him or her slightly different forms of a language he or she
already knows.
Labov points out that a major problem for speakers of
nonstandard dialects at school is the mutual ignorance of
teachers and pupils of each others' language. Teachers have
often no systematic knowledge of the nonstandard forms which
contrast with SE. And some teachers do not believe that
nonstandard dialects are systematic although this has been
shown in detail. They attribute what is a different, but rule-
governed and systematic dialect to 'sloppiness' or 'mistakes'.
Such mutual ignorance is not surprising since few people have
the opportunity of hearing a wide range of speakers in a wide
range of social situations. A teacher's evaluation of a pupil's
language is typically made on the basis of his or her restricted
response to a hostile classroom or test situation. Whereas a true
picture of a child's verbal capacity could only come from
78
studying his or her language across the range of social situations
in which it developed.

4.4 Labov and Bernstein

It is part of the myth about linguistic deprivation to believe that
Labov has produced a definitive refutation of Bernstein's
theories. As an American researcher has expressed it to me:
T h e idea is widespread in the USA that Labov has discredited
Bernstein and the strong implication is made that if you don't
agree with that, you're a racist.' There is also confusion over this
in Britain. For example, Tough asserts that 'there have been
many criticisms of Bernstein's work, most notably by William
Labov' (1977, p. 31). This is inaccurate.
Labov and Bernstein have commented only very briefly on
each other's work. Labov (1969, 1970) accuses Bernstein of
failing to provide a proper linguistic specification of the central
concept of code, and of failing to relate the theory to actual data
on the use of language in context. He also believes that what
Bernstein treats as code differences are merely stylistic prefer-
ences between speakers. Bernstein (CCC3, p. 28) accuses much
of the American work on sociolinguistics of being conducted at a
trivial theoretical level, in so far as it is not related to wider
problems of the socialization of the child, cultural transmission
and change. He accuses the American work on BEV (including,
presumably, Labov's) of being limited to relatively surface
concepts of'context' and 'language variety', and of neglecting to
analyse the deeper, underlying problems of how educational
knowledge is transmitted. Both sides of the debate are, it seems
to me, making valid points. It is a pity that there has so far been
no detailed discussion of the issues by those directly involved in
the debate. It would be illuminating for all of us.
It is clear that much of Labov's work implicitly questions
Bernstein's theory. But direct comparison is difficult. Bern-
stein's work is experimental and/or abstract and speculative,
rather than being based on details of observed language in use.
Labov's most important work, on the other hand, is closely
79
based on detailed linguistic analyses of language recorded dur-
ing fieldwork in natural social situations. It is also not clear how
far observations of the language of WC Negroes in the USA can
be extrapolated to a British context.

4.5 West Indian children in British Schools

Very important work has, however, been done in Britain with
West Indian children aged seven to nine in city schools, and this
British work confirms several principles which have emerged
from Labov's work in the USA. This work started as a Schools
Council Project at the University of Birmingham in 1967 (Wight
and Norris, 1970; Wight, 1971,1975; Sinclair, 1973).
There is considerable variation in the language of West Indian
children in Britain, depending on their social class, on which
part of the West Indies they come from, and on whether they are
recent immigrants or were British-born. Several different terms
are used for such varieties. One term is 'Jamaican Creole5, and
although Jamaicans in Britain are high in numbers, there are
different varieties used in other Caribbean islands. 'West Indian
English5 is another term, although the term 'West Indies5 is not
normally used by the speakers themselves. They often refer to
their language as 'Patois5, although this can also refer to French-
based Creoles (for example, from St Lucia). A neutral term
might be 'Caribbean Creoles5. Le Page (1981) and Sutcliffe
(1981) provide succinct descriptions of the varieties involved.
At the risk of over-generalization, the home language of many
of the children is an English-based Creole (see Ch. 4.2) which can
be regarded as an extreme dialect of English. That is, Creole and
SE lie at opposite ends of a dialect continuum whose extremes
are mutually unintelligible language varieties. The differences
between the language variety spoken in the home and the SE
used in schools can cause educational problems for such chil-
dren, but only indirectly. In the infant school, some creole-
speaking children may initially be unintelligible to the teacher.
But by the age of 5, most children are bidialectal in Creole and a
classroom dialect which approximates to SE. And by the age of
80
7, most children have an impressive command of SE. Inter-
ference from Creole is not therefore found to be a major source of
children's comprehension problems. Nor does the child's lan-
guage often impede spoken communication.
The project concluded that the children did not need much
remedial English teaching at all, since they acquire the school
dialect of their own accord with speed and skill. It was decided
that all they require is direct teaching of a few linguistic features
necessary to the conventions of standard written English.
Creole, for example, does not use inflexions in forms like he
come. Such differences produce no problems in spoken com-
munication, but teaching is needed to bring the children's
writing into line with conventional usage.
The most important sociolinguistic factor in the education of
West Indian children is therefore the attitude of the teacher. It is
crucial, for example, that initial minor difficulties in compre-
hension are correctly attributed to dialect interference and not
to lack of intelligence. The main requirement here is patience
and tolerance of dialect differences on the part of the teacher.
The most interesting discovery was that the problems in
language development which many of the children have are not
due to dialect differences at all, but are shared by numerous
children whose native dialect is SE. And language materials
which were originally developed for West Indian children have
been found to be useful for all children. These materials,
published by E. J. Arnold as Concept 7-9, aim at developing
language effectiveness and communication competence, rather
than spending time on superficial features of dialect, which were
found to be scarcely educationally relevant. Wight (1975) points
out that a central theoretical weakness in the language depriva-
tion concept is that it is derived from an undefined notion of
language proficiency: it is such proficiency that the materials
aim to develop. The Concept 7-9 materials are in the form of a
beautifully produced package of games, designed mainly to be
played between the children themselves. One game among
many, for example, provides one child with different shapes
drawn on cards, say:
81
 A X
A

A
X X
The child has to describe the shapes to his or her partner who has
to draw them working only from the verbal description, not
being allowed to see the original. This game can be made quite
difficult: readers might try describing this shape so that some-
one else could draw it exactly , maintaining the same shape and
size of the spirals and numeral forms.

3 4 3
The main strategy of the games is thus to get the children
talking to each other within well-defined communication tasks.
And the idea of the games is to force the children to explain,
describe, inquire, classify and differentiate, as accurately as is
required for the game. As Sinclair says:
It is fascinating to watch children using the communication
materials. They are so intent on solving the problem that they
squeeze their language to the last drop of meaning, flash from
one tactic to another, try new angles with all sorts of risks
involved. (1973)
The work therefore confirms the general finding quoted above
(see Chs. 3.2, 4.3) that measures of a child's linguistic compe-
tence in a potentially threatening or embarrassing adult-child
situation are simply no measure of his or her true capacity. A
major feature of the Concept 7-9 materials is that many of the
games are designed to be played between children with no adult
intervening, after an initial explanation of the rules. The project
and subsequent work have produced large amounts of materials
for use in schools.
V. K. Edwards's (1979) work does, however, suggest some
modifications to Wight's position. Edwards argues that there
are some differences between standard British English and
82
Caribbean Creoles which may cause interference problems
between the dialects, and therefore comprehension problems
for children in reading. She gives many such examples. One is
that Caribbean Creoles do not necessarily distinguish between
active and passive sentences. For example, a sentence such as
The chickens eat might be ambiguous to a child who might
interpret it as either 'The chickens eat' or as T h e chickens are
eaten'. Nevertheless, she also emphasizes the range of other
non-linguistic factors which may contribute to educational
problems: the children and their parents may have different
educational expectations from White MC British people, they
may be under severe social and economic pressures, and, as
always, the teachers' attitudes are crucial.
In a revised version of his earlier articles, Wight (1979) still
maintains that children of Caribbean origin do not find learning
to read more difficult because of differences between their home
dialect and SE. The important factor, he maintains, is the
teachers' attitude to such differences. The debate is therefore
not yet resolved.
Wight (1979) also points out that the dialect situation for
children of Caribbean origin may be more complex than that I
described in Chapter 2.3. For example, a child in London may
have two different prestige standard varieties of English to aim
at in formal situations, as well as a British nonstandard dialect
and distinct varieties of Caribbean Creole and London Jamaican
(now distinct from Jamaican Creole) to move towards in less
formal situations. According to the social setting, he or she will
move in a stylistic and dialectal space defined by at least five
norms:
Standard Nonstandard
Standard British English Nonstandard London dialect
Standard Jamaican English London Jamaican
Caribbean Creole
The Rampton Report on West Indian children in British
schools has asserted that 'schools should value the language
which all children, including West Indians, bring to school'
83
(HMSO, 1981). Since the Bullock Report (HMSO, 1975) this
has become almost an official litany. However, it is probably
much more often asserted than put into practice or even under-
stood. I have already shown in this book that to understand what
is meant requires an understanding of the complex relations
between informal spoken language, nonstandard dialects and
the standard language expected in formal educational settings.

4.6 A pseudo-problem?
The large, complex and imposing literature on the importance
of language in education (the present book being a very small
drop in a rather rough ocean) must not blind us to simple
wtfttlinguistic reasons why WC (and immigrant) children tend to
fail at school more often than MC children. For example, a child
may fail at school because he does not share the school's ideas of
what is important: he just has different values and commit-
ments. One nonlinguistic explanation of failure at school is that
you can take a child to Euclid but you can't make him think.
Alternatively, a child may appear 'uneducable' because the
school is insensitive to his culturally different forms of language
and thinking, and insists on treating such differences as deficits.
A second explanation of educational failure is that you have to
start teaching from where the child is: there is nowhere else to
start - but with different children we have to start in different
places. Nor must we assume in any case that there has to be a
single, tidy cause of educational failure: a single, magic predictor
of educational progress. In general, there is increasing evidence
for a range of nontraditional (including nonlinguistic) effects on
cognitive development, including the child's confidence in him-
self and hopes for the future, and the teacher's expectations of
him or her (Rosenthal and Jacobson, 1968).
The reader is warned, therefore, not to be taken in by the
large number of studies which discuss whether there is a direct
causal relationship between language and educability. Large
amounts of academic ink have often been wasted on pseudo-
problems. In the Middle Ages, scholars used to debate how
84
many angels could dance on the point of a needle: but debating a
problem does not make it meaningful in real world terms.
Several widely read books (e.g. Creber, 1972; Wilkinson, 1971;
Flower, 1966; Her riot, 1971) on language in education simply
take for granted that 'linguistic deprivation' is a meaningful
concept, and that it can be used to 'explain' some children's
educational problems. The concept, however, is severely in
dispute, and many linguists question whether the notion has any
validity whatsoever. In fact, a counter literature has now sprung
up out of necessity to point out some of the myths created by
social-science research in education. Keddie (1973) attacks the
'myth' of cultural deprivation; Labov (1969) attacks one aspect
of this myth in what he calls the 'illusion of verbal deprivation';
and Jackson (1974) attacks the 'myth' of elaborated and re-
stricted codes. It is unfortunate that one group of social scien-
tists is now having to try and clear up some of the confusion
caused by another group: unfortunate but necessary, since these
are not mere debates between academics, but live issues affect-
ing teachers and their pupils.
The general message of the last two chapters, then, is that a
great deal of the literature on the relation between language and
educational success is rather beside the point. There is, as yet,
no proven causal relationship between a child's language and
his cognitive ability, and it is not even clear what kind of
evidence would be required to demonstrate such a relationship.
Statements which assert that certain (dialectal) varieties of
language are 'deficient', and therefore cause cognitive deficien-
cies, are demonstrably incoherent. A major logical fallacy de-
rives therefore from seeing language as a cause of educational
success or failure. Since no clearcut relation can be demon-
strated between forms of language and forms of cognition, one
is left with a mere correlation: two groups of children, say
WC and MC, use different varieties of language and also (as a
statistical tendency) perform differently at school. But such a
correlation can never, in itself, be a demonstration of causality.

85
A sociolinguistic problem
What emerges, however, is a complex sociolinguistic rela-
tionship between a child's language and his success at school.
There is no doubt that different social groups use different forms
of language in comparable social situations. That is, they have
different norms of appropriate language use. This is shown both
by fieldwork in natural settings, and also by experimental work
such as Bernstein's. Teachers and schools may find the language
used by certain children stylistically inappropriate to the con-
ventions of the classroom situation; although the child's lan-
guage may be quite adequate to any cognitive demands made on
it. Further, teachers may react negatively to low-prestige
varieties of language, and, in extreme situations, may even
misunderstand the child (although neither party may realize
precisely what is happening). Even if the teacher goes out of his
way to accept the child's language as different but equally
valuable, his own language is likely to be noticeably different
from the child's in the direction of the standard, prestige
language variety. And the child will be aware that the teacher's
form of language is the one supported by institutional authority.
Children may then be caught in a double bind. They may
recognize that to get ahead they must adopt the teacher's style of
language, but to do this will separate them from their friends. A
nonstandard dialect may have low social prestige for schools,
but serve the positive functions of displaying group loyalty for
its speakers. And the peer group is always a much stronger
linguistic influence on children than either school or family.
Statements about language in education must always take into
account, therefore, the power of speakers' attitudes, beliefs and
perceptions of language. There is almost no one in Britain who is
not now constantly exposed to models of SE through education,
radio, television and the press. But there is no evidence that local
speech varieties are dying. Pressures to conformity are always
offset by pressures of loyalty to the local speech community.
This is as true of speakers of nonstandard Negro English as of
Glaswegians (Macaulay, 1978).

86
 Educational disadvantage may be the result of people's
ignorance or intolerance of cultural and linguistic differences.
But such a disadvantage is not a deficit. I would thus reject the
over-simple and dangerous catchphrase 'educational failure is
linguistic failure', and substitute for it the more guarded state-
ment: 'Educational failure often results from sociolinguistic
differences between schools and pupils.' My own view is there-
fore as follows: (1) schools and classrooms depend on language,
since education, as we understand it in our culture, is inconceiv-
able without the lecturing, explaining, reading and writing
which comprise it (see Chs. 1.2 and 7.5). So, (2) if a school
defines a pupil as 'linguistically inadequate' then he or she will
almost certainly fail in the formal educational system. But this a
tautology: (2) follows directly from (1), and merely raises the
question of what linguistic demands schools make on pupils.
One of the linguistic demands made by the school may be that
SE is the appropriate language for the classroom. If linguistic
competence is thus equated with the ability to use standard
dialect forms, this means that speakers of nonstandard dialects
are by definition 'linguistically deficient'. But such a definition
is, of course, a circular and empty one, and has no basis
whatsoever in linguistic fact.
Since I am arguing that sociolinguistic breakdown often
occurs between pupils and schools, the rest of this book must
look at ways of studying the language used by teachers and
pupils in classrooms.

87
5

The need for studies

of classroom
language

We are now half-way through this book without having discus-

sed directly how teachers and pupils use language in actual,
everyday classroom situations. This is a sad reflection on the
state of our knowledge about language in education, and on the
failure of most researchers, until recently, to go where the action
is: into classrooms.

5.1 Reasons for studying classroom language

There are important reasons for observing, recording and
studying teacher-pupil dialogue in the classroom.
The most fundamental reason is that, ultimately, the class-
room dialogue between teachers and pupils is the educational
process, or, at least, the major part of it for most children. Other
factors, such as children's language, IQ, social class and home
background, however important they may be as contributing
factors, are nevertheless external, background influences. Rel-
atively little educational research, paradoxical as this may
seem, has been based on direct observation and recording of the
88
teaching process, as it happens, in the classroom itself. There
was for a long time an assumption that the educational process
could be explained by looking at the external determinants of
educational success and failure, and outcomes (as measured by
tests and questionnaires), but without looking inside class-
rooms. On commonsense grounds alone, however, it would
seem that an understanding of teaching and learning would have
to depend, at least in part, on observation of teachers and
learners. There is 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.
It was emphasized above, for example, that in Bernstein's
work, although much of it is concerned with educability, no
attempt is made to examine how language is used in schools.
Bernstein's model proposes external determinants of education-
al success: social-class stratification is said to produce different
family types with different communication systems, which in
turn produce different codes and different cognitive orienta-
tions in children. He simply states, without evidence or illustra-
tions, that schools are predicated upon elaborated code, since
schools are concerned with 'making explicit and elaborating
through language, principles and operations' (CCC 1, p. 221)
and that 'the introduction of the child to the universalistic
meanings of public forms of thought. . . is education' (CCC 7,
p. 225, emphasis in original). It will become clear below, in
studying actual fragments of transcribed classroom lessons, that
classroom language is often highly constrained in some obvious
ways, rather than 'elaborated'.
It is also important to realize that these external, background
factors are inaccessible and highly interpretive abstractions
resting, at least in part, on social interaction. Consider, for
example, the concept of a child's IQ. This is not an absolute,
pure measure: it is a capacity which has to be measured in a social
situation, in the classroom or in a test, in which tester interacts
with pupil. It is now realized in general that learning is not a
89
purely cognitive or psychological process, but can depend
crucially on the social relationship between teacher and pupil
(see Chs 6.4 and 7.4). In exactly the same way, home back-
ground is partly physical environment (e.g. the number of
people to a room, or the number of books in the house), but it is
also quite inseparable from the social interaction in the family
(e.g. whether parents discuss homework with the children).
It is important, then, if our statements about education are
not to be vague, ungrounded generalizations, to tie these state-
ments down to analyses of actual, observable and recorded talk
and communication. It is all too easy to make generalizations
about 'classroom atmosphere' which are unrelated to what
actually happens in classrooms. I can illustrate this most easily
from a quote. Postman and Weingartner discuss how it is that
pupils get the message that the language used in school does not
have to satisfy the demands of a problem so much as the
demands of the teacher, although few teachers consciously
articulate such a message as part of the content of their lessons:
The message is communicated quietly, insidiously, relent-
lessly and effectively through the structure of the classroom:
through the role of the teacher, the role of the student, the
rules of their verbal game. . . . Each of these learnings (i.e.
what is communicated by the structure of the classroom) is
expressed in specific behaviours that are on constant display
throughout the culture. (1969, pp. 32-3)
I tend to agree with Postman and Weingartner, but I would
want to be much more specific. Given that many messages are
conveyed by teachers to pupils, just how are they communi-
cated? By what 'structure'? What are the rules of the verbal
game? What specific behaviours? If they are specific, then they
can be specified.
5.2 Our ignorance of classroom language
Our ignorance of what actually happens inside classrooms is
spectacular. We are often prepared to make broad generaliza-
tions purporting to relate children's language to their potential
90
educability, yet we lack basic descriptive information about how
pupils and teachers communicate. In a sense, of course, we all
know what classrooms are like: we have spent long enough in
them as pupils and teachers. But such intuitive, remembered
knowledge is no substitute for a conceptually adequate analysis
of classroom life based on recording and description of the
classroom routine which takes up thousands of hours of a pupil's
life. People often hold firmly entrenched views on the language
and education debate, often arguing more from prejudice than
from carefully considered observations and evidence. In any
case, a major problem in studying classroom behaviour is that it
takes a tremendous effort to really see what is happening: rather
than simply taking the scene for granted and interpreting it in
terms of conventional categories.
The teaching process has not yet been adequately described in
sociolinguistic terms. If one talks to teachers about their class-
room experiences, one discovers immediately that there is
simply no vocabulary of descriptive concepts for talking about
teaching. Despite the vast complexity of second-by-second
classroom dialogue, the discussion will be conceptually crude
and oversimplified. It is time that teachers had an adequate
descriptive language for talking about their own professional
behaviour.

The need for sociolinguistic study of classroom language

Sociolinguistic study of teacher-pupil dialogue, observed and
recorded in the classroom, would therefore begin tofilla serious
gap left by previous research on language in education.
The assumption that all the important determinants of class-
room performance lie outside the classroom is no longer so
widely held, and classroom research is developing fast. How-
ever, most research which has been based on direct observation
of teacher-pupil interaction in the classroom has been done ex-
clusively according to techniques of systematic observation. In
using techniques of this type, an observer sits in the classroom
and uses a set of pre-prepared categories to 'code' what teachers
91
and pupils say, usually every few seconds on a time-sampling
basis. Coding categories might include, for example, 'teacher
lectures', 'pupil asks question' and 'teacher justifies authority'.
Over seventy coding schemes have been published (Simon and
Boyer, 1967, 1970) and literally hundreds of studies have been
done with such systems. But results for research have been
disappointing, and no clear trends have emerged, for example,
between teachers' verbal styles, as defined by the schemes, and
measures of teaching effectiveness. This type of work is covered
by several comprehensive reviews (e.g. Medley and Mitzel,
1963). More recent articles have thoroughly documented
reasons for disillusionment with the technique as a research
tool, though not necessarily as a teacher-training method
(Nuthall, 1968; Walker and Adelman, 1975b; Delamont, 1983).
The most important shortcomings of this type of study, very
briefly, are as follows. Since the classroom talk is generally not
recorded but 'coded' by the observer on the spot in real time, the
actual language used by teachers and pupils is irretrievably lost.
Such a technique can therefore at best provide an overall,
average measure of classroom climate or atmosphere, without
being able to study the details of the actual talk which create this
climate. In other words, there is no study of how hearers
(researchers as well as pupils and teachers) interpret classroom
language: it is assumed that the coder can do this unproblemati-
cally. Hence, the data for study are not in fact the classroom
language at all, but the researcher's codings of it. In general, the
technique focuses in a fragmentary way on a succession of small
bits of behaviour, through pre-specified categories which allow
no adaptation or development. It deals only with what is
immediately observable to the researcher, ignoring qualitative
factors in favour of what is easily quantifiable (Delamont and
Hamilton, 1976).
Given that relatively little analysis has been based on direct
observation of classroom behaviour, it would in any case be
premature to remain committed to one narrowly defined type of
interaction analysis. However, an increasing amount of 'anthro-
pological' or 'ethnographic' research is now being done in
92
schools. Such work itself covers a wide range of styles, as befits
its often innovatory and exploratory aims. In general, it is based
on long-term or intensive fieldwork, that is, on some form of
participant observation. The researcher may spend several days
or weeks actually in the classroom, observing, taking notes,
recording, talking to teachers and pupils and getting to know
them: the aim being to produce a report resembling an ethno-
graphic description of a social setting, similar in some way to the
report an anthropologist might write after fieldwork with an
exotic tribe! (See Jackson, 1968; Stubbs and Delamont, 1976;
Delamont, 1983; and many other references in these works.)
Most such studies inevitably discuss aspects of classroom in-
teraction, but they are not centrally concerned to discuss
sociolinguistic behaviour.
So there are, as yet, very few studies of classroom life which
use sociolinguistic concepts to analyse teacher-pupil talk. This
lacuna is odd since socioli#guistics means simply the study of
language and how it is used in social settings: such as the
classroom. The lacuna results simply but sadly from the lack of
communication among experts in different aspects of com-
munication! Social interaction has been studied in fragmentary
fashion from within different academic disciplines, including
linguistics, anthropology, psychology and sociology. But these
disciplines have different methods and aims, and the insulation
of these academic areas from each other is almost complete. It
should be emphasized, then, that no single approach is widely
accepted in studies of classroom language. Different methods
are used to do fieldwork, to collect, analyse and present data,
according to different underlying objectives.
The tide does now seem to have turned, however, in favour of
classroom studies, which have now gained respectability as a
legitimate style of research. Various historical factors have
contributed to this development since the late 1960s: relatively
more money becoming available for educational research; a
wider social-psychological model of learning being proposed by
educationalists; an increasing emphasis on teacher training;
schools becoming more 'open' and therefore easier of access;
93
and the development of appropriate methods of participant
observation in sociology (Delamont and Hamilton, 1976).

5.3 The rationale for naturalistic studies

Theoretical reasons are also given, however, for neglecting
studies of classroom life, and we have to answer these. It is often
argued that studies of real, everyday behaviour are 'unscientific'
because real-life settings are vastly complex and contain many
uncontrollable factors, and because the same situation never
occurs twice and the studies are therefore unreplicable. Only by
isolating small bits of behaviour in controlled laboratory set-
tings, so it is argued, can we progress slowly but surely towards
the truth.
The short answer to these objections is simple: If we want to
know how people behave in classrooms, then we have to observe
them in classrooms. If we bring them into the psychological lab-
oratory at the local university, then we may discover how they
behave there. But this is unlikely to be of enduring interest!
One might rightly object, of course, that the behaviour of
teachers and pupils is unnatural even in the classroom in the
presence of an intrusive researcher with notebooks or tape-
recorder. However, people forget about the tape-recorders after
a relatively short period, say a few hours or a couple of days.
And, in any case, many of the things one is interested in
observing are simply not under conscious control.
A more powerful defence of naturalistic studies is that the
natural social pressure of interacting with a group will often
overcome the presence of a tape-recorder. This is particularly
true of children recorded in their own friendship groups, in
which the pressure of interaction with their peers typically
outweighs the presence of an observer, and helps approach the
unattainable ideal of recording 'normal' language (Labov,
1972a). Recording speakers away from their normal social
settings, on the other hand, causes severe problems of inter-
pretation and explanation. This is because there is no way of
'controlling' how people interpret and react to unfamiliar social
94
settings, such as interviews or tests. The crucial point is that the
psychological laboratory is not a neutral setting in which be-
haviour can be 'controlled'. It is itself a social setting, com-
prising perhaps, from a child's point of view, a large, strange
adult of high prestige, in unfamiliar surroundings, who is
testing the child in some unspecified way (see Chs 3.2 and 4.3).
One of the main points of Chapter 4 was that the social situation
is the strongest determinant of verbal behaviour. And it is for
this reason that I questioned the use of laboratory 'tests' in
investigating the verbal ability of children or the nature of
mother-child dialogue. Most children are, incidentally, most
helpful in such situations and will do their best to do what they
think the researcher wants them to do, however odd they may
find this. Most children have a remarkably high degree of
tolerance for the odd tasks which adults often ask of them!
What we are interested in is how people interpret classroom
language and attach social meanings and values to it (Hymes,
1972). And one cannot, in principle, 'control' how an adult or
child interprets his social environment. One might, for exam-
ple, ask the same form of question to a hundred people. One
might, in the style of some verbal tests, hold up an apple and
ask, 'What is this?' But there is no way to control how the hearer
interprets the question: Do you really not know that it is an
apple? Is it a trick question? Or a riddle? Or is it just a starter
question, leading into some more sensible and interesting ques-
tion?
Such points may seem obvious, but researchers have some-
times taken quite absurd steps to try and 'control' variables in
experiments in social behaviour. In one particularly artificial
study on how people communicate by gesture and posture,
experimental subjects were made to wear cardboard masks 'to
keep their minds off their facial expression', since the ex-
perimenter did not want to investigate this (Mehrabian, 1968).
Clearly, such procedures do not 'control' a variable (in this case,
facial expression). On the contrary, they probably made the
subjects particularly conscious of their facial expression, quite
apart from making them feel silly or embarrassed or restricting
95
their sight! Such experiments make the same mistake as the
oriental doctor who had discovered a wonderful medicine which
could cure his patients of any illness, but only if they did not
think of monkeys while they were drinking it. So he used to
warn them most particularly not to think about monkeys - and
wondered why the medicine never worked. The trouble is that
experimental subjects can never be prevented from thinking
and forming interpretations about the experiment. One must
therefore always take account of people's expectations of experi-
ments. Psychologists who believe that the behaviour of people
and rats is essentially similar are likely to construct experiments
for humans in which humans would be hard put to it to display
anything other than rat-like behaviour. Psychologists who be-
lieve, on the other hand, that human behaviour is essentially
different, are likely to set up experiments which demand be-
haviour which is qualitatively different from running through
mazes. People, including experimental subjects, generally do
what is expected of them: thus fulfilling the experimenters'
expectations.
Since the control over behaviour in laboratory settings is
largely spurious (one cannot control how people interpret test
instructions, for example) the notion that such experiments are
replicable is seriously weakened. A powerful form of replicabil-
ity is, however, possible for studies which attempt to study
natural language behaviour. If readers are given access to the
original data, in the form of transcripts of recorded speech, they
can study these directly and propose alternative interpretations
if necessary.
One reason, then, why researchers have fought shy of study-
ing classroom interaction is the notorious complexity of com-
municative behaviour. But complexity will not dissolve if we
ignore it. The drunk who loses his key along a dark stretch of the
street has no hope of finding it under a lamp post just because the
light is better there. The light may be bright in tests or experi-
ments in the psychological laboratory, but it may not illuminate
what we want to see. Complexity must therefore be admitted as
an essential feature of social interaction and studied in its own
96
right, with the help of appropriate concepts. In any case
teacher-pupil dialogue is less complex than many other types of
discourse, since it is often highly controlled in some fairly
obvious ways.

5.4 Teachers as researchers

The type of work on classroom language which will now be
discussed in the next two chapters is therefore very much work
in progress. On the other hand, it comprises observations and
analyses of the real, day-to-day classroom behaviour of teachers
and pupils. One reason why little educational research is read by
teachers is probably that hardly any of it is related in any obvious
way to events in the classroom jungle. Practising teachers are
clearly not in a position to do large-scale surveys or educational
experiments involving laboratory facilities. But they are able to
observe what happens inside classrooms. (Ch. 9 provides a list of
specific suggestions.) Classroom research therefore comprises
insights which any student or experienced teacher can verify,
amend or refute every time he or she is in a classroom. Indeed it
is theoretically important that practising teachers should do so.
Research on children and classrooms is usually done by outsid-
ers, but ultimately it is only the participants in a situation who
have full access to all its relevant aspects. Ultimately, a
sociolinguistic description of classroom language must come to
grips with the values, attitudes and socially loaded meanings
which are conveyed by the language, and only the participants
have full access to these values. So, as Hymes puts it, T h e
ethnography of a situation is not for a nonparticipant to say'
(1972), since aspects of the communication situation in a class-
room will often be quite opaque to an outsider (see Ch. 6.3).
Conversely, however, as I have argued throughout, intuitive
comments on language by insiders may well go astray, unless
related to well thought out sociolinguistic concepts.
To say, then, that teachers themselves have an important part
to play in studies of classroom language is not merely conven-
tional courtesy. Quite simply, if the study of classroom language
97
has to wait on outside researchers, it will never happen in most
classrooms. Hopefully, having read what follows, teachers and
student teachers will be able to record and observe classroom
lessons and notice new things. Linguists are rightly reticent
about telling teachers how to behave in classrooms. But it is the
task of linguists to provide educationalists with the means of
observing and describing classroom language, and to indicate
where they may look in language for educationally interesting
findings. Chapters 6 and 7 can be read, therefore, as a collection
of suggestions about how sociologists or teachers themselves can
study classroom language, and Chapter 9 will provide a list of
quite specific suggestions about how beginners to classroom
research might start collecting and analysing data on classroom
language.

98
 6

Studies of
classroom
language

In one sense teachers are teachers: that is their job. But a person
cannot simply walk into a classroom and be a teacher: he or she
has to do quite specific communicative acts, such as lecturing,
explaining, asking questions, encouraging pupils to speak and
so on. In other words, social roles such as 'teacher' and 'pupil' do
not exist in the abstract. They have to be acted out, performed
and continuously constructed in the course of social interaction.
It is only since about 1970 that descriptions of classroom
interaction have begun to appear. Such studies of classroom
language are all in different ways fragmentary, but they contain
interesting insights into the ways in which teachers and pupils
communicate in real classrooms. And it is these insights which
practising teachers can verify or amend from their own class-
room experiences and observations. The studies to be discussed
here are mainly small case-studies: of a single day's lessons or
just of a single lesson, perhaps. So, only guarded claims are
being made about teaching in general. On the other hand,
behaviour at this level is often highly repetitive and subject to
severe cultural constraints. And the state of the art at present
99
requires detailed study of small amounts of data, rather than
more superficial study of many hundreds of hours of classroom
lessons.
All such studies, then, assume that close and direct study of
classroom language will provide the most useful insights into
teaching and learning processes. They maintain, at least im-
plicitly, that general statements and theories of education will
ultimately stand or fall according to whether they can explain
how teachers and pupils communicate with each other in real
classrooms. For, despite the large literature on 'learning
theory', how and what children actually learn what they do in
school remains almost a total mystery. In discussing such
studies, it is well to bear in mind, however, that merely moving
closer to where the action is, does not necessarily lead to
understanding the action. For this, we need not mere surface
descriptions of behaviour, but descriptions related to a coherent
set of concepts: a theoretical framework in which to make sense
of our observations.

6.1 Commentaries on classroom dialogue

A well known early study of classroom language is by Barnes
(1969). It discusses extracts from tape-recordings of a day's
lessons of a first-year class in a comprehensive school. The
research is not systematic, but has many interesting points to
make, based closely on quoted dialogue from actual lessons.
The general theme of the study is the effect of teachers' language
on pupils, particularly in situations where a teacher's language
might be a barrier to learning, because the teachers use termi-
nology or an abstract style of language with which the pupils are
unfamiliar. The work is easily available, full of good examples of
classroom talk, and a good introduction to anyone interested in
the area.
One of Barnes's main arguments is concerned with a teacher's
overall style of language. He argues that many teachers, in
talking about their subject, use a specialist language of instruc-
tion which may be a barrier to pupils who are not used to it. This
100
specialist language, which may amount to a 'language of secon-
dary education', has different aspects. First, different academic
subjects have different technical terminologies associated with
them. Teachers are usually aware of difficulties caused by this
and will take care to present and define new terms, but some-
times such name-teaching seems to take on a value of its own.
Thus one science teacher insisted on using the term 'mortar' (as
in 'mortar and pestle'). Clearly, this term is not essential to
understanding the subject, but is simply part of a whole scien-
tific style of language in which the teacher feels at home. Such
language makes linguistic demands on pupils which are quite
extrinsic to the subject being taught. In other words, the style of
language may prevent the content from getting through, and
may prevent some pupils from contributing to the classroom
dialogue.
A more subtle point Barnes makes is that many teachers may
be unaware of a more pervasive language of secondary educa-
tion, which is not part of the language of any particular academic
subject and therefore likely never to be explained to pupils. One
quote is from a history teacher, talking about city states:
These states were complete in themselves because the terrain
between cities was so difficult that it was hard for them to
communicate.. . . Now because people lived like this in their
own cities they tended to be in tensely patriotic. . . .
The teacher uses the word terrain where land would do just as
well. But he also talks in abstractions likely to be unfamiliar to
first-form pupils: complete in themselves, communicate, tended,
leaving the pupils to fill in from their own experience what
might be meant in concrete terms. For example, 'tendency' is a
complex, abstract, statistical notion, but unlikely to be taught
by any teacher.
Because teachers are accustomed to such language, they may
not recognize a valid idea from a pupil if it is not expressed in this
abstract style (see Ch. 1.2). Thus, one science teacher had asked
how a chlorophyll stain might be removed from material, and a
pupil had suggested rubbing in shoe polish, then washing it all
101
off. As the pupil did not use (or know?) the term 'solvent' or the
name of a particular solvent, the teacher failed to see that he
had in fact grasped the idea of a solvent and rejected the
suggestion.
One general and very important point Barnes makes is as
follows. A teacher may see the language of his subject as having
an intellectual function of allowing concepts to be precisely
expressed. But the teacher's language will also have a socio-
cultural function of supporting his or her role as teacher. And,
from the pupils' point of view, each new term may have a
predominantly sociocultural function: it is 'the sort of thing my
teacher says'. Barnes is here pointing to a source of sociolinguistic
interference between pupils and teachers who have different
notions of stylistic conventions. That is, a teacher may use a
certain style of language, not because it is necessary for express-
ing certain ideas, but because it is conventional to use it. But the
pupils are unlikely to share these conventions. Conversely, the
teacher may reject a pupil's formulation, not because the pupil's
language is intrinsically unable to express an idea, but because it
does not accord with stylistic conventions.
A second topic of great: general importance is Barnes's pre-
liminary classification of the kinds of questions teachers ask. He
distinguishes four broad categories of questions: (1) factual
(what?) questions which demand a bit of information or a name
for something; (2) reasoning (how? or why?) questions which
might demand observation, recalling something from memory,
or more open-ended reasoning; (3) open questions, not deman-
ding reasoning; and (4) social questions, functioning either to
control the class or appealing to them to share in some experi-
ence. This part of the analysis is very loose, and Barnes admits
that the categories are not precise enough for use by others.
However, it usefully points to the difficulty of matching the
form of a teacher's question to the underlying intention or
function. Thus a teacher may ask a question which appears
superficially to be asking a pupil for his personal view, when the
teacher really has a particular answer in mind. Barnes calls such
utterances 'pseudo questions'.

102
 Barnes admits that his study is 'impressionistic' (p. 47), that
he has no objective way of describing the language of secondary
education (p. 53), and that his classification of different types of
question is vague. Since the study is intended to be of practical
use to teachers rather than a contribution to theory, this is not
necessarily an outright condemnation. As a whole, one could
call the method insightful observation. Note, however, the
positive points that the study is based on recording and observa-
tion of normal school lessons, and that Barnes presents extracts
from the records. Readers are therefore able, to some extent at
least, to study the data directly and to propose alternative
interpretations if they disagree with Barnes's intuitions.
In other work, Barnes (1971) has pointed to some implica-
tions of such descriptive studies of classroom language. He
begins from the statement that 'language is a means of learning'.
That is, we often learn, not only by passively listening to a
teacher, but by actively discussing, talking a question through,
defending our views in debate and so on. By studying teacher-
pupil interaction, one can therefore study how classroom lan-
guage opens and closes different learning possibilities to pupils.
Do they have to sit passively listening, providing answers on
demand? Or is there active, two-way dialogue between teacher
and pupil? These topics deserve a whole chapter, and will be
taken up again below, in Chapter 7.
A study by Mishler (1972) is similar to Barnes's in that it
comprises perceptive commentary on fragments of classroom
dialogue. But Mishler is more concerned with specifying which
particular features of language are indicators of different
teaching strategies. He shows that the particulars of actual
language used by teachers and pupils can be analysed in ways
which yield information about important aspects of the educa-
tional process; arguing that what teachers say and how they say
it creates a particular kind of world for pupils. The study is
based on recordings of three first-grade American teachers.
He argues first - an important theme of the present book -
that studies of language in use must present data in a form which
is open to reanalysis by readers. A minimal requirement for this
103
is a verbatim transcript of tape-recorded talk. Mishler's main
aim is then to specify how different teachers' cognitive strategies
are displayed (betrayed?) in the fine details of classroom dia-
logue: that is, to specify features of teachers' language which
indicate to pupils how information and concepts should be
organized, and therefore direct their attention to different forms
of order in the world. For example, a teacher's use of open-
ended questions (e.g. 'What could that mean?') may imply the
underlying pedagogic message that different answers are accept-
able: not an assumption made in all classrooms (see Ch. 7.3).
Conversely, he shows how a teacher may deny the legitimacy of
what a child says. A child asks about a film the class is going to
see:
C: What's it about?
T: I don't think I'm going to tell you.
In fact, the teacher has not seen the film, but she does not admit
this. She uses a type of exchange which denies the child access to
knowledge and maintains the teacher as the only person who
knows, and who controls what the pupil may know. (An English
lecturer at a Midlands university once told me with some glee
that, in a seminar on Jane Austen, he had asked 'Who does
Lydia Bennett marry?', knowing that this would be understood
by students as a test question - whereas he really asked the
question to fill in his own knowledge as he had not finished the
book and did not know the answer!)
Mishler's approach is one which could be of direct interest to
teachers. By close study of transcribed lessons he shows how
quite general teaching strategies are conveyed by the fine grain
of a teacher's use of language. Such a study can therefore begin
to throw a little light on how children learn what they do in
school. Only by close observation of how teachers and pupils
actually talk to each other can one discover how concepts are put
across, how some lines of inquiry are opened up and others
closed off, how pupils' responses are evaluated, and how their
attention is directed to the areas of knowledge which the school
regards as valuable.

104
 A study by Gumperz and Herasimchuk (1972) is similar in
style to Mishler's. It is based on a commentary on just two
tape-recorded lessons, chosen so as to be maximally different: a
teacher teaching a group of pupils, and an older child (aged 6)
teaching a younger child (aged 5). By close study of the record-
ings and transcripts, the authors show that the adult and child
use different means of communication. For example, the adult
teacher relies heavily on interrogatives to elicit answers from
pupils, and makes use of variation in choice of words with a
corresponding lack of variation in intonation. The child teacher
makes more use of intonational variety and repetition, especially
to distinguish questions, challenges and confirmations, and to
maintain an extraordinary degree of musical and rhythmical
relatedness with the pupil. That is, they show the child and
adult teachers using different means of communication.
They claim further that the adult and child differ in their
definition of the teaching task and of the social relationships
involved. But this does not seem to follow from the evidence
they present. They seem rather to have found the two 'teachers'
doing similar things (questioning, challenging, confirming) by
means of different linguistic devices.
It would be most useful to teachers if they could be made
more aware of the linguistic means by which children may com-
municate messages, especially where these differ from adult
usage. Gumperz and Herasimchuk place useful emphasis on
the precise linguistic signals which convey social messages
in the classroom, and show that these signs may not only be
the words or sentences used, but the way utterances are se-
quenced, and the paralinguistic signs such as intonation and
rhythm.
In work of my own (Stubbs, 1976) I have described one way in
which teachers in relatively formal chalk-and-talk lessons keep
control over the classroom discourse. One thing which charac-
terizes much classroom talk is the extent to which the teacher
has conversational control over the topic, over the relevance or
correctness of what pupils say, and over when and how much
pupils may speak. In traditional chalk-and-talk lessons, pupils
105
have correspondingly few conversational rights. This has often
been pointed out in general (e.g. by Barnes, 1969), but the
actual verbal strategies which teachers use to control classroom
talk have yet to be systematically described.
Teachers' talk in such lessons is characterized by the way the
teacher constantly explains things, corrects pupils, evaluates
and edits their language, summarizes the discussion and con-
trols the direction of the lesson. That is, a teacher is constantly
monitoring the communication system in the classroom, by
checking whether pupils are all on the same wavelength and
whether at least some of the pupils follow what the teacher is
saying. Such monitoring may actually comprise what we under-
stand by 'teaching'. It is useful to refer to such language as
metacommunication. It is communication about communication:
messages which refer back to the communication system itself,
checking whether it is functioning properly. Suppose, for exam-
ple, a teacher says, Now don't start now, just listen. He is not here
saying anything substantive, but merely attracting the pupils'
attention, opening the communication channels and preparing
them for messages to come. Or he might say, You see, we're really
getting on to the topic now, again not adding anything to the
content of the discussion, but commenting on the state of the
discussion itself. Or again, he might control the amount of
pupils' speech by saying, Some of you are not joining in the studious
silence we are trying to develop. Teachers exert control over
different aspects of the communication system in the classroom.
They control the channels of communication by opening and
closing them: OK now listen all of you. They control the amount
of talk by asking pupils to speak or keep quiet: Colin, what were
you going to say? They control the content of the talk and define
the relevance of what is said: Now, we don't want any silly
remarks. They control the language forms used: Thafs not
English. And they try to control understanding: Who knows
what this means?
All the italicized examples are taken from recorded classroom
lessons. The metacommunicative remarks can be generally
formally recognized by the use of metalinguistic terms which
106
refer to the ongoing discourse itself, e.g. listen, topic, say,
remarks, etc.
Such talk is characteristic of teachers' language: utterances
which, as it were, stand outside the discourse and comment on it
comprise a large percentage of what teachers say to their pupils,
and comprise a major way of controlling classroom dialogue.
Use of such language is also highly asymmetrical: one would not
expect a pupil to say to a teacher: Thafs an interesting point. Such
speech acts, in which the teacher monitors and controls the
classroom dialogue are, at one level, the very stuff of teaching.
They are basic to the activity of teaching, since they are the acts
whereby a teacher controls theflowof information in the classroom
and defines the relevance of what is said (see Atkinson, 1975;
and Ch. 7.5).

6.2 The structure of classroom dialogue

Detailed commentary on small fragments of teacher-pupil dia-
logue is clearly a necessary step in work on classroom language.
But one also wants to go beyond commentary on the details of
small excerpts torn from their conversational context, and to
make more comprehensive statements about how classroom talk
works as a system of communication. There is a clear need, then,
to pay close attention to the fine details of what teachers and
pupils actually say. On the other hand, it is not enough to make
insightful comments on short extracts, selecting, according to
intuition, extracts which seem interesting and ignoring the rest.
It is necessary to work towards generalizations about class-
room discourse. The next two studies discussed attempt to do
this.
Rather than focusing on the fine details of teacher-pupil
dialogue, Bellack et al. (1966) argue that much teacher-
pupil talk has a characteristic underlying structure and pattern
which teachers and pupils follow with remarkably little devia-
tion. Their analysis is based on a coding of transcribed audio-
recordings of about sixty classes. They start from Wittgenstein's
notion of a language game, in which speakers follow rules and
107
conventions. They propose four pedagogical moves as basic
units of classroom dialogue: structuring moves indicate the direc-
tion the speaker thinks the discussion should take; soliciting
moves serve to elicit a response from another speaker; respond-
ing moves bear a reciprocal relationship to soliciting moves (e.g.
a pupil's answer to a teacher's question); and reacting moves
modify or clarify a preceding utterance. These moves are a
preliminary definition of how teacher-pupil discourse works:
they define possible sequences of teacher-pupil talk.
The moves are analysed as building up into repetitive
teaching cycles; cycles build up, in turn, into sub-games, and
sub-games into games. Overall rules which they propose for the
classroom dialogue are: that the teacher is the most active
player; that, in general, the game is played within the teacher's
structure; that the teacher's primary role is solicitor while the
pupil's is respondent. They discovered, for example, that fifteen
teachers made 50 per cent more moves than 345 pupils, that the
core of the game was the solicit-response pattern (this was
slightly more than three-fifths of all moves made), and that the
pupils were mainly confined to responding.
This study is therefore a preliminary attempt not simply to
describe what teachers and pupils say and how they say it, in
terms of individual utterances: it is an attempt to formulate a
hierarchical structure for classroom discourse, an abstract and
general model to which the actual classroom talk conforms in
varying degrees. The work by Bellack et al. is thus a preliminary
attempt to study the overall structure of teacher-pupil dialogue.
This approach is taken much further by Sinclair and Coulthard
(1975).
Sinclair and Coulthard are linguists and primarily interested
in studying types of linguistic patterning in long texts collected
by tape-recording spontaneous conversation between several
speakers. Their work is one of the few attempts to provide an
analysis of the underlying structure of classroom dialogue and
there is much in it of interest to educationalists.
The easiest way to explain some of their main ideas is to begin
with a piece of their data.
108
 T: What makes a road slippery? (1)
P: You might have rain or snow on it. (2)
T: Yes, snow, ice.
Anything else make a road slippery? (3)
P: Erm,oil. (4)
T: Oil makes it slippery when it's mixed with
water doesn't it? (5)

There are five utterances in this teacher-pupil interchange, but

intuitively we feel that there is an obvious boundary in the
middle of (3), corresponding to the line division after ice.
Further we might feel that there are just two conversational
units: T asks question - P responds - T evaluates, and a repeat of
this unit. Sinclair and Coulthard propose that this exchange
structure is a typical one in many classrooms. They propose that
it is a basic type of teaching exchange, and label its constituent
parts: (teacher's) initiation, (pupil's) response and (teacher's)
feedback, or IRF for short. The constituent parts they call
moves: that is, exchanges consist of moves. Another type of
teaching exchange is illustrated by data such as:

T: Finished Joan? I
P: (Nods) R
T: Good girl. F
And Miri? I
P: Yes. R
T: Good. F
Finished? I
P: Yes. R

Here a teacher initiation is followed by a pupil response, and

sometimes by a teacher feedback. So the exchange structure is
IR(F), where brackets indicate an optional item. A different
exchange structure occurs when a pupil asks a question. Here
the structure is IR, with no F, since pupils do not generally
overtly evaluate teachers' answers! In many classrooms (but not
all) pupils' initiations are largely restricted to procedural mat-
ters such as, 'Can I leave the room?' or 'Do we have to use
109
coloured pencils?' Sinclair and Coulthard propose several other
types of teaching exchange, but these examples will suffice to
illustrate the type of analysis.
So, exchanges consist of moves (I, R and F). But moves are
often further analysable into acts. Thus it seems inadequate to
call this teacher's utterance an initiation and leave it at that:
T: A group of people used symbols in their writing,
they used pictures instead of words. (1)
Do you know who those people were? (2)
I'm sure you do. (3)
Joan. (4)
The proposed structure here is: (1) starter, (2) elicitation, (3)
prompt, (4) nomination; where these labels name acts which
comprise the teacher's opening move in an exchange.
As well as teaching exchanges, one also comes across utter-
ances such as,
T: Well, today I thought we'd do three quizzes.
Here the teacher is providing an opening boundary to the talk:
not yet teaching anything, but announcing that the lesson is
under way and indicating the direction it will go. So, Sinclair
and Coulthard propose a boundary exchange which has the move
structure: frame-focus. Frames function to indicate boundaries
in lessons and are realized by a small number of words: well,
right, good, OK, now. A focus functions to indicate where the
lesson is going.
Boundary exchanges therefore indicate that lessons are fur-
ther structured into larger units, which Sinclair and Coulthard
call transactions. They are therefore proposing a hierarchic and
structural analysis of teacher-pupil talk. The lesson consists of
transactions, marked off by boundary exchanges. Transactions
consist of exchanges, an initial boundary exchange, then
teaching exchanges of different types. Exchanges consist of
moves which consist of acts.
These units of discourse are functional units which specify
what a speaker is using language for: e.g. to mark a boundary in
110
the discourse or to evaluate a pupil's answer. The relationship
between such language functions and the language forms which
realize them is very complex. Sinclair and Coulthard provide a
good example of the type of complexity involved. A class have
been listening to a tape-recording of a speaker with a 'posh'
accent. One of the pupils laughs, and the teacher says: 'What are
you laughing at?' The pupil takes this as a criticism and as a
command to stop laughing, whereas the teacher intends it as a
genuine question, an opening move to explore the children's
attitudes. That is, the pupil has misunderstood a particular
language form, a what interrogative, as a command instead of a
question.
Fuller accounts of Sinclair and Coulthard's work are provided
by Burton (1980), Stubbs, Robinson and Twite (1979), Stubbs
(1983), Willes (in press) and by articles in French and MacLure
(1981).

6.3 The classroom as a sociolinguistic setting

Classroom dialogue requires study as a linguistic system. This is
likely to be the area in which linguists will make the largest
contribution. But it must also be remembered that classroom
talk is not merely a linguistic system, but a sociolinguistic
system.
Walker and Adelman (1975a and b, 1976) regard classrooms
as intense and complex social settings. They have been particu-
larly interested in different types of social organization in class-
rooms and in different kinds of social control and personal
relationships which the language between teachers and pupils
can sustain. They point out that much research (e.g. the studies
discussed earlier in this chapter) has concentrated on 'formal'
classrooms in which the teacher stands at the front of the class
and has the attention of the whole class. But there is little work
on 'informal' or 'open' classrooms where children are working
in small groups, the teacher passing from group to group or
talking to individual children. They see such situations as
crucially different.
Ill
 Walker and Adelman's research in such classrooms has
opened up important issues for the theory and methodology of
observing life in classrooms. In particular, they emphasize the
dangers of too narrow a notion of classroom language. If one
transcribes teacher-pupil dialogue from a formal chalk-and-talk
lesson, the transcript characteristically looks like a conventional
play script: speakers down the left and relatively well-formed
language on the right. Such transcripts are generally easy to
understand, even for the reader who has no information about
the context of the recording. (See, for example, most of the
dialogue extracts in the present book.) However, sound-
recordings from informal contexts are typically not completely
comprehensible without a visual record of the classroom. In
transcript, the talk appears fragmented, incomplete, full of
hesitations, interruptions and ambiguities. It is difficult to know
who is talking to whom, and what is being talked about. But this
is a mere surface description: these characteristics do not appear
as defects to the participants, who can fill in their understanding
from gestures, facial expressions, postures and the surrounding
situation.
One way of putting this is to say that in informal classrooms
the talk is strongly context bound. That is, its understanding
depends on a knowledge of the context in which it was recorded,
and, ultimately, on the culture of the class. An example will
make this clear. A teacher has been listening to a pupil read a
rather skimpy piece of work:
T: Is that all you've done?
P: Strawberries! Strawberries!
(Walker and Adelman, 1976). This exchange only makes sense
if we know that the teacher has previously said to the pupils that
their work was 'like strawberries - OK as far as it goes, but it
doesn't last long enough'. It is simply not possible to retrieve
such meanings from the transcript alone.
Walker and Adelman are making a crucial point: the mean-
ings of classroom language are often not as simple as they seem.
The talk often cannot be taken at its face value or commonsense
112
 meaning. They are thus criticizing the naive and impoverished
concept of classroom language, which is implicit in some stu-
dies, by emphasizing in particular the inherent complexity of
meanings which may develop between speakers over long
periods of time. Such meanings are part of the shared culture of
the class and may be hidden from a casual observer. Most
important, they point out that talk may have quite different
functions in formal and informal classrooms. In formal class-
rooms, the talk may be primarily concerned with transmitting
information. But in informal classrooms, the talk also has to
sustain complex social relationships during intimate, small
group discussions or one-to-one teacher-pupil talk.
Methodologically and theoretically, Walker and Adelman are
concerned to develop new ways of thinking about teachers and
pupils in classrooms. Their most important innovation here is a
system of stop-frame cinematography developed for filming in
schools. Using this system, photographs can be taken at inter-
vals of, say, every 2 or 3 seconds, and synchronized with a sound
recording. The timing of the photographs is not arbitrary,
however. In primary classrooms it is typically necessary to take
photographs more frequently, approximately every second, in
order to preserve enough visual information to allow a viewer
to make sense of the sound recording. That is, the recording
technique, in itself, tells us something about the activities in
different classrooms.
From the perspective of the other studies discussed in this
chapter, Walker and Adelman leave their linguistic data under-
analysed. But they warn workers in the area of several pitfalls.
They show how talk can sustain radically different concepts of
social relationships. They question the view that all classrooms
have a peculiar common structure and ask whether this view
simply means that research has only looked at a narrow range of
classrooms. And they emphasize the complexity of meanings
underlying classroom dialogue, many of which may be hidden
without an understanding of the classrooms in which the dia-
logue is recorded.

113
6.4 Studying social processes in classrooms
Other studies of classroom language could be reviewed, but the
reader will by now have a sufficient idea of the kind of work that
has been done.
Studies of classroom language are, as yet, a mixed collection
of exploratory work on a relatively narrow range of classrooms,
and some general limitations of such work will be discussed in
Chapter 8. Nevertheless, several important findings keep turn-
ing up in different studies. Much work, for example, has
pointed to the highly assymmetrical control which teachers
often maintain over classroom dialogue, dominating the talk
both by the amount of their own talk, and also by the use of
certain discourse sequences (e.g. IRF). Many studies also com-
ment on teachers' characteristic use of 'questions' which are not
genuine requests for information. These are variously called test
questions (by Labov), pseudo questions and closed questions
(by Barnes), and convergent and guess-what-I'm-thinking
questions (by Postman and Weingartner). It is worthwhile
pondering the effect on classroom dialogue when some teachers
rarely ask questions because they want to know something!
These two findings point to the highly artificial nature of much
teacher—pupil dialogue, compared to, say, casual conversation
between social equals.
One thing which is clear from studying teacher-pupil interac-
tion is just how constrained it often is by cultural rules. The
question-answer pattern, for example, has been found to have
been stable over the past fifty years (Hoetker and Ahlbrandt,
1969) and across different countries (Bellack, 1973), although it
has regularly been criticized by educational theorists. Work on
classroom language therefore begins to make explicit some of
the sociolinguistic demands made on pupils, and to give further
substance to the general finding cited earlier (Ch. 4.3) that the
social situation is the strongest determinant of verbal behaviour.
The most fundamental aspect of work on classroom language
is therefore as follows. Out of the vast body of educational
research, only a small fragment inquires into the social processes
114
which occur in schools and classrooms. Research has often
reflected educators' definitions of education. It has therefore
taken as its problem to discover how to teach pupils better or
faster, taking for granted underlying assumptions about the
aims of education. But the question of what is learned has often
been bypassed by research which has thus been designed to
measure pupils before and after some predefined teaching pro-
cess. Direct analysis of the teaching process itself can, however,
enable the nature of the process to be studied. One can study, for
example, how social control and discipline are maintained in
classrooms: not taking for granted that they should be main-
tained, but studying just how the trick is done (Torode, 1976).
Or one can study how different forms of teacher-pupil dialogue
inevitably imply sociocultural relations between teacher and
pupil as well as conveying intellectual messages. In a word, one
can begin to study how children are transformed into pupils
(Willes, in press).
Note a point which is often misunderstood. We are here
concerned with analysing what goes on in classrooms, and with
discovering some of the sociolinguistic pressures at work there.
We are not concerned with prejudging what goes on as either
good or bad. The argument for or against the value of question-
answer techniques should be considered separately.

The sociolinguistic situation in classrooms and tests

We can, however, point to the effect of certain dialogue situ-
ations on the verbal behaviour which is likely to be observed.
Work on classroom language is thus immediately relevant to the
earlier discussion (see Ch. 4.3) of how children's verbal compe-
tence is to be judged. We now have more precise sociolinguistic
reasons why such teacher-pupil (and, more generally, adult-
pupil) dialogue is typically a quite misleading context in which
to measure a child's linguistic capacity. The rules of classroom
dialogue are often quite distinct from conversation between
social equals, and the pupil often learns, for example, to give
short answers to discrete questions and not to initiate discus-
115
sion: in other words, he often learns a predominantly passive
role. In many classrooms (but not all) this is what it means to
acquire the role of 'pupil'. In such situations, pupils will
therefore not easily be encouraged to drop well-learned conver-
sational roles and to display the kinds of linguistic competence
which they may well typically display in quite different
sociolinguistic situations with their social peers. Evidence from
several independent studies (e.g. Labov, 1969; Wight, 1975;
Barnes and Todd, 1977) has indicated that teacher absence can
often lead to productive and complex discussion among
children.
The artificial nature of much teacher-pupil dialogue (in
which pupils tell the teacher what he or she already knows) is
analogous to the lack of any real communicative purpose in
many test situations in which a child may have to tell a story
about a series of pictures (see Ch. 3.2). Different types of work
have, however, shown that children's language often becomes
more complex and effective when they have to deal with real
communicative tasks: either as part of communication games
(Wight, 1975; see Ch. 4.5) or in talking about situations which
are important to them (Labov, 1969).
It is also clear that a child will be unable to display his total
verbal competence if he is restricted to a passive response role,
sandwiched between the teacher's initiation and feedback. The
child must also have the opportunity to initiate discussion. It is
therefore clear that if the status relations between adult and
child are highly asymmetrical, the child's language will charac-
teristically be much less complex than in conversation with
social equals.
We must therefore be very aware of the sociolinguistic press-
ures which operate in traditional teacher-pupil or tester-pupil
situations, and thus of the precise limitations on assessing
children's linguistic competence in such settings: limitations
which result from widely shared norms of speech behaviour in
our culture. To summarize: if we want to begin to tap something
approaching a representative range of a child's verbal capacity
we must construct test situations which comprise at least the
116
following factors. The language tasks must have real communica-
tive purpose. (Ideally we would have to observe the child in real,
that is naturally occurring, social situations outside the class-
room.) The rules which govern adult-child talk will often place
the child in a responding role. He will therefore only be able to
display certain verbal skills in a situation with his social equals, in
which he can initiate discussions as well as respond. If we do not
take such factors into account, verbal 'tests' will simply generate
verbal incompetence in children by repressing precisely those
aspects of linguistic proficiency they claim to measure.

117
7

Teaching and talking:

the hidden curriculum
of classroom talk

There is no reason why educationalists should be interested in

classroom language for its own sake. But analyses of classroom
discourse become interesting when they can be shown to be
sensitive to educationally relevant issues. It is important,
however, not to expect such educational relevance to lie too near
the surface. It would be naive, for example, to expect compara-
tive studies of different types of teacher-pupil dialogue to reveal
that one single type of classroom interaction produces better,
faster or more efficient learning than others. Research has never
revealed a clear correlation between styles of classroom dialogue
and 'teaching effectiveness'. Nor is this surprising, since all
teachers know they have to change their style of teaching
according to such imponderable factors as the topic of the lesson
and the mood of the class.
It is possible, however, to study classroom discourse in terms
of the meanings it conveys to pupils, and thus to investigate
much more fundamental questions: what knowledge is trans-
mitted from teachers to pupils? and what counts as educational
knowledge?
118
7.1 The hidden curriculum
Recent work in the sociology of knowledge (e.g. Young, 1971)
has emphasized the need to study how knowledge is selected,
organized and transmitted in schools, since it is clear that what
counts as knowledge is not decidable a priori. School examin-
ations and timetables define for the pupil what is legitimate,
examinable, educational knowledge: this might include, for
example, literature but not cinema, Greek myths but not British
folk culture. But little is known about how what counts as
knowledge is defined in schools and made available to pupils.
One way of studying just how knowledge is organized in the
classroom is to study classroom discourse in some of the ways I
have illustrated in the previous chapter. By studying the se-
quencing and structure of teacher-pupil dialogue, we can study
in empirical detail how teachers divide knowledge into discrete
'topics'; how they relate one topic to another; how they present
knowledge as discrete 'facts' or as more open-ended suggestions
and hypotheses; how they evaluate pupils' contributions as
correct or appropriate; how they control what it is relevant to
talk about in classrooms, and so on. In fact, only by studying
teacher-pupil dialogue directly can we fully study the mecha-
nics of how knowledge comes to be defined and transmitted to
pupils. There are, of course, other aspects of how knowledge
comes to be defined: one might study from this point of view the
implementation of new curricula (Hamilton, 1976).
The term 'hidden curriculum' is used by some authors
(Jackson, 1968; Snyder, 1971) to mean the tacit values and
attitudes concerning appropriate pupil behaviour which all
pupils must learn if they are to be successful at school: values
concerning what is appropriate educational knowledge, what
are appropriate pupil responses to teachers' questions, and so
on. Many such messages are transmitted to pupils, but they are
rarely transmitted explicitly in the content of what teachers say.
Neither, however, are they transmitted mystically or by osmo-
sis. When people are asked how it is that children acquire values
and attitudes, they often say that such values are 'absorbed'
119
 from parents and friends: I usually refer to this as the blotting-
paper theory of culture. It is not, however, a real theory, since it
leaves totally unexplained how values are absorbed, or from
where. It should be clear from the last chapter that many tacit
messages are transmitted by the form and structure of teacher-
pupil dialogue.
Few writers on the classroom have emphasized how regularly
the teacher defines and redefines the classroom situation. Jack-
son (1968) points out that children spend over a thousand hours
per year in school: which amounts to some ten thousand hours
by the time they leave, at least. And for most of this time the
teacher may be talking! In addition, school classrooms are often
rather standardized and routine places: a constant, ritualized,
stylized environment. It would be strange indeed if the very
organization of all this teacher talk did not hammer home time
after time taken-for-granted assumptions and expectations con-
cerning appropriate teacher and pupil behaviour. The medium
has ten thousand hours to convey its message.
For example, Barnes (1969) found that three (English, his-
tory and religious education) teachers used many more ques-
tions demanding facts than reasoning. That is, there were many
questions of the type, 'What books did Homer write?', and few
questions requiring pupils to think things out for themselves.
Barnes points out that a covert message being transmitted to
pupils is that information is more important than original
thought. The teachers never say this explicitly: the message is
not conveyed in the content of what they say. The message is
implicit in the form of the whole teacher-pupil dialogue. In the
relative proportion of different types of questions asked is a
covert message about the nature of the subject being taught. Of
course, the teacher may be quite unaware of this message, but
this does not prevent pupils from receiving it.
We can phrase this point slightly more generally as follows.
By studying the details of teacher-pupil dialogue it is possible to
get some insight into the participants' ideas about how educa-
tional knowledge should be transmitted, and therefore their
conceptions of what 'teaching' consists of. A teacher's use of
120
language in the classroom will serve to maintain a definition of
the situation, not only by maintaining social control and under-
pinning social relations, but also by maintaining a specific
concept of what constitutes valid knowledge and how this know-
ledge should be put across to pupils. In fact, there is no way in
which maintaining social control and transmitting knowledge
can be strictly separated (Young, 1971). In the classroom, we
have a quite specific case where 'knowledge is power'. (Con-
sider, in this connection, how revealing is the ambiguity in the
expression 'academic discipline'!)
In other words, we can regard teaching, first, as a speech
event with specific rules and expectations concerning appropri-
ate teacher and pupil behaviour. Second, we can study these
rules for what they reveal about underlying assumptions about
how knowledge should be selected, organized and transmitted
to pupils. This is, therefore, one way of providing a powerful
definition of what 'teaching' means to teachers and pupils. The
definition is powerful because it is grounded in detailed observa-
tions of the actual classroom speech of teachers and pupils. It is
not a notion which has been thought up in the abstract, indepen-
dent of actual teaching. Nor is it a notion derived, say, from
interviewing teachers and/or pupils: that is, from second-hand
accounts. Rather, it is possible to observe the details of actual
teaching and to recover from such data certain assumptions
about education.

7.2 The framing of educational knowledge

In Chapter 31 criticized Bernstein for failing to relate his theory
to observations of language in use, particularly in the classroom.
Would it be possible to relate some of Bernstein's concepts to the
kind of analysis of classroom language that I have proposed? It
would seem important to try, for the latest collection of his
papers. Class, Codes and Control, volume 3 (CCC 3), sets out to
provide a framework for a 'theory of educational transmission'.
This theory is seen as part of a more general 'theory of cultural
transmission': a theory of how cultural norms and values are
121
passed on from one generation to the next, from parents and
teachers to children. One of Bernstein's criticisms of American
sociolinguistic work on education is that it has failed to provide
any systematic theory of the transmission of educational know-
ledge (CCC 3, p. 29, and see Ch. 4.4 above).
In a major paper on this topic, Bernstein (1971b) discusses
how educational knowledge is selected, classified, distributed,
transmitted and evaluated. The way a school timetable is di-
vided into different academic subjects, and how these subjects
are presented and paced, is often taken for granted. But, of
course, there are many different ways of selecting what should
be taught, when and how. Innovations, such as Nuffield
science, integrated or interdisciplinary studies, have often
brought into question how educational knowledge itself is
defined. Bernstein proposes three 'message systems' through
which educational knowledge is transmitted: curriculum defines
what counts as valid knowledge; pedagogy defines what counts as
valid transmission of knowledge; and evaluation defines what
counts as valid realizations of this knowledge on the pupil's part.
In the last chapter, I discussed, from my own point of view,
pedagogy.
There are several concepts in Bernstein's paper which are
beyond the scope of discussion here, although they are very
important in their own right. The concept of Bernstein's most
relevant to my discussion here is frame. This refers to the actual
relationship between teachers and pupils, to the 'strength of the
boundary between what may and may not be transmitted', and
therefore to the range of choice teachers and pupils have over
what is to be taught. Bernstein's discussion is entirely abstract,
but an aspect of this control over knowledge could be studied in
transcripts of teacher-pupil dialogue. According to Bernstein,
'strong frames reduce the power of the pupil over what, when
and how he receives his knowledge'. Bernstein gives no
examples at all to relate his concepts to observed classroom
interaction, although he proposes that the concepts can be used
at the level of classroom encounters (CCC 3, p. 8). But I have
already illustrated in detail how a study of classroom dialogue
122
can investigate the specific mechanisms by which knowledge is
transmitted and paced. A teacher, for example, whose predomi-
nant teaching strategy involve strings of IRF exchanges would
be a very strong framer (see Ch. 6.2).
Bernstein has referred to his work as a 'theory of the structure
of cultural transmission'. One way in which he explains the
success of many MC children at school is to refer to the
'continuity of culture' between home and school. But precisely
what is transmitted and how, and just what this 'continuity'
comprises, is never made entirely clear. There are many levels at
which such 'continuity' might exist and could be investigated.
Continuity of culture (just restricting ourselves to cognitive/
verbal culture) might refer to many things: to the presence of
books in the home; to the use of the same standard dialect at
home and in school; to the use of the same elaborated code; or to
playing the same discourse games. Consider this extract of
mother-child dialogue. The child is 4 years old, and they are
cooking together:
M: Right, now - what do you think the next instruction is
because that's what I've got to do?
C: Put it in the baking tin.
M: Yes.
Well, first of all we've got to grease it though - why, do you
think? Why do you grease it Tommy?
C: So the pastry doesn't stick.
M: Right.
The discourse structure here is identical to some of the teacher-
pupil dialogue discussed above. The mother is asking 'test
questions' to check if the child knows certain things, and the
conversation is quite incidental to the cooking. (Imagine the
absurdity of this conversation between husband and wife,
instead of between child and mother!) It is a purely verbal game,
a knowledge game, which fits the IRF pattern. Clearly, this
child will have no difficulty in recognizing the same game if it is
played at school. This may then identify one specific level of
communicative behaviour at which culture is transmitted.
123
Adults talking to children often ask test questions to which they
know the answer. Thus the following interchange overheard in a
cafe, between an adult and a girl aged about 4 years:
A: What's my name?
C: (Smiles)
A: Is it Brian?
C: No.
A: What is it?
C: David.
A: Yes!
The child may well have been smiling politely at the lunacy of
the question. Again imagine the absurdity of this interchange
between two adults. Our society has characteristic ways of
talking to its children.
I do not wish here to make premature generalizations. On the
contrary, I am saying that such topics have hardly been investi-
gated. There is an almost total lack of field studies of how
educational knowledge is transmitted (although Barnes, 1969;
Keddie, 1971; Furlong, 1976; Gannaway, 1976; Delamont,
1983 are a start). But it is possible to study, for example, some
specific ways in which the culture of home and school may be
continuous or not, and to relate such high-level concepts as
'strength of frame' to actual observed and recorded interaction.
Wells et al. (1981) is the first volume of an important series
which studies children's language at home and at school on the
basis of a large amount of tape-recorded data collected over
some ten years.

7.3 Discourse structure and assumptions about teaching

What we are concerned with, then, is messages which may be
transmitted by the structure and sequencing of teacher-pupil
discourse, and the ways in which teaching and learning roles are
defined.
As a specific example, consider some possible implications of
the discourse structure IRF: teacher initiates - pupil responds -
124
teacher gives feedback (Sinclair and Coulthard, 1975; and Ch.
6.2). In dialogues of this IRF structure, the pupil's role is
passive: he must respond. It is the teacher who initiates, and
then evaluates the response before asking another closed ques-
tion. This conversational structure gives the teacher almost
complete control over initiating the topic, and over evaluating,
accepting or rejecting the pupil's contribution, and thus over
closing the exchange. Anything the pupil says is sandwiched
between whatever the teacher says. Note just how basic are the
assumptions underlying such a discourse structure: the assump-
tion that it is the teacher who has control over who talks when;
and that education consists of listening to an adult talking, and
answering his or her questions. / / classroom dialogue was
constructed primarily from such exchanges, a general tacit
message which might be transmitted by the three-part IRF
structure is something like:

Classroom knowledge consists of strings of short answers

which can be individually evaluated. Classroom knowledge
is therefore essentially closed, not open-ended. All questions
have correct answers. Teacher-pupil talk is effectively a
monologue with the pupil supplying short answers on de-
mand to contribute to the teacher's train of thought.

These are the messages which pupils might receive from the
form or structure of the dialogue, quite independently of the
content of the dialogue. Note that the only other common
conversational structure of the form question-answer-evalua-
tion is a riddle! Some teacher-pupil dialogue is quite literally
composed of little riddles where the pupils have to guess what
particular word or expression the teacher is thinking of. Thus,
an English teacher discussing a poem - his pupils have failed to
use a particular word which he wants them to:

T: Well, I'm going to help you - a word beginning with A.

P: Attitudes.
T: Yes - now answer again using the word 'attitudes'.
125
Or, another example, the teacher has been discussing how one
speaks to different people: in this case the pupils have failed to
guess another particular word the teacher 'had in mind', as we
often put it.
T: 'Respect' - this is the word I wanted to come out earlier
but it didn't - you have to speak 'respectfully' to a headmas-
ter.

Or consider, more generally, the message pupils may get from

teachers who never ask questions because they want to know
something, but because they want pupils to display their know-
ledge. The message here might be: teachers know everything or,
at least, everything that pupils know, and the pupil therefore
has nothing new to tell the teacher. Of course, the teacher may
not believe this, but this may be the message that the pupil
receives, transmitted by the structure of the classroom dialogue.
All teachers know that what they teach is not always what pupils
learn!
Atkinson (1975) discusses the mutual pretence involved when
teachers ask pupils questions to which they already know the
answers. He suggests viewing such conversational situations as
information games: they are simulations or 'mock-ups' in which
the teacher is asking about his own knowledge, and organizing
the classroom dialogue so that little bits of knowledge are
allowed to emerge when the teacher considers it appropriate. He
concludes that: 'The maintenance of reality based on the princi-
ple of discounting, suppressing or covering previously acquired
knowledge may turn out to be a fundamental feature of instruc-
tional situations.'
Suppose, on the other hand, that teacher-pupil dialogue
characteristically takes this form:

(A discussion about corporal punishment has been underway

for about 10 minutes.)
T: You don't think corporal punishment is, er - in a school
- you think corporal punishment is all right at home - but
er, not in a school.
126
PI: No, I don't say that. I said until a certain level the cane I
am against.
T: 'Until a certain level' -I don't understand you.
P1: Ah yes, I explained 10 minutes ago.
T: Well, I still don't - 'until a certain level', I don't -I don't
quite understand what you mean.
P1: The cane I am against, slaps I am for.
T: Oh yeah-I see.
P2: I can't agree - if, er, a smack can do nothing.
T: A slap?
P2: A slap can do nothing if, er -I don't know - a text to learn
by heart can do nothing.
T: You think that a text is just the same thing - thing to give,
er - something like, em - lines - to write out or to learn - it's
just the same thing?
P2: It's not the same thing -I don't say that - it has no more
effect.
T: It has no more effect.
(The discussion continued with P2 telling a story about a
friend of a friend who had committed suicide after being
corporally punished in school. The teacher brought the dis-
cussion to a close as follows.)
T: Would you like to, er, say - sum up what you think about
corporal punishment in general?
P1: In general?
T: Like to sum up, yeah - what you think now after this
discussion - in a few words to say - what you think.
P1: I am still of the same opinion. I am against.
T: You're against corporal punishment.
P1: Yes.
T: And,er.
P1: There are too many bad consequences in the future for -
P2: But I keep the same opinion as the, er
T: You have the same opinion.
P2: Yes, because what you said - what you said - what you
told us, it's nothing. I have destroyed - for me. I think that
127
 - it seems to me that - it seems for me that with the last
example that I gave you, all your opinions are com - all
your, em -
T: Arguments.
P2: Arguments are completely destroyed.
T: For you.
P2: Yes, I think so.
T: Well, I think we'll leave it at that.
In this dialogue there is no IRF structure. The pupils not only
initiate exchanges, but question the teacher's interpretation of
what they have said. This is rather more what we mean by
genuine discussion. And it implies quite different concepts of
knowledge, and a different concept of teacher and pupil roles.
Again, note how this approach to classroom knowledge shows
the impossibility of separating cognitive and social aspects of
learning.

7.4 The social construction of children's ability

It is not only knowledge which is (partly) defined through
teacher-pupil interaction. A pupil's ability must be constructed
and defined through classroom dialogue.
Hammersley (1974) discusses how it comes about that some
pupils are judged 'intelligent' by teachers. He points out that
such judgements are made on the basis of pupils' behaviour,
particularly their ability to answer teachers' questions correctly.
That is, such classifications of pupils are made (partly) on the
basis of their communicative competence in participating in
classroom discourse. In the kind of formal chalk-and-talk class-
rooms Hammersley studied, pupils have to demonstrate intelli-
gence by getting the teacher to hear their answers and accept
them as 'correct'. In order to do this, they must recognize the
teacher's sole right to organize classroom interaction, to choose
topics of discussion, and to demand and evaluate their answers.
Pupils who challenge these rights are liable to have their answers
ignored. The teacher might say, for example, 'Don't shout out!'

128
Answers must be appropriately presented in order to be accept-
able.
Mehan (1973) takes this idea further by showing how a child's
'ability' is not an absolute quality, but the outcome of a social
encounter: the 'test'. Test results are often regarded as the
product of a passive, standardized, routine record made objec-
tively (mechanically?) by the tester. But such results, and
therefore the child's IQ, are rather the outcome of how the child
interprets the questions and how the tester interprets the answers.
Thus one question in a language-development test instructs the
child to choose the 'animal that can fly', from a bird, an elephant
and a dog. Many young children choose the elephant along with
the bird. If they are asked why, they explain 'That's Dumbo',
i.e. Walt Disney's flying elephant. A 'wrong' answer here does
not therefore indicate a lack of ability, but an alternative inter-
pretive schema: in this case they are not separating what adults
would regard as reality and fantasy.
We have already seen (Ch. 6.4) that the social relations in the
classroom or test can crucially affect estimates of children's
linguistic capacity. Now, we see also that the test is itself a
communicative encounter, a social dialogue of questions and
answers, which can itself be studied as discourse. Children do
not 'possess' an IQ: their ability is constructed by the very social
situation in which it is measured. It is thus through the organiza-
tion of teacher-pupil or tester-pupil dialogue that children's
ability comes to be constructed by teachers and testers. Pupils
are judged bright or dim according to whether they interact
appropriately, and as adults expect, in particular sociolinguistic
situations. Much work has now shown how children's perform-
ance may differ radically in different social situations. (See
Labov, 1969, discussed in Ch. 4.3; Philips, 1972; and Dumont,
1972, discussed below.)

7.5 Teaching as talking: some cross-cultural data

Consider, finally, some broader ways of thinking about what
educational assumptions and conceptions of knowledge under-
129
lie classroom language. Many educationalists have put forward
versions of the view that pupils learn through talking. One often
comes across statements such as:
Not only do we learn by doing, but we also learn by talking
about an experience. . . . Our pupils will learn most by
reading, writing and talking about the experiences they meet
and through this will in time come to terms with subject
disciplines. (Barnes et al., 1969, p. 126)
Many teachers certainly hold the view that pupils learn by
expressing things in their own words. This is why we distrust a
pupil's work if he or she has simply copied it from a book.
Flanders (1970) takes it for granted that it is a good thing if
pupils can be encouraged simply to talk more in lessons. Clearly
he is interested in quality as well as quantity, but simply a larger
amount of pupil-talk is assumed to be a good thing. Much of my
discussion of teacher-pupil dialogue has centred on the fact that
teachers are often concerned to elicit talk from pupils, to get
them to answer questions at appropriate moments.
Consider this teacher with a class of 12 year olds:
T: David, about time I heard your voice this morning - so
wake up, it's not very difficult this, for even you. Come on,
David, show some sparks of life.
P: (Answers question)
T: What? Pardon? Well, speak up, don't speak to your hand,
your hand is not very interested in this - we are . . .
This teacher is making fairly explicit some of his assumptions
about the place of language in the classroom: that it is a good
thing just to 'hear a pupil's voice'; and that public talk should be
the norm. This teacher clearly assumed that an important part
of learning is a public, verbal display of knowledge. Often, we
tend to evaluate teachers according to how freely they can 'get
pupils to talk'.
One might question the view that it is a good thing if a pupil is
merely standing on his feet, composing talk in real time in front
of an audience. Some control over style and content is necessary
130
at some point. But there is a prima-facie case, if only from
common experience, that one can clarify ideas by 'talking them
through' and defending them in dialogue with others. There
are, of course, things that have to be demonstrated or practised
in order to be learned. A pupil cannot demonstrate knowledge
of technical drawing just by talking about it. But, in general,
schools are particularly concerned with publicly communicable
types of knowledge, and, in the past at least, 'education' has
often been equated with literacy.
Note, however, that this view of education is culture specific,
That is, this equation of learning with talking is widely held in
our society, but is by no means universal. Philips (1972) and
Dumont (1972) discuss the problem of the 'silent' 'American
Indian child. Both authors start from the frequent complaint of
White teachers with classes of Indian children that their pupils
'won't talk' in class - although they are observed to be highly
verbal in other social situations. They both explain this by
showing how the Indian groups in question (Cherokee and
Sioux) have quite different sociolinguistic norms and values
concerning the use of speech. In particular, Philips shows that
the Indian groups believe that learning occurs through observa-
tion, supervised participation and self-initiated testing. Speech
is minimal in this process.
Such ethnographic data from different cultures shows up our
society's taken-for-granted equation of language, written or
spoken, as the primary channel through which educational
knowledge should be transmitted. This is not a necessary or
natural feature of learning. Our society values highly the role of
language in education and tends therefore to assume that
children with radically different language from their teachers
are cognitively different, or even cognitively deficient (see Ch.
4).
In summary: very general sociolinguistic norms may be
conveyed by the fine grain and overall structure of classroom
dialogue. Teachers who ask sequences of closed questions, or
university lecturers who talk for 50 minutes without stopping,
or teachers who try to get pupils to talk more, are all displaying
131
in the form and structure of their talk implicit theories of the
relation of teaching and learning to language, and of what it
means for students to be 'intelligent'.
The study of classroom language thus brings out very clearly
that the act of teaching is culturally defined. Teaching IS
sociolinguistic behaviour, and by studying it as such one can
recover some of the educators' assumptions which underlie it.
Different levels of meaning in this chapter's title, Teaching and
talking', should now be evident. On the one hand, teacher-
pupil talk is just talk and is open to sociolinguistic description as
any other type of discourse is. But also, our culture assumes that
teaching and learning are somehow necessarily dependent on
language. This culture-bound assumption about the role of
language in education is, in turn, revealed in the underlying
discourse structure of the classroom talk itself.

132
 8

Towards a
sociolinguistic
analysis of language
in education

Throughout, we have seen the complexities of the relationships

between language and the social contexts in which it is used, and
we have therefore seen that any simple causal model purporting
to relate superficial aspects of language directly to educational
processes will be oversimple. 'Superficial' aspects of language in-
clude accent, grammatical differences between standard and
nonstandard dialects, the proportion of grammatically complex
sentences a speaker uses, and so on. There is no evidence
whatsoever that such features of language are related, for
example, to thought processes. We have also seen that certain
concepts (such as restricted code and verbal deprivation), which
are being taken for granted by many educationalists and psycho-
logists, are being seriously questioned by linguists as having
little or no basis in linguistic fact. We must therefore now try to
reach some conclusions about the state of research in this area,
and about what would constitute a sociolinguistically adequate
statement of the role of language in education. These will be
tentative conclusions, it is only since around 1970 that any
concentrated amount of work has been done on the role of

133
language in education. Much of the work has been done by
psychologists and sociologists, and only comparatively recently
have linguists begun to show an interest: Hymes (e.g. 1972) and
Labov have led the way here.
Certain principles do, however, seem clear. I have argued that
an adequate analysis of language in education must be based at
least partly on (a) a close analysis of real language, observed and
recorded as far as possible in natural social situations, especially
in the classroom itself; and (b) adequate sociolinguistic concepts
to handle the complex relations between language, attitudes to
language and the social contexts of language use. We require
also (c) a systematic framework which describes the languages in
use in the wider community beyond the school and in the
country as a whole. In Britain, this means different accents and
dialects of English, whether native to Britain (for example,
Cockney) or spoken here due to recent immigration (for
example, Caribbean Creoles); and also languages other than
English, whether native to Britain (for example, Welsh) or
spoken here due to recent immigration (for example, Punjabi).
A full discussion of this aspect of language diversity is beyond
the scope of this book, but is discussed in more detail in the
companion reader.

8.1 Language as evidence for educational statements

The work discussed in this book comprises attempts, in one way
or another, to use aspects of language as evidence for educational
statements. That is, aspects of children's or teachers' language
are recorded or observed, and these linguistic data are used as
'indicators', 'markers' or 'indices' of learning and teaching
processes. Sometimes, aspects of language are said, further, to
be the causes of educational phenomena, for example the deter-
minants of teachers' effectiveness or pupils' educability. It is
revealing to consider just what is related to what in such state-
ments.
It becomes clear, first, that different studies take language as
134
evidence of a very mixed collection of educational processes,
including: the cognitive orientation of pupils; possible barriers
to pupils' learning and understanding; teaching strategies, e.g.
'open' versus 'closed'; methods of social control in the class-
room; different types of classroom organization, e.g. 'formal'
versus 'informal'; taken-for-granted concepts of classroom
knowledge. This list could be extended.
Second, different studies pick out a very mixed collection of
linguistic items as 'indicators' of such educational concepts,
including: teachers' differing use of pronouns (e.g. Mishler,
1972; Torode, 1976); the complexity of noun groups (cf. Ch.
3.2); intonation and paralinguistic clues (cf. Ch. 6.1); the overall
structure of teaching cycles (cf. Ch. 6.2). This list could also be
extended. In general, different researchers seem to feel justified
in selecting as evidence any feature of language which strikes
their intuition as interesting. More particularly, the linguistic
items are selected, apparently, according to the whim of the
researcher, from different levels of language, including lexis
(i.e. individual words), syntax (i.e. grammatical structure),
semantics (i.e. meaning) and discourse (i.e. overall conversa-
tional structure).
Such linguistic items are then often related directly to educa-
tional statements. Together, then, these points mean that stu-
dies of language in education often attempt to relate concepts
which are at quite incompatible levels of abstraction and gener-
ality. Often isolated, surface features of language (e.g. use of
pronouns or frequency of complex noun groups) are taken as
indicators of highly abstract educational concepts (e.g. 'open'
teaching style or educability). Many studies of classrooms, for
example, characteristically select for quotation short extracts of
recorded classroom talk, and select particular features of the
extract as evidence for some nonlinguistic statements. It is often
assumed that details of what teachers and pupils say can serve
unproblematically as illustrations of the educational process.
Thus readers may be expected to be able to recognize a particu-
lar teacher-pupil interchange as an instance of, say, a 'democra-
tic' teacher or of a pupil who is 'thinking for himself. But such
135
claimed relationships, between surface features of language and
educational concepts, may lack any principled basis.

8.2 Language is organized

On the face of it, there is no reason why superficial or isolated
linguistic items should be interpretable as direct evidence of
educational processes. For example, the cognitive complexity of
an argument can never be computed directly by calculating the
average length of its sentences, or by counting the number of
subordinate clauses, rare adjectives or complex noun groups it
contains. Such crude linguistic measures have been used in
some studies as indices of cognitive variables, and such studies
have been cited above. But it should be clear that there is no
necessary connection between complex or unusual linguistic
structures and complex cognitive processes. Profound argu-
ments may be stated in simple language, and trivial thoughts
may be dressed up in superficially complex language.
The relation between language and social or cognitive proces-
ses is much less direct than this - and much more interesting.
Consider, for example, the type of relationship which
sociolinguists have found between language and social struc-
ture. Labov's work can provide a precise example. In a major
article, Labov (1973) studies the language of pre-adolescent
Negro gangs in Harlem. He found the boys' language to be a
sensitive index of how closely the boys were involved in the street
culture. That is, he identified core members and marginal
members ('lames') of the gangs by nonlinguistic methods. He
then found that the language of the core members differed
systematically from the lames' language. But the important
point is that their language did not differ in the use of particular
words or isolated items. What differed was the whole system.
Labov's finding was that the core members' language was more
consistent in its use of the rules of BEV. The index of street
culture is therefore provided by the overall consistency of the
whole linguistic system.
136
 The main point is that if language is to be used as evidence of
social structure and processes, then it must be examined as a
system, not as isolated items. Since relations are found between
linguistic systems and social structures, the language data must
be studied for their own systematic linguistic organization.
Another way of phrasing the criticism of the last section is
therefore to say that language is often used as data for educational
statements as though the language had no organization of its own.
However, one of the fundamental principles of this book has
been that we will never fully understand the role of language in
education unless we take into account how language itself works.
Linguistics treats language as a highly organized phenomenon,
tightly structured and patterned at many levels, including the
levels of sound, word structure, meaning, sentence structure,
discourse and social context.
The principle of treating language as an organized system has
been illustrated in several ways. Detailed examples were given
above of why standard and nonstandard dialects must both be
treated as systematic wholes, with their own internal organiza-
tion. It is quite misleading, and counter to all the linguistic
evidence, to treat nonstandard dialects as erratic deviations
from the standard language: they are highly organized systems
in their own right (see Ch. 4.1). It is similarly misleading to treat
spoken language as a series of isolated utterances, rather than as
connected discourse. My criticism of Hawkins's experiment
(Ch. 3.2) was that he failed to take account of the conversational
organization of his data, but tried instead to relate isolated
linguistic items (frequency of noun groups) directly to the
cognitive orientation of children. That is, Hawkins proposed a
cognitive explanation of his data, whereas I would propose a
sociolinguistic explanation, as Hawkins (1977) himself does in
later work.
As an example of the need to treat language in terms of its
discourse organization, and not as single, isolated utterances,
we might return to the question of assessing children's verbal
capacity. A book on Bernstein's theories (Lee, 1973) has the
following cover design:
137
 Q: What is a dog?
A: A dog is furry and barks.
Restricted code? Elaborated code?
(This exchange is shown against a photograph of a little boy who
is looking studiously at a book.)

Q: What is a dog?
A: You have a dog at home.
Restricted code? Elaborated code?

(This exchange is shown against a photograph of another little

boy, who somehow looks rather scruffier, and who is gazing at
nothing in particular, turned round in his chair.)
This is hypothetical data, of the crude kind rejected earlier,
comprising the cover design of a book; but for that reason it
catches the attention and is worth comment. In this area, it is
often the crude stereotyping of linguistic behaviour which does
damage. I take it the implication is that answer (1) is elaborated
code, because it is an explicit definition; and that answer (2) is
restricted code, because it is implicit and would not define a dog
for someone who had no idea what a dog was. On the other hand,
regarded as discourse, exchange (1) is rather odd and stilted.
Whereas exchange (2) shows the kind of oblique answer to
questions which commonly occurs in normal conversation.
Answer (2) would, in many circumstances, be a more helpful
type of answer: it is the kind of answer which refuses, sensibly,
to assume that the questioner really has no idea what a dog is,
and thus refuses to treat discourse as an artificial game. The
moral is that utterances have to be seen within their system of
discourse, and not as isolated events. And in general therefore,
linguistic items have to be related to linguistic systems before
being used as evidence for cognitive or sociological statements.
A well-known analogy for this point in linguistics is as
follows. Suppose you want to analyse the meaning of a move in a
game of chess. It would be pointless to try and explain a move by
direct reference to the intelligence of the players, or other such
factors. The point of a particular move can only be understood
138
by relating it both to the state of play on the board and also, of
course, to the rules of chess. For someone observing a game of
chess with no knowledge of the rules, any isolated move would
literally have no meaning at all. A move only has meaning within
the system of the game, and has no absolute value in itself. This
analogy holds for any game. In card games as a whole, the five
of spades, say, has no meaning. Its value depends (a) on the
particular game being played; and (b) on what context it occurs
in within the game, e.g. along with three other fives, or in a run
of spades. It is similarly misleading if a literary critic tries to
relate an incident in a novel directly to, say, the social context in
which the book was written; an event in the author's childhood;
or some underlying characteristic of the author's personality.
An incident in a novel must first be understood in relation to the
structure of the novel as a whole: it might, after all, be required
simply to advance the plot one stage.
Such analogies could be multiplied. Although analogies can
be misleading if pushed too far, they are often helpful in
grasping a concept. In a book of this sort, it is possible only to
suggest a couple of analogies to the reader and to refer back to
the previous discussion of the ways in which language is org-
anized at the levels of dialect (Ch. 2.2 and 4.1), discourse (Chs
6.2 and 7.3) and grammar (Ch. 2.4).
It is, then, only sensible to try and account for bits of language
in terms of linguistic systems wherever possible, before pro-
ceeding to non-linguistic explanations. Again, this has been a
general, but so far largely implicit, principle in the present
book. One of the general arguments throughout has been that it
is more plausible to account for the effect of language on
educational success in terms of people's attitudes to language,
rather than to propose that different language varieties have
different cognitive effects. That is, I have constantly proposed a
sociolinguistic account of language in education, rather than
make direct links between language and cognitive ability. Lan-
guage and thought are related in some way, but no one has yet
shown precisely what this link is, and it is demonstrably not a
simple, causal link. Similarly, one can account for the poor
139
performance of many children in classroom or test situations by
pointing to precise sociolinguistic constraints on such occasions.
Again, one can propose a sociolinguistic explanation of differ-
ences in language use between different social groups, without
resorting to the assumption that one group is cognitively in-
ferior to the other.
The general theoretical point at issue here can be summarized
as follows. When we are dealing with linguistic data (audio-
recordings, fieldnotes, language elicited in tests and so on) we
must study how linguistic items are related in linguistic and
sociolinguistic systems (e.g. within dialects, in connected dis-
course and in norms of usage in different social contexts). It is
not sufficient to relate linguistic items directly to non-linguistic
(e.g. cognitive, educational, sociological) categories. They must
first be related to the linguistic and sociolinguistic systems in
which they are terms.

8.3 Criteria for studies of language in education

The demands which one has to make for work on language in
education are therefore as follows. The work should be based
primarily on naturalistic observation and recording of language in
real social situations: mainly in the classroom itself, but also in
the home, and in the peer group, which is the most powerful
linguistic influence on children. The work must be based on a
linguistically adequate analysis of what is said. This means both
being explicit about the relation between language forms and
functions and also analysing language as linguistic systems. It is
not enough, however, for the analysis to be rigorous in a
mechanical way: what is required is an analysis of the social
meanings conveyed by language and an analysis of people's
attitudes to language. Finally, if we are to understand the general
principles underlying the sociolinguistic forces at work in
schools, the analysis of language in educational settings must be
related to what we know of sociolinguistic behaviour in other
settings.
These demands are stringent, and it will be clear to the reader
140
that no work, including the present book, yet satisfies them on
all counts. Much work fails, however, to meet any of the
demands, while other work meets some of the criteria. And the
demands must be made if we are to progress beyond linguistic
folklore, personal opinions, unsupported speculation, pseudo-
theories and an accumulation of interesting anecdotes; none of
which, in themselves, can lead to real understanding.

141
9

Some topics for

investigation

This book has tried to provide the reader with the basic concepts
necessary to understand the continuing debate over the place of
language in education. Most importantly, perhaps, it has tried
to illustrate how various researchers have begun to investigate
how language is used in schools and classrooms. Rather than
summarize the argument, I will devote this last chapter to listing
some topics which readers could now go on to study for them-
selves. Suggesting topics for investigation will also point to
many aspects of language which this book has not dealt with,
largely because we know very little about them.
There is no reason at all why student teachers, for example,
should not investigate, possibly in a quite informal way, one of
the topics suggested below. Students on teaching practice often
have to spend considerable periods of time just 'observing'.
What this means depends very much on the school they find
themselves in. But students often find themselves sitting in
classrooms, unsure of what or how to 'observe'. Yet armed with
a few ideas, a notebook and possibly a tape-recorder (if teachers
and pupils have no objection to being recorded) the observation
142
period can be exciting and informative. So here are several
topics on which readers could collect badly needed information.

Different types of classroom language

We still know very little about what actually happens in class-
rooms, between teachers and pupils, and have little basic
descriptive information about teacher-pupil dialogue in differ-
ent teaching situations. The extracts of teacher-pupil talk in
this book have been drawn from a very narrow range of class-
rooms, largely from rather formal chalk-and-talk lessons.
Readers of this book will have experience of many types of
teaching which have not even been touched on. We lack basic
descriptive information about the kinds of teacher-pupil in-
teraction in small rural schools, in bilingual classrooms in
Wales, in open-plan Nuffield science lessons, in remedial read-
ing classes, in team-teaching sessions and so on. Primarily, we
need to be able to relate generalizations about language in
education to a wide range of observational data on different
types of classrooms. The most important research will certainly
be done in the classroom, and such information can be provided
only by observers in the classroom. Walker and Adelman
(1975a) provide excellent practical advice on doing such
observation.

Learning the rules of classroom discourse

Chapters 6 and 7 discuss the expectations which teachers and
pupils have about appropriate classroom discourse. But little is
known about how children learn the rules of the verbal game.
Classroom discourse is characteristically different from other
discourse types, and somehow children must learn the
sociolinguistic behaviour expected of them. Some children may
learn one type of classroom game from their mothers (see Ch.
7.2). Other rules may be taught explicitly by teachers who tell
children, 'Don't shout out!', 'Don't all talk at once' or 'Put your
hand up before your answer'. Much information could be
143
collected on this by observing playgroups, or in particular, the
first few days a child spends at primary school or at a new school
or with a new teacher. Willes (in press) provides a model for
such a study.

The language environment of a lesson

The total language environment of a lesson consists only partly
of the teacher-pupil dialogue, although in most lessons this is
likely to be the major part. It consists also of textbooks,
worksheets, handouts, charts or posters on the wall, whatever
the teacher writes on the board, pupils' writing, and possibly
radio programmes or films. As well as recording public talk, it
might be possible to collect copies of worksheets or similar
material; to xerox relevant pages of books or pupils' own
writing; to copy down blackboard notes or to photograph them
along with relevant wallcharts. It may also be possible to collect
observations on pupil-pupil talk in lessons. Often this is not
public and is consequently difficult to observe, but a great deal is
possible by informal observation.
In this way, one might try to construct, as far as practically
possible, the total language environment of a single lesson.
When such material has been collected, it can be compared in
various ways, to study, for example, how written and spoken
materials on the same topics differ in style and in the linguistic
demands made on pupils.

The uses of written material

Part of the language environment of a classroom is written
material. However, it is rare to find observational studies of how
pupils actually use books, worksheets, wallcharts, notes on the
blackboard, and so on. How much of their time is spent reading
and writing? Do they read aloud or silently? Do they discuss
written material with a neighbour or in groups? Do they learn it
by heart or rewrite it in their own words? In reading a book, do
they start at page one and keep going, or do they use the contents
144
page and index to move around in the book to find what they are
immediately interested in? Do such reading strategies differ
with different kinds of books, for example fiction versus non-
fiction? Do they use reference books such as dictionaries and
encyclopedias? Lunzer and Gardner (1978) discuss many such
issues.

The language environment of a school

Alternatively, one could collect examples of the range of
different kinds of language used throughout a school. What
kind of hidden curriculum is conveyed, for example, by the
language in different textbooks? the language of the school
rules? the language of morning assembly or prize day? the
language of the school motto? the language of the school or
class magazine? the language of examination papers? and
soon.
The term 'language climate' is sometimes used to describe the
very different kinds of linguistic demands made on pupils by
such uses of language. Often this term implies a vague and
woolly horticultural theory of education: that learning 'thrives'
best in a favourable 'climate'. But it is possible to specify
precisely what constitutes such a climate by collecting speci-
mens of language and considering just what implicit messages
they convey.

Child-child talk
Although the most important research will be done in class-
rooms, we also require descriptions of how children talk to other
children when no adults are present. For it is a well-documented
sociolinguistic fact that a child's friends are a more powerful
influence on his or her language than both family and school.
Almost all the research on children's sociolinguistic behaviour
has focused on adult-child talk, usually mother-child or
teacher-pupil. We know very little about how children talk, in
the playground or outside school, when no adults are around.
145
Clearly there is a paradox in trying to observe how children
interact when the adult observer is not present. But a lot is
possible by informal observation of snatches of conversation
caught in passing, or by tape-recording groups of children (with
their permission!) when no adult is in the room. The pressure of
normal social contacts with friends will often outweigh, at least
partly, the presence of the tape-recorder. Labov (1973) provides
an important technical study of the powerful influence of gang
membership on children's language.
A good way to collect data on children's language is to take
advantage of any opportunities which arise to observe children
when they gather together in places with no adults in control: in
adventure playgrounds, at school bus-stops, and the like. If it is
possible to observe children without attracting their attention,
good data can be collected informally on how they interact
amongst themselves. The conversational rules are very different
from adult-child or adult-adult dialogue! McTear (1981) pro-
vides interesting analysis of conversation between two young
children.

The languages used in a school

In multilingual classrooms it might be possible to investigate
which languages children know. This would have to be
approached very carefully and tactfully, since such questions
will almost inevitably raise social and ethnic attitudes, and
whilst some children may be proud of their knowledge of
different languages, others may well be ashamed of what they
think are 'not real languages' or what they know well are
low-prestige varieties. The questions rapidly become complex.
Just what is meant by knowing a language? Can languages just
be counted? Do children have a passive understanding of a
language other than English, or do they actively use it at home?
Do they use it only with particular groups of people? Do they
read and write it as well as speak it? Does the language have
particular cultural or religious functions? Rosen and Burgess
(1980) discuss both the difficulties in defining what is meant by
146
'a different language' and also the tact and sensitivity required
by anyone who plans to do some investigation in this area.

Sociolinguistic barriers between schools and pupils

The argument of Chapters 3 and 4 was that, rather than talk
about pupils' 'linguistic failure', it is more useful to talk about
sociolinguistic barriers between schools and pupils, and that one
aspect of such barriers is teachers' attitudes to children's lan-
guage. But we know very little about precisely what attitudes
teachers have towards nonstandard dialects or towards what
constitutes 'good English', or how these attitudes are transmit-
ted by colleges of education to teachers and by teachers to
pupils.
Such attitudes may be informally investigated in various
ways. In the classroom one may be able to observe directly what
kinds of language teachers correct as being 'wrong', and
whether they give pupils any rationale for such judgements.
Alternatively, one might talk to teachers about the importance
they attach to 'good' English, and the ways they have, if any, of
encouraging it. Similarly, pupils often have much better de-
veloped ideas on such topics than they are generally given credit
for, especially if the discussion can be based on recordings of
other people's speech (see Ch. 2.1). Collecting such data from
various sources could give a fairly good picture of a school's
attitudes to language. Milroy and Milroy (in prep.) provide a
detailed sociolinguistic analysis of prescriptive attitudes in
schools.
One particular kind of sociolinguistic breakdown can occur
between teachers and pupils who have different language sys-
tems: for example, different social dialects. Labov (1972a) has
shown that teachers and pupils are often mutually ignorant both
of each other's dialects and also of each other's attitudes and
values about language. It would be possible to interview
teachers and pupils separately about their attitudes to local
dialect versus SE. Do they, for example, share the same value
system? Do they recognize the same features as being non-
147
standard or stigmatized? Alternatively, one might talk to
teachers to discover how they think their pupils speak, which
nonstandard forms they use, and so on. And then one could
compare this with tape-recordings of pupils' actual speech.
What one characteristically finds is that people perceive speech
in categorical, all-or-nothing terms, while speakers' actual
speech is variable (Labov, 1972b). Thus a teacher might per-
ceive pupils as using double negatives all the time, whereas they
actually use the standard form in some contexts.

It is theoretically unsound to make generalizations about

language in education which cannot be related to how language
is actually used in schools and classrooms. But it is equally
unsound to make intuitive inferences on the basis of isolated
observations and anecdotes about language use, without con-
sidering the complexities of the sociolinguistic context of
schools and classrooms. I hope that at least some readers of this
book will use some of its concepts and suggestions to observe
and study their own classroom experiences. Only in this way can
we ever really discover the role that language plays in education.
When language is seriously studied in its own right, it
becomes clear just how awesomely complex is the sociolinguistic
competence of all speakers. And when such a position is
reached, it becomes impossible to maintain a position of
linguistic prejudice or intolerance.

148
 Further
reading

1 Companion reader
Michael Stubbs and Hilary Hillier (eds) (1983) Readings on Language,
Schools and Classrooms. London: Methuen.
This contains articles on topics discussed here, such as classroom
language and Caribbean Creoles, and also articles on other topics only
mentioned briefly here if at all: for example, accents and dialects of
British English, ethnic-minority languages in Britain, reading and
writing, and listening comprehension.

2 General and introductory books on linguistics in education

(Particularly relevant to Chapter 1)
David Crystal (1976) Child Language, Learning and Linguistics: an
Overview for the Teaching and Therapeutic Professions. London:
Edward Arnold.
An accurate title.
Ronald Carter (ed.) (1982) Linguistics and the Teacher. London:
Routledge & Kegan Paul.
Ten articles on the value of linguistics to teachers, making precise
proposals about the place of linguistics in both teacher training and
in the classroom.

149
3 The sociolinguistic perspective
(Particularly relevant to Chapters 2, 3 and 4)
Gordon, J. C. B. (1981) Verbal Deficit: a Critique. London: Croom
Helm.
A short, clear book on verbal deficit theories in general, and
Bernstein's work in particular.
William Labov (1969) The logic of nonstandard English. In Language
in the Inner City. Oxford: Blackwell, 1977 (and reprinted in many
other places).
A very important and influential article: highly recommended.
Michael Stubbs (1980) Language and Literacy: the Sociolinguistics of
Reading and Writing. London: Routledge & Kegan Paul.
A discussion of the relations between spoken and written language,
including spelling and the different functions of speech and writing.
Peter Trudgill (1975) Accent, Dialect and the School. London: Edward
Arnold.
Short and precise. Very much influenced by Labov's work. Argues
that accent and dialect cause educational problems only because of
people's intolerance of linguistic diversity.

4 Classroom language
(Particularly relevant to Chapters 5, 6 and 7)
Douglas Barnes et al. (1969, revised ed. 1971) Language, the Learner
and the School. Harmondsworth: Penguin.
Informal and discursive, but well illustrated by many examples of
classroom talk. A very influential study.
Sara Delamont (1976, revised ed. 1983) Interaction in the Classroom.
London: Methuen.
In the same series as the present book. A short introduction to a
sociological theory of classrooms and teacher-pupil interaction.
Complementary to the present book since it covers some of the same
topics from a different theoretical position.
Michael Stubbs and Sara Delamont (eds) (1976) Explorations in Class-
room Observation. London: Wiley.
A collection of articles on classroom behaviour, all based on direct
observation and recording in real classrooms.
Mary Willes (in press) Children into Pupils: a Study ofLanguage in Early
Schooling. London: Routledge & Kegan Paul.
A very readable observational account of children's first days and
weeks in school. Based on Sinclair and Coulthard's work (see Ch.
6.2).

150
 References and
name index

The numbers in italics after each entry refer to page numbers within this
book.

Atkinson, P. (1975) In cold blood: bedside teaching in a medical

school. In Chanan and Delamont (eds). 107,126
Atkinson, P. (1981) Bernstein's structuralism. Educational Analysis, 3,
1.66
Barnes, D. (1969) Language in the secondary classroom. In Barnes et
al. 100-3,106,114,120,124,130, I50
Barnes, D. (1971) Language and learning in the classroom. Journal of
Curriculum Studies 3,1. 101
Barnes, D., Britton, J. and Rosen, H. (1969, revised ed. 1971)
Language, the Learner and the School. Harmondsworth: Penguin.
Barnes, D. and Todd, F. (1977) Communication and Learning in Small
Groups. London: Routledge & Kegan Paul. 116
Bellack, A. (ed.) (1973) Studies in the Classroom Language. New York:
Teachers College Press. 114
Bellack, A. A., Kliebard, H. M., Hyman, R. T. and Smith, F. L.
(1966) The Language of the Classroom. New York: Teachers College
Press. 107-8
Bernstein, B. (ed.) (1971a, 1972,1973,1975) Class, Codes and Control,
151
 vols. 1,2 and 3. London: Routledge & Kegan Paul (references to vol.
1 in the text are to the revised ed. of vol. 1, 1973, London: Paladin).
47-50, 55-9, 62, 64-6, 79,121-3
Bernstein, B. (1971b) On the classification and framing of educational
knowledge. In Young (ed.) (1971). 122-3
Boggs, S. T. (1972) The meaning of questions and narratives to
Hawaian children. In Cazden et al. (eds). 77
Burton, D. (1980) Dialogue and Discourse. London: Routledge &
Kegan Paul. 111
Cazden, C , John, V., Hymes, D. (eds) (1972) Functions ofLanguage in
the Classroom. New York: Teachers College Press.
Chanan, G. and Delamont, S. (eds) (1975) Frontiers of Classroom
Research. Slough: NFER.
Cheshire, J. (1982) Variation in an English Dialect. Cambridge: Cam-
bridge University Press. 12
Creber, P. (1972) Lost for Words. Harmonds worth: Penguin. 85
Crystal, D. (1976) Child Language, Learning and Linguistics: an Over-
view for the Teaching and Therapeutic Professions. London: Edward
Arnold. 23,149
Crystal, D. (1980) Introduction to Language Pathology. London: Ed-
ward Arnold. 12
Delamont, S. (1976, revised ed. 1983) Interaction in the Classroom.
London: Methuen. 92, 93,124,150
Delamont, S. (1983) Readings on Interaction in the Classroom. London:
Methuen.
Delamont, S. and Hamilton, D. (1976) Classroom research: a critique
and a new approach. In Stubbs and Delamont (eds). 92, 94
Dittmar, N. (1976) Sociolinguistics. London: Edward Arnold. 65
Dumont, R. V. (1972) Learning English and how to be silent: studies
in Sioux and Cherokee classrooms. In Cazden et al. (eds).77, 129,
131
Edwards, J. R. (1979) Language and Disadvantage. London: Edward
Arnold. 65
Edwards, V. K. (1979) The West Indian Language Issue in British
Schools. London: Routledge & Kegan Paul. 12, 68, 82
Flanders, N. (1970) Analysing Teaching Behaviour. London: Addison-
Wesley. 17,130
Flower, F. D. (1966) Language and Education. London: Longman. 85
French, P. and MacLure, M. (eds) (1981) Adult-Child Conversation.
London: Croom Helm. 111
Furlong, V. (1976) Interaction sets in the classroom: towards a study of
pupil knowledge. In Stubbs and Delamont (eds). 124

152
Gannaway, H. (1976) Making sense of school. In Stubbs and Delamont
(eds). 124
Gazdar, G. (1979) Class, 'codes' and conversation. Linguistics, 17. 56
Giles, H. (1971) Our reactions to accent. New Society, 14 October. 26
Gordon, J. C. B. (1981) Verbal Deficit: A Critique. London: Croom
Helm. 60, 65,150
Greene, J. (1975) Thinking and Language. London: Methuen. 19
Gumperz, J. J. and Herasimchuk, E. (1972) The conversational analy-
sis of social meaning: a study of classroom interaction. In R. Shuy
(ed.) Sociolinguistics. Georgetown Monograph Series on Language
and Linguistics, 25.105
Hall, R. A. Jr (1972) Pidgins and Creoles as standard languages. In
Pride and Holmes (eds). 72
Hamilton, D. (1976) The advent of curriculum integration: paradigm
lost or paradigm regained? In Stubbs and Delamont (eds). 119
Hammersley, M. (1974) The organization of pupil participation.
Sociological Review 22, 3, 355-68. 128
Hawkins, P. (1969) Social class, the nominal group and reference. In
Bernstein (ed.) (1972). 50-3,137
Hawkins, P. (1977) Social Class, the Nominal Group and Verbal
Strategies. London: Routledge & Kegan Paul. 53-4,137
Herriot, P. (1971) Language and Teaching. London: Methuen. 85
Hess, R. D. and Shipman, V. C. (1965) Early experience and the
socialization of cognitive modes in children. In A. Cashdan and E.
Grudgeon (eds) (1972) Language in Education. London: Routledge &
Kegan Paul. 74-6
HMSO (1975) A Language for Life. Report of the Bullock Committee.
London: HMSO. 11, 21-3, 84
HMSO (1981) West Indian Children in Our Schools. Interim Report of
the Rampton Committee of Inquiry into the Education of Children
from Ethnic Minority Groups. London: HMSO. 83-4
Hoetker, J. and Ahlbrandt, P. A. (1969) The persistence of recitation.
American Educational Research Journal 6, 2. 114
Hughes, A. and Trudgill, P. (1979) English Accents and Dialects.
London: Edward Arnold. 12,37
Hymes, D. (1967) Models of the interaction of language and social
setting. Journal of Social Issues 23, 2. 39
Hymes, D. (1972) Introduction. In Cazden et al. (eds). 29, 95, 97,134
Jackson, L. A. (1974) The myth of elaborated and restricted code.
Higher Education Review 6, 2. 62, 65, 85
Jackson, P. W. (1968) Life in Classrooms. New York: Holt, Rinehart &
Winston. 93, 119-20

153
Keddie, N. (1971) Classroom knowledge. In Young (ed.). 18,124
Keddie, N. (ed.) (1973) Tinker, Tailor . . . The Myth of Cultural
Deprivation, Harmondsworth: Penguin. 85
Kochman, T. (1972) Black American speech events and a language
programme for the classrooms. In Cazden et al. 62, 74
Labov, W. (1969) The Logic of Nonstandard English. Washington, DC:
Center for Applied Linguistics. (Also in Keddie (ed.) and Labov
(1972a). 43, 62, 65, 68-71, 76, 79, 85,116,129,150
Labov, W. (1970) The study of language in its social context. Excerpt
in Pride and Holmes (eds). In Labov, 1972b. 79
Labov, W. (1972a) Language in the Inner City. Philadelphia: University
of Pennsylvania Press. (Also published in 1977 by Blackwell, Ox-
ford.) 67, 74, 94,147
Labov, W. (1972b) Sociolinguistic Patterns. Philadelphia: University of
Pennsylvania Press. (Also published in 1977 by Blackwell, Oxford.)
39, 67,148
Labov, W. (1973) The linguistic consequences of being a lame. Lan-
guage in Society 2, 1. 136,146
Lambert, W. E. (1967) A social psychology of bilingualism. In Pride
and Holmes (eds). 73
Lee, V. (1973) Social Relationships and Language. Milton Keynes:
Open University Press. 137-8
Le Page, R. B. (1981) Caribbean Connections in the Classroom. London:
Mary Glasgow Trust. Extract in Stubbs and Hillier (eds). 12,150
Lunzer, E. A. and Gardner, K. (eds) (1978) The Effective Use of
Reading. London: Heinemann. 12,145
Macaulay, R. K. S. (1978) Language, Social Class and Education: a
Glasgow Study. Edinburgh: University Press. 12, 28,41, 67, 71, 73,
86
McTear, M. F. (1981) Towards a model for analysing conversations
involving children. In French and MacLure (eds). 146
Medley, D. M. and Mitzel, H. E. (1963) Measuring classroom be-
haviour by systematic observation. In N. L. Gage (ed.) Handbook of
Research on Teaching. Chicago: Rand McNally. 92
Mehan, H. (1973) Assessing children's school performance. In H. P.
Dreitzel (ed.) Children and Socialization, Recent Sociology 5. Lon-
don: Macmillan. 129
Mehrabian, A. (1968) Inference of attitudes from the posture, orienta-
tion and distance of a communicator. In M. Argyle (ed.) (1973)
Social Encounters. Harmondsworth: Penguin. 95
Milroy, L. (1980) Language and Social Networks. Oxford: Blackwell.
72,67

154
Milroy, L. and Milroy, J. (in prep.) Authority in Language: a Socio-
linguistic Analysis of Prescriptivism. London: Routledge & Kegan
Paul. 28,147
Milroy, J. and Milroy, L. (1974) A sociolinguistic project in Belfast.
Queen's University, Belfast, mimeo. 28
Mishler, E. (1972) Implications of teacher-strategies for language and
cognition: observations in first-grade classrooms. In Cazden et al.
(eds). 103-4,135
Mitchell-Kernan, C. (1972) On the status of Black English for native
speakers: an assessment of attitudes and values. In Cazden et al.
(eds). 73
Mittins, W. H. (1969) What is correctness? Educational Review 22, 1.
42
Nuthall, G. A. (1968) A review of some selected recent studies of
classroom interaction and teaching behaviour. In J. Gallagher et al.
Classroom Observation. Chicago: Rand McNally. 92
Philips, S. (1972) Participant structures and communicative compe-
tence: Warm Springs children in community and classroom. In
Cazden et al. (eds). 77,129,131
Postman, N. and Weingartner, C. (1969) Teaching as a Subversive
Activity. New York: Delacorte. (Also published in 1971 by Penguin:
Harmondsworth.) 90,114
Pride, J. and Holmes, J. (eds) (1972)Sociolinguistics. Harmondsworth:
Penguin.
Rosen, H. (1973) Language and Class: A Critical Look at the Theories of
Basil Bernstein. Bristol: Falling Wall Press. 65
Rosen, H. and Burgess, R. (1980) Languages and Dialects of London
Schoolchildren. London: Ward Lock. 12,146
Rosenthal, R. and Jacobson, L. (1968) Pygmalion in the Classroom.
New York: Holt, Rinehart & Winston. 84
Seligman, C. R., Tucker, G. R. and Lambert, W. E. (1972) The effects
of speech style and other attributes on teachers' attitudes towards
children. Language in Society 7, 1. 26
Sharp, D. (1973) Language in Bilingual Communities. London: Edward
Arnold. 29
Simon, A. and Boyer, E. B. (eds) (1967, 1970) Mirrors for Behaviour.
Philadelphia: Research for Better Schools. 92
Sinclair, J. M. (1973) English for effect. Commonwealth Education
Liaison Committee Newsletter 3,\l. 80, 82
Sinclair, J. M. and Coulthard, R. M. (1975) Towards an Analysis of
Discourse: The English Used by Teachers and Pupils. Oxford: Oxford
University Press. 108-11,124-5

155
Snyder, B. (1971) The Hidden Curriculum. New York: Knopf. 119
Stubbs, M. (1975) Teaching and talking: a sociolinguistic approach to
classroom interaction. In Chanan and Delamont (eds).
Stubbs, M. (1976) Keeping in touch: some functions of teacher-talk. In
Stubbs and Delamont (eds). 105-6
Stubbs, M. (1980) Language and Literacy: the Sociolinguistics of Reading
and Writing. London: Routledge & Kegan Paul. 13, 31, 37, 65, 73,
150
Stubbs, M. (1983) Discourse Analysis. Oxford: Blackwell. 111
Stubbs, M. and Delamont, S. (eds) (1976) Explorations in Classroom
Observation. London: Wiley. 93,150
Stubbs, M. and Hillier, H. (eds) (1983) Readings on Language, Schools
and Classrooms. London: Methuen. 13,149
Stubbs, M., Robinson, B. and Twite, S. (1979) Observing Classroom
Language. Block 5, PE232. Milton Keynes: Open University Press.
111
Sutcliffe, D. (1981) British Black English. Oxford: Blackwell. 12, 68,
80
Torode, B. (1976) Teacher's talk and classroom discipline. In Stubbs
and Delamont (eds). 115, 135
Tough, J. (1977) The Development of Meaning. London: Allen &
Unwin. 79
Trudgill, P. (1974) Sociolinguistics. Harmondsworth: Penguin. 29
Trudgill, P. (1975a) Accent, Dialect and the School. London: Edward
Arnold. 30, 34, 64,150
Trudgill, P. (1975b) Review of B. Bernstein Class, Codes and Control,
vol 1. Journal of Linguistics 11, 1.62,65
Trudgill, P. and Hannah, J. (1982) International English: a Guide to
Varieties of Standard English. London: Edward Arnold. 37
Turner, G. J. (1973) Social class and children's language of control at
age 5 and age 7. In Bernstein (ed.) vol. 2. 54
Walker, R. and Adelman, C. (1975a) A Guide to Classroom Observation.
London: Methuen. 111-13,143
Walker, R. and Adelman, C. (1975b) Interaction analysis in informal
classrooms: a critical comment on the Flanders system. British
Journal of Educational Psychology 45, 1.92, 111
Walker, R. and Adelman, C. (1976) Strawberries. In Stubbs and
Delamont (eds). 111-12
Wells, G. et al. (1981) Learning through Interaction: vol. 1, Language at
Home and at School. Cambridge: Cambridge University Press. 124
Wells, J. C. (1982) Accents of English. 3 vols. Cambridge: Cambridge
University Press. 12

156
Wight, J. (1971) Dialect in school. Educational Review 24,1.80
Wight, J. (1975) Language through the looking glass. Ideas, Curricu-
lum Magazine, Goldsmiths College, London, 31. 52, 80-1,116
Wight, J. (1979) Appendix: dialect and reading. In Supplementary
Readings for Block 4, PE232. Milton Keynes: Open University Press.
83
Wight, J. and Norris, R. (1970) Teaching English to West Indian
Children, Schools Council Working Paper 29. London: Methuen. 80
Wilkinson, A. (1971) The Foundations of Language. Oxford: Oxford
University Press. 85
Willes, M. (in press) Children into Pupils: a Study of Language in Early
Schooling. London: Routledge& KeganPaul. 13, 111, 115,144,150
Young, M. F. D. (ed.) (1971) Knowledge and Control. London: Collier-
Macmillan. 119,121

157
Subject index

accent, 25-6, 32 correctness, 41-2

American Indian children, 29, 77, 131 Creoles, 72-3, 80-1
appropriateness of language, 39-42, curriculum, 119-21
51-2, 72, 86
attitudes to language, 25-30, 36,
deprivation, linguistic and cultural,
72-3, 147-8
44—5, 68-87, 133
see also llnguiitic stereotyping
dialect, 20, 25, 32-4, 36, 43, 64, 69,
137, 147
classroom, as sociolinguistic setting,
discourse structure
54, 111-13, 115-17, 143-4
analyses of, 107-11
classroom language
messages conveyed by, 104, 119,
importance of, 88-90
122-8
methods of analysis of, 100-13
neglect of, 90-3
transcriptions of, 104, 108-10, English
125-8 Black English Vernacular (BEV),
see also discourse structure 43, 68-9, 71, 73, 136
codes, restricted and elaborated, Standard (SE) and nonstandard,
46-64, 123, 133, 138 20, 32-7, 40-1, 72-3, 86
definition, early, 49-50 teaching of, 17, 28
definition, recent, 56-60 see also Creoles, dialect, pidgins
existence questioned, 61-2 ethnography, 93, 97, 131
evidence for, 61-2 experiments on language, 50-4, 65,
lack of linguistic data for, 61, 79 74-6
competence, 37-9, 81-2, 87, 148 see also tests
framing, 121-3 recording
audio, 103-4, 112
IQ, 20, 70, 88-90, 128-9 visual, 113

language
and thought (cognition), 18-19, 56, Scots, 28-9, 32, 71, 73
64, 68-9, 85, 133-5 social class, 20, 25, 46-66, 67, 88-9,
functions (uses) of, 37-40, 51-4, 123
101-2, 105-6 sociolinguistic
structure of, 37-8, 44 explanations, 86-7, 138-40
style of, 17-18, 33-5, 39-42, 62, interference, 102
71, 100-1 sociolinguistics, definition of, 24, 93
variation (diversity), 39-40, 46 statistical statements, 62, 75
written versus spoken, 25, 32-6, system, concept of, 31, 39, 69-71,
80-1 107, 136-40
linguistic
failure, 19, 24, 87, 147 teacher-training, 21-3, 93, 142
folklore, 25, 141 teaching
stereotyping, 25-8, 138 as sociolinguistic behaviour, 99,
121, 130-2
myths cultural definitions of, 17, 120,
about codes, 85 130-2
about primitive languages, 30-1, 70 tests
about verbal deprivation, 68, 74-8, as sociolinguistic settings, 51-2, 77,
85 89, 96, 115-17, 128-9
see also llnguiitic folklore of linguistic ability, 74-6, 81-2,
115-16
observation see also experimentss ,IQ
naturalistic, 66, 91, 94-6, 103, 135, theory
142-8 causal and correlational statements,
participant, 68, 93 46-7, 49-50, 65-6, 85, 133, 139
systematic, 91-2 criteria for, 140-1
see also recording forms of, 59-60, 134-5, 140
relation to data, 23, 56, 59-62, 79,
peer group, linguistic influence of, 86, 93, 123, 133-41, 148
136, 145-6 see also statistical ltatements
pidgins, 72-3

questions, types of, 102, 114, 120, Welsh, 29

123-6 West Indian children, 42, 80-3

160