Reproduced with written permission from the rights-holder Elsevier

(now owner of Academic Press) and the editor Peter Trudgill.

Applied Sociolinguistics
 Applied Language Studies
Edited by David Crystal

Directions in Applied Linguistics, Crystal, D. 1981

The Nature of Phonological Disability in Children, Grunwell, P. 1981
Information and Meaning in Child Communication, Lloyd, P. and
Beveridge, M. 1981
Perspectives in Communicative Language Teaching, Johnson, K. and
Porter, D. 1982
Current Developments in Language Testing, Hughes, A. and
Porter, D. 1983
Lexicography: Principles and Practice, Hartmann, R. R. K. 1983
Analysis of Structural Learning, Jeeves, M. A. and Greer, B. 1983
Applied Sociolinguistics, Trudgill, P. 1983
Applied Sociolinguistics

Edited by

Peter Trudgill
Department of Linguistic Science
University of Reading

1984

ACADEMIC PRESS, INC.

Harcourt Brace Jovanovich, Publishers
London Orlando San Diego
Austin New York Toronto Montreal
Sydney Tokyo
Cf 50599925
'301.il -3,oc.o ·44-
ACADEMIC PRESS INC. (LONDON) LTD.
24/28 Oval Road
LondonNWJ

United States Edition published by

ACADEMIC PRESS INC.
(Harcourt Brace Jovanovich, Inc)
Orlando. Florida 32887

Copyright <' 1984 by

ACADEMIC PRESS INC. (LONDON) LTD.

All Rights Reserved

No part of this book may be reproduced in any form by photostat, microfilm, or

any other means, without written permission from the publishers

British Library Cataloguing in Publication Data

Applied sociolinguistics. -(Applied language
studies)
l. Sociolinguistics
I. Trudgill, Peter II. Series
401 '.9 P40

ISBN 0-12-701220-6

LCCCN 83- 71853

PRINTED IN THE UNITED STATES OF AMERICA

85 86 87 88 9 8 7 6 5 4 3 2
 Contributors

Allan Bell
4/6 Parau Street, Mt. Roskill, Auckland 4, New Zealand
Jenny Cheshire
Department of Applied Linguistics, Birkbeck College, University of London,
Malet Street, London WCI, UK
John Edwards
Department of Psychology, St. Francis Xavier University, Antigonish, Nova
Scotia, Canada
Ralph Fasold
Department of Linguistics, Georgetown University, Washington D. C. 20057,
USA
Howard Giles
Department of Psychology, Bristol University, Berkeley Square, Bristol BSB
JHH, UK
James Milroy
Department of Linguistics, Sheffield University, UK
Lesley Milroy
Department of Speech, The University, Newcastle-upon-Tyne, UK
Michael Stubbs
Department of Linguistics, Nottingham University, University Park,
Nottingham, NG7 2RD, UK
Peter Trudgill
Department of Linguistic Science, Faculty of Letters and Social Sciences,
University of Reading, Whiteknights, Reading, RG6 2AA, UK
Alastair Walker
Nordfriesische Worterbuchstelle, Neue Universitiit Kiel, Olsenhousenstrasse
40-60, 2300 Kiel, West Germany
 Foreword

Sociolinguistics, more than any other branch of linguistics, has in recent

years been simultaneously pulled in several different directions. Most of its
practitioners are primarily concerned with the theoretical question of how to
explain the linguistic variation which society manifests, and with the meth-
odological question of how to explore it. But, because their field encounters
the social implications and repercussions of language use much more
immediately than other branches of linguistics, sociolinguists have regularly
found themselves called upon to participate in language decision-making:
evaluating or planning linguistic proposals related to a wide range of local or
international issues. Some sociolinguists have avoided the publicity which
often ensues from such encounters, preferring that their research be directed
towards the publication of statements along conventional academic lines.
Others have willingly become involved in social encounter, feeling that the
application of their skills in the public domain is a responsible and desirable
step. Yet others have taken this step with reluctance, believing that the
application of sociolinguistic knowledge in this domain is premature, given
the limited theoretical and empirical development of the subject to date.
I believe that there have now been sufficient moves in these various
directions to permit a clearer analysis of the subject-matter of sociolinguistics
than has hitherto been possible or desirable; and that the distinction between
"general" and "applied", which has proved so useful in other fields, might
prove to be a realistic and fruitful model in the present case also. The general
field is well-represented by the several textbooks and monographs; but there
has been no corresponding attempt to bring togther the various topics that
could be construed as applications: in effect, to address the question, "What
problems can sociolinguistics help to solve?". It might be that there are
parallels in the way sociolinguists have approached these topics, such that
the problems enountered in one area might be illuminated by the findings of
another. It might be that a systematic consideration of what is involved in
viii FOREWORD

'"applying" sociolinguistics might bring to light neglected topics which could

usefully be studied along such lines. At the very least, I would hope that the
juxtaposition of topics will lead to the discovery of correspondences in
approach which are theoretically or methodologically interesting. And if the
existence of a field of applied sociolinguistics can be usefully demonstrated,
it may thereby provide a perspective within which the concerns of the general
field can be more coherently defined. The present book is a pioneering
work, in its intention and scope, and one whose appearance in a series on
applied language studies I warmly welcome.
Reading David Crystal
October 1983
 Preface

Sociolinguistics is the science which deals with the relationship between

language and society. It is therefore a very broad topic, encompassing the
study of social dialects, language attitudes, stylistic variation, conversational
interaction, multilingualism, language change, and much more. This book is
a reflection of the fact that a great deal of work in this field is of direct or
indirect practical relevance in a number of different spheres of life, including
education and politics.
The book brings together for the first time work by experts in particular
branches of sociolinguistics and is intended to demonstrate the range of
real-world activities in which the findings of sociolinguistic studies can be of
some assistance and interest. Sociolinguistics is an area of study that
concerns itself with the ways in which human beings actually use language in
social interaction in their everyday lives. It is therefore no surprise that it
turns out to be of concern for workers in a wide range of fields. Applied
Sociolinguistics should be of interest to sociolinguists, language teachers,
social psychologists, educationists, media specialists, lawyers, and just about
anybody else who speaks a language.

Reading Peter Trudgill

October 1983
 Contents

Contributors v
Preface VII

1. Introduction
by Peter Trudgill

Social Issues
2. Comprehension and Context: Successful Communication
and Communication Breakdown
by Lesley Milroy 7
3. The Relationship between Language and Sex in English
by Jenny Cheshire 33
4. Sociolinguistic Methodology and the Identification of
Speakers' Voices ;n Legal Proceedings
by James Milroy 51
5. Good Copy - Bad News: The Syntax and Semantics of
New Editing
by Allan Bell 73

Educational Issues
6. Applications of the Social Psychology of Language:
Sociolinguistics and Education
by John Edwards and Howard Giles 119
7. Applied Sociology of Language: Vernacular Languages
and Education
by Alastair Walker 159
8. Applied Discourse Analysis and Educational Linguistics
by Michael Stubbs 203
xii CONTENTS

9. Variation Theory and Language Leaming

by Ralph W. Faso Id 245
Index 263
 1
Introduction

Peter Trudgill

As is well-known, the term "Applied Linguistics" has acquired, in the

literature, both a narrow and a broad interpretation. In its narrow inter-
pretation, the term has come to refer to the application of the results of
theoretical linguistic research to the teaching and learning of foreign and
second languages. In its broad interpretation, on the other hand, "Applied
Linguistics" refers to the application of linguistic research to the solution of
practical, educational and social problems of all types. In the title of this
book, Applied Sociolinguistics, (used here, we believe, for the first time),
the term "applied" is employed in the second, wider sense. This book is an
attempt to present information on the ways in which sociolinguistic findings
and sociolinguistic insights can be and have been of assistance in the tackling
of real-world problems.
As a field of study involving the interaction of language and society,
sociolinguistics has naturally been a form of linguistics which has typically
been more "applied" than most. We believe, nevertheless, that it is per-
fectly legitimate to point to a distinction between "pure" or "theoretical"
sociolinguistics, on the one hand, and its applications, on the other. Socio-
linguistics is a science which has among its objectives (some would say its
only objective) the achievement of a greater understanding of the nature of
language, as well as (others would say), and at least in some of its manifesta-
tions, a greater understanding of the nature of society. At its most linguistic,
sociolinguistics is a way of doing linguistics, a methodology referred to by
William Labov as "secular linguistics". And even at its most social and least
linguistic, sociolinguistics is a subject which has some linguistic objectives.
"Applied sociolinguistics", however, is clearly to be interpreted as re-

APPLIED SOCIOLINGUISTICS Copyright© 1984 by Academic Press, London

ISBN: 0-12-701220-6 All rights of reproduction in any form reserved
2 PETER TRUDGILL

ferring to activities in which the research of sociolinguists is employed (very

often by sociolinguists themselves) in other fields where it can be of some
value. In some cases sociolinguistic research (unlike, say, research in
theoretical syntax) has been carried out with an avowedly practical purpose.
In other cases, sociolinguists have found themselves involved in "applying"
their findings, if not exactly against their will, at least without initially
intending that they should be used in this way. One area of sociolinguistic
work of which this has undoubtedly been true is "secular linguistics".
Following Labov's lead in this field, large amounts of often very exciting
work have been carried out in the form of empirical studies of language as it
is spoken in its social context. Research of this type has been aimed
principally at improving linguistic theory, and acquiring a better under-
standing of the nature of language variation and the sources of linguistic
change. However, the results of Labov's work, and that of his followers,
were very rapidly, from the 1960s onwards, applied in a highly significant
way to two different but related educational problems: the debates con-
cerning non-standard (especially American Black English) dialects in edu-
cation; and so-called "verbal deprivation" or "language deficit". In the first
of these debates, it was the sociolinguists' increased understanding of
linguistic variation and their phonological and grammatical descriptions of
non-standard dialects that became important. In the second, on "verbal
deprivation", one crucial factor proved to be the large amounts of spon-
taneous, relatively unmonitored, casual speech recorded by sociolinguists
concerned to overcome the "observer's paradox".
In the present volume, the chapter by J. Milroy, on "Sociolinguistic
methodology and the identification of speakers' voices in legal proceedings"
exemplifies a new use for quantitative studies of the Labovian type: the
employment of sociolinguistic findings in a field we can perhaps call "forensic
sociolinguistics". And in another examination of the uses of work of this
type, Fasold, in his chapter on language learning, looks at the application of
variation theory, a relatively recent theoretical development out of Labov-
style secular linguistics. His chapter uvariation theory and language learn-
ing" starts from attempts to incorporate sociolinguistic findings concerning
the structure and probabilistic nature of linguistic variation into theoretical
models, and links these to foreign language acquisition processes.
This book, however, is not only concerned with the applications of secular
linguistics. We are concerned also with the Jess than purely linguistic forms
of sociolinguistics, such as the social psychology of language; the ethnography
of speaking; the sociology of language; and discourse analysis. The social
psychology of language is an area of study which deals with attitudes to
varieties of language, and with the way in which speakers interact with each
other through conversation. Clearly, its range of potential applications is
 INTRODUCTION 3

enormous, and it already has a very respectable applied history, notably for
example in the study of teachers' attitudes to their children's language. In
this volume, Edwards and Giles examine the applications of the social
psychology of language to the sphere of education in general.
The ethnography of speaking is that area of sociolinguistics which looks at
norms for using language in different cultures. It is therefore of obvious
importance in foreign language teaching, and also clearly crucial in the
increasingly important field of cross-cultural communication. The chapter
by L. Milroy on "Comprehension and Context" deals, in part, with this
topic.
The applications of the sociology of language to the solution of social and
other problems is already a very well-known field with an established label of
its own: language planning. "Language planning" has come to refer to
governmental or quasi-governmental activities, particularly in multilingual
situations, designed to influence or solve linguistic problems. Walker's
chapter in this book, "Applied Sociology of Language", deals with language
planning, but with particular reference to the issue of vernacular languages
in education.
Discourse analysis looks analytically, in various ways, at texts and con-
versational interaction with a view to achieving a greater understanding of
textual cohesion and coherence, and rules for carrying out and interpreting
conversation. It, too, is increasingly coming to be recognized as an area
which has important applications as in, for example, the study of doctor-
patient communication. In this volume, Stubb's chapter "Applied discourse
analysis and educational linguistics" deals with mother-tongue and foreign-
language teaching, while Bell actually applies the techniques of textual
analysis to an examination of what happens to the language of news reporting
in his chapter "Good Copy - Bad News".
Finally, we may note that one of the most important and best publicized
areas of sociolinguistic research in recent years has been the study of
language and sex. Work in this area has a number of components and involves
work in a number of different aspects of sociolinguistics. Researchers in this
field have variously dealt with sex differences in the usage of particular
linguistic features, sex differences in the use of conversational strategies,
attitudes to language use by men and women, and sexism in language. In her
chapter on the "Applications of research on language and sex", Cheshire
discusses some of the implications of this type of research.
The extent to which work in the area of sociolinguistics can be of practical
value has been discussed by Labov in his article "Objectivity and commit-
ment in linguistic science: the case of the Black English trial in Ann Arbor"
(Language in Society, 1982). In this paper, Labov brings two principles to
the attention of researchers in sociolinguistics. They are:
4 PETER TRUDGILL

The principle of error correction

A scientist who becomes aware of a widespread idea or social practice
with important consequences that is invalidated by his own data is
obligated to bring this error to the attention of the widest possible
audience.
The principle of the debt incurred
An investigator who has linguistic data from members of a speech
community has an obligation to make knowledge of that data available
to the community, when it has need of it.
The chapters contained in this book are all, to varying degrees, motivated
by these principles, and are concerned with putting them into practice. It is
our hope that the book will help to bring to the attention of those concerned
the practical value of sociolinguistic research in all the disciplines dealt with
here. We hope, too, that it will stimulate further research, and further
applications.
Social Issues
 2
Comprehension and context:
successful communication and
communicative breakdown

Lesley Milroy

Introduction

The major theoretical issue to which this paper addresses itself is the
manner in which hearers use a combination of linguistic knowledge, knowl-
edge of a heterogeneous range of factors which broadly might be called
"contextual", and various perceptual strategies when they interpret utter-
ances in context. This issue is approached using methods familiar in socio-
linguistics; that is, the data base is real utterances as they are spoken in their
social context rather than constructed data intuitively interpreted. It seems
likely that both intuitive and empirical approaches are necessary (and a
combination of the two) if we wish to advance our knowledge of how
speakers comprehend what they hear in everyday situations. The problem is
tackled initially by looking at a number of communications which have gone
wrong; as Gumperz and Tannen (1979) point out: "by studying what has
gone wrong when communication breaks down, we seek to understand a
process that goes unnoticed when it is successful" (p. 308). A similar
approach to speech production may be found in Fromkin (1973 and 1981).
Most of the miscommunications examined here are between people from
different dialect backgrounds: that is people whose internal grammars are
different in some specifiable way. This limitation is important methodo-

APPLIED SOCIOLINGUISTICS Copyright© L984 by Academic Press, London

ISBN: ll-12-701220-6 All rights of reproduction in any form reserved
8 LESLEY MILROY

logically as it helps us to assess the role of specifically linguistic knowledge in

comprehension. and it seems likely that conclusions on this wider matter can
be extended to communications between people with the same dialect
background.
Although it may be seen as a contribution to a major theoretical issue. this
chapter is primarily intended as an exercise in applied sociolinguistics: the
insights of sociolinguistics are used to analyse the manner in which com-
municative breakdown occurs between speakers in everyday situations and
the consequences of this breakdown are considered. These nvo issues. the
theoretical and the applied. will be considered in parallel throughout the
discussion.

A miscommunication may be said to take place when there is a mismatch

between the speaker's intention and the hearer's interpretation. When two
persons do communicate successfully. it is clear that much more is involved
than the mapping of internal structures (or linguistic rules) on to external
sequences. or conversely (from the listener's point of view). mapping
external sequences on to internal structures (Bever. 1970. p. 286). A
number of perceptual strategies or .. short-cuts .. appear to be implemented
by a listener in decoding utterances. and it is probably most sensible to adopt
for the moment Aitchison's notion of a person's grammar as a ''linguistic
archive··. available and ready for consultation in interpreting utterances. but
not necessarily consulted. Thus. a sentence like the following is difficult to
comprehend. although it is perfectly grammatical:
The pig pushed in front of the piglets ate all the food.
This is because the perceptual strategies adopted by the listener are to
assume that the first noun will go with the first verb in an NP-VP (actor-action)
sequence as part of the main clause ... So he understandably makes the wrong
guess when he hears the words the pig pushed . ... especially as his knowledge of
the world tells him that pigs arc not usually pushed they generally do the
pushing. (Aitchison. 1976. p. 203).
Thus. the relation between internal grammars and comprehension is
indirect. especially since comprehension in natural settings (as opposed to
the experimental settings from which much of the psycholinguistic work
discussed by Aitchison derives) is often assisted by such factors as the
following:
I. Natural language contains a great deal of redundancy (see Lyons.
1977. p. 44 for a discussion of this).
JI. .. Context"' and "shared background knowledge" assist interlocutors
 COMPREHENSION AND CONTEXT 9

in interpreting utterances (Smith and Wilson. 1979). I refer to both

linguistic and extra-linguistic context here, and will return to a dis-
cussion of it shortly.
111. Interlocutors in unplanned discourse have a range of strategies for
monitoring comprehension, for clarifying and for repairing mistakes
and misunderstandings as they follow the facial, gestural and inter-
jectional cues which accompany interactions. Constant repetition
and repair are important components of communication in context
and are themselves highly systematic (Schegloff et al., 1977~ Giv6n.
1979, p. 230).
IV. It appears that out of consideration for each other's and their own
••face··. speakers often make remarks whose ambiguity has a clear
social motivation (Brown and Levinson, 1978, p. 74). Speakers also
appear to be willing to allow a great deal of vagueness and ambiguity
to pass, on the assumption that meanings will become clear as the talk
proceeds (Cicourel, 1973, p. 54).

Thus, when miscommunication does take place. a whole range of factors

may be responsible. It is probably the indirectness of the relationship
between .. linguistic knowledge" (in Chomsky's sense) and language use
which accounts for the fact that speakers with partly different grammars
understand each other much of the time in natural settings; they rely on a
wide and varied range of comprehension strategies. However, when com-
prehension between such speakers is tested experimentally (i.e. when
factors I to IV above are controlled), they appear to be unable to associate
sentences not generated by their own grammar with an appropriate semantic
structure (see Labov, 1972a; Trudgill, 1981). Generally speaking. argu-
ments for a ··polylectal grammar" as a reflection of a speaker's pan-dialectal
competence have not been supported by experiments such as Labov's or
Trudgilrs which require subjects either to select a paraphrase for sentences
generated by other dialect grammars. or to judge whether such sentences
are ··possible" English sentences (see also Ross, 1979 for a different, but
related, approach). The gap between these experimental results and results
of the observational research reported in this paper is sometimes quite large,
and will be discussed below. Meantime. it is perhaps worth noting Plutchik's
comments on the value of observational (as opposed to experimental)
research in a field where knowledge is limited (Plutchik, 1976, p. 23).
While then it is certainly true that speakers with markedly different
grammars are frequently able to understand each other in context, it does
not seem quite justifiable to adopt uncritically the common view cited by (for
example) Smith and Wilson that differences between the grammars of
dialects are essentially trivial, and that potential misunderstandings can be
10 LESLEY MILROY

resolved by the implementation of perceptual strategies. or by context

(Smith and Wilson. 1979. p. 197). One purpose of this chapter will be to try
and isolate the conditions under which differences in speakers' grammars
may produce misunderstandings in natural settings; that is when the various
comprehension strategies available to a speaker have failed to work and he is
forced to consult his .. linguistic archive .. to assist him in interpreting
utterances.
Since Labov's initial demonstration of the .. limits" of individual gram-
mars. it has been generally accepted (at least by sociolinguists) that pan-
dialectal competence, active or passive, is extremely sharply constrained.
However, there is an implicit tendency to assume that it is competence in
non-standard grammars which is limited. It is significant, for example, and
quite characteristic of work in this area, that all the sentences discussed in
TrudgiH's 1981 paper are designed to test knowledge of non-standard
grammars. In an earlier publication, Trudgill expressed the view more
explicitly that non-standard speakers had at least a passive competence in
standard English (Trudgill, 1979) and Trudgill's view does not seem at all
uncommon.
In relation to this issue. it is worth noting Lyons' observation (made
generally, rather than with reference to cross-dialectal communication) that
misunderstandings are probably rather frequent during communicative acts,
and that this likelihood should be allowed for in any theoretical model of
communication (Lyons, 1977, p. 33). It is only rarely that speakers are
provided with feedback in any consistent way, and when this does happen
(as when students write down in examinations what they think has been said
to them) the level of misunderstanding is frequently revealed to be quite
horrifying (Aitchison, 1976, p. 197). If we take comments such as those of ·
Lyons and Aitchison into account, as well as the complex of knowledge and
strategies underlying comprehension, I think we have to treat the idea that
non-standard speakers "understand" standard English as a not very clearly
defined and quite unsupported assumption, rather than a self-evident truth
(in fact, some data will be discussed below which suggests that the assump-
tion requires some modification). A more interesting question may well be
concerned with the extent to which miscommunications are associated with
cross-dialectal communication.
The approach taken in this chapter will be to examine and analyse instances
of miscommunication against the extremely complicated but still by no
means comprehensive background which has been outlined so far; an attempt
will be make to bring together material from a number of relevant academic
perspectives. The main focus of interest will be in identifying as precisely as
possible the linguistic and contextual factors contributing to miscommunica-
tion, but there will also be some brief discussion of the consequences of
 COMPREHENSION AND CONTEXT 11

miscommunication. This will be both in terms of speakers' responses, and in

more general social terms. However, preliminary to analysis of data two of
the issues already raised are looked at in a little more detail. These are, first,
the relationship between standard and non-standard English (this will be a
brief section), and next the generalized notion of "context" to which
linguists frequently appeal when they are attempting an explanation of how
speakers comprehend utterances in real situations.

Standard and non-standard English

It is important to remember in a study like this one, which looks primarily at
cross-dialectical communication, that any account of a language-system like
English is extremely idealized and abstract and inevitably based on the
standard form of the language (a form routinely used, it should be noted, by
only a small minority of speakers). Although the notion of a standard is
notoriously complicated and hard to pin down, it is useful here to ignore the
social dimensions of standardization and adopt Lyons' view of it as one kind
of idealization of the data in which the linguist ignores differences within a
system and Hdiscounts all but the major systematic.variations in the language
behaviour of the community" (Lyons, 1977, p. 587). Standard English then
is the form of the language from which most linguists cite their examples,
and it is usually believed that structural differences between standard and
non-standard varieties are relatively superficial and can be accounted for in
terms of rule addition or rule loss. Some examples of syntactic differences
within British English are cited by Hughes and Trudgill (1979) which include
differences in tense and aspectual systems, and in the formal distribution of
individual verbs such as have and do.
The question of genuine communication problems arising from structural
differences of this kind is not often seriously discussed, the assumption
generally being that non-standard speakers can "understand" standard
English presumably in the sense that in some hypothetical context-free
situation they are able to assign the same semantic structure as a standard
speaker to a given phonetic string. The reason for this assumption (which is
noted here mainly to encourage readers to consider whether they them-
selves hold it) is probably the constant dissemination of spoken and written
standard English through the news media and the educational system. The
possibility that some speakers do not "understand" standard English
appears for the most part to be mooted only when their systems are so
divergent that they may, like West Indian creole-speakers in England, or
Black English-speakers in the United States, be thought of in some un-
satisfactorily specified way as not being English speakers at all (Edwards,
12 LESLEY MILROY

1979; Nelson and McRoskey, 1978). However, in a recent extensive study of

the Hiberno-English perfect (associated with a larger study ofnon-standard
English) John Harris refers to frequent misinterpretations by Hibemo-
English speakers of such standard English utterances as
How long are you staying here?
This type of structure is often interpreted as being equivalent to standard
English
How long have you been staying here?
Harris goes on to argue convincingly that there is no semantic isomorphism
between any given set of Hiberno-English and standard-English sentences
which exemplify the range of perfect constructions. For example, the three
sentences
(a) Joe has sold the boat
(b) Joe has just sold the boat
(c) Has Joe sold the boat?
exemplify the simple perfect tense-aspect form in Standard English. How-
ever, they are all quite anomalous in even educated Hiberno-English and
would be .. translated" as follows:
(a) Joe has the boat sold
(b) Joe is just after selling the boat
( c) Did Joe sell the boat?
The non-isomorphism is semantic as well as formal. since, for example
Hiberno-English (c) is not only a translation of Standard-English (c) but is
also equivalent to the same standard English string. Thus, Hiberno-English
cannot distinguish between .. Has Joe sold the boat?" and .. Did Joe sell the
boat?,, However, an action completed in the recem past is expressed by a
construction of the (b) type. Cross-dialectal miscommunications are very
commonly reported which seem to be located in the disparity between the
two versions of sentence (b) and these may be seen as symptomatic of a deep
structure disparity between the grammars of the two dialects. The two
dialects cannot be related satisfactorily simply by applying slightly different
transformational rules to derive them from the same deep structure (Harris,
1982). This suggests that dialect grammars may sometimes be more different
than is commonly supposed, and justifies a somewhat closer examination of
how far the communicative competence of a non-standard speaker includes
the capacity to assign the same semantic structure as standard speaker to a
given phonetic string.
 COMPREHENSION AND CONTEXT 13

The notion of context

We have already noted that comprehension involves very much more than
knowledge of phonological, syntactic and semantic rules. So much redundant
information is introduced into communicative situations by a combination
of linguistic and real world knowledge that it is often possible to produce an
appropriate response to an utterance· without any knowledge at all of the
structural rules underlying its production. For example, during a recent
visit to a rural area of Austria, I was able to comprehend a request (from a
shopkeeper) to shut the door even though I was quite unable (despite my
knowledge of standard German) to recognize in the utterance any familiar
syntactic patterns or even a single lexical item. A combination of cold
weather, gestural cues, and limitations on the kind of message likely to be
transmitted in a service encounter of that kind facilitated successful com-
munication; that is, an appropriate response was produced to the utterance.
It is of course a truism to assert that actual utterances are always interpreted
in context, and instances like this one seem to have encouraged linguists to
exaggerate the capacity of context to resolve ambiguity, while at the same
time not defining particularly closely what is meant by context. It is still
commonplace to find the term used freely without any more careful examina-
tion of its scope. Yet, context-boundedness is an important inherent charac-
teristic of all natural languages and arguably should be accounted for
explicitly in any comprehensive linguistic theory, rather than being used as a
"terminological immunization" (Klein and Dittmar, 1979, p. 4) to cover up
problems of fitting data to theory.
That linguists are in general aware of this problem is evidenced by the
major and long-running controversy within generative semantics as to
whether the illocutionary potential of an utterance should in some way be
incorporated into a formal grammar (Ross, 1975) or whether the interpreta-
tion of utterances in context should be accounted for by separate sets of
conversational rules (Sperber and Wilson, forthcoming; Smith and Wilson,
1979; Gordon and Lakoff, 1971). More recently, attempts have been made
to resolve empirically the issue of how far the social and situational context
affects understanding and judgement (Ervin-Tripp, n.d.). Preliminary find-
ings suggest that some, but not all, listeners rely heavily on routine non-
Iinguistic knowledge, and on what might be described as "pragmatic intel-
ligence''. Stimulus utterances are decoded by applying linguistic rules only if
something unusual or discordant occurs (cf. the notion of the "linguistic
archive", discussed above).
It is useful here to note Lyons' distinction between the layman's intuitive,
pre-theoretical concept of context to which speakers frequently make
14 LESLEY MILROY

appeal when they discuss interpretations of utterances, and on the other

hand context as a theoretical construct which has been analysed from
various points of view by different traditions of scholarship. Lyons' working
definition of context (in the second sense) is as follows:
a theoretical construct in the postulation of which the linguist abstracts from the
actual situation and establishes as contextual all the factors which, by virtue of
their influence upon the participants in a language event, systematically deter-
mine! my italics] the form, the appropriacy and the meaning of utterances.
(Lyons, 1977,p.572)
We look now at two different approaches to the systematization of context,
both of which are relevant to the discussion here.
Hymes's discussion of the nature of '"communicative competence"
attempts to provide a theoretical framework for describing components of
context which have just the kind of systematic effect referred to by Lyons on
the production and comprehension of particular phonological, syntactic and
lexical options within the language system (Hymes, 1972). They are mani-
fold and complex, and the reader is referred to Hymes' own paper for a
comprehensive account and to Coulthard (1977, p. 30) for a brief but useful
discussion. Examples of the kind of knowledge which forms part of com-
municative competence in any real society are knowledge of status systems
and social roles, and of the language appropriate to various settings and
topics. Many linguists regard communicative competence as part of a
speaker's linguistic knowledge in just the same way as his knowledge of
syntactic and semantic rules. For example, Lyons suggests that a French
speaker's knowledge of when to select the intimate or the polite second
person pronoun is as much part of his linguistic competence as his knowl-
edge of the French language-system (Lyons, 1980, p. 248); others would
argue that this is an example of encyclopaedic, rather than strictly linguistic,
knowledge.
A different approach to the systematization of context follows the tradi-
tions of Searle (1975) and Grice (1975). Scholars working along these lines
are generally concerned with specifying those contextual cues, logical pro-
cesses and other conditions which enable interlocutors to interpret utter-
ances in context appropriately. Particularly interesting attempts to apply
analyses of this kind to real speech events have been made by Labov and
Fanshel (1977) and Gumperz and Tannen (1979). But on the whole, the
most sophisticated analyses of utterances in context are applied to isolated
pieces of constructed data, and have not yet been challenged by being tested
in re~I conversations. In the sections which follow, an attempt will be made
to assess the relative importance of these various systematic components of
context, of perceptual strategies, and of internal grammars in enabling
successful communication. Particular attention will be paid to those instances
 COMPREHENSION AND CONTEXT 15

where a miscommunication seems to be located in a mismatch between the

grammars of the speaker and the hearer.

Some miscommunications: an analysis

First, a brief note of some terminological and methodological points. For

reasons which will become clear, it is important to try and maintain a
distinction between two different kinds of miscommunication. What will
be described as misunderstandings involve simple disparity between the
speaker's and the hearer's semantic analysis of a given utterance. However,
misunderstandings are not perceived as interrupting communicative effi-
ciency, and it is probable that many more misunderstandings take place than
are ever discovered. Communicative breakdown on the other hand occurs
when one or more participants perceive that something has gone wrong.
This "something" may or may not be analysed as linguistic; if it is, the
breakdown may or may not be repaired.
All the data discussed below are taken from real conversations. Only (1) is
tape-recorded, having cropped up by a happy chance during a long con-
versation between myself and an inner-city informant in Belfast. The rest of
the data were recorded in notebooks by myself and colleagues, and the
researchers were usually able to report their own reaction to a miscommuni-
cation. By its nature, material on miscommunications is hard to collect in
any systematic way, since unlike some of the classic linguistic variables
described in the quantitative literature they occur unpredictably and some-
times infrequently. I tackled this problem simply by searching out, over a
period of time, analysable examples of miscommunication which seemed to
be located in disparate individual grammars. These were transcribed on the
spot, and where necessary contextual factors were noted in as much detail as
seemed appropriate. An analysis was then carried out to try and locate the
source of the miscommunication.
This post hoe type of procedure is very similar to that of Gumperz (see
especially Gumperz and Tannen, 1979) who deals with those miscommuni-
cations which result from disparities in the inferences drawn by conversa-
tional participants from utterances in context. Both pieces of research follow
the principles of careful observation in the tradition of social anthropology
pioneered by Evans-Pritchard, and more recently in linguistic research
recommended by Ferguson (1977) for dealing with phenomena about which
little is known. Careful observation of this kind is of course an essential
preliminary to the development of a predictive and testable theory.
16 LESLEY MILROY

Youse wash the dishes

This heading is part of a narrative told to me during a fieldwork session by a
Belfast working-class woman. The rest of it, which may serve as an illustra-
tion of vernacular second person pronoun usage, is as follows:
(I) /So I says to our Trish and our Sandra/youse wash the dishes/and I
might as well have said you wash the dishes/for our Trish just got
up/and put her coat on and walked out/
In much of Scotland and Ireland the second person pronoun can be marked
for number as illustrated here, giving the plural form youse. This extract is
quoted here because it shows clearly that the speaker considers you an
inappropriate choice for a plural referent. This fact is relevant to analysis of
subsequent miscommunications between myself and some Belfast vernacular
speakers.
An examination of the sociolinguistic distribution of youse in Belfast
reveals rather more complexity than might at first be assumed. First of all, as
we might expect, many relatively standardized speakers have categorical
you for both singular and plural pronouns. For these speakers, the youse
form is high Iy stigmatized (despite the obvious usefulness of the distinction),
and is often singled out for overt criticism. As a consequence of this stigma
many speakers who do have the you/youse distinction in their grammars
alternate between the marked and the unmarked form of the plural pronoun
according to various contextual (and possibly syntactic) constraints. For
example, one local post-graduate student was observed to address the class
as you while giving a seminar paper, but to revert to youse as the class broke
up and the topic changed. Thus the pronoun appears to be a classic social
and stylistic variable rather similar to, for example, the double negative. It
has of course quite a different sociolinguistic distribution from the T and V
pronouns studied by Brown and Gilman.
The really surprising fact to emerge during a lengthy and systematic study
of the language of several very low-status social groups in Belfast was that
many speakers categorically distinguished you (singular) and youse (plural);
that is. evidence began to build up which suggested that they never alter-
nated between the marked and the unmarked form of the pronoun.
This sociolinguistic distribution might be diagrammed as in Table I, with
three different groups of speakers distinguished, which are likely to cor-
respond to hierarchical social stratifications. What is of interest here is the
communicative consequence when a Group I speaker and a Group Ill
speaker interact. Do categorical differences in this portion of their grammars
result in communicative difficulties? Or do Group I I I speakers have the kind
 COMPREHENSION AND CONTEXT 17

Table I. The sociolinguistic distribution of •'youse".

singular plural
Group I you you
Group II you you/youse
Group III you yo use

of passive competence in standard English which is often assumed to exist?

In fact, the speaker of (1) was almost certainly a Group III speaker and the
addressee (myself, as the listener to the narrative) a Group I speaker. It is
perhaps worth emphasizing that the social contexts in which speakers from
these two socially distant groups interact are likely to be quite limited, and to
be in situations where Group I speakers have a measure of control over the
proceedings and may be interacting on a one-to-one or one-to-many basis.
Examples are classroom, legal, medical and bureaucratic situations. In
other words, as a consequence of the social norms governing cross-group
communications, occasions on which the categorical nature of the mismatch
between the two grammars might emerge are likely to be quite limited.
It has been suggested by James Miller (personal communication), correctly
I think, that as speakers grow older '-nd become more firmly established in
educationally and occupationally determined statuses, they interact less and
less on a symmetrical basis with socially distant persons. This constitutes a
social problem which vitiates systematic knowledge of non-standard dialects
and gives rise to unsupported claims which tend to exaggerat~ not only the
similarities between the grammars of these dialects and of Standard English
but also the knowledge which non-standard speakers have of Standard
English. Thus, evidence of how efficiently people from socially distant
groups decode each others' utterances is extremely difficult to collect. The
fieldwork situation which involved a Group I speaker in prolonged and
relatively relaxed conversation with Group III speakers is rare, and so is a
valuable source of data. The evidence and conclusions which emerged are as
follows.
First, it is fairly clear from (1) that the speaker takes the you/youse
distinction for granted to the extent that she comments on its communicative
importance to a socially distant addressee without any apparent awareness
of the stigma attached to it, or that the addressee might not use such a
distinction. Secondly, one consequence of this categorical distinction was
that during the field-work period I consistently encountered problems when
I used the unmarked form of the plural. For example a conventional greeting
such as "How are you?", addressed to two or more persons usually elicited a
zero response (silence) and an exchange of puzzled glances. Since I found it
18 LESLEY MILROY

hard to believe that the youse rule was responsible for these breakdowns,
they initially went unrepaired. However, the sequence ··How are youse
doing?" always elicited an appropriate response such as .. Not too bad, what
about yourself?" Further examples can be cited to illustrate the categorical
nature of the distinction for some speakers, the most spectacular of which is
a letter to two researchers who are consistently addressed as youse. Other
examples of youse in formal written communication have been reported. It
seems that the communicative competence of these speakers is limited to a
knowledge of the choice appropriate within their own social group. The
same can probably be said of more standardized speakers.
It is also clear that the usage exemplified in (1) bears strongly on the issue
discussed above, that is the kind of linguistic knowledge which non-standard
speakers may be said to have of standard English. At the very least we may
conclude that there is a great deal which we simply do not know about this
matter. Although the assumption (made usually by Group I speakers) that
all speakers have at least a passive competence in standard English seems on
the face of it perfectly reasonable, the youse evidence suggests that this may
be correct for Group II speakers but not for the more socially distant Group
III speakers. In a similar vein, Miller shows that Scottish speakers who do
not use the will/shall distinction have it in their passive competence only in
the limited sense that they recognize it as belonging to some variety of
English. When pressed to state the basis of the distinction, they either confess
to bafflement or manufacture "rules" which bear little relation to linguistic
reality such as "Shall is more polite than will" (Miller, 1981, p. 87). All in all,
Labov's argument that linguists derive many of their facts about low-status
vernaculars from upwardly mobile and relatively standardized speakers with
whom they have a measure of social affinity seems to be depressingly near
the mark (Labov, 1972, p. 287).
In view of our general ignorance both of non-standard grammars and of
the limits of communicative competence, I do not think that we can afford to
dismiss the breakdowns located in the you/youse disjunction as isolated and
unrepresentative of cross-dialectal interactions; the extent to which break-
down located in other sociolinguistically distributed grammatical elements
might occur is at the moment simply a matter of speculation.

Whenever he came in he hit me

The set of data discussed in this section has a sociolinguistic distribution
quite different from that of the youse variable. The temporal conjunctions
whenever and when do not contrast even in educated varieties of Hibemo-
English. Thus, Hiberno-English whenever has a meaning quite different
from the whenever of standard British English, and need not be interpreted
 COMPREHENSION AND CONTEXT 19
as indefinite or as referring to more than one occasion in the past (Milroy,
1978). 1 The following utterances are taken from four separate conversations:
(2) Whenever I saw her I fell for her.
(3) Whenever he came in he hit me.
(4) My husband died whenever I was living on the New Lodge Road.
(5) Whenever Chomsky wrote Syntactic Structures there was a revolu-
tion in linguistics.
It is clear that Irish and English listeners are likely to map these utterances
on to different grammars, and so produce different interpretations. How-
ever, (2) and (3) differ from (4) and (5) in that (2) and (3) are semantically
acceptable for both sets of speakers. But, for the standard English speaker,
(4) and (5) simply do not make sense, because the interpretation suggested
by his grammar conflicts with real world knowledge. What appears to
happen then in real situations (but not under experimental conditions, as we
shall see shortly) is that he concludes correctly that Hibemo-English when-
ever is equivalent to standard English when probably because this is the
easiest way of extracting relevance from the utterance. This feature of
Hiberno-English grammar is in fact quite frequently the subject of overt
comment by standard English speakers, and where an interpretation is
impossible for them in accordance with their own grammars does not appear
to be a source of miscommunication.
The same cannot be said of utterances like (2) and (3). While the temporal
clause in (2) was misinterpreted by a standard English speaker as referring to
several different occasions, the misunderstanding did not have the effect of
producing communication problems and was not noticed by either of the
interlocutors. Only direct questioning after the exchange revealed the dis-
parate interpretations, and it is very likely that a great many unimportant
misunderstandings of this kind go both unresolved and unnoticed. A similar
case is documented by Berdan (1977, p. 15), where in cross-dialectal
communication speakers may communicate successfully with each other
despite the fact that both are mapping quite different grammars on to
utterances containing got as a verb of possession. In the terms which we have
adopted here, the.ambiguity of (2) resulted in misunderstanding but not in
communicative breakdown.
The misunderstanding resulting from (3), which passed unnoticed at the
time, was potentially more serious simply because of the propositional
content of the utterance. The addressee was myself, and the topic of
conversation was the unreasonable behaviour of the speaker's husband.
Being at that time unfamiliar with this part of the Hibemo-English system, I
interpreted the reference of the utterance as a large number of assaults upon
the speaker. Only many weeks later did it become apparent, on reflection,
20 LESLEY MILROY

that I had drawn an entirely wrong series of inferences from the miscom-
munication.
If these observation!> on examples (2) to (5) are taken together, we can
begin to isolate some structural and pragmatic conditions which appear to be
necessary before syntactically located cross-dialectal miscommunications
take place. First, the utterance must be semantically acceptable to both
speakers; otherwise it is simply odd, like (4) and (5) rather than ambiguous
{but see further below). Second (and this is a matter of social norms) the
reference must be of some "real world" importance. Otherwise, as with (2),
the interaction will proceed without the interlocutors necessarily being
aware of the misunderstanding. In contrast, the misunderstanding following
utterance (3), although not noticed at the time, was identified retrospectively
because in our society a man who repeatedly beats his wife attracts more
notice than one who repeatedly falls in love with the same woman. Thus, it
appears that successful cross-dialectal communication in context depends
on a fairly complicated combination of structural and pragmatic factors.

It is worth drawing attention here to an experimental study by Trudgill

where subjects are asked to select an interpretation of a range of sentences
generated by non-standard English dialect grammars {Trudgill, 1981). The
whenever sentence selected (whenever it was born Ifelt ill) is semantically
acceptable only if whenever is interpreted as equivalent to when, and so is of
the same type as utterances (4) and (5) above. Contrary to what appears to
happen in real contexts, where such utterances do not seem to result in
communication problems, Trudgill found that "the grammatical constraints
of their own dialects forced a majority of subjects to select [an] anti-
commonsense interpretation". Subjects often regarded the sentence as
complete nonsense or, in some cases, attempted a contextualization of their
semantically odd interpretation. This disparity between experimental and
observational results is instructive, and is probably best viewed as showing
the unwillingness of speakers in real social situations to regard any utterance
as complete nonsense. Relevance will always be assumed (in Grice's sense)
and a likely interpretation will be sought, while a subject in an experiment (a
constructed situation) will not be under the same social pressures to locate
the speaker's intention. Additionally, he may not assume that the stimulus
sentence is relevant.
An interesting parallel to this mismatch between experimental results and
the results of naturalistic observation may be found in recent psycholinguistic
work on the basis of linguistic intuitions (Carroll et al., 1981). Here, the
authors describe the manner in which judgments of "borderline" sentences
as acceptable vary in quite a systematic way, as the judges' mental states are
 COMPREHENSION AND CONTEXT 21

manipulated towards either "objective self-awareness" or "subjective self-

awareness". These states may well be parallel to states in experimental as
opposed to real life situations.

How long are youse here?

This utterance is the opening sequence of (6). A is a native of South West
Donegal; B and C are both standard English speakers (and linguists).
(6) A: (i) How long are youse here?
B: Till after Easter
(A looks puzzled; a pause of two seconds follows)
C: We came on Sunday
A: (ii) Ah, Youse're here a while then.
The miscommunication here is located in an extremely complicated dif-
ference between the aspect the systems of Standard English and Hibemo-
English, which may for the present purposes be outlined as follows (but see
also p. 12 above).
A standard English speaker will usually interpret an utterance like A (i) as
equivalent to How long will you be here for? Hibemo-English generally
avoids the use of the have auxiliary as a marker of the perfect, and has a
range of structures, including the A (i) type, which appear to be equivalent
to the standard English perfect (see Harris, 1982 for further discussion).
Utterance B was therefore an inappropriate response to A (i), and speaker
A was quite unable to interpret it. For him, A (i) and B simply constituted an
ill-formed exchange. The breakdown was, after a time lag, noticed and
repaired. It is by no means certain that a linguistically unsophisticated
speaker such as A (both B and C were linguists) could have located the
breakdown and effected the repair.
This is an extremely clear example of a cross-dialectal breakdown which
can be located specifically in a disparity between the grammars of speaker
and hearer. The difference between the two semantic structures associated
with A (i) involves the existence of a state in time either up to a certain point,
or from a certain point. The somewhat intangible nature of this difference,
together with the absence of further relevant contextual cues, appear to rule
out comprehension strategies which involve contextual factors or real world
knowledge. The claim made by, for example, Smith and Wilson (1979,
p. 197) that perceptual strategies of various kinds facilitate cross-dialectal
communication, to the extent that differences in internal grammars are
communicatively unimportant, seems to be contradicted by this example.
Moreover, the fact that speakers B and C knew (in the encyclopedic rather
than the linguistic sense) of the disparity between this portion of the two
22 LESLEY MILROY

grammars was not enough to prevent the breakdown in the first place,
although it was eventualJy noticed. It may be concluded then that there are
specifiable situations where grammars (as opposed to strategies) are crucial
to effective communication between speakers of different dialects.
Finally, it is worth noting the reaction of the participants to the break-
down. During the two-second pause, B and Cat least had an extremely
unpleasant sense of simply not knowing what was going on and being quite ·
unable to analyse it; this is very reminiscent of the reactions of puzzlement
and frustration reported by Gumperz (1976) when breakdown involving
differences in the kind of conversational inferences drawn by speaker and
hearer takes place. Thus, the possibility that breakdowns of this kind may
not be analysed as linguistic at all is another reason for caution in assuming
that they are rare.
The communicative breakdown exemplified in (6) may be compared with
(7), a successful communication between an Ulsterman and an Englishman
working as a temporary lecturer in Northern Ireland.
(7) A: How long are you here?
B: Just this term.
Here the misunderstanding was not noticed (although the same Englishman
had experienced problems in understanding similar constructions) appar-
ently because the response can be construed as "length of time during
which", i.e. unlike 6B it does not contain a temporal item referring to either
future or past. Alternatively we may assume ellipsis, and restructure 7B as
either 8 or 9:
(8) Just (until the end of) this term.
(9) Just (since the beginning of) this term.
Since neither past nor future time were overtly marked in the response by a
temporal conjunction, each speaker was able to associate the response with
a different underlying semantic structure. A comparison with (6) and (7)
suggests that ellipsis may play a part in masking miscommunications-.

A somewhat surpnsmg insight into the role of linguistic context in

successful cross-dialectal communication is provided in an important experi-
mental study by Berdan (1977). (In view of the facts reported above, we
should be cautious however of drawing over-general conclusions about
comprehension-in-context from experimental studies.) Berdan presented to
two groups of American students (one black and one white) a number of
potentially ambiguous sentences and, as a means of determining their
interpretation of these sentences, asked them to construct tag questions for
 COMPREHENSION AND CONTEXT 23

them. The rule of copula deletion in Black English was the locus of the
ambiguity.
Where contracted is appears on the surface, it is often phonetically
indistinguishable from the plural morpheme. Thus, the sequence [frenz) in
(10), with contracted is, is indistinguishable from plural friends in (11),
although each is derived from a different underlying sentence:
(10) Her best friend is playing jump rope.
(11) Her best friends arft playing jump rope.
On hearing the resulting identical phonetic sequence, speakers of a dialect
which does not allow copula deletion would allow, Berdan considered, only
the interpretation shown in (10) while speakers of a dialect which does allow
copula deletion would find both ( 10) and (11) possible interpretations.
In fact, for the sequence Her best [frcnz] playing jump rope, all of the
white and all but two of the black students provided a singular tag, showing
that they understood the sentence as (10). However, Berdan noted that it
appeared to be possible to change semantic and other contextual cues in an
ambiguous sentence to make one reading at first sight more plausible than
another without making either reading ungrammatical. He suggested, for
example, that a superlative like best seemed to facilitate a singular inter-
pretation, although plurals with best, such as (12) clearly do occur:
(12) They are my best friends.
He also suggested that a singular possessive pronoun like her might facilitate
a singular reading, while a plural pronoun like their might increase the
possibility of a plural reading. To explore the effects of these contextual
cues, the sequences (13) to (16) were included among the test items:
(13) Her best [frenz] playing jump rope.
(14) Her [frenz] playing jump rope.
( 15) Their best [fn:nz] playing jump rope.
(16) Their [frenz] playing jump rope.
As predicted, singular interpretation frequencies decreased progressively
from (13) through to (16), most spectacularly among black students. How-
ever, one unexpected finding was a significant number of plural readings of
(15) and (16) even among the white students whose grammars were thought
not to allow the deletion rule.
Berdan's results confirm the extremely indirect relationship between
comprehension and grammatical rules. They add to our knowledge of
comprehension strategies in that they show that for some white informants,
logically irrelevant linguistic cues are sufficient to override interpretations
suggested by the application of grammatical rules. However, the magnitude
24 LESLEY MILROY

of the contextual conditioning was smaller for the white than for the black
students; therefore it seems that the capacity of contextual cues to encourage
persons to employ particular perceptual strategies is greater when the cues
do not conflict with linguistic rules, than when they do conflict with them.
Although Berdan 's findings confirm the importance of perceptual strate-
gies in skewing the interpretation of ambiguous sentences, it appears that
quite simple lexical changes such as those which he manipulates may
influence comprehension one way or another. However, these cues do not
operate independently of a speaker's internal grammar or, as has sometimes
been suggested, take precedence over it. Rather, grammatical rules and
perceptual strategies operate in conjunction with each other (and, as we
have seen, with extra-linguistic contextual factors) in an orderly way in
influencing the interpretation of utterances.

Conversational inferences

Since much reference has been made in this chapter to contextual and
pragmatic cues, it is appropriate to note that the locus of a miscommunica-
tion may be specifically "pragmatic" rather than "structural". By this I mean
that it is located in a disparity between the inferences which conversational
participants draw from a given utterance, rather than in a disparity of the
kind discussed here so far between the semantic structures from which they
derive that utterance. This section relies heavily on the work of Gumperz
and his associates on cross-cultural communicative breakdowns (Gumperz,
1976, 1977(a), 1977(b);Gumperzetal., 1979;GumperzandTannen, 1979).
Fundamentally, Gumperz argues that difficulties in inter-ethnic communi-
cation arise not only as a result of generalized prejudice, but through
differences in various kinds of shared knowledge and in culture-specific
communicative preferences. These differences affect the kind of inferences
drawn in conversation to interpret a given utterance. For example, ethnic
groups differ in the uses which they make of a given prosodic pattern: the
tune appropriate for polite questions in one culture may signal an emphatic
statement in another. Many miscommunications are documented which
may be located specifically in disparate interpretations of prosodic cues
(Gumperz, 1977(b); Gumperz and Tannen, 1979, Gumperz et al., 1979).
The culture-specific nature of different levels of loudness and different
proxemic behaviours in a given situation has been shown elsewhere to be the
source of miscommunication (Watson and Graves, 1966; Hall, 1963; see also
Brown and Levinson, 1978, p. 258). Similarly, culture-specific differences in
"the distribution of required and preferred silence" (Hymes, 1972) have
been documented as a source of miscommunication (Philips, 1976; Milroy,
 COMPREHENSION AND CONTEXT 25

1980). In Milroy (1980) the differences were between norms appropriate to

different status groups rather than different ethnic groups in the usual sense.
A source of miscommunication which may be labelled as more specifically
pragmatic may be located in disparate interpretations of the illocutionary
force of indirect speech acts. For example, an utterance such as (17) can in
some situations be interpreted as a directive, and in others as a simple
comment with no particular illocutionary force, while (18) might be inter-
preted either as a request for information, or a request for action:
(17) It's dinner time and I'm hungry.
(18) Can you play the Moonlight Sonata?
In the absence of clear prosodic cues, the speaker's communicative intent is
interpreted in relation to certain presuppositions which are influenced by
aspects of the (extra-linguistic) context such as setting, topic, the speakers'
knowledge of each other and of their mutual rights and obligations. Currently
there is a great deal of interest in the presuppositions and logical processes
which underlie the interpretation of such indirect speech acts; but as
Gumperz points out, what is meant by context and the manner in which it
enters into the interpretation process has been discussed only in general
terms (Gumperz, 1976; but see Labov and Fanshel, 1977 for an attempt at a
more precise formulation). More seriously, it seems to be commonly
assumed, though not explicitly stated, that conversational participants
generally share a kind of pragmatic competence which enables them to
interpret contextual factors in much the same way (Gumperz, 1976, p. 281).
This assumption is particularly clear in an early article by Lakoff (1971 ), and
more recently Smith and Wilson (1979) refer to utterances-in-context as
containing uitems of non-linguistic knowledge shared by speaker and hearer,
and a set of shared inference rules" (p. 174). Leaving aside for a moment the
question of how much relevant non-linguistic knowledge persons with
different cultural norms might be said to share, it is clear that the "shared
inference rules" are often insufficient to prevent misinterpretation of the
illocutionary force of indirect speech acts even between persons who know
each other well; sometimes these misinterpretations are repaired. Many
examples of exchanges such as the following could no doubt be provided by
the reader:
(19) Wife: Will you be home early today?
Husband: When do you need the car?
Wife: I don't, I just wondered if you'd be home early.
It may in fact be more accurate to talk of "inferential preferences" of a
probabilistic nature, rather than "inference rules", so as to allow for the fact
that such misinterpretations are quite frequent (cf. the comments of Lyons
26 LESLEY MILROY

and Aitchison above, and see Klein and Dittmar, 1979 for a good discus-
sion of probabilistic models). Tannen (1981) has recently made a very
similar suggestion, viz. that in explaining how persons interpret indirect
speech acts, the notion of patterns of interpretation is more appropriate than
that of rules of discourse. These patterns of interpretation depend, in a
complex way, on "context, individual and social differences and inter-
personal dynamics" (p. 484).
While taking the non-categorical nature of the inference process into
account, Gumperz suggests tentatively, on the basis of a number of docu-
mented conversations and direct questions put to judges, that a predisposi-
tion to interpret contextual cues in a given manner may vary systematically
between speech communities. For example, he considers that American
speakers are more likely than British speakers to interpret interrogatives of
a pragmatically ambiguous kind (such as (18)) as requests for information.
Interestingly, the miscommunication between himself and an American
salesman reported by Trudgill (1981) seems to support this view. More
recently, Gumperz and Tannen have examined in some detail a number of
miscommunications, some but not aU of which may be located in the
systematically different types of inference likely to be drawn by a number of
American ethnic groups. Further, Tannen reports that in a study of cross-
cultural differences in inferential preferences, more Greeks than Americans,
when presented with a sample of conversation, interpreted why utterances
(such as Why are you here?) as indirect speech acts (Tannen, 1981).
One point made by Gumperz repeatedly is that communicative break-
downs are seldom analysed as linguistic. Speakers typically react with a
sense of frustration and hostility; frequently each participant accuses the
other of perverse and wilfully difficult behaviour, and if confronted with a
re-run of the conversation, claims that his own interpretation of a given
prosodic pattern, or a given utterance, is the only reasonable one. In the case
of cross-ethnic communication,. this kind of breakdown appears to con-
tribute to hostile stereotyping, and one concern of Gumperz has been to
devise training programmes for those whose work involves them in com-
municating with other ethnic groups. This would appear to be an important
application of any theory of communication which could account elegantly
and systematically for the frequent miscommunications which do un-
doubtedly take place. It seems likely, moreover, that the kind of culturally
distributed problems and reactions to these problems which Gumperz
describes are not limited to inter-ethnic communication. They almost
certainly occur in a similarly systematic manner in conversations between,
for example, Englishmen who, like the Group I and Group III speakers
discussed in relation to youse wash the dishes, are socially distant from each
other.
 COMPREHENSION AND CONTEXT 27

Summary and conclusion

By focussing on miscommunications, the purpose of this chapter has been to
explore the manner in which people communicate successfully. Compre-
hension is a complicated matter, requiring a multidisciplinary approach; it
seems that the perspective offered by sociolinguistics, concerned as it is with
the analysis of speech events in their social context, might be a helpful one.
Analysis of a number of miscommunication has shown that, despite the
complex range of strategies involved in comprehension, it is possible to
locate some problems specifically in the mismatch between the dialect
grammars of speaker and hearer.
The first of the miscommunications examined (altogether three relevant
areas of Hiberno-English grammar were isolated) was located in a syntactic
disparity which also had a clear sociolinguistic distribution; that is, the
you/youse contrast. Analysis suggested that the non-standard speaker's
general communicative competence, as well as his passive competence in
standard English, might be more limited than is commonly supposed.
The analysis focussing on Hiberno-English whenever suggested that
structurally located miscommunications were likely to occur and be noticed
under specifiable conditions: the ambiguous utterance had to be semantically
acceptable to both speaker and hearer; and the ambiguity had to be located
in a message which was of "real world" importance to the hearer. However,
it was noted that under experimental conditions, subjects appeared to
misinterpret even semantically unacceptable sentences. This may be because
experimental subjects do not search out relevance in "odd" sentences
whereas conversationalists do. In other words, Grice's maxim of relevance
may not apply in test conditions. A further analysis which focussed on the
Hiberno-English perfect suggested that, probably because of the abstract
nature of the reference, non-linguistic context was often of little use in
helping conversationalists avoid breakdowns. However, ellipted responses
sometimes concealed misunderstandings. The question of how specifically
linguistic context influenced comprehension was pursued further. and some
experimental results were discussed which suggested that in cross-dialectal
communication perceptual strategies worked together with linguistic rules
in an orderly manner when persons interpreted ambiguous sentences.
Since consideration of all the data suggests that it is possible to specify
conditions under which syntactic mismatches are likely to produce mis-
communications, we may consider these observations as a modest contribu-
tion towards a more explicit and testable theory of communication.
Finally, some breakdowns located in disparate inferential strategies were
briefly discussed, and it was suggested that the notion of inferential preferences
28 LESLEY MILROY

of a probabilistic kind might be easier to work with than that of inference

rules, so as to allow theoretically for frequently occuring miscommunications.
One particularly interesting (if negative) conclusion is that there is much
about cross-dialectal communication which we simply do not know, although
the general assumption in standard linguistics textbooks is that there is very
little worth knowing. I do not think that we can afford to dismiss the
instances discussed here as .. marginal" in the sense that Hiberno-English
has an unusally exotic grammar. There are many subtle differences between
even Anglo-English dialects and standard English (such as those associated
with the temporal conjunction while in Northern England) which might, if
investigated, be found to be the locus of communicative problems.
Linguists who consider themselves to be socially responsible may also feel
that the possibilities of applying work of this kind to situations where
cross-dialectal communicative problems have unpleasant social consequences
are considerable. Preliminary attention has already been given to the analysis
of problems which are known to arise in language-testing situations (see for
example, Labov, 1972(a); Taylor, 1977; Milroy and Milroy, forthcoming). It
is hoped that this chapter will stimulate readers to consider further specific
applications of a detailed, dispassionate and accountable analysis of cross-
dialectal communication problems. These are problems which are of practical
concern to many, but only a linguist is capable of carrying out such an
analysis; iflinguists do not do the job, it will be done, but less well, by others.

Note

Much of the work reported in this chapter was carried out in the course of
research projects HR 3771 and HR 5777, funded by the Social Science
Research Council. This support is gratefully acknowledged here.
I also acknowledge the contribution of my co-workers, James Milroy,
John Harris and Linda Policansky, all of whom have been struggling to cope
with the vagaries of non-standard syntax which sometimes seems to defy
analysis. Thanks are due particularly to James Milroy for his initial insights
into the very great differences between Hiberno-English and Standard
English syntax. The following have been of great help in providing data or
commenting on earlier drafts, and I thank them: Greg Brooks, Alison
Davis, Michael McTear, John Harris, James Milroy, James Miller and
Michael Stubbs and Margaret Deuchar.
 COMPREHENSION AND CONTEXT 29

References

Aitchison, J. (1976). ''The articulate mammal: an introduction to psycholinguistics",

Hutchinson, London.
Berdan, R. (1977). "Polylectal comprehension and the polylectal grammar'', In
(Fasold and Shuy, eds), pp. 12-29.
Bever, T. G. (1970). The cognitive basis for linguistic structures. In "Cognition and
the development of language", (J. R. Hayes, ed.), Wiley. New York.
Brown, P. and Levinson, S. C. (1978). Universals in language usage: politeness
phenomena. In Goody, (ed.), pp. 56-289.
Carroll, J. M., Bever. T. G. and Pollack, C. R. (1981). The non-uniqueness of
linguistic intuitions. Language, 57.2, 368-381. .
Cicourel, A. ( 1973). "Cognitive Sociology", Penguin, Harmondsworth.
Coulthard, M. ( 1977). "An introduction to discourse analysis", Longmans, London.
Edwards, V. K. (1979). ''The West-Indian language issue in British schools: chal-
lenges and responses", Routledge, London.
Ervin-Tripp, S. (n.d). "Structures of control", (mimeo, Department of Psychology,
Berkeley).
Fasold, R. W. and Shuy, R. W. ( 1977). ''Studies in language variation", Georgetown
University Press, Georgetown.
Ferguson, C. A. (1977). Linguistics as anthropology. In Saville-Troike (ed.). pp.
1-12.
Fillmore, J., Kempler, D. and Wang, S. Y. (1979). "Individual differences in
language ability and language behaviour", Academic Press, London and New
York.
Fromkin, V. (ed.) (1973). "Speech errors as linguistic evidence", Mouton, The
Hague.
Fromkin, V. (ed.) (1981). "Errors in linguistic performance'', Academic Press,
London and New York.
Givon, T. (1979). "On understanding grammar", Academic Press, London and New
York.
Goody, E. N. (ed.) (1978). "Questions and politeness", C.U.P., Cambridge. ..
Gordon, D. and Lakoff, G. (1971). Conversational postulates. In "Papers from the
seventh regional meeting of the Chicago Linguistic Society", CLS, Chicago. .
Grice, H.P. {1975). Logic and conversation. In "Syntax and Semantics: speechacts'',
(P. Cole and J. Morgan, eds), Academic Press, London and New Y9rk'.'-'
Gumperz, J. J. (1976). Language, communication and public neg6tiatioil'., In
"Anthropology and the public interest: fieldwork and theory",·(P: Sanday_; ed.),
Academic Press, New York and London. ·
Gumperz, J. J. (1977a). "The conversational analysis of interethnic communication",
Mimeo: University of Berkeley. ·· ··
Gumperz, J. J. (1977b). Sociocultural knowledge in conversational inferences. In
Saville-Troike (ed.), pp. 191-211. ·· ·
Gumperz, J. J., Tupp, T. C. and Roberts, C. (1979). "Crosstalk: a study of .cross-
cultural communication", National Centre for Industrial Language Training,
Southall. · ·
Gumperz, J. J. and Tannen, D. (1979): Individual and social differences in language
use. In Fillmore et al. (eds), pp. 305-326. ' · · ·
30 LESLEY MILROY

Gumperz, J. J. and Hymes, D. (1972). "Directions in Sociolinguistics", Holt,

Reinehart & Winston, New York.
Hall, E. T. (1963). A system for the notation of proxemic behaviour. American
Anthropologist, 65, 1003-26.
Harris, J. (1982). The Hiberno-English verb phrase: variation, change and dialect
maintenance. In Belfast Working Papers in Language and Linguistics, 6.
Hughes, A. and Trudgill, P. (1979). "English Accents and Dialects", Arnold,
London.
Hymes, D. (1972). Models of the interaction oflanguage and social life. In Gumperz
and Hymes (eds), pp. 35-71.
Klein, W. and Dittmar, N. (1979). "Developing grammars'', Springer, Berlin.
Labov, W. (1972a). Where do grammars stop? In "Sociolinguistics: current trends
and prospects", (R. W. Shuy ed.), Georgetown University Press, Washington
D.C.
Labov, W. (1972b). "Language in the inner city", Pennsylvania University Press,
Philadelphia.
Labov, W. and Fanshel, D. (1977). "Therapeutic discourse", Academic Press, New
York and London.
Lakoff, G. (1971). Presupposition and relative well-formedness. /n "Semantics", (D.
Steinberg and D. Jacobovits, eds), Cambridge University Press, London.
Lyons, J. (1977). "Semantics" (2 vols), Cambridge University Press, Cambridge.
Lyons, J. ( 1980). Pronouns ofaddress in Anna Karenina: the stylistics of bilingualism
and the impossibility of translation. In "Studies in English Linguistics", (S.
Greenbaum, G. Leech and J. Svartvik, eds), Longmans, London.
Miller, J. (1981). ••spoken language, written language and competence", (Mimeo,
Department of Linguistics, University of Edinburgh).
Milroy, J. (1978). Stability and change in non-standard English in Belfast. Journal of
the Northern Ireland Speech and Language Forum, 72-82.
Milroy, L. (1980). "Language and Social Networks", Blackwell, Oxford.
Milroy, L. and Milroy J. (forthcoming). "Authority in language: a sociolinguistic
study of prescriptivism", Routledge and Kegan Paul, London.
Nelson, N. W. and McRoskey, R. L. (1978). Comprehension of standard English at
varied speaking rates by children whose major dialect is black English. Journal
of communication disorders, 11, 37-50.
Philips, S. (1976). Some sources of cultural variability in the regulation of talk. Lang.
Soc. 5, 81-95.
Plutchik, R. (1976). (2nd ed.). "Foundations of experimental research", Harper &
Row, New York.
Ross, J. R. (1975). Where to do things with words. In "Syntax and Semantics 3:
speech acts", (P. Cole and J. L. Morgan, eds), Academic Press, New York and
London.
Ross, J. R. (1979). Where's English? Jn Fillmore et al. (eds), pp. 127-164.
Saville-Troike, M. (ed.) (1977). "Linguistics and Anthropology", Georgetown
University Press, Washington D.C.
Schegloff, E. A., Jefferson, G. and Sacks, H. (1977). The preference for self-
correction in the organisation ofrepair in conversation. Lar,guage, 53, 361-382.
Searle, J. (1975). Indirect speech acts. In "Syntax and semantics 3: speech acts", (P.
Cole and J. L. Morgan, eds), Academic Press, New York and London.
Smith, N. and Wilson, D. (1979). "Modern Linguistics", Penguin, Hannondsworth.
Sperber, D. and Wilson, D. (forthcoming). "Language and relevance; foundations of
pragmatic theory", Oxford, Blackwell.
 COMPREHENSION AND CONTEXT 31
Tannen, D. (1981). Review of Therapeutic Discourse by W. Labov and D. Fanshel,
Language,51,2,481-486.
Taylor, 0. (1977). The sociolinguistic dimension in standardized testing. In Saville-
Troike (ed.), pp. 257-266.
Trudgill, P. J. (1979). Standard and non-standard dialects of English in the United
Kingdom: problems and policies. lnter11ational Journal of the Sociology of
Language,21,9-24.
Trudgill, P. J. (1981). On the limits of passive "competence": sociolinguistics and the
polylectal grammar controversy. Jn "Linguistic controversies: Festschrift for F.
R. Palmer", (D. Crystal, ed.), Arnold, London.
Watson, 0. M. and Graves, T. D. (1966). Quantitative research in proxemic
behaviour. American Anthropologist, 68, 971-985.

1
It is also distinct, semantically, from Scottish whenever, which is equivalent to "as
soon as".
 3
The relationship between
language and sex in English

Jenny Cheshire

Many sociolinguistic studies have shown how social divisions between

speakers, such as age, socio-economic class or sex, are reflected in their
language. In some tribal societies where men and women lead relatively
separate lives, there are often clear phonological, lexical or syntactic
differences in their speech: this was true, for example, of many Amerindian
and Australasian languages. Where the sexes are less segregated, as in
modern Western societies, language differences are more subtle (for
examples, see Trudgill, 1974a; Chapter 4). More recently, research on
language and sex has gone beyond the analysis of sex-differentiated varieties
of language, to investigate the way that language reflects and helps to
maintain social attitudes towards women and men.
In Britain and the USA, for example, it has been shown that our con-
ventional use of English treats men and women unequally, rather than
simply differently. This paper will briefly collate some of the more important
research findings in this area, and will discuss their implications. It will also
document some examples of changes in the use of English, which reflect an
increased social awareness of the way in which English discriminates against
women. These changes result in part from the dissemination of research
findings (Miller and Swift, 1977 and 1981, for example, are written for the
general reader rather than for a purely academic audience), but they can
also be seen as a natural result of the changing position of women in British

APPLIED SOCIOLINGUISTICS Copyright© 1984 by Academic Press, London

ISBN: 0.12-701220-6 All rights of reproduction in any fonn reserved
34 JENNY CHESHIRE

and American society, and of growing interest and support for some of the
issues raised by the feminist movement.
The first two sections of this paper discuss research into sexism in English,
and the way in which our use of language maintains stereotyped images of
the sexes. They also discuss some of the implications of the research
findings, and describe some efforts that have been made, by both individuals
and group organizations, to avoid discriminatory language themselves and
to encourage others to avoid it also. The following two sections point out
some potential applications of research into two other aspects of language
and sex: the analysis of sex differences in language use, and the evaluation of
men's and women's speech.

Sexism in English

A great deal of research into sexism in English has focussed on the third
person singular pronoun forms, which force speakers to specify the sex of
the person to whom the pronoun refers. As Conklin (1974) points out, this
means, among other things, that it is possible to write a recommendation for
a job that avoids discrimination on the grounds of race, nationality, or
religion; but it is not possible to avoid discrimination on the grounds of sex,
since it is difficult to avoid using pronouns.
Severa) attempts have been made to introduce a neutral third person
singular pronoun into English, one (thon) dating from as far back as 1859.
This early proposal is, in fact, the one that has had the greatest impact; it was
listed in Funk and Wagnell's New Standard Dictionary in 1913, and it was
still sufficiently recognized in 1959 to be included in the second edition of
Webster's International Dictionary. Other proposals have had more limited
success. Co is used in some communes in the USA, particularly in Virginia
and Missouri, and it is routinely used in the magazine "Communities". It has
also been used in a book on radical therapy, published by Harper and Row in
1973 (see Miller and Swift, 1977, p. 130). Some recent novels have used na
(Arnold, 1973) and person or per (Piercy, 1979) and a supervisors' guide
issued by a division of American Management Associations uses hir
(Killian, 1979). Other suggestions include e. tey, hesh. po, re, xeandjhe, but
none of these has been widely accepted, perhaps, as Lakoff {1973) suggests,
because of the difficulty of artificially introducing a new item into a closed
linguistic system. A solution that might meet with more success is to extend
the function of an item that is already in the system. The plural pronoun
they. in fact, has been used in spoken English for centuries in phrases such as
"everyone must do their best" (see Bodine, 1975). Prescriptive grammarians
argue against this on the grounds that they is inaccurate in terms of number,
 LANGUAGE AND SEX IN ENGLISH 35
and recommend he instead as a "sex-indefinite" pronoun. This, of course, is
equally inaccurate, in terms of sex, but their insistence has meant that he is
the form that is naw generally used as a "neutral" pronoun in formal written
English.
One important application of work on language and sex has been the
experimental demonstration that "neutral" he is interpreted as referring not
to both males and females, but to males only (see, for example, Martyna,
1978; Moulton et al., 1978). Mackay (1979) found that this form occurs so
often in university textbooks that educated Americans must be exposed to
more than ten million occurrences during their lifetimes, which means that
this cannot be dismissed as a trivial phenomenon. Furthermore, it is not only
the intended "neutral". pronoun that is misinterpreted: generic man, in
phrases such as stone-age man or no man is an island, is also interpreted as a
masculine noun, by both children and adults (see Nilsen, 1973; Harrison,
1975; Schneider and Hacker, 1973). The main implication of these findings is
that the use of these terms excludes women from our thinking and our
culture. Most school and university textbooks, in fact, do exclude women
and women's achievements, as Hoffmann {1981) points out. Another
important implication is that the use of "false generics" in surveys, and of
male-orientated thinking generally, may lead to inaccurate results in many
areas of research. Goot and Reid {1975) suggest that most of the received
wisdom about the political attitudes of women is inaccurate, because it stems
from ambiguous questionnaires that ask about "the ordinary man", or "the
man with high ideals". And Delamont (1980) points out that political
scientists often talk of "democracies" that deny the vote to women and that
cannot, therefore, be democracies. It seems probable that culturally con-
ditioned sexist thinking has affected our .. knowledge" in other areas of
enquiry also (for some attempts to remedy this, see Roberts, 1981; Spender,
forthcoming).
"Thinking male" is not confined to academic research, however. The
following quotations, from everyday life, show that it is not only nouns of
masculine gender that are assumed to have male reference. (They also,
incidentally, provide examples of the way in which women are often
considered to be the possessions of men.)
My ambition is to have a show in London with the same sort of reputation that
the Crazy Gang had. It would be glamorous, spicy. but above all a family show.
People would bring their wives, mothers, and children. (Ken Dodd, in Woman
magazine)
What causes most distress to the residents is the kerbcrawlers, molesting their
womenfolk. (James Hill, MP, on This Week in Westminster, Radio 4).

Lack of a neutral singular pronoun and the use of "generic" nouns of

36 JENNY CHESHIRE

masculine gender are the most obvious examples of sexism in English; there
are many other linguistic features, however, that reflect the social status of
women (for examples, see Lakoff, 1973; Miller and Swift, 1971; 1981).
Recognition of the existence and of the implications of these features has
led to conscious attempts to eliminate them from the language. These moves
originated in the USA, but awareness of the issues is now spreading to
Britain. Several American publishing companies issue guidelines for their
authors and editors that suggest ways of avoiding sexist language. Scott,
Foresman and Co .. for example, issued "Guidelines for improving the
image of women in textbooks" in 1972, and McGraw-Hill's "Guidelines for
equal treatment of the sexes in McGraw-Hill book Publications" has been
widely distributed since 1974 to individual writers as well as to government
agencies, schools and universities, and the media. In Britain, the Women in
the Publishing Industry Group drew up the "Non-sexist code of practice for
book publishing" in 1976, and the British Edition of "The handbook of
non-sexist writing" (Miller and Swift, 1981) contains examples of writing
that is unintentionally sexist, from The Sunday Times and The Observer, and
suggests ways of avoiding offending forms.
Some writers make a conscious effort to avoid sexism: Aitchison (1981),
for example, uses both she and he as "sex-indefinite" pronouns in order, as
she writes in the Preface, "to help conquer the all-pervading sexism which
exists in the English language". Use of the written forms/he is becoming
quite widespread, and the order of the nouns and pronouns in phrases like
"'he and she", "men and women", •'mother and father" is sometimes
purposely reversed. The reference book "Baby and Child" (Leach, 1977)
uses she as a "neutral" pronoun throughout. 1 This is an important step, for
the book is widely used, and although it is ultimately, of course, as sexist as
using only he, it should attract the attention of readers who might otherwise
not have been aware of the issue. Books on childcare appear to be leading
the attack on the use of "neutral" he. Dr Spock has promised to alternate she
and he in the next edition of his popular "Baby and Child Care", and Salk
has already done so in his standard reference book for parents (Salk, 1974).
Professional organizations are also taking steps to encourage change in
the language. The American Anthropological Association, for example,
passed a resolution in 1973 urging its members to "become aware in their
writing and teaching that their wide use of the term •man' as generic for the
species is conceptually confusing" (reported in Miller and Swift, 1977,
p. 129). The American Library Association resolved in 1975 to avoid using
sexist language in all future publications and official documents, and to
change existing publications when they were revised. And in 1976, library
cataloguers in the USA initiated a campaign to revise the use of sexist
language in subject headings and in catalogue descriptions.
 LANGUAGE AND SEX IN ENGLISH 37
Several religious bodies are also changing the language used in their
publications. In the USA, the General Synod of the United Church of Christ
announced in 1973 that it would eliminate sex and race discrimination in all
areas of its teaching, and it has, since then, been revising all its printed
materials, including hymn books, service procedures and journals, to ensure
that the language used is deliberately inclusive. The Jewish prayerbook
"Gates of Prayer", was also revised in 1975, and it acknowledges in the
Introduction to the Revised Edition a similar commitment to equality of the
sexes. The revisions include substitutions such as '"God of all generations"
for "God of our fathers", and additions such as "God of our mothers, God
of Sarah, Rebekah, Leah ... " as a parallel to "God of our fathers, God of
Abraham, Isaac and Jacob ... " And the influential inter-denominational
Journal of Ecumenical Studies devoted an editorial to "linguistic sexism" in
1974 (Volume XI, no. 2, Spring edition).
Although religious bodies in Britain are not renowned for equality in the
treatment of the sexes (witness, for example, the refusal until last year to
admit women into the Ministry, and the persistence of all-male choirs in the
Anglican church), an initiative has come from the Methodist movement.
The revised version of the Methodist Hymn Book, which is due to be
published in 1983, omits some hymns that are considered to be blatantly
sexist ("Rise up, 0 men of God", for example, has been excluded), and
contains adapted versions of others. In many cases the adapted hymns are
historically more accurate than the original versions: "O God our help in
ages past", for example, contained the lines

Time, like an ever-rolling stream

Bears all its sons away

but the revised version, which substitutes "Bears mortal flesh away" for the
second line, above, bears more resemblance to the wording of the psalm on
which the hymn is based. The Methodist Hymn Book also makes some
attempt to avoid male personification of God, by addressing the Deity
directly as Thou, rather than indirectly as He. It continues, however, to use
masculine imagery, referring to God as a father, king, shepherd and lord.
Research into the language of religion has shown how the predominance of
masculine imagery in religion results from the orientation of Western culture.
Female imagery that.was present in the Hebrew scriptures, for example, was
often changed to masculine imagery during the process of translation (see
Miller and Swift, 1977, Chapter 5; Spender, 1980, pp. 165-171). And until
about AD200, Christian writers portrayed God as androgynous or feminine
more often than as masculine (see Pagels, 1976). The use of masculine
pronouns to refer to God must have some effect on the way that our culture
38 JENNY CHESHIRE

conceptualizes a deity; and the effect will inevitably be greater for children,
as the following "letter to God" from "Sylvia" suggests:
Dear God, Are boys better than girls'? I know you are one but try to be fair (from
Marshall and Hample, 1966)
Children are particularly susceptible, of course, to discriminatory
language. Although the majority of the school books used in Britain still
contain sexist language, some teachers and educators are now pointing out
the social implications to their colleagues and their pupils. The 1980 Special
Issue of the journal Women in Education, for example, provides a checklist
designed to be used by teachers for assessing the language used in school
books. It warns that few books will be free of sexism, but gives suggestions
for overcoming its effects, such as through class discussions or projects
involving writing to publishers. There are also books designed for use in the
classroom that point out the way in which the position of women in society is
reflected in language. Though most of these are primarily concerned with
sexual stereotyping, some also deal with sexism in English (see, for example,
Adams and Laurikietis, 1976, 3, Unit 1: 3).
The changes in the use of English that have been discussed in this section
are relatively minor when seen in isolation, but together they reflect an
increasing awareness of the way in which language often discriminates
against women, and of the need to change the way that we use language, if
we are to change the way that society views women.

The linguistic maintenance of sexual stereotypes

It is sometimes difficult to separate linguistic features that are sexist from

linguistic features that help to maintain sexual stereotypes. For example, the
choice of he as the sex-indefinite pronoun is sexist because it excludes
women, but it also perpetuates the idea that women are of secondary
importance.
Sometimes language is used to make sexual divisions, but simply as one
aspect of a more general sexual discrimination. Several schools, for example,
use gender differences between pupils as a convenient way of dividing the
class. Teachers list girls and boys separately in their registers; they may also
play one gender off against the other, hoping to encourage the class to finish
their work quickly (see Delamont, 1980). Language is involved here, but it is
not the only way that a sexual division is enforced: girls and boys may be
made to sit separately in assembly, and even to enter school by separate
doors (again, see Delamont, 1980). This is not sexual stereotyping, but it
prepares the way for it by encouraging children to strongly identify with their
 LANGUAGE AND SEX IN ENGLISH 39
own sex and to view the opposite sex as completely distinct from themselves.
Although language plays a part in this, it is not the language that needs to be
changed, but the divisive practices of the schools.
Sometimes, however, language is more directly involved in making un-
necessary gender distinctions. Some universities list male students with their
surname and initials (as in J. A. Smith) but list female students with their full
name and marital status (Mrs Jane A. Smith) (see Acker, 1980); and most
make the same distinction in their lists of academic staff.
There are several cases, though, where language clearly reflects stereo-
typed sex-roles. Research in this area has been able to point to those features
of language that need to be changed. For example, some "pairs" of words
like to mother and to father are parallel in form (both are verbs derived from
nouns) but not in meaning. As Lakoff (1973) points out, the phrase "she
mothered the child'' implies a psychological as well as a biological relation-
ship, reflecting the fact that traditionally it is the mother who is responsible
for the upbringing of children, whereas "he fathered the child" implies only
a limited biological act. Some writers have made conscious efforts to dispel
the stereotypes implied by pairs of this kind; Dodson (1975a) and Parke
( 1981) use to father as a semantic parallel to the verb to mother in their books
"How to Father" and "Fathering", and Dodson introduces a neutral term in
a second book entitled "How to Parent,, (Dodson, 1975b).
Dictionary definitions often reveal the existence of sexual stereotypes
and, of course, perpetuate them. The Concise Oxford Dictionary ( 1976), for
example, defines manly in terms of virtues said to be possessed by men:
"having a man's virtues, courage, frankness, etc." But if manly is used to
describe a woman, qualities rather than virtues are involved: "(of woman)
having a man's qualities". Womanly, in contrast, is defined not in terms of
inherent virtues but in terms of unspecified qualities that are considered
suitable for women (one wonders by whom!): "(of woman or her feelings,
conduct, etc.) having or showing the qualities befitting a woman". The
Shorter Oxford English Dictionary (1973), which gives fuller definitions,
adds independence and uprightness to the list of manly virtues, and gives
gentleness and devotion as examples of qualities of women.
One application of the analysis of language and sexual stereotyping has
been to avoid these kinds of definitions in the compilation of the American
Heritage School Dictionary. This uses examples that assign to women virtues
that traditionally have been attributed to men (for example, "she has brains
and courage"). Similarly, characteristics that are usually considered to be
feminine are attributed to men, in sentences such as "tears welled up in his
eyes" or "striving to attain mastery over his emotions". Job stereotyping is
also corrected, by using sentences like "'he teaches kindergarten" and "he
studies typing at night" (see Graham, 1975). The dictionary is designed to be
40 JENNY CHESHIRE

used by schoolchildren, which means that it will play an important role in

educating the next generation towards a less sexually divisive society.
Another result of work in this area has been the publication of The Feminist
English Dictionary (Todasco et al., 1973), which takes material from
established dfotionaries to show how their definitions embody sexual stereo-
types.
Work on sexual stereotyping in children's books has led to attempts to
correct it. Research has shown that the vast majority of children's books
portray stereotyped sex roles; what is more, they do not merely fail to
prepare children for a more egalitarian society, but they even fail to depict
life as it is at present (see Hoffman, 1981). The reading schemes that are
used most commonly in British schools have twice as many male characters
as female characters, and show the male characters taking part in a wider
variety of roles and activities than the female characters (see Lobban, 1974;
1975). Furthermore, the books that are read most widely by children are
frequently the ones that are the most guilty of stereotyping: Little Women,
for example, is among the five books most often read by children over ten
(see Whitehead et al., 1977); and some of the worst offenders have been
specifically recommended by the Schools Council and by literary critics (see
Hoffmann, 1981). School textbooks also perpetuate outdated stereotypes:
junior science books, for example, show experiments conducted by boys,
while girls look on (see Austerfield and Turner, 1972), and textbooks on
other subjects usually portray stereotyped sex roles, or else omit women
entirely (again, see Hoffmann, 1981).
In Britain the Equal Opportunities Commission (1979, 1980) and the
National Council for Civil Liberties (1978) have published practical sugges-
tions for avoiding inadvertent sex discrimination in schools, that include a
discussion of stereotyping in children's books. Some writers of children's
fiction have purposely made girls the central figures of their books (for
example, Lindgren, 1971; Gamer, 1979; Sutcliff, 1967); others portray girls
as tough and daring, as well as compassionate (Bawden, 1967; Kemp, 1977),
and boys as sensitive and caring, but still tough (Southall, 1971; Byars,
1976). Other writers attempt to dispel outdated stereotypes by using as
characters a working mother (Gripe, 1973), a politically active mother
(Avery, 1978), a mother active in the Women's Movement (Perl, 1978), and
even a mother who is a pirate (Mahy, 1972). A useful list of children's books
which are free of sex bias (and also of class and race bias) can be found in
Dixon ( 1977). Collections of fairy stories that avoid sexual stereotyping have
also been published, some in their original form (Phelps, 1981), others
specially written (Williams, 1979) or rewritten (Merseyside Women's
Liberation Movement, n.d.). Lists of books that are free of sexual stereo-
typing are provided by a number of publishers and organizations (for details,
 LANGUAGE AND SEX IN ENGLISH 41
see Hoffmann, 1981; and Women in Education, Special Issue, 1980), and
checklists and guidelines for assessing the sexist content of children's books
are produced by the Centre for Urban Educational Studies, the Equal
Opportunities Commission, the National Union of Teachers and several
other organizations. Discussion material for use in schools is also available,
on fiction (see, for example, Cadogan and Craig, 1976; Whyatt, 1980), on
women's issues (for example, Fys0n and Greenhill, 1979), and also on ways
of avoiding sexism and sexual stereotyping in teaching history, science and
other school subjects (for example, O'Faolain and Martines, 1979; Snail et
al., 1981). Excellent annotated bibliographies and resource lists can be
found, again, in the Special Issue of Women in Education (1980) and also in
Spender and Sarah (1980).
In the USA the problem of sexual stereotyping in education has been
taken seriously for a very long time. Not only have writers and professional
organizations taken steps to correct stereotyping, but government funds
have been set aside for intervention programmes for teachers, both as
in-service and pre-service training courses (for details, see Ekstrom, 1979).
Language also maintains (and exaggerates) sexual stereotypes in advertis-
ing. The advertisement below, used in the early 70s in the USA by Parker
Pens, is a good example:

You might as well give her a gorgeous pen to keep her checkbook unbalanced
with. A sleek and shining pen will make her feel prettier. Which is more
important to any girl than solving mathematical mysteries. (quoted in Komisar,
1971).

Nowadays stereotyping tends to be more subtle, and often results from the
pictures rather than the wording of the advertisements (for examples, see
Goffmann, 1979). Language is still sometimes involved, though, as in the
current advertisements for TWA airlines, which have the words "Fly me"
accompanying photographs of pretty air hostesses. In both the USA and
Britain, stickers can be obtained with the messages "This ad insults women"
or "This exploits women", and these have been used on advertisements in
public places. In Britain the Women's Monitoring Network studies the
representation of women in advertising and in the press generally, and relays
its findings to offending companies.
Advertisements portray male stereotypes as well as female stereotypes, of
course. In the USA some advertising companies are beginning to reverse
traditional stereotypes in an attempt to correct them (for examples, see
Komisar, 1971). In Britain a few advertisers appear to be aware of the issue:
a recent television commercial for Sony, for example, plays on stereotyping
by saying that since women may have been offended by their commercials,
this one will deal only with the technical details that women are interested in.
42 JENNY CHESHIRE

And a current television commercial for British Rail makes a half-hearted

attempt to avoid stereotyping by addressing "Business men and business
ladies" - half-hearted because the choice of the word "ladies" rather than
"women" gives the phrase an ironic air (for discussion of the different
connotations of lady and woman, see Lakoff, 1973). These advertisements,
of course, reinforce traditional stereotypes by making fun of the issue, but
they may nevertheless attract the attention of people who had not thought
about stereotyping before. And, as Komisar (1971) suggests, the images of
men and women that appear in advertisements are so exaggerated and
ludicrous that they may influence people to stop acting out sexual stereo-
types.
Sex-role stereotyping also exists in job descriptions. Job titles have been
officially revised, first in the USA, and then in Britain to comply with the Sex
Discrimination Act of 1975: air hostesses, to give one example, are now
officially referred to as flight attendants. Changing the titles of jobs will help
to avoid sex-role stereotyping, but the language that is used in job descrip-
tions needs to be changed also. The third person singular pronoun often
indicates which sex employers have in mind: a recent advertisement in The
Guardian (I8th December, 1981), for instance, read "Our clients ... are
looking for a Secretary who can use her initiative". Careers Guides often
perpetuate stereotyping in the same way: the "Daily Telegraph Careers
A-Z'' guide, which is aimed at school leavers, writes under "Accountant":
""he may work in public practice" (p. 7).
Some books designed to counteract sexual stereotyping in schools contain
examples of this, and suggestions for discussion. Adams and Laurikietis
( 1976, I) for example, has a useful chapter on "Your choice of career". At
least one Careers guide makes a conscious effort to dispel sexual stereo-
typing in career possibilities: "'Equal Opportunities: a careers guide" (Miller,
1978) uses the feminine third person singular pronoun throughout, and
explains in the introduction that its aim is to "chivvy girls off the tramlines"
and .. encourage girls to widen their career choice and speed their progress
towards equality in the job market".
Language, as a social phenomenon, inevitably reflects social attitudes
towards women and men. But it also influences and to some extent moulds
the views of its speakers, as Kress and Hodge (1979) point out. This means
that the changes in language use that have been outlined in this section
should lead to changes in the way that society treats men and women, which,
in turn, will lead to further changes in language. Language change and social
change, in other words, are mutually reinforcing.
 LANGUAGE AND SEX IN ENGLISH 43

The formal analysis of linguistic differences

in male and female speech

Sociolinguistic surveys of American and British English usually include the

sex of speakers as a sociological variable. Where linguistic features can occur
in both a standard English and a non-standard form, the surveys have found
that female speakers tend to use more of the standard English forms than
male speakers. This is true for both phonological and morphosyntactic
features (see, for example, Labov, 1966; Trudgill, 1974b). One reason may
be that women are more conscious of the social significance of different
linguistic features, so that they use more of the socially prestigious speech
forms; another reason is, perhaps, that non-standard working-class speech
has masculine connotations of "roughness" and "toughness" in Western
society, so that men choose to use more non-standard forms (see Trudgill,
1974a). Conklin (1974) suggests that women are simply more sensitive to the
constraints of different social contexts, so that they are less likely than men
to use their most relaxed speech style (where the maximum use of non-
standard forms would occur) when talking to a linguistic investigator, par-
ticularly if, as has usually been the case, the investigator is male. In fact,
some recent studies where the investigator was female found that differences
in the use of non-standard forms depended not only on the sex of speakers
but also on the degree to which they were integrated into the local culture
(see Milroy, 1980; Cheshire, 1982a). Trudgill (1972) found that covert
prestige is attached to non-standard features by younger working-class
speakers of both sexes, and suggested that this reflects adherence to a
subculture that is distinct from the mainstream value system of our society.
The analysis of sex differences in language has been useful within
linguistics, by helping to explain some of the social mechanisms that are
involved in language change (see, for example, Trudgill, 1972; Milroy,
1980). But it also has important implications for education. Children who
speak non-standard English are at a disadvantage at school, because their
variety of English is not the same as the variety used by the teachers and in
school reading books. They have to choose, perhaps at an unconscious level,
whether to use the standard English forms that are linked with mainstream
culture and the school, or the non-standard English forms that symbolize
solidarity with the vernacular culture. It is sometimes thought that boys are
more likely than girls to reject standard English and the values of the school,
because "female" values and female teachers predominate there (see, for
example, Shuy, 1969). In fact, however, girls who are integrated into the
peer group vernacular culture are equally likely to reject the language and
44 JENNY CHESHIRE

the values of the school (see Cheshire, 1982b), and any attempt at teaching
standard English should take this into consideration.
Research has also shown that some linguistic features function differently
for female and male speakers (see Milroy, 1980; Cheshire, 1982a). One
example of this is the non-standard past tense verb form come, which occurs
in sentences such as .. I come here last night". Adolescent boys in the town of
Reading, in Berkshire, invariably use this non-standard form in both formal
and informal speech; adolescent girls, on the other hand, use the form
intermittently, varying it with the standard English past tense form came.
We would expect, therefore, that boys will have more difficulty in replacing
the non-standard form with the standard form, in their school writing. This
kind of differentiation, then, also needs to be given careful consideration by
educationists and teachers.
Some researchers have carried the analysis of sex differences in language a
stage further, by looking at female and male roles in conversations. It has
been suggested, for example, that in conversations between men and
women, it is women who initiate the conversation and encourage men to
speak, while men control the topics. It has also been claimed that men
interrupt more often than women (see Zimmerman and West, 1975; Acker,
1980). As yet there is no valid empirical confirmation, but if these dif-
ferences do exist, the knowledge could have some useful social applications,
perhaps in counselling or therapy, or even, as Kramer et al. {1978) suggest,
as an unobtrusive measure of sexual inequality.

Evaluation of speech

Sex differences in the evaluation of speech have been analysed from the
point of view of the hearer (in other words, analysing the way in which men
and women evaluate speech) and from the point of view of the speaker
(focussing on the way in which judges evaluate men's speech and women's
speech). Elyan et al. (1978) found clear differences in the way that women
and men evaluate speech. Women rated speakers who used Received
Pronunciation more highly than men, in terms of status, intelligence,
independence and egoism; and they gave lower ratings than men to speakers
with regional accents. These findings have implications for education, for
they imply that female teachers might form different stereotyped views of
their pupils than male teachers. The way that children speak is thought to
affect teachers' evaluation of their academic potential (see Seligman,
Tucker and Lambert, 1972) and this, in tum, can affect academic success
(Rosenthal and Jacobson, 1968). There could be serious consequences in
 LANGUAGE AND SEX IN ENGLISH 45
other areas, too: in trial by jury, for example, in job interviews, and in
politics.
As Smith (1979) points out, it is difficult to analyse how women's and
men's speech is evaluated, because judges may be reacting to the sex of
speakers, rather than to their language. It has been suggested that women's
speech is characterized by a number of linguistic features, such as a more
frequent use of tag questions, fillers, intensifiers and "empty" adjectives (for
example, divine or lovely) (see Lakoff, 1973), but this has not been con-
firmed by empirical studies. It is now thought that the linguistic charac-
teristics described by Lakoff typify a more general "powerless" variety of
speech, used by speakers who have little social prestige (see Lind and
O'Barr, 1979), and, therefore, associated more often with women than with
men. Lind and O'Barr set up simulated legal hearings, and found that jurors
of both sexes rated witnesses using the "powerless" variety of speech as less
competent, trustworthy, convincing, socially attractive and socially dynamic
than witnesses who did not use this variety, irrespective of their sex.
Although these findings do not bear directly on sex differences in language
use, they do show that using certain linguistic features can affect the way in
which speakers of both sexes are evaluated, and are relevant in all areas
where language plays a central role. Those speakers who have low social
prestige may reveal this in their speech, and be evaluated negatively in job
interviews, legal proceedings and other important social situations. Thus,
although the initial impetus for this research came from an interest in sexual
inequality, the results have wider implications that could affect all sections
of society that are treated unequally.

Conclusion

Changes that take place in society are reflected in language, though language
change tends to lag behind social change. Sex roles in the USA and in Britain
have been changing during the course of this century, and we would
normally expect these social changes to be accompanied by gradual changes
in language. We have seen, however, that research into the relationship
between language and sex has led to conscious attempts to change dis-
criminatory and stereotyping language. These practical applications are
accelerating the rate at which language is changing and this should, in tum,
accelerate the rate at which society is changing.
We have also seen that research in this area has implications for education
and for legal proceedings and job interviews. It has potential applications for
many other areas also: in the field of language pathology, for example,
46 JENNY CHESHIRE

where it seems that sex-role stereotyping could account for some language
disorders in male speakers (see Kramer, 1974).
The main application of work on the relationship between language and
sex has, of course, been in attempts to change our use of discriminatory
language, in order to remove sexual inequality from society. It has wider
applications also, however, that have been only briefly mentioned here: it
helps our general understanding of the way in which language reflects and
maintains social divisions, and of the way in which our thinking is often
unconsciously moulded by our language. An understanding of this will help
to eliminate not only sexual inequality, but inequality in all areas of social
life.

References
Acker, S. (1980). Women, the other academics. British Journal of Sociology of
Education, 1, (1), 81-91.
Adams, C. and Laurikietis, R. (1976). "The gender trap: a closer look at sex roles",
Book 1: Education and work. Book 2: Sex and Marriage. Book 3: Messages and
Images. Virago, London.
Aitchison, J. (1981). ''Language change: progress or decay?", Fontana, London.
Arnold J. (1973). "The Cook and the Carpenter", Daughters, Houston, Texas.
Austerfield, V. and Turner, J. (1972). What are little girls made of! Spare Rib,
3 September.
Avery, G. (1978). "Huck and hertime machine'', Armada, London.
Bawden, N. (1967). "A handful of thieves", Gollancz, London.
Bodine, A. (1975). Androcentrism in prescriptive grammar: singular 'they', sex-
indefinite 'he', and 'he or she'. Language in Society, 4 (1), 129-146.
Byars, B. (1976). "The Midnight Fox", Penguin, Harmondsworth.
Cadogan, M. and Craig, P. (1976). "You're a brick, Angela: a new look at girls'
fiction 1839-1975", Gollancz, London.
Cheshire, J. (1982a). "Variation in an English dialect: a sociolinguistic study",
Cambridge University Press, London.
Cheshire, J. (1982b). Dialect features and linguistic conflict in schools. Educational
Review, 34 (1), 53-67.
Conklin, N. (1974). Toward a feminist analysis of linguistic behaviour. University of
Michigan Papers in Womens Studies, 1(1),51-73.
"Daily Telegraph Careers A-Z", (1981). William Collins, London.
Delamont, S. (1980). "The Sociology of Women", George Allen and Unwin,
London.
Dixon, B. (1977). "Catching them young: sex, race and class in children's fiction",
(Vol. 1), Pluto Press, London.
Dodson, F. (1975a). "How to parent", Star books, London.
Dodson, F. (1975b). "How to father", New American Library, New York.
Ekstrom, R. B. (1979). Inteivention strategies to reduce sex·role stereotyping in
education. In "Sex·role stereotyping: collected papers", (0. Hartnett, G.
Boden and M. Fuller, eds), Tavistock Publications, London.
 LANGUAGE AND SEX IN ENGLISH 47
Elyan, 0. et al. (1978). R.P. accented female speech: the voice of perceived
androgyny? In "Sociolinguistic patterns in British English", (P. Trudgill, ed.),
Edward Arnold, London.
Equal Opportunities Commission, {1980). "Ending sex-stereotyping in schools".
Equal Opportunities Commission, (1979). "Do you provide equal educational
opportunities?"
Fyson, N. L. and Greenhill, S. (1979). •'People talking", Macmillan Educational,
London.
Garner, A. (1979). "The Stone Book", Armada, London.
Goffman, E. (1979). "Gender advertisements", Harvard University Press, Cambridge,
Mass. __
Goot, M. and Reid, E. {1975). "Women and voting studies", Sage, London.
Graham, A. (1975). The making of a nonsexist dictionary. In "Language and sex:
difference and dominance", (B. Thorne and N. Henley, eds), Newbury House
Publishers, Rowley, Mass.
Gripe, M. {1973). "The Night Daddy", Chatto and Windus, London.
Guttenag, M. and Bray, H. (1976). "Undoing sex-stereotypes: research and resources
for educators", McGraw Hill, New York.
Harrison, L. {1975). Cro-magnon woman - in eclipse. The Science Teacher, April
1975, pp. 8-11.
Hoffman, M. {1981). Children's reading and social values. In .. Language in School
and Community", (N. Mercer, ed.), Edward Arnold, London.
Kemp, G. (1977). "The turbulenttenn of Tyke Tiler", Faber, London.
Kidd, V. ( 1971 ). A study of the images produced through the use of the male pronoun
as the generic. Moments in contemporary rhetoric and communication, I (1),
25-30.
Killian, R. A. (1979). "Managers must lead!", AMACOM, USA.
Komisar, L. {1971). The image of woman in advertising. In "Woman in sexist
society", (V. Gornick and B. K. Moran eds), Basic Books, New York.
Kramer, C. {1974). Women's speech: separate but unequal? Quarterly Journal of
Speech, 60 {February), 14-24. Reprinted in Thome and Henley, 1975.
Kramer, C., Thorne B. and Henley, N. {1978). Perspectives on language and
communication. Signs: Journal of women in culture and society, 3 (3), 638-651.
Kress, G. and Hodge, R. (1979). "'Language as ideology", Routledge and Kegan
Paul, London.
Labov, W. (1966). "The social stratification of English in New York City", Center for
Applied Linguistics, Washington, D.C.
Lakoff, R. (1973). Language and woman's place. Lanugage in Society, 2 (1), 45-79.
Leach, P. (1977) ... Baby and Child", Michael Joseph, London.
Lind, E. A. and O'Barr, W. M. {1979). The social significance of speech in the
courtroom. In •'Language and social psychology", (H. Giles and R. N. St. Clair,
eds), Blackwell, Oxford.
Lindgren, A. (1971). HPippi Longstocking'', Oxford University Press, Oxford.
Lobban, G. (1974). Presentation of sex-roles in reading schemes. Forum for the
discussion of new trends in Education, 16 (2), Spring, pp. 57-60.
Lobban, G. (1975). Sex-roles in reading schemes. Educational Review, 27 (3), 202-9.
Mahy, M. (1972). "The Man whose mother was a pirate", Dent, London
Marshall, E. and Hample, S. (1966) ... Children's letters to God", Pocket Books,
New York.
Martyna, W. ( 1978). What does "he" mean? Use of the generic masculine. Journal of
communication, 28 (1), 131-138.
48 JENNY CHESHIRE

Mackay, D. G. (1979). On the goals, principles and procedures for prescriptive

grammar: Singular they. Language and Society 9, 349-367.
Mackay, D. G. and Fulkerson, D. (1979). On the comprehension and production of
pronouns. Journal of verbal learning and verbal behaviour.
Merseyside Women's Liberation Movement. (n.d.). "Once and Future Tales", (Snow
White, Red Riding Hood, The Prince and the Swineherd). Available from
53, Sandown Road, Liverpool.
Miller, R. (1978). "Equal opportunities: a careers guide", Penguin, Harmondsworth.
Miller, C. and Swift, K. (1977). "Words and Women", Gollancz, London.
Miller, C. and Swift, K. (1981). "The handbook of non-sexist writing for writers,
editors and speakers", The Women's Press, London. ·
Milroy, L. (1980). Language and social networks. Blackwell, Oxford.
Moulton, J., Robinson, G. M. and Elias, C. (1978). Sex bias in language use:
"neutral" pronouns that aren't. American Psychologist, 33 (11), 1032-1036.
National Council for Civil Liberties. (1978). "Sex discrimination in schools: how to
fight it".
Nilsen, A. P. (1973). Grammatical gender and its relationship to the equal treatment
of males and females in children's books. Unpublished Ph.D. Thesis, University
of Iowa.
O'Faolain, J. and Martines, L. (1979). "Not in God's image: women in history",
Virago, London.
Pagels, E. H. (1976). What became of God the mother?: conflicting images of God in
early Christianity. Signs: Journal of women in culture and society, 2 (2), 293-303.
Parke, R. (1981). "Fathering", Fontana, London.
Perl, L. (1978). "That crazy April", Armada, London.
Phelps, E. J. (1981). "The maid of the North and other folk tales' heroines", Holt,
Rinehart and Winston, New York.
Piercy, M. (1979). "Woman on the edge of time", The Women's Press, London.
Roberts, H. (1981). "Doing feminist research", Routledge and Kegan Paul, London.
Rosenthal, R. and Jacobson, L. (1968). "Pygmalion in the classroom", Holt,
Rinehart and Winston, New York.
Salk, L. (1974). "Preparing for parenthood", David McKay, New York.
Schneider J. W. and Hacker, S. L. {1973). Sex role imagery and the use of the generic
'man' in introductory texts: a case in the sociology of sociology. The American
Sociologist, 8 (1), 12-18.
Seligman, C.R., Tucker, G. R. and Lambert, W. E. (1972). The effects of speech
style and other attributes on teachers' attitudes towards pupils. Language in
Society, 1(1),131-142.
Shuy, R. (1969). Sociolinguistic research at the Center for Applied Linguistics: the
correlation of language and sex. Giornata internazionali di sociolinguistica,
Palazzo Baldassini, Rome.
Smith, P. (1979). Sex markers in speech. Jn "Social markers in speech'', (K. P.
Scherer and H. Giles, eds), Cambridge University Press, Cambridge.
Snail, M., Kelly, A. and Whyte, J. (1980). "Girls into science and technology",
available from 91 Didsbury Park, Manchester 20.
Southall, I. (1971). "Josh", Penguin, Harmondsworth.
Spender, D. (1980). "Man made language", Routledge and Kegan Paul, London.
Spender, D. (forthcoming). "Men's studies modified: the impact of feminism on the
academic disciplines", Pergamon, Oxford.
Spender, D. and Sarah, E. (eds) (1980) ...Learning to lose: sexism and education'',
The Women's Press, London.
 LANGUAGE AND SEX IN ENGLISH 49
Sutcliff, R. (1967). "The chiefs daughter", Hamish Hamilton, London.
Thorne, B. and Henley, N. (eds) (1975). "Language and sex: difference and
dominance", Newbury House Publishers, Rowley, Mass.
Todasco et al. (1973). "An intelligent woman's guide to dirty words: English words
and phrases reflecting sexist attitudes towards women in patriarchal society,
arranged according to usage and idea", Vol. I of "The Feminist English
Dictionary", Feminist Writers Workshop, Loop Center YWCA, Chicago.
Trudgill, P. (1972). Sex, covert prestige and linguistic change in the urban British
English of Norwich. Language in Society, 1, (1), 179-195.
Trudgill, P. (1974a). "Sociolinguistics'', Penguin, Harmondsworth.
Trudgill, P. (1974b). "The social differentiation of English in Norwich'', Cambridge
University Press, London.
Whitehead F. et al. (1977). "Children and their books", Macmillan Education,
London, for the Schools Council.
Whyatt, B. (1980). "Myths and legends for young feminists, and for old feminists the
original version", available from B. Whyatt, 90, Plane Avenue, Wigan, Lanes.
WN59PT.
Williams, J. (1979). "The Practical princess and other liberating fairy tales", Chatto
and Windus, London.
"Women in Education", Special Issue 1980. Autumn, no. 20: Women and Education
Newsletter Group: 14, St. Brendan's Road, Withington, Manchester20.
Zimmerman, D. H. and West, C. (1975). Sex roles, interruptions and silences in
conversations. In "Language and sex: difference and dominance", (B. Thorne
and N. Henley, eds), Newbury House Publishers, Rowley Mass.

1
In the 1979 and 1982 editions, however, Leach reverts to using he to refer to babies
that may be either male or female, on the grounds that most readers prefer this. She
remains in the text of most charts on the book, and in the captions to illustrations.
 4
Sociolinguistic methodology
and the identification of speakers' voices
in legal proceedings 1

James Milroy

The work that has so far been carried out on linguistics and the law has
been, broadly speaking, of two kinds. The first kind has been concerned with
the difficulties experienced by the ordinary citizen in understanding the
language used by lawyers and judges in court and is represented by the work
of O'Barr (see O'Barr, 1981 and references cited therein). Active research
into this field has been carried out mainly in the USA (e.g. in the Law and
Language Project of Duke University directed by O'Barr) and may have
been largely motivated by the acknowledged cultural pluralism of that
country and the problems of minority linguistic groups. In Britain there does
not seem to have been as much active research into this particular area,
although similar problems of cultural and linguistic diversity are now present
(see e.g. Furnborough et al., 1982).
Research into the language attitudes of the professions (law, medicine,
dentistry, civil service, journalism) would, however, be of the greatest
interest. In some professions there is an ancient tradition of linguistic
obscurity that can be justified on various grounds, usually those of technical
precision and clarity. Legal proceedings, for instance, were conducted in
French until after 1362, even though the majority of the people of England
had never been native speakers of French. Writers on medicine before 1600
who dared to write in English rather than Latin still had to preface their work

APPLIED SOCIOLINGUJSTJCS Copyright© l984 by Academic~. London

ISBN: 0-12-701220-6 All rights of reproduction in any form reserved
52 JAMES MILROY

with an apologia, as it was thought that forbidden knowledge of private

bodily functions should not be imparted to the lewd and vulgar populace. A
famous work on paediatrics (Thomas Phaire's The Boke of Chy/dren, 1545)
strongly and memorably attacks the Latinism of medicine:
... how long wold they haue the people ignorant? why grutche they phisyke to
come forth in Englyshe? Woulde they haue no man to know but onely they?
(ed. Neale and Wallis, 1955)

In subsequent centuries, the professions have continued to build up an

extensive vocabulary based on Greek and Latin that is not immediately
accessible to the layman. This can be seen as part of the functional elabora-
tion in vocabulary involved in the long-term process of language standard-
ization. It is, however, clear, in the law at least, that some difficulties may
arise from this historical fact: O'Barr recommends that a training in language
matters should be part of the curriculum of law-schools, and Gumperz
( 1976) reports that efforts have been made in the USA to implement this
kind of recommendation.
The second area in which linguistics has been relevant to the law is in the
identification of voices on tape-recordings. In the USA, controversy has
raged on the matter of "voice-prints", i.e. spectrograms or other visually
interpretable print-outs of a person's speech. It has been claimed that
"voice-prints" are as individually identifiable as finger-prints (indeed the use
of the term "voice-print" instead of "spectrogram" is intended to imply
this). Without wishing to deny that phonetic instruments may be of value in
forensic matters, most linguistics and phoneticians are very sceptical of the
strongest claims that have been made for "voice-prints", and generally feel
that such claims take advantage of the wishful thinking and credulity of those
who are not experts in phonetics and linguistics. One expert, Peter Ladefoged,
has appeared in a large number of court cases and has examined thousands
of spectrograms: he states (1975, p. 187) that claims for the accuracy of
voice-prints are greatly exaggerated. One reason for this is that (as Bolinger,
1980, p. 112, points out) .. voiceprint identification is not refined enough to
establish guilt" (although it may help to establish innocence): indeed visual
print-outs merely substitute visual expertise for aural expertise on the part of
the analyst. The phonetician is likely to be wrong part of the time whether he
uses visual or aural techniques, and there is at present no obvious way of
determining absolutely whether instrumental methods of voice-identification
constitute an advance on the refined ears of phoneticians. A second reason,
which is partly connected with the first, is ultimately statistical. No one can
define a finite population from which the data-sample to be analysed is
drawn; therefore, there is no way of knowing how many individuals have
similar vocal characteristics. If "voice-prints" were as reliable as finger-
 SOCIOLINGUISTIC METHODOLOGY 53
prints, this might not be a serious problem. As it is, this is far from the case.
One prominent forensic phonetician (Hollien, as reported in Bolinger,
1980, p. 112) claims that voiceprinting is "a fraud being perpetrated upon
the American public and the courts of the United States".
It has been necessary to consider voice-printing briefly, even though the
court-case that forms the basis of this chapter made no use of spectrographic
evidence. The claims made in favour of voice-prints exemplify very clearly
the natural human desire to be absolutely certain about matters in which
certainty may not always be possible. They also demonstrate a tendency to
be naively optimistic about advances in instrumental techniques. The most
important point for the linguist, however, is that these claims, together with
more impressionistic claims made by witnesses in court, demonstrate the
general ignorance (shared by the public and some experts alike) of the
complexity of language. To consider voice-quality alone, it is known that the
voice-quality of an individual can vary from time to time for a number of
reasons, from the effect of emotional or sexual arousal, throat-infections,
hangovers and menstrual periods, for example (see, in general, Laver,
1980). This takes no account of a host of other ways in which an individual's
speech can vary. But this general gullibility seems to have the effect of
motivating people to try to establish guilt rather than to prove innocence (or
at least fail to establish guilt). In the case to be discussed, the linguistic
evidence fails to establish guilt; yet, the accused was convicted on the
unsubstantiated opinions of witnesses on voice-identification (the expert
evidence that was offered was set aside). The procedures used by the police
and in court in this case may also seem rather surprising to the reader: the
law may well reply that these procedures were perfectly correct in law; yet, it
can also be said that in this case certain evidence taken into account should
have been ruled inadmissible for reasons known to field linguists and
phoneticians, but not (apparently) to the law.

Dialect variation
The case to be discussed (Regina vs Mullan, tried in Belfast in 1980)
depended on voice identification by witnesses. Positive identification was
taken to suggest guilt, and the jury's decision (by a majority verdict) that the
accused was guilty was based mainly on the fact that several witnesses had
sworn under oath that the voice (or voices) heard on several tape-recordings
was/were one and the same voice, and that the voice was that of the accused.
Ladefoged writes (1975: 189), on the basis of his forensic experience:
it is completely irresponsible to say, as I have heard witnesses testify in court,
54 JAMES MILROY

"The voice on the recording is that of the accused and could be that of no other
speaker".
We cannot know how many other speakers in the population may have
voices so similar to that of an accused person that they sound virtually
identical on rather poor tape-recordings.
For the linguist, however, the Mullan case does not involve only the
identification of voices by differentiating voice-qualities: it raises quite
clearly the question of dialect and accent differences between speakers.
Expert dialect evidence has been in the news in recent years, especially in
the notorious "Yorkshire Ripper" case. At one stage, the voice on a tape
sent to the police was identified by the dialectologist, Stanley Ellis, as being
from the Sunderland area. The fact that this tape turned out to be a hoax
should not blind the police and public to the potential value of expert
evidence in such cases.
It is common knowledge that there is considerable dialectal diversity
within the British Isles, a much greater diversity than within the United
States. There can often be some clear differences between the speech of two
places that are only a few miles apart, and of course gross and obvious
differences over a distance of forty miles or so (compare Liverpool with
Manchester, Leeds with Teesside). It is obvious that in a legal case where
two accents are grossly different (say, a suspect's London accent and a
Manchester accent on a tape-recording), the police would not normally
pursue the case, as they would know that no one could be persuaded to
believe that the two accents were those of the same speaker. However, when
two accents sound broadly similar to the layman (but showing consistent
differences that can be specified by the expert), it may be that some people
do not notice the differences. On the other hand, they may be aware of
differences in a vague way. In the Mullan case we are dealing with two
different accents, and the case raises the rather interesting issue that people
may be aware of an accent different (defence counsel seems to have been
well aware of it), but may be prepared in court to dismiss this possibly
subliminal awareness for various reasons.

The MuUan Case

Seamus Mullan, of Garvagh, Co. Londonderry in Northern Ireland, was

sent for trial in the summer of 1980 accused of blackmail. He was convicted
and sentenced to ten years in prison. The circumstances leading to his
conviction read rather oddly, but the conviction depended on two vital
pieces of evidence:
(1) that the voice of Mullan was the same as that of an armed and masked
 SOCIOLINGUISTIC METHODOLOGY 55
man who broke into the home of the Barton family of Kilrea (near
Garvagh) on the evening of 7 September 1979, that the masked man's
voice was the same as that of Mullan overheard being questioned in a
police station, and that Mullan's voice (and that of the masked man)
was/were also the same as that of a telephone caller or callers (tape-
recorded at the Barton's home, later in September) who made five
threatening phone-calls demanding a large amount of money (£45 OOO)
from the Barton family:
(2) that, apart from evidence on voice identification, Mullan was the man
who made the fifth telephone call, as he was alleged to have been
observed apparently making a call from a telephone-box in Draperstown,
Co. Londonderry, at the same time as the fifth threatening phone-call
was received by the Barton family {about 3 p.m. on 21September1979).
It is not clear how Mullan became a suspect in the first place. Although the
case was not defined as a terrorist case (it was tried by an ordinary jury, as
terrorist intimidation of juries and witnesses was not considered to be an
issue), there were sectarian overtones. Repeatedly, according to the tran-
script and in the tape-recordings, the Barton family are told that they have
been singled out for this treatment because they are Catholics and that
co-religionists of theirs have been held responsible for killing members of
the security forces. The odd thing about this is that Mullan is also a Catholic.
However this may be, the Barton family reported to the police the
intrusion by the masked man. On 23 September (after the threatening
telephone-calls had also been received), members of the family were invited
to attend the police station, where they overheard a man being questioned
on a different matter in the next room. The man was Mullan, and without
seeing him, Mrs. and Miss Barton in response to the policeman's question
identified his voice as that of the masked man. At the same time, the police
had made a tape-recording of their interrogation of Mullan, and this
recording was subsequently made available for linguistic analysis.
Given that there have been severe misgivings about visual identification of
suspects (even when they are presented in an identification parade), it is odd
that evidence on voice identification of the above kind can be admitted. One
might suppose that visual identification is actually easier than voice identi-
fication, as many people have voices of similar timbre. The least that one
might expect in such circumstances is that an array of similar voices should
be presented and that witnesses should be asked to pick out the voice of the
miscreant from that array. Clearly, two distressed people might have been
predisposed to identify Mullan's voice as that of the masked man. They
would be anxious to help in apprehending the criminal. Unfortunately for
the defence, Mrs Barton was the first witness called at the trial. She swore
under oath that the accused's voice was the same as that of the masked man.
56 JAMES MILROY

Although the defence barrister (R. A. Ferguson, Q.C.) suggested that she
might have been mistaken, he could not go too far in his cross-examination,
as the witness was very distressed. A strong cross-examination might well
have increased the jury's sympathy for the witness and alienated any
possible sympathy for the accused.
Miss Maureen Barton was then questioned, and her examination included
details of the five threatening telephone calls which were received between
18 and 21 September, more than ten days after the intrusion by the masked
man. The important point to note here is that Miss Barton testified that the
voice on call 1 was the same as the masked man and that all five telephone
calls were by the same person {the masked man). With reference to the first
call, the court transcript reads thus:
Q. Did you recognize the voice in that?
A. After a very, very short conversation with the telephone caller I
immediately recognized his voice as the man who had come into
our kitchen armed and masked. I have no doubt in my mind that
it was exactly the same voice, exactly.
Miss Barton was also certain that the voice on call 2 was the same as the voice
on call 1.
It was exactly the same voice as the voice of the caller who had rung me four
hours earlier, exactly the same voice. I have no doubt in my mind, the fear and
the terror that that voice put into me.
Later, prosecuting counsel asked the witness what she would say if it were
suggested to her that the phone calls were by more than one person.
Although court transcripts do not indicate changes in voice quality, it is clear
that Miss Barton's response was highly emotional.
Listen the voice that was in our kitchen and the voice on those telephone calls is
from the same person. That voice it has haunted me. It does haunt me and I
know that for the rest of my life I will never forget that voice. The fear and the
terror I will never forget it. It was the same voice.
The tape recordings, when analysed, indicate that Miss Barton is mis-
taken in certain important particulars: the speaker on call 2 can hardly be the
same as the speaker on call 1 (see further below). It is obvious, however, that
it would have been extremely difficult for the defence to cast much doubt on
her evidence without alienating a jury's sympathy. In fact, the defence
believed that they had a trump card. Hoping, no doubt, that expert voice
identification might help to establish guilt, the police had called in Dr J. R.
Baldwin of the Department of Phonetics and Linguistics, University College
London. Dr Baldwin had analysed the tape-recording of Mullan in the
police station and the recordings of the telephone calls. He was prepared to
 SOCIOLINGUISTIC METHODOLOGY 57
testify that the voice on calls 1, 3, 4 and 5 differed from the one in call 2, and
that only the caller in call 2 had a voice similar to Mullan 's. In other words,
Baldwin's evidence tended to establish innocence, and not guilt. The police
made this evidence available to the defence, who hoped that it would be
sufficient to cast doubt on Mullan's guilt, as indeed it is.
Before Baldwin was called, however, many witnesses had testified that
there was only one telephone caller and that his voice was that of Mullan.
Detective Inspector Houston, for example, had listened to the two calls
recorded on 18 September. In answer to the question "Did you recognize
the voice of anyone on that tape?", the officer replied that he recognized
the voice of Mullan. Furthermore, the prosecution then established that
Detective Inspector Houston was already familiar with that voice: thus, it
was implied that his opinion was reliable. Other witnesses replied that
Mullan 's voice and that of the caller were "one and the same", or in words to
that effect.
There is another peculiar circumstance, to which I have already referred.
At the time of phone call no. 5 (about 3 p.m. on 21 September), Mullan
had been observed apparently making a call from a telephone kiosk in
Draperstown. A Mr Carron, who was a soldier in plain clothes, had been
deputed by the police to observe Mullan 's movements on that day and
testified to this effect. Oddly enough, the voice on call no. 5 appears to be the
same as the voice on calls 1, 3 and 4, and that is the voice that cannot
reasonably be shown to be that of Mullan (as we shall see).
The police had in fact consulted two expert witnesses, Dr Baldwin, and Dr
Peter Roach of Leeds University. For some reason, Dr Roach was not called
as a witness, but it is difficult to believe that his opinion could have been
substantially different from that of Baldwin. The latter testified that only the
voice on call no. 2 was similar to Mullan's, and that he was not sure that even
that voice was Mullan's, In cross-questioning, the prosecution managed to
suggest that Baldwin's evidence was not as reliable as those of the other
witnesses, and the jury apparently gave it little weight in reaching their
verdict.
There are two reasons why it is a very serious matter that Mullan should
have been convicted on evidence that is based solely on the unsubstantiated
opinions of witnesses on voice identification. First, it is not as easy as
the layman believes to identify voices on the telephone or on poor tape-
recordings. Second, apart altogether from differences of voice-quality,
there may be accent differences between speakers. These differences are
systematic and not random. In this case, it is odd that the witnesses had not
noticed that there were two different accents on the telephone calls: it is of
course possible that they had noticed subliminally, but had underestimated
the importance of this difference.
58 JAMES MILROY

Identifying voices on the telephone and on recordings

Any field-linguist who has collected and analysed large amounts of con-
versational talk from different people (as in our own research reported in
Milroy, 1978 etc.) will know that the limitations of human memory are such
that it is difficult to identify who is speaking at any given point. People who
are all from the same area and social group speak in a similar accent and also
have broad similarities of voice quality and intonation. For instance, the
habitual posture of the larynx is similar in a local population and varies in a
general way from region to region: some accents have "raised larynx" voice,
others may have a lowered larynx. Some accents (e.g. Belfast) have nasal-
ization, others (e.g. Liverpool) are denasalized (see, e.g. Wells: 1982,
pp. 91-3).
In sociolinguistic fieldwork and analytic practice (in which 50-100 voices,
or more, are involved}, it is important for the linguist to listen to his tapes
immediately after they have been collected and make a note of who is
speaking at different points on the recordings, solely because it is impossible
to distinguish reliably on voice quality alone. In our own research (J. and L.
Milroy, 1978; L. Milroy, 1980; J. Milroy, 1981), we have followed this
practice, as (by definition) all speakers, being from the same area and of the
same social group, have broadly similar accents. Part of the point of the
research was to show that within this homogeneous population, quite subtle
but consistent differences of accent could nevertheless be demonstrated. In
prior identification of speakers, however, we could not quantify accent
differences, and we could not rely on voice quality (except in differentiating
male from female, old from young), as we knew that speakers of similar age
and the same sex would usually have similar voice-qualities. When (as was
usual} five or six voices were recorded on the same tape, we relied, not on
voice-quality, but on short-term memory of the field-work itself, memory of
the content of the conversations and the views expressed by the different
speakers. If we had delayed in this work of identification, long-term memory
of who was speaking at different points would have become less reliable, and
memory of differing voices would only have helped in rare cases of extreme
and obvious differences (the lady who had no teeth, for instance, was always
easily identifiable).
In any situation in which visual cues are absent, listeners rely on a large
number of factors besides voice-quality. In particular, listeners may identify
a telephone caller by their knowledge of the kind of business that the caller is
engaged in, by memory of topics of conversation enjoyed by the caller and
by shared knowledge in general. In practice, however, telephone callers
usually identify themselves immediately. There are good reasons why a
 SOCIOLINGUISTIC METHODOLOGY 59
ritual of identification is customary in telephoning. The visual cues by which
we normally identify people are absent, and many people share similarities
of voice quality and accent. Furthermore, there may be distortion on the
line, and many distinctive ac~ustic features of voice are cut out on the
telephone. A person's voice does not sound the same on the telephone as it
does in real life. Thus, it is normal for the caller to identify himself/herself
immediately. If, for any reason he does not do so, it is quite likely that the
listener will enquire as to his/her identity with some question such as "Is that
George speaking?" If it were really a simply matter to identify a telephone
caller, there would be no need for these ..opening sequences".
However, as Brown (1980) points out, listeners are conditioned by their
expectations as to who is likely to be telephoning much more than by
idiosyncratic differences of voice. When the telephone rings, the caller may
literally be anyone in the world, but everyone knows that there are greater
or lesser probabilities as to who is calling. Thus, if the telephone rings in my
office there is an immediate presumption that the caller is telephoning on
some matter of professional business. He or she is therefore most likely to be
someone in the university who has academic business to discuss with me; he
is less likely to be a professional colleague from another university, and I
shall be very surprised if he turns out to be someone (now living in Canada)
with whom I was friendly twenty or thirty years ago. In the Mullan case, it is
clear that the listener's expectations had been narrowed down to such a
degree that they were convinced that the caller could only be one person: the
armed and masked man who had broken into their house. Their expectation
would naturally be that there would be only one caller identical to the
masked man: there would be no reasonable expectation that anyone else
would telephone to blackmail them. Thus, it should be understood by courts
that witnesses' expectations in such cases will lead them to answer as these
witnesses did. It would be surprising if they actually testified otherwise. As it
is, their testimony is only of value if their opinions are backed up (or at least
not contradicted) by objective analysis of the recordings.
Before we proceed to our linguistic analysis, it should be noted that the
witnesses' identification of Mullan's voice "live" in the police station is also
subject to the same reservations. Here they had been led to expect that there
was quite a high probability that the person being questioned was the
masked man: they would reasonably presume that the police would not ask
them to identify the voice of anyone not believed to be the man. It is clearly
unfair to narrow down the possibilities in such a case to only one voice, as no
one can define the size of the population who may have similar voices (there
could be tens of thousands of people who have voices and accents similar to
Mullan). To hold a voice identity parade which narrows down the pos-
sibilities to, say, ten voices might be judged a fairer procedure provided that
60 JAMES MILROY

all ten voices were similar in voice quality and accent. It is extremely
worrying that voice identification evidence of this kind, obtained surrepti-
tiously and without the accused's knowledge or permission, should be
admissible in any court of justice. There is a very high probability of wrong
identification.

Linguistic analysis of the evidence

The tape-recorded evidence consists of two recordings made of Mullan 's
voice at the police station and five recordings made of the anonymous
telephone calls. Of these, the police station recordings are quite short
(amounting to five or six minutes in all) and of poor quality. They are not
good enough for instrumental analysis. The first three telephone calls are
much longer (about eight to ten minutes each); call 4 is shorter and call 5 is
the shortest of all. Dr Baldwin analysed all these tapes: in response to a
question by defence counsel (R. A. Ferguson, Q.C.) he answered as follows:
... I came to the conclusion that the suspect's speech was least like the caller on
the four telephone calls (viz. calls I, 3, 4, 5) and most like the speaker on call 2,
but that there were significant differences between the suspect and the speaker
on call 2 ...
In this, Baldwin was absolutely correct, and his conclusions should have
been enough to cast considerable doubt on the purely impressionistic views
of other witnesses. However, it does not seem to have been made clear in the
questioning that Baldwin's views were based on facts about systematic
differences in voice and accent that had been analysed by an expert
phonetician, not on unsubstantiated opinion. No substantive analysis of
these differences seems to have been presented at the trial, nor does the jury
seem to have been instructed that objective expert evidence based on facts
cannot be lightly set aside.
Cross-questioning by the prosecution concentrated on suggesting that
Baldwin's opinion was no more reliable than other witnesses, and the
prosecution seems to have convinced a majority of the jury. Unfortunately,
at one point, Baldwin answered a particular question in a way that the
prosecution was able to exploit. In answer to the question: "And the accent
was similar in all of the telephone messages?", he said "Yes, that is correct".
A preferable answer would be: "No, that is not correct". The witness could
have explained that, to a layman, that the accents might appear to be similar,
but that - despite these similarities- there are clear, consistent and systematic
differences between call 2 and the other calls, which are obvious to a
 SOCIOLINGUISTIC METHODOLOGY 61
phonetician, dialectologist or sociolinguist, and particularly obvious to a
linguist familiar with Northern Irish accents.
My analysis of the recordings used the methods pioneered by Labov
(1966) in New York City, and now familiar in British sociolinguistics (used
by Trudgill, 1974, in Norwich; Macaulay, 1977, in Glasgow; Milroy and
Milroy, 1978 etc., in Belfast). These methods, as is well known, crucially
involve isolating from the great mass of talk a series of recurrent linguistic
variables. These are usually phonological or morphological (e.g. as in
Trudgill's study of deletion of the 3rd person singular inflexion in present
tense verbs: he go as against he goes). A number of variants of particular
variables are then identified. Sometimes, two variants are sufficient, as in
the Belfast variable ( th): this measures presence or absence of the consonant
[ <J ] in a set of words of the type: mother, rather, bother; in other cases,
several variants can be identified, as in the Belfast variable [a], which ranges
from low-front [a] to central (a) to low-back (a] to mid-back-round [::>],in
words like that, grass, have, man (Milroy and Milroy, 1977, 1978).
The Labov methodology also depends on quantification. Large numbers
of tokens are counted, and it is usually found that the incidence of particular
variants differs for individual speakers and groups of speakers (classified
according to social class, region, age or sex). Most people associate Labov's
methodology with differentiating speakers according to social class; how-
ever, it has been successfully used to differentiate individuals and groups
within the same social class and region. Milroy and Milroy (1977, 1978; see
L. Milroy 1980; J. Milroy, 1981) studied a population in three close-knit
areas of inner-city Belfast, in which it could be presu,med that all persons
sampled had roughly the same regional and social background (i.e. to the
layman their accents would appear to be very similar indeed, if not identical).
Even in this project, clear and significant differences were repeatedly shown
between groups of speakers classified according to age, sex and area of the
city: furthermore, all individual speakers had different quantitative scores
on each of the ten variables studied, except in cases where a few speakers
shared a 100% score on one variable out of the ten (L. Milroy, 1981, pp.
204-5).
Quantitative work of this kind depends on a knowledge of linguistics that
the average person does not possess. In general, it depends on the knowledge
that a language is a highly complex phenomenon that is sytematically
structured, and that variation within a language is also largely systematic and
not random. It is, of course, better known that phonological differences are
indexical (Abercrombie, 1967); in particular, everyone knows that specific
phonological differences are indices of different regional accents, and that
you can guess where a person comes from by his accent: that is to say that
certain variables mark regional differences (as in the Mullan case). How-
62 JAMES MILROY

ever, phonological differences can also be markers of social class, age and
sex of speaker, or of variation according to the situational context in which a
person is speaking.
In a country where regional dialect differences are marked and where
specific phonological variables can readily be identified (e.g. the glottal stop
in Glasgow-Macaulay, 1977; /hi-dropping in Norwich-Trudgill, 1974), the
quantitative methodology is highly promising as a means of demonstrating
that, in some particular case, two recorded voices are probably not those of
the same speaker. For example, if over a series of variables, one recorded
voice consistently had scores markedly different from another voice, it
might well be unreasonable to conclude that they were the same speaker. In
the Mullan case, the differences on prominent variables as between Mullan's
voice and that of telephone caller I, 3, 4 and 5 is enormous: the incidence is
100% m one case to zero in the other. This establishes as a matter of
verifiable fact (not a matter of opinion) that the two accents are completely
different. I shall return to this point below.
I am not aware that the quantitative methodology has been used before in
cases of this kind. In the present state of our knowledge, it would appear to
be more capable of being applied forensically than are instrumental tech-
niques. Notice that even if instrumental techniques were more advanced
than they are, it would still be possible for one analyst to select from the data
certain "voice-prints" that might tend to suggest that two speakers were the
same, and another analyst to select from the same data other "voice-prints"
that might indicate the opposite. The quantitative methodology, however,
requires that, for given variables, all the relevant data must be accounted
for: the analyst is not permitted to ignore relevant data that might not suit his
argument.
Finally, the selection of variables depends crucially on the linguist's prior
knowledge of regional variants in the area to be studied. In a case like the
present one, this means that the linguist should know before he begins what
features are likely to be diagnostic in differentiating Ulster accents; he
should have a good professional knowledge of Ulster dialect. We consider
this point more fully in the next section.

Ulster dialect

Quantitative sociolinguistics of the type pioneered by Labov has always

depended on isolating a small number of variables to be studied from a very
large number of possible variables. In New York, for example, Labov
studied post-vocalic (r), i.e. presence or absence of /r/ in all words of the
types represented by guard, water; {th), i.e. variation between /d/ and /fj/ in
 r'
(
I
r·
__ .,..,.,· r·~·"""
,../
r'
i
\
64 JAMES MILROY

the first consonants of words like this, that, there; and three other variables.
Depending on his prior knowledge of New York speech, he selected those
variables that he believed might be diagnostic of social differences and which
were also likely to occur frequently on the tapes. There might have been
many other "sounds" in New York speech that would either not occur
frequently or not show as much variation. My selection of variables in this
case also depended on similar factors:
( 1) (most importantly) prior knowledge of Ulster dialects, which enabled
me to isolate those features that would be diagnostic of regional
differences and ignore those features that might be more similar as
between two regional dialects;·
(2) relative frequency of incidence of particular variables. Thus, varia-
tion between the glottal stop and (t] was ve'ry frequent (in words like
not, later)
As I have already indicated, the differences of accent between the tape-
recordings of Mullan and the anonymous speaker on telephone calls 1, 3, 4
and 5 are considerable and they are easily identified by a trained analyst. It is
clearly desirable, however, that the analyst should have prior knowledge of
the features that are likely to distinguish different Ulster accents: an expert
from outside Ulster may, in the first place, notice those features that do not
differentiate these accents (e.g. fronted pronunciations of /u/, which are
common to almost all Ulster dialects, but not found in most of England); it
may therefore take such an expert some time to identify the variables that
are diagnostic, and he may have less confidence in his conclusions than an
analyst familiar with variations in Ulster.
Dialectologists (Gregg, 1964, 1972; Adams, 1964) have traditionally dis-
tinguished two broad dialect areas in the northern counties of Ireland. The
first, which comprises North Down, Central and North Antrim, North Co.
Londonderry and a large part of East Donegal, are Ulster Scots in type: the
Ulster Scots areas in fact extend more widely than is indicated on the
map on page 63 (based on Gregg, 1973) and include the North-East corner
of Ulster. In the core Ulster Scots areas, the most extreme rural dialects are
barely distinguishable from those of Galloway and Ayrshire (see Gregg,
1964; J. Milroy, in press). Most of the area west and south of the Antrim and
Derry Scots areas (as marked on the map) is known as the mid-Ulster
(Adams, 1964) or the Ulster Anglo-Irish (Gregg, 1972) dialect area.
Although most Ulster dialects are affected by Scots, there is a sharp dialect
boundary between Ulster Scots and Mid-Ulster dialects, and many features
of pronunciation distinguish them. These are discussed by Adams and
Gregg, and those relevant to Belfast English are listed in J. Milroy (1981,
p. 25). The following major differences are relevant in this case:
 SOCIOLINGUISTIC METHODOLOGY 65
( 1) The historic short vowels /e, a, 'J/, before voiceless stops (as in pet,
pat, pot) tend to be long or half-long in Ulster Scots areas, clearly
distinguished from one another. In Mid-Ulster they are noticeably
short, and may merge or overlap in various patterns. In particular, the
Ulster Scots vowel in words like cot, pot, stop, rock, is ['J'] whereas the
mid-Ulster vowel is unrounded and advanced to [a].
(2) Before velar stops and the velar nasal, there is a marked tendency for
historic short /a/ to be raised to [e] in Ulster Scots. Thus back, bag,
bang sound like beck, beg, beng. This is found to a certain extent in
Belfast, but is not generally characteristic of Mid-Ulster, west of
Lough Neagh.
(3) The glottal stop for post-vocalic [t) is very characteristic of Antrim
and Derry Scots, and has a much lower incidence outside of Ulster
Scots areas. Sometimes, in Ulster Scots, the [t) is present, but has
strong glottal reinforcement
In other possible situations, other variables might be relevant: e.g. Belfast
dialect and many of those to the south and west fail to distinguish between
/w/ as in wine and /hw/ as in whine. It is not known at present how far north
of Belfast this merger extends. In the tapes of the Mullan case, all speakers
recorded distinguish between /w/ and /hw/; therefore, there may also be
quite fine phonetic differences that we have not studied. It is not necessary to
go this far, as the differences between Mullan 's speech and that of telephone
caller 1, 3, 4, 5 are perfectly clearly established on two variables which show
gross phonetic differences. Mullans' speech is Mid-Ulster, that of telephone
caller 1, 3, 4 and 5 is Ulster Scots.
Recall that Mullan is from Garvagh and the Barton family from Kilrea.
Both these places are at or near the dialect boundary as defined by Gregg. It
would be unwise to be too precise about this boundary, but it is quite clear
that telephone-caller 1, 3, 4 and 5 is from a place within the Ulster Scots
area. My opinion is that he is from an area well within this boundary:
somewhere east or north of Kilrea, possibly from the area around Ballymena,
Ballymoney and Coleraine. Mrs and Miss Barton have mid-Ulster accents,
but Miss Barton's brother Joe (who answered some of the calls) has an
accent that is more Scots in type. It is, however, less heavily Scots than that
of the caller. It happens that the mid-Ulster accent is spreading at the
expense of Ulster Scots (Gregg, 1972), and this sex differentiation within the
Barton family demonstrates very nicely the tendency for male speakers to be
more conservative in accent and for females to adhere to more innovative
forms.
66 JAMES MILROY

The variables
The most obvious diagnostic in this case is incidence of the glottal stop [?]or
glottalized /t/, [?t], in three positions:
( 1) in final position in words of the type: not, it, what;
(2) medially between vowels in words of the type: later, butter;
(3) in other post-vocalic positions, e.g. before the consonant in the fre-
quently occurring word Patsy (the Christian name of Mr Barton
senior).
The statistics in this case are easy to report. The incidence of glottalization in
these positions in the accent of caller 1, 3, 4, 5 is 100% (about 120 tokens); in
the accent of Mullan (15 tokens) the incidence is zero; in the accent of
telephone caller 2 (40 tokens), the incidence is also zero.
Typical examples (from call no. 1) are [pa.?tsi, gel, 1?, we.?t, n:> .?t, n:> 1,
hwa.?] for Patsy, get, it, wait, not (twice), not, what (twice). All tokens of /t/
that can be glottalized or replaced by a glottal stop are glottalized or replaced
by a glottal stop. Mullan, as recorded in the police station never uses
glottalization or a glottal stop in these environments: not and what are [nat,
hwat ], except that final /t/ is sometimes flapped (as would be expected in
mid-Ulster speech), when it is followed by a vowel in connected speech,
thus: (hwar]. The same applies to caller no. 2. Caller 1, 3, 4, 5 never uses the
flap: being Ulster Scots, it is not to be expected that he would.
In articulatory, auditory and acoustic terms, the phonetic difference
between /t/ and Ill is gross. In articulatory features, they can be said to differ
in all features but one (+stop). It is inconceivable that a single speaker can
consistently maintain such a difference over a stretch of talk.
Among the several vowel differences that differentiate the speakers, the
most frequently occurring diagnostic variable is (:.')), measuring the incidence
of [:>·] as against [a] before voiceless stops in words of the type what, clock,
stop. Again, the phonetic differences are numerous:
]
r~ can be described as:
+back
+low
+mid
+round
+long
whereas [a], a low central to front vowel, is:
-back
+low
-mid
-round
-long
 SOCIOLINGUISTIC METHODOLOGY 67

The two vowels can be said to differ in four features out of five.
Again the statistics are decisive. Mullan and telephone caller 2 always
use [a], whereas caller 1, 3, 4, 5 always uses [6). Thus Mullan has [hwat, stap,
klak] for what, stop, clock; caller 2 has [nat, hwat] for not, what; caller 1, 3, 4,
5, however, has [st~p, ddp, n ~ 1, g~Z~ hw:i?) for stop, drop, not, got, what.
Mullan uses the word stopped four times in quick succession: it is always
[stapt].
These variables alone are convincing, as they both have scores of 100%
for caller 1, 3, 4, 5 against scores of zero for Mullan and for caller 2. They
demonstrate that we are dealing with two quite different accents. But there
are other distinguishing variables also - of a kind which are predictable to
the student of Ulster dialects. Telephone speaker 1, 3, 4, 5 consistently
raises I a/ to [a!] or [e] before velar consonants: thus, he says [kantrckt~r,
k~ntckt] for contractor, contact. Telephone speaker 2 also uses relevant
words (e.g. contact, attack), and does not raise /a/ to [re) or [c]. Mullan has a
palatal glide after initial [kJ in the word Kelly's (repeated): this is a well-
known mid-Ulster feature. Caller 1, 3, 4, 5 has no such glide in relevant
words such as Catholic. Caller l, 3, 4, 5 also has Ulster Scots lengthening
of most short vowels before /p, t, k/; Mullan does not have this lengthening;
Caller 1, 3, 4, 5 has a relatively high {Scots) monophthong [e ·],in words like
wait, days; Mullan has a lower vowel. There are other differences.
It is not a matter of opinion that these two accents are different: the
differences are audible, verifiable and consistently maintained. When these
differences are demonstrated, a court should be willing to accept that since
the accent difference between Ulster Scots and Mid-Ulster is of the same
order as differences between Newcastle and Leeds, and much greater than
the differences between Leeds and Sheffield, then the likelihood that
Mullan is the same speaker as caller l, 3, 4, 5 is so minimal as to be not worth
considering. It is not even remotely likely that a brilliant mimic could
maintain these differences consistently (there would be variation and some
"mistakes''). In any case, this hypothetical mimic would presumably have
no reason to alter his accent completely in call 2.
An experienced phonetics teacher will know that speakers who do not
normally use glottal stops cannot easily be persuaded to produce them in
isolated words, let alone use them in prescribed environments in a long
stretch of talk. It would probably be a simple matter to demonstrate that a
speaker like Mullan is unable to utter the sentence "What a lot of little
bottles" with glottal stops in the "right" places. This kind of evidence could
be obtained~ but in this case it was not.
68 JAMES MILROY

Conclusions

My two major conclusions concern the status of linguistic evidence in cases

of voice identification, and the potential uses of linguistics in forensic
matters.
As we have seen, Miss Barton testified first that telephone caller 1
was the masked intruder who had threatened the family. It is quite possible
that she was correct in this. There is no recording that is known for certain
to be that of the masked man; therefore, his speech cannot be analysed and
compared with that of Mullan and the telephone speakers. However, if
caller 1 was the masked man, then Mullan could not be the masked man, as
caller 1 cannot reasonably be said to be Mullan. Miss Barton further testified
that telephone caller 2 was the same speaker as telephone caller 1, and
furthermore that all the telephone calls were by the same speaker. It is
inconceivable that any linguistic analysis could support her opinion. In
short, whereas Miss Barton expressed an unsubstantiated opinion influenced
by certain expectations, a linguistic analysis is based on audible and verifi-
able differences between accents. Furthermore, it shows that the differences
are not merely random variations in pronunciation of occasional words, but
that they are consistent and systematic. The details analysed point to
structural differences between dialects that are of a high order of generality
and abstraction: for example, the Ulster Scots vowel-length system (as
evidence in calls 1, 3, 4, 5) differs from the Mid-Ulster one not merely in one
particular vowel, but throughout the whole vowel system. Abstract system-
atic differences of this kind are never obvious to ordinary speakers; they
can be described only by trained phonologists who have carefully observed
the different systems.
It would appear, however, than an expert witness can still be said to be
expressing an "opinion" if he concludes that the demonstrable accent
difference implies that different speakers are involved. But this opinion is of
much greater value than an unsubstantiated opinion, as the probability that
an ordinary speaker can switch from one non-standard accent to a perfect
imitation of another is infinitesimal. Clearly, in such a case, it should be
incumbent upon the prosecution to prove that the accused is a mimic of
unparalleled talents. As for caller no. 2: his accent is similar to Mullan's.
There are some general indications that he is not Mullan, and a closer
phonetic analysis might tend to support this.
We must conclude that caller 1, 3, 4, 5 is not Mullan, even though Mullan
was observed to try to make a telephone call at the same time as call no. 5
was received. In call no. 5, glottal stops and other Scots characteristics are as
prominent as they are in calls 1, 3 and 4. It is a great pity that only one
 SOCIOLINGUISTIC METHODOLOGY 69
linguistic expert was called in this case, and that the prosecution managed to
imply that his evidence was not conclusive. If others had been called in
addition, the linguistic evidence - compelling as it is - could hardly have
been ignored. Indeed longer recordings could have been made of Mullan's
speech and linguistic tests administered to him. These too could have
helped.
It follows from all this that linguistic evidence of various kinds may be
more helpful in legal proceedings than has been realized, but specifically in
differentiating voices, not in proving identity of voices. We cannot entirely
blame the legal profession or the juries for being unable to weigh up the
linguistic evidence if we are slow to come forward to help them, even though
juries have been willing to believe quite extraordinary things, e.g. that
virtually inaudible tape-recordings are recordings of the voice of some
accused person (J. Anthony, personal communication). Similarly, it is odd
that (as in this case) very doubtful voice identification procedures can be
used by the police.
In the United States the role of expert linguistic evidence has been more
readily recognized than in the UK. In one recent case, the "Black English
case" reported by Labov {1982), there has been a significant breakthrough
as far as applied sociolinguistics is concerned. A number of linguists of
different persuasions were all able to testify to the effect that BEV (Black
English Vernacular) differed radically in a number of systematic ways from
Standard English. Their evidence was in this case decisive: probably for the
first time, jurists have recognized that dialects differ from one another at an
abstract systematic level, and, further, that these differences have specifi-
able social and educational consequences. The Mullan case, it seems to me, is
no less significant than the Black English case. It is important that the
consequences of conscious or unconscious discrimination against non-
standard dialects should be understood (as in the "Black English case"), but
in a different way, it is equally important that an individual should be
protected from conviction when that conviction is mainly based on mis-
identification of his voice. However we view it, there must be many potential
legal cases where linguists of both the "armchair" and the "field" varieties
can be of more help than they have been to date. It is somewhat unedifying
for linguists to debate so-called general principles (e.g. whether or not
linguists should consent to be expert witnesses) when it can be specified
exactly how useful they can be in cases like the present one. It should
certainly be emphasized that they can often show that speakers are different,
but cannot necessarily show that the voices on two recordings are the same.
70 JAMES MILROY

Postscript

At the time of writing (September-October 1982) Seamus Mullan has

served two years of his ten-year sentence. Believing that he had been
wrongly convicted, Mullan attempted to draw public attention to his case by
going on hunger-strike in the autumn of 1980. Unfortunately, a number of
other persons in Northern Ireland started a hunger-strike at about the same
time for purely political reasons, and the publicity given to their political
strike drew attention away from Mullan. Mullan's lawyers (Mr F. McNicholl
and Mr R. A. Ferguson, Q.C.) promised to apply to the House of Lords for
leave to re-open the case, and Mullan ended his strike. At the same time,
being convinced that there had been a miscarriage of justice, the lawyers
appealed publicly for assistance in uncovering new evidence that might help
to cast doubt on Mullan's guilt. I offered to analyse the tapes, fully prepared
to find that the differences between them were perhaps not decisive. I was
horrified to discover that the taped speakers had what I know to be two
radically different accents, and that the Ulster Scots accent was already
evident in the first full sentence of the first call analysed (no. 1). I was
completely convinced that at least two different speakers were involved. My
report to the lawyers was submitted in January 1981. It does not seem to
have been regarded as "new evidence", and leave to appeal to the House of
Lords was refused.
Thus, although it is clear that Mullan was convicted and sentenced mainly
as a result of misidentification of his voice, his only hope now is that he may
be granted a pardon.
Whether or not Mullan actually made phone-call no. 2, at least one very
unpleasant blackmailer remains at large, and the "guilty" verdict on Mullan
will stand for ever. 3 ·

References
Abercrombie, D. (1967). "Elements of general phonetics', Edinburgh University
Press, Edinburgh.
Adams, G. B. (1964). Ulster dialects. In "Ulster dialects: an introductory symposium",
Ulster Folk Museum, 1-4, Holywood, Co. Down.
Bolinger, D. (1980). "Language: the loaded weapon", Longman, London.
Brown, R. (1980). The role of the listener's expectations in speaker recognition.
"Work in progress", (Department of Linguistics, Edinburgh University) 13,
72-8.
Furnborough, P., Jupp, T., Munns, R. and Roberts, C. (1982). Language, dis-
advantage and discrimination: Breaking the cycle of majority group perception.
 SOCIOLINGUISTIC METHODOLOGY 71
In ••Language and Ethnicity", Linguistic Minorities Project and British Associa-
tion for Applied Linguistics, London.
Gregg, R. J. (1964). Scotch-Irish urban speech in Ulster. In "Ulster dialects: an
introductory symposium'', Ulster Folk Museum, 162-91, Holywood, Co.
Down.
Gregg, R. J. (1972). The Scotch-Irish dialect boundaries of Ulster. In Patterns in the
folk-speech of the British Isles, pp. 109-39. (W. F. Wakelin, ed.), Athlone Press,
London.
Gumperz, J. J. (1976). Language, communication and public negotiation. In
"Anthropology and the public interest: fieldwork and theory", (P. Sanday,
ed.), Academic Press, New York and London.
Labov, W. ( 1966). "The social stratification of English in New York City", Center for
Applied Linguistics, Washington D.C.
Labov, W. (1982). Objectivity and commitment in linguistic science: The case of the
Black English trial in Ann Arbor. Language in Society, 11, 165-201.
Ladefoged, P. (1975). "A course in phonetics", Harcourt, Brace, Jovanovich, New
York.
Laver, J. (1980). 'The phonetic description of voice quality", Cambridge University
Press, Cambridge.
Macaulay, R. K. S. ( 1977). •'Language, social class and education: a Glasgow study",
Edinburgh University Press, Edinburgh.
Milroy, J. (1981). ••Regional accents of English: Belfast'', Blackstaff, Belfast.
Milroy, J. (1982). Some correspondences between Galloway and Ulster Speech.
Scottish Language, I.
Milroy, J. and Milroy, L. (1977). Speech community and language variety in Belfast.
Report to the Social Science Research Council, London.
Milroy J. and Milroy, L. ( 1978). Belfast: change and variation in an urban vernacular.
In "Sociolinguistic patterns in British English", (P. Trudgill, ed.). Edward
Arnold, London.
Milroy, L. (1980). "'Language and social networks", Basil Blackwell, Oxford.
Neale, A. V. and Wallis, H. R. E. (eds) (1955) ...The boke of chyldren", by Thomas
Phaire. E. and S. Livingstone, Edinburgh.
O'Barr, W. M. (1981). The language of the law. Jn ··Language in the U.S.A." (C. A.
Ferguson and S. B. Heath, eds), Cambridge University Press, Cambridge.
Trudgill, P. (1974). •'The social differentiation of English in Norwich", Cambridge
University Press, Cambridge.
Wells. J. C. (1982). ••Accents of English", 3 vols. Cambridge University Press,
Cambridge.

Notes
Details of the case reported in this chapter are from transcripts and tape-recordings
of Regina vs Mullan. I am extremely grateful to Mr Francis J. McNicholl, solicitor,
and Mr R. A. Ferguson, Q.C., for drawing my attention to the case, and for making
these materials available. In details of the case. I have used real names, except that I
have used a fictitious name for the threatened family, to save them possible embarrass-
ment. I am also grateful to Lesley Milroy for her assistance and criticisms at all points.
72 JAMES MILROY

~ As a result of intimidation (and in some cases murder) of juries and witnesses, the
Government introduced the "Diplock Courts" for terrorist cases in 1976. Such cases
are similar to the Special Courts in the Republic of Ireland except that they are tried
by a single judge, sitting alone. It has been suggested that in this case, Mullan would
not have been convicted by a Diplock Court, owing to the inconclusive nature of the
evidence.
·' The Mullen case was re-opened by the Secretary of State for Northern Ireland. In
September 1983 the appeal was rejected on the purely technical grounds that the
defence should have sought expert evidence before the trial commenced (R. A.
Ferguson, Q.C., pers. comm.). Although this evidence cast doubt on Mullen's guilt,
it was not considered relevant by the appeal court.
 5
Good copy - bad news
The syntax and semantics of news editing

Allan Bell

Editing is the process by which one text is transformed into another text
which is different in form but congruent in meaning. Editing is a very
common language activity. Probably most written language undergoes
editing, the only obvious exception being personal letters. Editing may be
done by oneself, the original writer, or by others. It may range from the most
cursory re-reading of a text, resulting in a few minor alterations, to the
multiple re-workings which a poet gives to a work. Editing is the intra-
language equivalent of translation or interpreting, a process by which a text
in one language is transformed into another language.

The study of editing

The application of linguistic analysis to editing may serve three purposes.

First, it can provide insights into the nature of linguistic structure. The
editing process often results in minute changes which put language structure
under the microscope in a way which usually occurs only in the arm-chair
linguist's intuitions. This motivation is purely internal to the discipline of
linguistics. Although it applies linguistic analysis to an object to which it is
not normally applied, it is not applied linguistics in the sense of addressing
issues beyond the discipline itself.

APPLIED SOCIOLINGUJSTICS Copyright ·© 1984 by Academic Press, London

ISBN: 0-12-701220-6 All rights of reproduction in any form resen·ed
74 ALLAN BELL

Secondly, linguistic analysis of editing may provide information on issues

relating directly to language use or ability. Comparison of the edited text
with the original reveals any stylistic differences between the two. We may
then draw inferences concerning the motivations of author and editor (or
even their identities if these are unknown) or concerning the kinds of
readerships for which the texts were intended. Already such a study moves
away from applied (unhyphenated) linguistics to applied socio-linguistics,
since we seek here a link between linguistic facts and extra-linguistic
factors.
This kind of analysis may serve, for instance, to test the authenticity or
authorship of a text. In much biblical or Shakespearean scholarship, the
received text is compared with possible sources, or competing versions are
compared with each other, in order to establish a canonical text. Editorial
overlays are peeled away to reveal the presumed original text. If authorship
is in doubt, stylistic comparison with other texts is used to support arguments
for or against various contenders. Where the purpose for which a text was
edited is known, analysis of input and output can establish whether that
purpose has been achieved. If a text has been edited for a particular reader-
ship, comparison with the original can show whether the edited version is in
fact more suitable for the readership than the original. Texts are often
re-written for learners, whether foreigners or children, according to read-
ability formulas. The study by Davison et al. (1980) indicates how writing to
such formulas does not necessarily produce a text suitable for the target
readership.
Editing analysis constitutes a description of what copy editors do in their
professional role (cf. Crystal, 1981, p. 18). Editors focus on language
options. Changes they make in the original result from a focussed choice,
rejecting the alternatives taken by the author. It seems justified to impute
considerable significance to social factors which may correlate with an
editor's choices - more than when the choice is among open alternatives, as
it was for the original author.
A third purpose of editing analysis is to identify meaning differences
between the original and the edited versions. Here the focus is on the
content of the text rather than language and its use. We may compare, for
instance, the final and rough drafts of a poem. The forms which a poet has
tried but rejected can throw into sharper relief the expression finally chosen.
Its significance might have been obscure or even unnoticed without the
contrast of the poet's deliberate alteration.
Usually our interest in comparing the content of texts is to establish
whether the edited version is a faithful rendering of the original. We
acknowledge authors' rights to change the meaning of what they originally
wrote, assuming they have changed their minds on the matter. Outside
 SYNTAX AND SEMANTICS OF NEWS EDITING 75
editors are not given the same licence. They are expected to transmit
faithfully the meaning of the original text. In many cases, it may not matter if
an editor takes liberties with the content of a text. If details of a story are
changed in re-writing for children, it is unlikely to matter. However, more
substantial changes may, even in fictional writing, remove the edited version
unacceptably far from reality or from the author's original. Adding and
deleting information or restructuring connections in a text can produce a
misleading version (Davison et al., 1980).
Most editing goes on in the realm of "facts" not fiction. In diplomacy,
government and commerce, inaccurate editing (and especially inaccurate
inter-language translation) can cause misunderstanding and conflict.
Reports and documents are edited and summarized for busy decision-
makers who lack time to read the full version. If the edited text does not
faithfully represent the original, decisions may be taken on mistaken
information.
The classic case of editing is news. Millions of words of news pass daily
through the hands of many copy editors throughout the world. A single news
text may be edited five or more times. If news editors do not faithfully pass
on the content they receive, the consumers of news, both the public and
decision-makers/governments, will be misinformed. While the basic analysis
here is a linguistic comparison of texts, the findings take on a pattern only in
the light of social factors. Any inaccurate editing which we find will be
accountable largely in terms of the social systems and ideologies in which the
editors work.
This chapter serves the second and third purposes, and the first only
incidentally. Although the following editing analysis puts a number of
linguistic facts in an unusual light, I will focus on the pattern of syntactic
changes, and inaccurate semantic changes, made in the course of news
editing. The analysis provides evidence on issues of news style and accuracy
which both news professionals and society at large treat as significant.

Sample
The data used in this study are drawn from a large random sample of news
collected in New Zealand in 1974. All news broadcast on the six radio
stations in Auckland, New Zealand's largest city, was recorded on five days
(approaching 400 OOO words). Over one third of the analysable sample was
international news, which the stations received from a common source. At
this period, all staple international news entered New Zealand at Wellington
(the capital) on the Australian Associated Press-Reuter wire from Sydney,
Australia. AAP received most of its copy from Reuters, some from Associ-
ated Press and United Press International, and a little from Agence France
76 ALLAN BELL

Presse, Tass and other agencies. In Wellington, AAP copy was received by
two separate internal wire services: the New Zealand Press Association, and
the General News Service of the public corporation, Radio New Zealand
(Fig. I). 1 The two public radio news networks which broadcast from
Wellington, YA and ZB, received their AAP copy direct, selected by the
General News Service editor (who works in the same newsroom). This
editor also edited the copy for transmission on the GNS wire to Radio New
Zealand stations outside Wellington, including ZBR and ZM in Auckland.

- Transmits AAP-Reuter output

Selects Sydney

../V'-- Selects a edits

GNS input
Wellington

GNSw1re NZPA radio

ootput wire output

Fig. 1. The flow of international news to New Zealand radio stations:

AAP Australian Associated Press-Reuter wire service, Sydney, Australia;
GNS General News Service, Wellington, New Zealand-wire service of the public
corporation, Radio New Zealand;
PA New Zealand Press Association, Wellington;
YA National Programme of Radio New Zealand, prestigious news service;
ZB Radio NZ Community Network news - both YA and ZB are relayed through
Auckland stations, but originate in Radio NZ's Wellington studios;
ZBR Local Radio NZ community station;
ZM Radio NZ rock music station - both ZBR and ZM broadcast from the
Auckland studios of Radio NZ;
XA Radio Hauraki, private rock music station, Auckland;
XI Radio i, private community station, Auckland.
 SYNTAX AND SEMANTICS OF NEWS EDITING 77

The Press Association edited AAP copy for transmission to privately owned
radio stations, including XA and XI in Auckland. Figure 1 diagrams the flow
of international news from the AAP-Reuter wire to the six radio stations
heard in Auckland. (On the structure of international news agencies and
flows, see Boyd-Barrett, 1980, p. 73 ff.)
The main selection of news is made by the GNS and PA editors. In the
two-stage flow, GNS and PA also do the basic rewrite, and the recipient
stations (ZBR, ZM; XA, XI) make usually minor alterations. I obtained
much of the AAP copy for the five sample days, and rather less of the copy
transmitted by GNS and PA. Actual edited wire copy for many stories was
available from all but one station. ("Story" is used throughout in its
journalistic sense of one report/dispatch on one topic).

The editing process

News editing is the archetypal editing situation. A television or radio news-

caster is only the last of several people who have handled the news copy
which is read out. Between the original journalist and the newscaster there
may be five or more copy editors. 2 These are language professionals with a
threefold function:

Select
From the stories that come off the external wires or from internal journalists,
the copy editors select some few for the next edition. The copy editor limits
the volume of news largely by rejecting entire stories. For broadcast news,
from which I draw my data in this study, the selection process is drastic.
At BBC news, the "copy taster" discards 90% of news agency material
(Schlesinger, 1978, p. 60). In New Zealand, the main overseas wire carries
about 100 OOO words daily, which would occupy ten hours of continuous
radio broadcasting. Again, about 10% gets through the gate.

Cut
A second way the copy editor limits news output is to abbreviate the·stories
which are accepted. This is particularly necessary in broadcast news. The
average radio news item is less than 100 words long, but international wire
services are intended mainly for newspapers, so some stories reach 2000
words. The copy editor has to cut, usually by accepting the first page of copy
and throwing the rest away.

Alter
It is the copy editor's third operation that is really interesting to the linguist.
78 ALLAN BELL

Selection leaves the form of accepted stories untouched. Cutting deletes

matter, but does not otherwise change what remains. But alteration changes
the actual language form of news accepted for broadcast. So while the object
of much content analysis of news is what gets left out (e.g. Cutlip, 1954), the
focus of this study is the fate of what is left in. Editing alterations serve two
purposes. They reduce the amount of news copy even further, because
editors delete individual words and phrases as well as the complete sentences
rejected in cutting a story. The second purpose is stylistic. The linguistic form
of copy originating both outside and inside the news organization must be
made to conform to "house style". Most obviously, house style is codified in
the "style book" which newspapers and broadcast stations issue to news-
workers. However, such manuals tend to deal with only the most obvious
points of style, and copy editors in any case rarely consult them. For broadcast
news, editing is also intended to restyle written language to a form more
suitable to be read aloud. Both radio newsworkers and the official style books
lay great stress on the differences between "writing for the eye" and "writing
for the ear".

Methodology
The news editing process was analysed in four steps. This methodology is
applicable, with appropriate adjustments, to analysis of all types of editing. In
many cases, there is no problem in identifying what was the original version on
which an edited text is based (step 1). Steps 2 and 3 are the syntactic and
semantic analyses of editing changes: the heart of any such study. At step 4,
the precise methods used to identify patterns in editing changes will depend on
the kind of questions which the study addresses.

Step 1.
Establish which agency story is the input to a given output story.In the world
of news, identifying the source text is a major problem. International wire
systems are massive and complicated, and the copy is hard to obtain (cf.
Boyd-Barrett, 1980, p. 103). Figure 1 shows only a fragment of how news
flows in one small country. Newsworkers and managers usually know only
their small link in the system. And the systems are constantly changing. The
only foolproof method of identifying what is the input to a given output story
is when the input copy bears the editor's actual markings. However, editing
increasingly occurs at video display terminals. Copy is fed in on tape,
displayed on a screen, and edited on the tape without necessarily ever being
printed out. The spread of this technology is rapidly eliminating blue-
pencilled wire copy and the sure external evidence of identity which it offers.
Physically edited copy was available for many stories analysed from my
 SYNTAX AND SEMANTICS OF NEWS EDITING 79

sample (Figure 2 is a page of wire copy with a recipient editor's markings.)

For most of the other stories used, the input was identified through internal
evidence. These were once-off stories, or reports containing idiosyncratic
phrasing or information (such as figures). In other cases, it is impossible to
decide whether an output story is in fact derived from a certain input wire.
Big stories may have 10-20 continually updating wires, with a broadcast
news item which is a rewrite combining several wires. The form of a few
input wires was reliably reconstructed by comparing several output versions.
In all cases, the necessary principle followed was: when in doubt, omit.
When the output is broadcast news, we use the form actually read out by the
newscaster and tape-recorded, which often differs from the edited script.
The eventual data for this study included 290 international stories, with both
input and ouput copy.

AAP ZEA453

GNA ZEA CNS QRH

MIDBAST - FORCB

TMFAOIV, PY]§ N?Mr·{AK ISRAELI )ofILITARY FORCE~

CROSSED THE BORDER INTO LEBAIOK AND B~11tUP THREE HOUSES llf 11981 1!11
..;..,.af~
VILLAGE.. [AN ARMY SPOJ:E~llCBD AtRi!i, AGiffC! PRXWCE Pitt'SSE
"l!Plm'ft:D, I
~ /Jwt'• .•
'IE '§AID THE HOUSES WERE USED BY ~ OPERATING AGAINST
ISRAEL.

lTRE ~ FORC! RETURNED SAFBf't TO BASE 1 ......,_,

Fig. 2. A page of AAP-Reuter wire copy, with markings by a recipient New Zealand
editor. Note the slip in line 3, failing to delete village.

Step 2.
Describe the syntactic and lexical changes by which the output copy can be
derived from the input. What copy editing does is put up for re-negotiation
the syntactic and lexical choices made by the original journalist (or previous
editor). The editor can reverse these choices, or take options which the
journalist refused. Linguistically, the editing process can be described by
80 ALLAN BELL

means of rules, which are formally similar to the ordinary rules of linguistic
theory. Editing rules cover all types of linguistic operations, and often
perform major surgery on the syntax of the input copy. They take as input
one well-formed sentence and turn it into another, semantically congruent,
well-formed sentence of news English.·1 There is thus a Syntactic Condition
on the application of editing rules:
After all editing rules have been applied in a sentence, the resulting
structure must be syntactically we/I-formed.
Each step by which the output form is derived from the input must be
specified precisely by a separate rule or group of rules. As many as 100 rules
may be needed to derive an output version several sentences long from its
input form. Below I sketch something of the work which editing rules
perform, and the extra-linguistic factors which influence their application.
Step3.
Examine all editing rules to identify any which have introduced inaccuracy.
As well as meeting the Syntactic Condition, copy editors must also satisfy
the Semantic (or Accuracy) Condition:
No editing rule may be applied if its effect is to make the meaning of
the output story in any way non-congruent with that of the input.
Editing inaccuracy is an incongruence of output copy with input. We do not
require that editors retain all the information contained in the original: that
would mean no editing whatever. Rather, the output must faithfully represent
the content of the input. There may be gaps in the output information, but
no mismatch with the input. Any change made in output copy requires a
warrant in the input copy. Our standard of accuracy is therefore the input
copy. For the purposes of editing analysis, the input story is treated as an
accurate and adequate representation of the real-world facts. This is of
course an idealization, which the results presented below seriously question.
Nevertheless, we are concerned here only with incongruencies introduced at
a given editing step. Each successive copy editor has to work on the same
principle: that the input copy received is accurate. The researcher likewise
exercises this suspension of disbelief for the editing stage being scrutinized.
We test for inaccuracy by turning the edited version into a question, in the
frame Is it the case that X?. If this can be answered Yes, it is the case that X
from the original story, then the Accuracy Condition has been successfully
met. If the answer is No, then the edited version contains something which is
incongruent with the original.
 SYNTAX AND SEMANTICS OF NEWS EDITING 81

Step 4.
Look for patterns in, and propose explanations of, the distribution of editing
rules and editing inaccuracies. At this stage, we have a detailed analysis of
editing rules, and of the meaning inaccuracies which some rules have
introduced. We now proceed to see if there is a pattern in the syntactic or
semantic aspects of editing.
On the syntactic side, the editing process allows us to see a language
variety in the making, actually on the production line. We identify what
kinds of structures are typically affected by editing rules, and what opera-
tions the rules perform. We can see whether the application of any rules is
influenced by extra-linguistic factors. Do certain editors, or certain radio
stations, favour paticular rules? Are editing rules merely optional, or are
they variable rules? The next section deals with these questions.
In the semantic analysis, I identified some 150 editing incongruencies in
the sample. These grouped under five classes of inaccuracy, and we see how
specific editing rules tend to result in certain kinds of inaccuracy. We can
move from this description to prescriptive guidelines on what changes
editors should and should not make. Analysis of which stories were in-
accurately edited shows that certain categories of news suffer more than
others. We can quantify the seriousness of inaccuracies, and relate these to
various extra-linguistic factors. Do certain individuals or stations edit less
accurately? Are certain kinds of news more liable to inaccurate editing? I
present evidence on these .questions in the last section of this paper.

Syntactic surgery: the art of editing

The scale of operations which a copy editor pedorms on the sudace of a news
item can be major (Bell, 1977, p. 227 ff.). Here is the original text of an item
of sports news transmitted by AAP-Reuter from Sydney, followed by the
version as edited by the New Zealand Press Association (the arrow denotes
"is rewritten as''):
(1) AAP/Sl The waterlogged conditions that ruled out play yesterday
still prevailed at Bourda this morning, and it was not
until mid-afternoon that the match restarted.
S2 Less than three hours' play remained, and with the West
Indies still making their first innings reply to England's
total of 448, there was no chance of a result.
S3 At tea the West Indies were two for 139.
--+ PA/Sl Waterlogged conditions ruled out play this morning, but
the match resumed with less than three hours' play
remaining for the final day.
82 ALLAN BELL

S2 The West Indies are making a first innings reply to

England's total of 448.
S3 At tea the West Indies were 139 for two, but there's no
chance of a result.
The Press Association preserves most of the information, but attacks the
syntax on a broad front. Bits of sentences are amputated and re-attached
elsewhere; core constituents are transplanted or transformed. The output of
the operation is syntactically very different but still well formed, and
remains semantically congruent with the input. The main editing rules
needed to describe the derivation of PA/S 1 are:
(a) Place adverbial (at Bourda) deleted.
(b) Time adverbials (yesterday, still) deleted.
(c) Main verb (prevailed) deleted.
(d) Relative clause (that ruled out play yesterday) raised to main clause.
(e) Relative pronoun (that) and definite article deleted.
(f) Clefting reversed (the match did not restart until mid afternoon).
(g) Time adverbial deleted (not until mid-afternoon).
(h) Lexical substitution made (resume for restart).
(i) Alternative co-ordinator substituted (but for and).
(j) Main clause of S2 (less than three hours' play remained) embedded into
S1 under with.
(k) Finite verb goes to non-finite (remaining).
( m) Time adverbial (for the final day) inserted from another version.
All these changes are entirely typical of the editing process, and their effect
on the syntactic structure of the input is considerable. We can distinguish
three broad categories of editing rules:
Information deletions remove information from a sentence (rules a, b, c, g
above, plus information addition by rule m).
Lexical substitutions replace one or more lexical items with another
congruent item or group of items (h, i above).
Syntactic rules perform various kinds of operations, with minor or major
effects on the structure of a sentence (d, e, f, j, k).

Information deletions
Cutting the length of stories remains the broadcast editor's main task even
with the severely reduced volume of copy which the selection process
retains. The copy editor makes the main additional cut by deleting entire
sentences. The news agencies are press-orientated, and often transmit
 SYNTAX AND SEMANTICS OF NEWS EDITING 83
stories containing 40 or more sentences. Some stories which are only 2-5
sentences long may be broadcast without sentence deletions. Long wires
with more than a dozen sentences run to three or more pages of teleprinted
copy, and usually the first page only is used. For medium-length wires, often
the first 2-3 sentences are used, together with one or more sentences from
the middle or end of the story. Most sentence deletions can be performed
without requiring any changes in the remaining sentences of a story. The
most common consequent change occurs when the deleted sentence contains
the antecedent of a pronoun which occurs in a later sentence and which must
then be depronominalized.
Another type of deletion takes out constituents within a sentence. It is
unusual for any sentence, except in the shortest wires, to pass the copy editor
without some constituent(s) being deleted. The editor treats detailed in-
formation as superfluous, or at least unwarranted in the few sentences
avaffable to tell a story. Often omitted are personal details of age, occupa-
tion, nationality, or even the person's name if the focus of a news item is
what happened rather than who it happened to. One of three defendants in a
murder trial was referred to as: a third girl, Josephine Kona Burton, 20,
masseuse. In broadcast news, description plus name plus age plus occupa-
tion is too much in a row for both newscaster and listener, so the copy editor
left simply: a third girl aged twenty.
The simplest deletions are those, like the above, which drop a single node,
and any dependent structure, without requiring other changes to structure.
Time and place adverbials are often deleted (as in rules a, band g above),
especially from the lead sentence of a story. Also common is the deletion of
non-head items within an NP - adjectives, numerals, embedded PPs, rela-
tive clauses:
(2) 16 per cent of the 143 166 smallpox cases
--+ 16 per cent of the smallpox cases
(3) the image of the title she won in November
--+ the image of the title
In co-ordinate structures, one or more of the co-ordinated constituents can
be deleted, provided that one element remains to keep the sentence well-
formed. In this example, the copy editor was able to take out all the detail
and link up the final clause without even having to insert and:
(4) They say she died after being brutally beaten around the head,
kicked, and stabbed with a pair of scissors, and finally strangled.
--+ They say she died after being brutally beaten and strangled.
All these are pure deletions, which leave behind structures unaffected by
the removal of optional constituents. There is a second class of constituent
84 ALLAN BELL

deletions where the structure assigned to the input copy differs from the
output in more than the mere absence of the deleted constituent(s), yet the
linear surface of the sentence remains unruffled. In example (1) above, the
radical structural changes effected by rules a-g result solely from judicious
deletion by the editor:
(5) .Jht!" waterlogged conditions ;)Mr( ruled out play ye~y JtHr
pr~ed_,,at'B_9llrtfa this morning and.ft"~_J}0f'yPri(.JlHtf-
af~on;)Mf(the match restarted ...
This is an unusually complex instance of a very common phenomenon.
Higher nodes of a derivation are deleted, which automatically raises and
re-attaches the lower constituents without requiring any further changes.
This occurs particularly in complex NPs (6), or with time or place adverbials
embedded in a subordinate clause (7):
(6) for pickpocketing offences
~ for pickpocketing
(7) who was killed last week while practising in South Africa for the
Grand Prix
~ who was killed last week in South Africa
The motive for these deletions is abbreviation. But the form which the
deletions take is governed by the physical process of copy editing. Editors
favour changes which are easy to make on a page. Re-ordering material is
disfavoured; minor insertions are more acceptable; and words or phrases
that can be crossed out and still leave a good sentence are a gift to the
overworked copy editor. There is probably no time to retype the copy before
the next deadline. If something can just drop out and the linear surface be
rejoined, it is much faster than shifting bits of sentences around.
In studying processes like editing or translation, the practical limits and
strengths of the technology used need to be remembered, as well as linguistic
factors. For example, the advent of video editing may well be changing the
patterns of deletion. What is easiest to delete by pen on paper may not be so
easy on video screen. 4 Explaining editing changes in grammatical depth
must be complemented by examining the surface of the text on which a copy
editor works.
In the third type of constituent deletion, the sentence which remains is
ungrammatical and needs repair. Such repairs may require minor structure-
mending rules, as when the tense must be changed in reported speech, or a
determiner reinserted in an NP:
(8) Scotland Yard solicitor Neil Denison
~ *Scotland Yard solicitor ~
~ a Scotland Yard solicitor
 SYNTAX AND SEMANTICS OF NEWS EDITING 85

When the deletions are major, they may require surgery before they meet
the Syntactic Condition. Subject NPs are deleted, main verbs disappear
without trace, transitive verbs are left without objects. The copy editor then
has two options: transplant the surviving structure into another sentence or,
less probably, rebuild the original sentence until it is a well-formed structure
again. Contrary to popular belief, broadcast news very rarely contains
ungrammatical sentences. In a transplant, the recipient sentence normally
remains almost untouched, with only a preposition or conjunction inserted
under which the new structure can be embedded. The donor sentence,
however, often undergoes quite radical changes to fit it easily into the
recipient structure:
(9) ZB/Sl In London, the soccer star George Best has been
remanded on bail on charges of stealing a fur coat
and other items from the former Miss World,
Marjorie Wallace, because she hadn't arrived from
America to give evidence.
The ZB editor keeps all of AAP's first sentence (down to ... Marjorie
Wallace). But the clause because she hadn't arrived from America to give
evidence is all that survives of AAP/S2:
(10) AAP/S2 Scotland Yard solicitor Neil Denison had sought the
adjournment from the court because Miss Wallace,
the main witness, was still in the United States:
In the earlier example (1), rules j and k voluntarily transplant the main
clause of AAP/S2 into PA/Sl as a subordinate clause. Because the editor
also transplants there was no chance of a result (into PA/S3), the remains of
S2 are a non-sentence. The editor applies structure-mending rules to rebuild
this into a full sentence:
( 11) AAP /S2 . . . and with the West Indies still making their first
innings reply to England's total of 448 ...
~ PA/S2 The West Indies are making a first innings reply to
England's total of 448.

Lexical substitutions
These rules replace one lexical item, or several items, with one or more
alternative items. We can group lexical substitutions according to the level
of semantic equivalence between the input and output items. Some items are
virtually equivalent as dictionary entries: resume for restart in (1) above,
lawyer for solicitor, in jail instead of behind bars. Other substitutions are
equivalent in the immediate sentence context:
86 ALLAN BELL

(12) completely ready for full nationhood

~ independence
Still others do not mean the same thing but prove to be true in the context of
the whole story:
( 13) she was later sacked
~ recently
On occasionsl substitutions are made which violate the condition of semantic
accuracy, so that the output story says something different from the input.
Inaccuracies such as (14) are the subject of the second half of this chapter:
(14) emergency measures were being taken to aid flood victims ...
emergency law is being enforced
Syntactic editing rules
Most editing rules are syntactic rules rather than information deletions or
lexical substitutions. I have already outlined how rules are applied to re-
pair ungrammatical structures after constituents have been deleted. Such
structure-mending rules are obligatory and often accomplish major re-
organizations of surface constituents. They are required particularly but not
exclusively after deletion rules. Depronominalization commonly becomes
obligatory after deletion of its antecedent; without such deletion, depro-
nominalization would be an optional reversal.
The kinds of operation by which structure is optionally reorganized in (1)
above are entirely typical of those used in the 290 stories I have analysed.
Variations on co-ordination and de-co-ordination, relativization and de-
relativization are particularly common. Attribution sentences have different
sentences embedded as their complements; an NP of one sentence is
embedded under a preposition as PP in another sentence; relative clauses
are de-embedded to become main clauses (15), or re-embedded into other
sentences; main clauses are embedded as relative or time clauses in other
sentences (16).
(15) Miss Wallace, the main witness, is still in the United States.
~ Miss Wallace is still in the United States and is the main witness.
(16) He said the funeral is to take place next Friday, and Miss Wallace
is attending.
---+ He said Miss Wallace is attending the funeral, which is to take
place next Friday.
In ( 16), the subject NP of the one co-ordinate clause (the fu.neral) is copied as
object of the second co-ordinate clause, and the first co-ordinate then
embedded under the NP as a relative clause.
 SYNTAX AND SEMANTICS OF NEWS EDITING 87
In copy editing, the syntactic choices made in the input text are re-
negotiated. Syntactic editing rules are either applications, reversals, or
alternatives. If the original version chose not to apply any particular rule of
grammar, the editor may reverse this choice and apply the rule. All the
optional rules of a normal grammar are potential editing application rules.
Particularly common are auxiliary and negative contractions, there inser-
tion, passivization, agent deletion, and relative clause reduction. Such rules
belong in any grammar of English, and all the copy editor does is take
options which were not applied previously.
A second group of rules reverses choices which have already been made.
Complementizer that is reinserted; co-ordinate sentences are de-co-ordinated;
sentences are de-passivized and agents reinserted; clefting is reversed, and
there re-deleted; NPs are de-pronominalized. A wide range of optional rules
is subject to editing reversal.
So-called "optional rules" are not the only syntactic rules which offer
choice. Obligatory as well as optional rules may produce several alternative
outputs. In editing, the choice is open to select an output that was not chosen
in the original. Such rules are among the copy editor's most common tools.
Lexical substitutions are in fact a choice of alternatives, as is the substitution
of one tense or aspect by another, and the choice of an alternative article or
complementizer.
The nature and range of editing changes which I have described in this
section appear to be common to all kinds of editing practice. The study by
Davisson et al. ( 1980) compared four texts re-written for US school children
with the original. Their independent categorization of editing changes is
strikingly similar to mine, particularly in the rules needed to describe the
'"splitting" and "merging" of clauses.

Patterned variability in editing

The analysis so far has been purely linguistic: applied to text comparison,
but nevertheless not seeking any social explanation of the linguistic facts.
The re-styling function of editing, however, can only be described in a
sociolinguistic framework. Editing rules are not simply optional: they are
variable rules. Like the standard variable rules of sociolinguistic theory (e.g.
Cedergren and Sankoff, 1974), they are constrained by linguistic and social
factors. I will examine briefly one variable rule and the social factors which
influence its application.
Determiner deletion is the rule which deletes the determiner in a des-
criptive NP which precedes a name NP in apposition to it: 5
 88 ALLAN BELL

( 17) the fugitive financier Robert Vesco

~ fugitive financier Robert Vesco
Copy editors can handle determiner deletion in three ways. They may leave
the copy exactly as they found it. That is, if the determiner is present in the
copy, leave it in; if it is absent, having been already deleted by a previous
editor, then leave it absent. As active alternatives, a copy editor may delete
a determiner which was retained in the input copy; or reinsert a determiner
previously deleted. We can formalize this process through two editing rules:
the editing rule of determiner deletion, which is formally identical to the
ordinary linguistic rule of determiner deletion; and the editing rule of
determiner re-insertion, which precisely reverses the deletion rule.
Table I presents data on determiner deletion for two New Zealand radio
stations. the AAP-Reuter news agency from which they received their
international news, and the NZPA agency which edited AAP for the private
station XA (cf. Fig. 1). In ··national" news originating within New Zealand,
both stations have low determiner deletion. Station YA, the prestigious
.. National Programme". permits no determiner deletion at all. Station XA,
a rock music station. has a level of 14% deletion. However, in inter-
national news the picture is different. AAP, the source of international news
for the other three, has 59. 7% deletion, a much higher level. To shift the
style of the international copy, YA, PA and XA start applying the editing
rule of determiner reinsertion, putting back in the determiners previously
deleted in AAP copy (Fig. 3). YA reinserts almost all the deleted deter-
miners it receives to reach a level of 4.5% in the international news it
broadcasts. The PA agency brings the AAP level down slightly to 53.5% in
the copy it sends out. That frequency is then the input to station XA, which
reinserts rather more determiners to reduce the level further to 40%.

Table /. Total determiner deletion on 2 news agencies and 2 radio stations. New
Zealand.
National International
news news
AAP-Reuter news agency 59.7%
Station YA 0% 4.5
NZPA news agency 53.5
Station XA 14.0 40.0

These shifts are influenced by one clear factor. The frequency of deter-
miner deletion in a station's own internal copy represents a target level.
Copy editors ••know" the frequency of deletion in external input copy,
 SYNTAX AND SEMANTICS OF NEWS EDITING 89
"know" the target level in their station's own copy, and apply editing rules to
shift the input towards that target. 6 The further a station's target level is from
the frequency in the input copy received, the higher will be the level of
application of the editing rule needed to adjust that frequency. In addition,
no editing rule of determiner deletion occurs in the sample. The shift is
always to reinsert the determiners, and never to delete more, because the
target is a low level of deletion. But the editor does not reinsert all the
deleted determiners, so that even after editing the style of external copy is
not identical to internally-originated copy. Editors are prepared to sanction
in external copy forms they would not accept from their own journalists.

100%

-
INPUT from AAP

~ INPUT from PA
-
i
OUTPUT by PA

OUTPUT by XA

OUTPUT by YA ---------
TARGET
t
0% - - - Te-B§.EJ- - -
YA PA XA

Fig. 3. Determiner reinsertion in international AAP copy by New Zealand editors

on 3 news media: station YA, agency PA, and station XA.

The relative frequency of application of variable rules in a station's news

language correlates only with the social composition of the station's audience
(Bell, 1982a). Variable editing rules thus function to shift the style of the
input text closer to the style which the station deems suitable for its kind of
audience. This shows how a number of divergent styles can be derived from
a single text. The same set of international news analysed here was received
and edited by two internal news agencies and six different radio stations in
90 ALLAN BELL

Auckland. These diverse stations have widely differing audiences, and

language styles to match. They take a common input and apply variable
editing rules to shape the style towards their own audience. We can formalize
the goal of copy editing for style in the following axiom:
If the probability of a lingui.stic variable on a given station differs
from its frequency in input copy, editing rules operate to approach
the input frequency to station probability.

Semantic butchery: the failures of editing

This research was begun as a purely formal study of how editors change the
language they work on. However, I was soon struck by the fact that, in the
process of editing news for length and style, copy editors often altered the
meaning of news to something different from the original. This second
section examines inaccurate editing, using the techniques outlined above in
steps 3 and 4 of the methodology.
News-makers, workers and consumers all have a warm, sometimes
passionate, interest in the accuracy of news. Accusations of bias or in-
accuracy in the news media are so commonplace in the political life of
Western countries that they need little documentation. In Britain, the
tensions have surfaced most recently over the 1982 Falklands/Malvinas war
with Argentina. The government criticized coverage of the war by the BBC
and some newspapers as "over-neutral". The Sun newspaper (editorial of
7 May 1982) accused the BBC, Daily Mirror, and the Guardian of treason.
Simultaneously, the Glasgow University Media Group was documenting the
BBC's coverage as overwhelmingly pro-British (Sunday Times, 16 May
1982).
Internationally, debate on this issue has become increasingly sharp in the
past decade. The nations of the "South" or Third World have accused the
media of the "North" of consistent bias in reporting about their countries.
The question of news accuracy is thus important and salient both within
societies and between nations.
Disputes over news bias are, however, rarely resolved satisfactorily.
Opposing parties bring their overt opinions and unexamined ideologies to
an interpretation of the news. The media traditionally maintain that criti-
cism by two opposing sides means they must be getting it right in the middle.
The issue in all such debates invariably turns on one point: What is a true
account of the situation? What is the standard against which particular
reports can be measured? Who is to judge what were the real-world facts:
source, reporter, researcher, an independent expert, a neutral panel of
judges?
 SYNTAX AND SEMANTICS OF NEWS EDITING 91

Studying inaccuracy
In mass communication studies, the investigation of news accuracy was
pioneered by Charnley (1936), who clipped stories from local newspapers
and sent them to sources mentioned in the report to identify inaccuracies.
This and subsequent studies have found on average about one error per
story, with usually about half the stories rated fully accurate. Most re-
searchers have been aware of the risk of having the accuracy of a report
judged by the person being reported on, and some have developed other
measures to identify inaccuracy. Lawrence and Grey (1969) interviewed the
reporter as well as sending questionnaires to the reported.
These developments reflect the problem of how to decide what is an error.
Step 3 of the methodology described earlier solves this by treating the input
copy as if it were fully accurate, and identifying any non-congruence in
output copy as bias or inaccuracy. We thus limit our field to the smaller
compass of copy editing stages (rather than a final report versus "the facts"),
but by that means sharpen our focus and strengthen our inferences.
A second strand of research emerges in recent work in Britain and
Australia. Here linguistic analysis is used to draw inferences concerning the
ideological biases of certain media. The most publicized research (e.g. The
Listener, 29 July 1982) is that of the Glasgow University Media Group (1976,
1980). The Group periodically gets into debate with BBC television over
bias in the BBC's industrial news, or alternatively, in the Glasgow Group's
views. The Group has noted (1976, p. 256) inaccurate editing changes
between successive news bulletins which are of precisely the kind present in
my data.
Research by Kress and Trew is more linguistically oriented. Trew (in
Fowler et al., 1979) analyses contrasting reports in different British national
dailies, and the development over time of reports and editorials about a
single event. Kress and Trew (1978) make independent use of a method-
ology similar to that developed here to study the London Sunday Times'
rewrite of a crucial industrial text. Kress (e.g., 1983) discusses the ideological
bases of journalistic rewriting of news, and compares different reports of
one story in two Australian papers. This research all involves close work on
the language of news reports, and is strongest when it contrasts actual texts
(Bell, 1977, p. 266 ff.). At its best, such analysis can show convincingly
how language is a vehicle of covert interpretation in supposedly neutral
reporting. 7
The third line of research relevant to our study deals with news flow and
selection patterns. White (1950) pioneered research on the "gatekeeping"
performance of copy editors. Some studies have directly questioned gate-
keepers on their reasons for accepting or rejecting stories (Gieber, 1956).
92 ALLAN BELL

Others (Hester, 1971) have combined interview with content analysis of

input and output news. Some have analysed only the news flow, comparing
what stories editors received with what stories are transmitted (Cutlip,
1954). From imbalances in selection, inferences are drawn concerning the
news values of gatekeepers or their institutions at any stage of the news flow
from wire service to newspaper. Such research has concluded, for instance,
that gatekeepers tend to reduce the proportion of news from the countries of
the South, and to select negative rather than positive news (Peterson, 1981).
This work focusses on patterns in the selection of entire stories. There is
little research on the significance of editing changes within stories accepted
for publication. McNelly ( 1959) stresses the importance of the copy editing
role of wire editors, and points out that the term "gatekeeping" is in-
adequate to describe this function. Gatekeepers who are not solely "'copy
tasters,, operate a filter as well as a gate. Once a story is through the selection
gate, it must then pass the cutting and altering filter, and will probably
emerge in a different form. Garrison's research (1979) on the effect of video
technology is the only major study of copy editing. He concentrated on
interviews and observation of copy editors, and compared input and output
copy of 15 stories. The kinds of copy changes were categorized, but too
broadly to bear comparison with my study.

Five types of inaccuracy

I have formalized the requirement for accuracy under step 3 of the method-
ology. Any meaning change requires a warrant in the input copy. The
Semantic Condition is that no editing rule may be applied if its effect is to
make the meaning of the output story in any way non-congruent with that of
the input. Such a condition corresponds to Grice's Co-operative Principle of
conversation, category of Quality (1975, p. 46): "Try to make your contribu-
tion one that is true". Accuracy is not strictly a goal of copy editing: the goals
are to cut and to re-style. Accuracy is rather a condition on the application of
editing rules, a sine qua non of editing. Most of the editing alterations
described in the previous section do not and need not falsify the content of a
news story. But some do, when applied injudiciously.
The inaccuracies identified in my sample fall into five categories. We can
state five sub-conditions of editing accuracy, each of which defines a type of
inaccuracy:
(a) No editing rule shall make the information in the output copy non-
congruent with that in the input copy (inaccuracy type: falsification).
(b) No rule shall increase the assertive strength of a linguistic unit beyond
the warrant of the input copy (over-assertion).
 SYNTAX AND SEMANTICS OF NEWS EDITING 93

(c) No rule shall redefine the scope of a unit beyond the warrant of the input
copy (over-scope).
(d} No rule shall reorder or delete information so that the focus of events
becomes non-congruent with that in the input copy (refocus).
(e) No rule shall add new information not contained in the input copy
(addition). 8
The input copy is thus treated as a canonical text. Potential violations of
the sub-conditions a-e are tested at successively higher levels of the input
copy: constituent, sentence, story; then other available copy, and real-world
knowledge. The basic information unit of news is the story rather than the
sentence. Apparent inaccuracies at the level of the sentence are frequent,
but usually they merely increase or reduce redundancy, or incorporate
information from other sentences of the input. Sometimes a change which
may still appear inaccurate in the context of the input story is warranted by
information in other wires on the same topic. On rare occasions, we must
seek a warrant in sources beyond the news, for instance to check on
geography.
Fa/sification
Falsification, the first type of inaccuracy, is non-congruence between in-
formation in the output and input copy. We test for falsification by turning
the edited version into a question and treating the input copy as its answer.
Any unit to which we must give the answer No, it is not (quite) the case that
X is a falsification:
(18) Jose Lopez Rega, an astrology follower
---+ Jose Lopez Rega, an astrologer
Questioning the output copy in the frame ls it the case that Lopez Rega is an
astrologer? we must reply No, it is not the case: Lopez Rega is an astrology
follower. Most believers are, after all, not themselves priests.
The most extreme case of falsification in my data saw not being deleted, to
leave the outgoing sentence saying the polar opposite of what came in. But
this occurred in ,interestingly mitigating circumstances. Two independent
editors were mi~led by the double negative in did not deny:
(19) AAP The spokesman could not say how many people had
died but did not deny a report by a journalist who
claimed more than 1500 had perished ...

{: ZB
PA
denied
hasdenied
Many "errors of fact" seem to be slips of the tongue or pen - by which 53
people became 83 people, and April 24 went to April 25.
94 ALLAN BELL

Time and place deictics are traps for unwary editors. In overseas news
entering New Zealand, today in the original wire story usually needs to
become yesterday. Copy editors have to resist the temptation to give these
stories a greater but unwarranted immediacy. One item that originated in
London about 6 a.m. New Zealand time was broadcast in Auckland at
6p.m.:
(20) AAP Nineteen of those victims - eight of them children -
still lay in hospital last night. (Sydney: 0951 N .Z.
time)
GNS Nineteen of the victims - eight of them children - are
still in hospital today. (Wellington: 1039) --+
ZBR late today {Auckland: 1800)
Successive changes were made as the day passed, first by the GNS wire
service, then at the next editing step by station ZBR. They wrongly claim
greater recency for an event which is in fact retreating into the past. The
accurate change would have been to early today. 9
Inaccurate approximations of time adverbials usually imply that an event
is more recent than it actually is. Copy editors often reduce detail by
rounding specific figures, or approximating them with quantifiers such as
many, several, some. Occasionally such time approximations are unaccept-
able, again claiming greater immediacy:
(21) Mrs Peron ... was a cabaret dancer 14 years ago
--+ some years ago'°
One of the most regular editing rules changes simple past tense to present
perfect, which increases the immediacx of an event, normally without
falsifying meaning. However, this rule operates on the presupposition that
the events described have only just taken place - within about the last 24
hours. In one interesting case this presupposition fails:
(22) AAP/S5 Mrs Peron granted salary rises to workers on her first
full day in office.
S6 General Peron's widow, as Vice-President to her
husband, assumed the chief executive's spot after
he died on July 1.

~
--+ ZB has granted
--+ GNS has granted
--+ PA has granted
This sentence occurs well down the wire {dated 8 July) describing Mrs
Peron's appointment of a secretary the previous day. A trap has been set in
the middle of this story, probably because at an earlier editing stage two
 SYNTAX AND SEMANTICS OF NEWS EDITING 95
wires have been combined. AAP/SS above is old news, since S6 repeats what
the copy editor should have known - that Mrs Peron became President on
1July1974. Her first full day was 2 July: a week before this story. The only
outlet which avoided this pitfall was station YA, which deleted the whole
sentence. ZB, GNS and PA (the other three outlets which received the AAP
copy direct) all changed simple past tense to present perfect.
Expressions of place are generally less of a problem than time. A New
Zealand copy editor clearly needs to change here in the lead sentence of a
wire datelined London or Addis Ababa. However, editors do sometimes
claim greater proximity of place. In a story about the siting of a nuclear
power plant, about 50 miles south of here [Boston] was edited to just south.
Just does not seem a congruent approximation of 50 miles in such a delicate
matter. In (23), a spatial at-relation becomes beyond, which reduces the
importance of the site of wreckage from a minor aircraft accident in Sydney,
Australia:
(23) over a wide area of the city's northern suburbs
~ over a wide area north of the city
It is a common feature of reference in news items that a first reference is by
name and second reference by description, with no surface marking of the
identity:
(24) The assassination attempt on the President in Washington, D.C.
today has shocked many people in the capital.
A correct reading depends on the knowledge that Washington, D.C. is a
capital city. With countries less well known than the United States, the
editor's geo-political knowledge may fail:
(25) AAP/Sl Rebel troops and police arrested top-ranking military
and civilian officials in Asmara today ...
S2 Armed mutineers led by junior officers held the
radio station, airport and key intersections in the
northern provincial capital of Eritrea, and manned
road blocks outside the town.
PA/Sl Rebel troops and police in the Ethiopian city of
Asmara have arrested top ranking military and
civilian officials ...
S2 Armed mutineers led by junior officers have seized
the city's radio station, airport and key inter-
sections.
S3 They have also seized control of the northern
provincial capital of Eritrea and are manning road
blocks outside the town.
 96 ALLAN BELL

P A/S2 correctly maintains the identity of Asmara and the city. However, the
editor transplants the place adverbial in the northern provincial capital of
Eritrea from AAP/S2 to become object NP of a new sentence. PA/S3 bears
the second, incompatible reading that the northern provincial capital of
Eritrea is a city other than Asmara (or Eritrea may be taken as the name of
the city rather than of the province, on the pattern of the city of Boston). A
two-city interpretation is made possible because the editor has split the
original AAP/S2 into two sentences. Insertion of also in PA/S3 then forces
the reading that this is a second city, different from that in PA/S2. This is in
spite of the form the city in P A/S2, a one-city interpretation of the original
source phrase which should have made the error of P A/S3 impossible.
For a recipient who knows that Asmara is the northern provincial capital
of Eritrea, the PA story becomes just uninterpretable. S3 repeats information
from Sl and S2, while claiming (by also) that it is new information.
Repeating identical information is highly deviant in a news story. In broad-
cast news, maxims such as "Be brief' and "Be informative" (cf. Grice, 1975)
are near-absolutes. At the next editing step, station XI received PA's
version and used it in successive news bulletins. The first three broadcasts
made no significant change, but by the fourth the XI copy editor noticed
something wrong and re-edited PA/S3 to:

(26) They have also seized control of a northern provincial capital and
are manning road blocks outside the town.

XI solves the non-sense by deciding that S3 is indeed new information and

editing to make the two-city reading fully consistent. But in the process, the
last remaining clue to PA's misreading disappears.
i One of the hidden traps of copy editing is that almost any syntactic change
may initiate a semantic change that is not immediately obvious (cf. example
22). XI's version seemed a perfectly legitimate piece of editing but, because
it operates on already inaccurate copy, it compounds the error. We have
been able to identify this case only because an earlier editing step was
available. International news heard on New Zealand radio has passed
through the hands of 3-5 gatekeepers. It is certain that earlier editors have
made inaccurate changes which are compounded at later editing stages.
Without access to the earlier versions, such errors cannot be identified by
later editors or researchers, let alone by the audience.

Over-assertion
The second class of inaccuracy occurs with changes which go beyond the
evidence of the input copy. Lexical items, constituents or whole sentences
may be intensified beyond the point warranted by the input. So we might
 SYNTAX AND SEMANTICS OF NEWS EDITING 97
rank items within the lexical field of walking, roughly from least to most
intense:
shuffle--+ amble--+ stroll-+ WALK--+ stride--+ rush--+ race.
It is generally illegitimate to replace one term of this set with another term to
its right. We identify over-assertion when we have to answer No, it is not
necessarily the case that X to a question about some editing change in the
output copy. A falsification is something contrary to the evidence. Over-
assertion is a change which further evidence could prove to be warranted,
but that warrant is lacking.
Over-assertion is yielding to the ever-present journalistic (and academic)
temptation to make things sound a little better than you know them to be. In
(27), it is not necessarily the case that a horde of newsmen was present:
(27) reporters keeping vigil
--+ a hoarde of newsmen lsic)
Two complementary types of change result in over-assertion. First, a copy
editor may insert an intensifier not warranted by the input copy. Mrs Peron
was reported to have granted salary rises to workers, which one editor
changed to all workers. Again, in the finely-balanced world of Middle East
diplomacy, inserting only into this story regarding the status of the PLO was
unwarranted:
(28) The joint communique declared that the PLO was "the legitimate
representative of the Palestinians, except Palestinians living in
the Jordanian Hashemite Kingdom".
--+ the only legitimate representative
Second, and more commonly, an editor may delete linguistic hedges: devices
which tone down the strength of what is asserted. The original wire in (29) is
careful to hedge its guess about the purpose of the patrol. The output version
strengthens a speculation to an assertion:
(29} Paratroops believed loyal to the government patrolled the air
force base at Debre Zeit, near Addis Ababa, in an apparent
attempt to prevent dissident airmen from joining in the revolt.
--+ to prevent airmen
As Rosenblum points out in his perceptive, insider's book about the work
of the international news agencies (1979, p. 114), correspondents may choose
their words very carefully to cover both what they know and do not know. If
copy editors find the account fuzzy, they may rewrite it into a clearcut but
distorted story. Here is the end of a wire story describing how a piece of wing
broke off an aircraft approaching Mascot airport in Sydney, Australia:
98 ALLAN BELL

(30) AAP Mr W. A. Norton of Dent Street, Lindfield, saw the

incident on his way to work.
He said the jet continued towards Mascot without any
apparent problem.
-+ ZB The jet continued to Mascot without any apparent
problem.
-+ GNS The aircraft apparently landed without any trouble.
-+ PA The jet landed without any trouble.
The original is a sourced eye-witness account that sticks scrupulously to
reporting what Mr Norton actually saw. It hedges this around with the words
continued, towards and apparent, and makes no commitment on whether the
aircraft eventually landed safely. But the three separate editors who used
that story started chopping out the hedges. ZB replaced towards with to.
GNS took it further and used landed for continued. The PA editor removed
the last remaining hedge, apparent, and stated bluntly that The jet landed
without any trouble.
This version goes well beyond the original wire, and has been made
possible only because the copy editors have removed all reference to the
eye-witness. The first AAP sentence, and consequently the attribution he
said in the second sentence, have been deleted. If the editors had tried to
make their changes without deleting the source, they would have seen that
the resulting sentence was unacceptable:

(31) Mr W. A. Norton of Dent Street, Lindfield, saw the incident on

his way to work.
He said that ...
? the jet continued to Mascot without any apparent problem.
* the aircraft apparently landed without any trouble.
** the jet landed without trouble.
Mr Norton was in no position to say anything of the kind, because he was ten
miles from the airport on the other side of Sydney harbour.
The most common device by which journalists hedge their assertions is
quotation and attribution. Wire stories frequently carry direct quotations
from spokespersons or public figures. Copy editors usually delete quotation
marks and often this does not matter. For read-aloud news, direct quotation
is obviously a problem. Frequently quotation marks remain in edited copy,
and it is then the newscaster's responsibility to pick this out through
intonation. Sometimes quotation is marked syntactically:

(32) Egypt and Jordan today agreed that the Palestine Liberation
Organisation (PLO) should attend the Geneva Middle East
 SYNTAX AND SEMANTICS OF NEWS EDITING 99
peace conference "at the appropriate stage" as an independent
body.
-+ at what they call the appropriate stage
The function of quotation markers is to dissociate the sender from direct
responsibility (cf. Kress, 1983). When they are deleted and not marked
by intonation, the broadcasting station adopts the expression as its
own:
(33) AAP They said the Israelis left behind leaflets warning local
inhabitants against cooperating with "terrorists".
ZB terrorists
YA the terrorists
Many news stories involve someone being quoted: an eye-witness, spokes-
person, observer, .or other source. As we have seen, copy editors often
take out these attributions. However, attribution deletion gives a false air of
impersonal authority to what is said. And it usually sets off a chain reaction
of deletion or alteration in later sentences of a news story, where it may be
even less acceptable than in the earlier sentence. A news story can be seen as
embedded under a stack of attribution sentences:

(i) Station XA reports in Auckland at 1200 that

(ii) PA reported in Wellington at 1100 that
(iii) AAP reported in Sydney at 1000 that
(iv) Reuters reported in London at 0900 that
( v) Correspondent PQ reported in Jerusalem at 0800 that
(vi) An Israeli military spokesman said in Tel Aviv at 0700 that ...

Sentence (i) is always implied in the very fact of station XA broadcasting the
story. We do not expect all the other editing steps to be explicitly acknowl-
edged, but source (vi) and agency (iv) should be.
When a statement is attributed to a source, the listener is explicitly alerted
to the fact that this is one individual's description and viewpoint. The
practice of deleting attributions becomes particularly undesirable when a
story involves sharply opposed groups. Here it is vital to know which side is
the source of information so we can make allowance for its viewpoint:
(34) Israeli naval units tonight raided three Lebanese ports sinking
about ten vessels in each of them, the Israeli military spokesman
announced.
-+ Israeli naval units have raided three Lebanese ports sinking
about ten vessels in each of them.
It is a common feature of war reporting that one event produces two
100 ALLAN BELL

irreconcilable versions, and sets of statistics, from the opposing sides. 11

Other (unused) reports gave the Lebanese account of this incident: 21 boats
sunk, not 30. Neither figure may be correct: certainly both cannot be.
Preferably both counts should be reported, but at least the statistic should be
anchored to its source.
On a higher level of attribution, most wire stories are sourced to the
originating news agency (iv above). This is almost invariably deleted on New
Zealand radio, for obvious reasons. Repeating the source would slow down
news presentation. Attributions like according to Reuters or Associated
Press reports can be easily removed without adjusting the rest of the lead
sentence or even altering its remaining structure. But by deleting all attribu-
tion of a story to its source, the editing station makes itself responsible for
the content of that item. It says We tell you here and now that X, and takes
responsibility for errors inherited from the wire copy. The significance of
this becomes clear in analysis of the two-step editing process from AAP to
PA to XA and XI. Over 80% of editing errors between the AAP wire and
XA/Xl's broadcast versions occur at the PA stage. But XA and XI vouch for
this previous editing, first by virtue of broadcasting it, and then by removing
any explicit disclaiming attribution.

Over-scope
In the third type of inaccuracy, the scope of information is mistakenly
broadened beyond the warrant of the input copy (or sometimes narrowed:
"under-scope"). If place or time adverbials are shifted or deleted, the scope
of the sentence or phrase may be over-extended or over-restricted. We
identify over-scope by questioning the time or place adverbial in the frame
of the output sentence.
An interesting group of scope redefinitions occurred in a story about
flooding in southern Brazil. Most of the details in the early sentences of the
original report concern only one state, and one city in the state, while the
floods also affected at least six other states of Brazil:
(35) AAP Serious flooding in southern Brazil was today reported
to have brought a heavy death toll in the town of
Tubarao, cut off by swollen river waters ...
Authorities in Santa Catarina State where Tubarao is
situated have declared the state a disaster area ...
The death toll could be over 1500.
~ ZB A possible death toll of more than a thousand is
reported from Brazil in the flooding in the south of
the country.
The area's been declared a disaster area ...
 SYNTAX AND SEMANTICS OF NEWS EDITING 101

In the attempt to abbreviate the story, specific reference to Santa Catarina

state and Tubarao city has been deleted from the early part of the output
version, but the other facts stand as if they referred to the whole situation. So
the floods appear less serious than they are: the death toll of more than a
thousand actually applies to one city of many hit by the floods. Later
paragraphs of the AAP wire show that large areas of northern, western and
eastern Brazil were also flooded, with great devastation. On the other hand,
Santa Catarina State has been quasi-pronominalized to the area, which then
appears in ZB 's version to refer to the whole of the south of the country. The
rule has been applied where there is no true identity, thus over-generalizing
the extent of the official disaster area. So ZB's alterations both over-restrict
and over-extend the scope of the place adverbials.
Other editors treated the story to imply that the flooding was more limited
and its consequences less disastrous than in fact. One version deleted all
specific place adverbials:
(36) PA More than 1500 have perished in the city of Tubarao
in southern Brazil.
~ XA More than 1500 people have perished.
We can often identify over-scope by questioning the output copy in a frame
such as:
(37) Is it the case that 1500 people have perished in general?
Using the input to (36) as answer, we must say that it is not the case in
general, only in Tubarao. Hardly a single New Zealander would know the
name of that city or state. South America is very remote from New Zealand,
not so much in distance (they are after all "neighbours" across 6000 miles of
the Pacific Ocean) as in consciousness. Yet, it would be unthinkable for a
copy editor to treat so casually the geography of more distant but better-
known regions such as the United States or western Europe.
Time over-scope occurred in the editing of a story about a visit to a British
regiment in Germany which Princess Anne made shortly after her car was
ambushed in London (March, 1974):
(38) AAP/S4 It was her first trip abroad since the abortive kidnap
attempt in London last week ...
S6 British army and West German police security men
worked together round-the-clock in providing
maximum security for Anne.
S7 A helicopter, plain-clothes detectives, German police
with tracker-dogs and British military police with
submachine guns were constantly on guard around
her.
102 ALLAN BELL

~ P A/S3 It was her first trip abroad since the abortive kidnap
attempt in London last week.
S4 A helicopter, plain-clothes detectives and military
police with submachineguns have been constantly
guarding the Princess since the kidnap attempt last
week.
Most of the really serious editing errors (e.g. 25) require the copy editor to
adopt a single-mindedly perverse reading and make a cluster of changes for
the interpretation to stick. This example is no exception. The offending rule
copies a reduced version of the time adverbial since the abortive kidnap
attempt in London last week from the first sentence on to the last, over-
generalizing the information in the last sentence. The original twice specifies
British army and West German police, and indicates explicitly that the
precautions described were in force only on Princess Anne's three-day trip
to Germany. To make its transference of the time adverbial acceptable, PA
has to delete AAP/S6 completely, and cut from S7 all reference to German
police. The resulting version gives the quite false (but rather more sensa-
tional) information that, since the time of the ambush, for the four days in
Britain as well as in Germany, Princess Anne was provided with this intense
security.

Refocus
Refocus and addition are the two less frequent types of inaccuracy. Refocus
involves the reordering or deletion of information so that the balance of a
story is no longer congruent with the input copy. Such distortion is classically
difficult to "prove" when looking at a report in isolation, but it is made far
more visible when editing actively refocusses a story. Refocus is the prime
cause of non-congruency in the following story:
(39) PA/Sl Argentina's new president - Isabel Peron - has named
her late husband's former aide as her personal
secretary.
S2 He's Jose Lopez Rega, an astrology follower, who
also managed Mrs Peron when she was a cabaret
dancer fourteen years ago.
S3 He was a constant companion of the late General
since 1965 ...
~ XA/Sl Argentina's president - Isabella Peron - has named
her personal secretary - an astrologer who managed
her career when she was a cabaret dancer some
years ago ...
S2 Isabel, who takes over in Argentina following the
 SYNTAX AND SEMANTICS OF NEWS EDITING 103
death of President Juan Peron last year - ah, last
week. [sic]
S3 The new personal secretary is Jose Lopez Rega, a
close companion of the late president and his wife
since 1965.
There are three refocusses in XA's version:
(a) The relative clause an astrology follower who also managed Mrs Peron
when she was a cabaret dancer fourteen years ago has been shifted from
the second sentence to the lead sentence.
(b) The information that Lopez Rega had been Mrs Peron's late husband's
former aide has been deleted from the first sentence.
( c) The pronoun he is replaced in S3 by the full noun phrase the new
personal secretary, repeating material from the first sentence.
Together with no less than seven other changes which affect the meaning, 12
these highlight the astrology and cabaret dancer motifs (a), and refocus
Lopez Rega's past relationship from General Peron to Mrs Peron, streng-
thening the innuendo of sexual liaison (b, c).These editing alterations have
a considerable cumulative effect on the tone of the story. Argentina is made
laughable: a country run by an astrologer and ex-cabaret dancer who share
an invitingly illicit degree of personal intimacy.

Addition
Lastly, information addition occurs surprisingly often, but only occasionally
is the editor's new information wrong. In (40) there is no warrant for the
insertion of police in either of the output sentences:
(40) Miss Brice's body was found ...
She had been brutally beaten.
--+ Police found the body ...
They say she died after being brutally beaten.
One editor, exasperated with the inconclusiveness of a five-day cricket
match in Guyana which was continually interrupted by rain, prejudged the
result and rewrote a story to say the match had ended:
(41) AAP Play was again held up at the start of the final day of
the Fourth Test between the West Indies and
England here today.
--+ PA In the West Indian city of Georgetown the final day of
the Fourth Test between the West Indies and
England has been washed out by rain, resulting in a
draw.
104 ALLAN BELL

However, the match did resume - had in fact already done so before PA
received this wire in Wellington. Two hours later the same editor unblush-
ingly put through an update wire saying that the game was underway. And
once again, station XI suffered from PA's bad editing, and compounded fhe
error unawares by repeating at the end of the inaccurate item: That result
again - a draw.

Inaccuracy and news values

The examples of inaccuracy cited include a few of the worst among some 150
editing errors in the sample of 290 stories. We now tum to identify what are
the patterns in these inaccuracies (step 4ofthe methodology). What kinds of
stories are poorly edited? Do any social or political factors seem to promote
inaccurate editing? What sorts of semantic effects do editing changes have?
Why do editors make inaccurate changes?
The semantic effect of inaccuracies can be largely described under one
heading: they make a story more newsworthy. I have already noted in
passing that many changes to time adverbials make a story sound more
recent than it really is. The concept of recency is just one news value or
factor which media researchers such as Galtung and Ruge (1965) suggest are
influences on the selection of news. In my data, we see these factors in-
fluencing the editing of accepted news, as Galtung and Ruge hypothesized
they would (1965, p. 71). In almost every example given above, the in-
accuracy clearly enhances the story's value on some news factor. Here I can
only sketch findings relevant to a few of the factors:
Frequency includes what I have called "recency". Certain inaccuracies serve
to make events conform to the established cycles of news work, e.g. by
prejudging the result of a cricket match (example 41) or claiming events as
more recent (20, 22).
Unambiguity is an asset which promotes the selection of a news story. This
has two repercussions on editing. Firstly, unambiguous, clearcut stories are
less liable to be badly edited. If a story is ambiguous, complicated, diverse, it
is more likely to suffer, as did the complex, updating wires on the Brazil
floods. And second, copy editors are likely to commit inaccuracies in the
process of making complicated stories less ambiguous. Maybe's become
certainties (29, 30), open sets are closed (28). Attribution deletion is the
classic case, since acknowledging a viewpoint, and particularly conflicting
viewpoints, necessarily makes a story ambiguous (33, 34).
Negativity of a news story is frequently increased during editing: from the
potential chaos of Peron's government (39), to the threat to Princess Anne's
 SYNTAX AND SEMANTICS OF NEWS EDITING 105

safety (38), to broadening the scope of the Ethiopian rebellion (25) or the
Brazilian floods (35).
Consonance of a story with preconceptions about a nation or region affects
which stories are badly edited. Many inaccuracies serve to shift stories
towards greater consonance with stereotypes, particularly for the Latin
American examples cited (14, 19, 39).
Recent research shows how the unwitting "frame" or ideology of news
workers affects the selection and form of the news (e.g. Tuchman 1978,
Glasgow University Media Group 1976). The findings presented here make
it clear that the frame, and the news factors which are part of it, influence the
transmission as well as the creation of news. Editing inaccuracies are not
unpatterned accidents. They serve to make news stories more newsworthy,
however unconscious copy editors may be of the errors and their effect.

News of the "South" and the New Information Order

The bias becomes even clearer when we quantify inaccuracies. I score the
seriousness of each editing error by assigning it a value on a scale from fully
accurate (at 0) to wildly inaccurate (at 6), in the manner of the semantic
differential (Osgood et al., 1957). A news story fully accurately edited
scores 0. A story with one very serious editing inaccuracy, or several smaller
errors, scores 5-6, with two very serious inaccuracies about 10-12. 13 We can
then examine the data on parameters such as originating source, topic
classifications, country of origin, or the news factors mentioned above. I
present data on only one dimension here: country of origin.
Grouping stories by country of origin, we sum the scores of editing
inaccuracies in news about each country, and divide the total by the number
of stories about that country. This yields a measure of editing inaccuracy per
story concerning these countries. Table II presents indexes of editing
inaccuracy for those countries with an adequate number of stories in the
sample. News about Brazil, Ethiopia and Argentina has been exceptionally
badly edited: 2.2-3.7 on the 0-6 point scale, the equivalent of one serious
error per story. Individual stories about these countries score very heavily:
12 for PA's editing of the Brazil floods, 14 Ethiopian revolution (PA), 10
Mrs Peron (XA). These scores represent the equivalent of two or more
extreme inaccuracies per story. The other countries were all Western except
the Soviet Union, and were all passably well edited.
Clearly, there has been a considerable difference in editing quality
between news about the nations of the North and of the South. This finding
accords with studies on the quality of gatekeeping of Third World news.
Hester (1971) has analysed how US copy editors select from the Associated
106 ALLAN BELL

Table JI. Indexes of editing inaccuracy per story for news from 12 countries (0-6
scale). A low score indicates relatively accurate editing, a high score inaccurate
editing.
Inaccuracy Number of
index stories
Brazil 3.7 9
Argentina 2.7 7
Ethiopia 2.2 9
France 1.4 9
United States 1.1 44
Great Britain 0.9 56
Australia 0.7 34
Soviet Union 0.6 14
South Africa 0.5 9
Canada 0.3 10
Netherlands 0.2 12
Rhodesia (now Zimbabwe) 0.1 8

Press wires they receive. "Developing" countries accounted for less than
one third of the news on the AP wire, and this proportion was further
reduced at each editing stage. The wire editors whom Hester interviewed
asserted that reader interest in Latin America and Africa was especially low.
Hester (1974) analysed the flow of news from Latin America to the United
States. The "criminal/violent" category of news made up 14% of what was
sent from Latin America, but 48% of the Latin American news that
Associated Press passed ·on to its client newspapers. Peterson (1981)
studied how The Times of London selects its foreign news. She concluded
that even this elite among newspapers tended to ignore nations of the South
unless conflict occurred there.
The North's reporting of the South has been the subject of increasingly
vigorous argument in international forums such as Unesco over the past
decade. Nations of the South have accused the Western media, especially
the "Big Four" international news agencies, of under-reporting their
countries and concentrating on negative news. Western media have in tum
accused their opponents of advocating censorship and destroying the free-
dom of the press. The debate resulted in the MacBride Report (1980)
presented to U nesco by the International Commission for the Study of
Communication Problems and in Unesco's adoption of a New World
Communication and Information Order.'"'
Editing inaccuracy is cumulative. We have seen how an error at one
editing stage may be compounded at later stages. Quantifying inaccuracies
for the New Zealand agencies/stations shows that the PA agency averaged
2.1 per story on the 0-6 inaccuracy scale. Radio station XA scored 0.9. Ifwe
 SYNTAX AND SEMANTICS OF NEWS EDITING 107
add the two scores, we have an inaccuracy index of 3.0 for these two
successive editing steps within New Zealand. To put this in perspective,
recall that New Zealand hears its international news only after it has passed
through 3-5 editing filters. If the earlier steps are no more nor less accurate
than the last two, the average international news item broadcast on New
Zealand radio may contain the equivalent of one very serious inaccuracy as a
result of copy editing alone, quite apart from the reliability of the original
report. If some wires are any indication, a few stories will contain as many
as three or four major editing errors. My findings on editing inaccuracy lend
support to the criticisms the South has made. The results are more specific
and persuasive than much of the evidence available on news proportions,
selection, etc. News about the South has been distorted during editing,
however 'unintentionally.
The editing steps I have analysed are only the last in the news flow. Here
inaccurate editing has patterned according to the same factors which bias
news selection. Since these factors have also been shown to affect selection
at the earlier stages of international news flow, it is likely that they influence
how the international news agencies edit their copy. From accounts such as
Rosenblum (1979, p. 113), the scale of rewriting at the agencies' own desks
in New York (AP, UPI), London (Reuters) or Paris (AFP) seems to be
greater than at the later editing stages I have analysed.

The contribution of sociolinguistics

We have applied (socio )linguistic analysis to news editing and found evidence
for the kind of inaccuracy of which news media are so often accused. What
can sociolinguistics do to help solve the problem? There are three com-
ponents to the problem of inaccurate editing: practical, technical/linguistic,
and social/ideological.
The practical problem is the time/space pressure on editors: the time
before the next deadline, and the space available for news. Radio editors
handle large amounts of copy under the pressure of often hourly deadlines.
They receive copy which is intended for newspaper use, and re-style it to be
read aloud. Above all, they have to abbreviate. Storiesof2000wordsarecut
to 50-100 words, complexities simplified, detail removed. Some errors
result from ignorance (e.g. of geography), others are pure slips of the pen.
These are matters of working conditions and ethos. Sociolinguists can do
nothing here except point to the evidence of inaccuracies apparently caused
by haste (e.g. example 19) or over-abbreviation (35).
On the linguistic component of inaccurate editing, linguists can offer
something. "Technical" failure presumably plays a part in all inaccurate
108 ALLAN BELL

editing, in some cases very obviously (e.g. 25, 38). We can move from
description to prescription, and propose linguistic guidelines by which copy
editors can avoid certain inaccuracies. A very small percentage of the
hundreds of changes made by copy editors result in inaccuracy. There are
few editing rules which are essentially bad, which in themselves necessarily
cause inaccuracy. Deleting or inserting negatives are such non-rules, for the
obvious reason that a statement and its negation cannot be compatible. And
few rules are obligatory, e.g. updating time adverbials.
Most inaccuracy results when legitimate rules are applied without regard
to the congruence of output copy with its input. From the analysis and
examples above, we can identify particular editing rules which typically lead
to one of the five types of inaccuracy. Under the five conditions for accuracy
(a-e, page 92), we can group a number of specific positive and negative
guidelines for copy editors (with examples where the guidelines were
violated):
To avoid falsification:
(i) Don't delete or insert negatives (as in example 19).
(ii) Do delete or update time adverbials, or convert to non-deictics (20).
Today in the lead sentence of a wire story must usually be updated (to
yesterday), converted to a non-deictic (on April 25), or deleted
altogether.
(iii) Don't approximate time and place adverbials with non-congruent
expressions (21).
(iv) Don't substitute non-congruent lexical items (18).
(v) Don't delete or insert expressed agents.
(vi) Don't reverse the transitivity of verbs.
Copy editors often change an active intransitive verb into its passive
transitive equivalent, or vice versa. This implies either the insertion or
deletion of an agent, which may not always be warranted by the input
copy. P A's version in (42) deleted the expressed agent: the legislative
acts which increased police powers in South Africa. The transitivity of
increasing is now ambiguous. XA's subsequent change to active in-
transitive implies that police powers increased by natural growth.
(42) AAP various new acts which had increased police powers
~ PA an increasing of police powers ~
~ XA increasing police powers
To avoid over-assertion:
(vii) Don't delete attributions (30, 34).
I have argued that copy editors should retain source and agency
attributions. It is most important that source attribution (eyewitness,
 SYNTAX AND SEMANTICS OF NEWS EDITING 109

spokesperson, etc.) should be retained. This reminds listeners that the

news has a specific, human source . .It is a report, just one version of
events. Agency attributions are less obviously vital, but in my view still
important. If agency attributions were always retained, audiences
would have the chance to recognize two facts: that most of the world's
news originates in the Big Four agencies, and that most overseas news
enters New Zealand on a single wire. In many other countries, both
news sources and outlets are more diverse. In New Zealand, the public
can read a story in the morning paper, hear a parallel account on radio,
and watch it again on evening television news. Multiple reinforcement
from apparently independent outlets makes the story and its details
seem unassailably authoritative. But many witnesses do not make one
truth when they are all second-hand accounts based on a lone wire
containing unknown errors of fact and interpretation. Attributing a
story to its source agency would at least make this situation explicit.
Finally, the problem of non-attribution is aggravated when there is
double deletion, of both agency and spokesperson. In (43), the Press
Trust of India has taken particular care not to give its own backing to
the statement:
(43) Thousands of Indians were reported homeless today as flood
waters continued to rise in the north and northeast, inundating
large areas of standing crops, the Press Trust of India said.
The Agency quoted officials as saying several thousand hectares
of jute and rice crops have also been destroyed ...
~ Thousands of Indians are reported homeless as flood waters
continue to rise in the north and north east.
Several thousand hectares of jute and rice crops have also been
destroyed.
Retaining attributions will slow down broadcast news presentation,
but it need not be repetitive. The copy editor has a wide choice of
lexical variants, and several syntactic options: foregrounding (PQ said
that X); backgrounding (X, said PQ); passivizing (PQ is said to be X);
relativising (PQ, who is said to be X); reduction to according to PQ. A
lead sentence can credit the agency, perhaps sometimes in headline
form: a report from Reuters.
(viii) Do mark direct quotations by syntactic devices or intonation (33, cf. 32).
As with attribution deletion, this is especially necessary in conflict
situations.
(ix) Don't delete hedges.
In pursuit of less ambiguity, editors are tempted to delete items like try
or attempt {29), or to describe incomplete actions as accomplished
110 ALLAN BELL

(41). Hedging devices tend to cluster, and they are often deleted in
clusters as in (30), thus compounding the over-assertion.
(x) Don't insert intensifiers (28).
(xi) Don't intensify lexical items beyond what the input story warrants (27).
To avoid over-scope:
(xii) Don't pro-form non-identical referents (35).
(xiii) Don't delete, shift or transplant constituents which define the scope of a
sentence or constituent (36, 38}.
To avoid refocus:
(xiv) Don't!
To avoid information addition:
(xv) Don't!
The scale of this study is too small to produce adequate guidelines for
accurate editing, yet following the guidelines given here would eliminate
many inaccuracies. The prime rule of thumb for working editors must be:
when in doubt, don't. As Rosenblum (1979, p. 112) notes, "any alteration
increases the chance for inadvertent error". Severe copy editing can be
compatible with accuracy. Station YA edited heavily but very accurately
(index 0.5 on 0-6 scale). But one PA editor who edited as heavily scored a
quite astonishing 4.2 for inaccuracy, an extreme error per story. Editors who
make few alterations give themselves less chance to produce an inaccurate
version.
Yet inaccurate editing is not mainly a practical or linguistic problem.
Reducing time/space pressures and upgrading language skills should reduce
the amount of inaccuracy over-all. But it will not redress the imbalance in
where the inaccuracies occur. If inaccurate editing were purely a practical or
linguistic matter, all types of news would be affected alike. Yet we have seen
this is not so: inaccuracies are not randomly distributed, but concentrate in
certain categories of news (e.g. about some countries), and distort news in
consistent ways (to make it more newsworthy). This can only result from the
frame or ideology within which the editors are working.
Here sociolinguistic analysis can make its second contribution. It provides
harder evidence than is possible in other methods of investigating inaccuracy.
Newsworkers are noted for receiving researchers and their findings with
hostile disbelief. 15 But editing inaccuracy constitutes a failure by the copy
editor's own professional standards. The copy editors who introduced them
would themselves condemn these inaccuracies if made aware of them in
retrospect. Pointing out the bias in where and how the inaccuracies occur
should stand a little better chance than most arguments of persuading
newsworkers that there is indeed a problem.
 SYNTAX AND SEMANTICS OF NEWS EDITING 111

The third contribution would be in further studies of news editing in a

wide range of media and places. The accuracy of international news editing,
especially by the major agencies, is important if nations are to have a fair
picture of each other. The bad editing of news about the nations of the South
is of particular urgency and significance for the issues surrounding Unesco's
New World Information and Communication Order. Research in this area
can promote justice in news reporting between the South and the North.
More studies can gather more evidence to present to newsworkers. And the
findings need to be publicized so that audiences are alert to the human
fallibility of news.

Conclusion
The methodology of this study limits its range to defined stages of news
transmission, concentrating on the paired texts of input and output copy. It
examines not just selection among arguably equal alternatives, but the
active, purposeful intervention of the copy editor in the form and meaning of
news copy. Copy editors do not merely chose among an open set of options.
They actively, if unconsciously, substitute alternatives already rejected,
reverse choices already made by the source journalist.
Because this process is one of active intervention, where copy alteration is
the subject of a focussed choice, the force of the patterns which emerge is
much more striking. In the syntactic analysis, we saw how editing can
re-style a text towards a target considered suitable for a particular audience.
In the semantic analysis, we saw the active substitution of meanings which
are incongruent with the original. Inferences of bias according to social/
political factors made on this evidence are specific and persuasive.
Any situation where one text is transformed into another can be subjected
to this kind of analysis. It can be applied to any stage of the news flow; to the
effects of news censorship; to the editing of other mass communication
content such as books or television scripts; to interpretation and translation
from one language to another in news transmission, diplomacy, business; to
re-writing texts for Ll or L2 learners; to the captioning of TV programmes
for the deaf. Such research depends on social and political conditions to
define its issues, on a sound and detailed linguistic analysis for its evidence,
on media sociology for its interpretation, and on newsworkers for its
application.
112 ALLAN BELL

References
Bailey, B. (1971). Jamaican Creole: can,dialect boundaries be defined? In "Pidgin-
ization and Creolization of Languages", (D. Hymes, ed.), pp. 341-8. Cambridge
University Press, Cambridge.
Bell, A. (1977). The Language of Radio News in Auckland: a Sociolinguistic Study of
Style, Audience and Subediting Variation. Unpublished Ph.D. thesis: University
of Auckland (University Microfilms, Ann Arbor, Michigan, 1979).
Bell, A. (1982a). Radio: the style of news language. Journal of Communication 32,
I, 150-164.
Bell, A. (l982b). This isn't the BBC: colonialism in New Zealand English. Applied
Linguistics, 3, 3, 246-58.
Berry, F. C., Jr. (1967). A study of accuracy in local news stories of three dailies.
Journalism Quarterly, 44, 3, 482-90.
Boyd-Barrett, 0. ( 1980). •The International News Agencies... Constable, London.
Burns, T. ( 1977). "The BBC: Public Institution and Private World", Macmillan,
London.
Cedergren, H. and Sankoff, D. (1974). Variable rules: performance as a statistical
reflection of competence. Language, 50, 2, 333-55.
Charnley, V. M. (1936). Preliminary notes on a study of newspaper accuracy.
Journalism Quarterly, 13, 4, 394-401.
Crystal, D. (1981). "Directions in Applied Linguistics", Academic Press, London.
Cutlip, S. M. (1954). Content and flow of AP news-from Trunk to TIS to reader.
Journalism Quarterly, 31, 4, 434-46.
Davison, A., Kantor, R. N., Hannah, J., Hermon, G., Lutz, R. and Salzillo, R.
(1980) ... Limitations of Readability Formulas in Guiding Adaptations of Texts"
(Technical Report No. 162). Center for the Study of Reading, University of
Illinois at Urbana-Champaign.
Fowler, R .• Hodge, B., Kress, G. and Trew T. (1979). "Language and Control",
Routledge & Kegan Paul, London.
Galtung, J. and Holmboe Ruge, M. (1965). The structure of foreign news.Journal of
Peace Research, 2, 1, 64-91.
Garrison. M. B. (1979). The Video Display Terminal and the Copy Editor: a Case
Study of Electronic Editing at the Milwaukee Journal. Unpublished Ph.D.
dissertation: Southern Illinois University at Carbondale.
Gieber, W. (1956). Across the desk: a study of 16 telegraph editors. Journalism
Quarterly, 33, 4, 423-32.
Glasgow University Media Group. (1976). "Bad News", Routledge & Kegan Paul,
London.
Glasgow University Media Group. (1980). "More Bad News", Routledge & Kegan
Paul, London.
Grice, H. P. ( 1975). Logic and conversation. In '"Speech Acts" ( ..Syntax and
Semantics". Vol. 3) (P. Cole and J. L. Morgan, eds), pp. 41-58. Academic
Press, New York and London.
Hester, A. ( 1971). An analysis of news flow from developed and developing nations.
Gazette, 17, 1, 29-43.
Hester, A. (1974). The news from Latin America via a world news agency. Gazette,
20, I, 82-98.
 SYNTAX AND SEMANTICS OF NEWS EDITING 113

Klima, E. S. (1964). Relatedness between grammatical systems. Language, 40, 1,

1-20.
Knightley, P. (1975). "The First Casualty", Andre Deutsch, London.
Kress, G. (1983). Linguistic processes and the mediation of "reality": the politics of
newspaper language. International Journal of the Sociology of Language, 40,
43-57.
Kress, G. R. and Trew, A. A. (1978). Ideological transformation of discourse: or
how the Sunday Times got its message across. Journal of Pragmatics, 2, 4,
311-29.
Lawrence, G. C. and Grey, D. L. (1969). Subjective inaccuracies in local news
reporting. Journalism Quarterly, 46, 4, 753-7.
MacBride, S. et al. (1980). "Many Voices, One World", (Report by the International
Commission for the Study of Communication Problems). Unesco, Paris, and
Kogan Page, London.
McNelly, J. T. (1959). Intermediary communicators in the international flow of news.
Journalism Quarterly, 36, 1, 23-26.
Osgood, C., Suci, G. and Tannenbaum, P. (1957). "TheMeasurementofMeaning",
University of Illinois Press, Urbana, Illinois.
Peterson, S. (1981). International news selection by the elite press: a case study.
Public Opinion Quarterly, 45, 2, 143-63.
Righter, R. (1978). "Whose News? Politics, the Press, and the Third World",
Burnett Books/ Andre Deutsch, London.
Rosenblum, M. (1979). "Coups and Earthquakes", Harper & Row, New York.
Schlesinger, P. (1978). "Putting 'Reality' Together: BBC News", Constable, London.
Smith, A. (1980). "The Geopolitics of Information", Faber & Faber, London.
Tuchman, G. (1978). "Making News: a Study in the Construction of Reality", Free
Press, New York.
White, D. M. (1950). The "Gatekeeper"; a case study in the selection of news.
Journalism Quarterly, 21, 4, 383-90.

Notes
Allan Bell was Leverhulme Visiting Fellow in 1982 at the University of Reading,
England, on leave from New Zealand, where he works as an editor and sociolinguist.
Parts of this paper were presented at the Second New Zealand Linguistics Con-
ference, Wellington, 1978; the Center for Applied Linguistics, Washington, D.C.,
1981; and the lOth NW A VE Conference, Philadelphia, 1981. Thanks to Walt
Wolfram and Joy Kreeft for their comments on earlier drafts. I am grateful for the
financial support of the Leverhulme Trust, and the hospitality of the Department of
Linguistic Science, University of Reading, especially PeterTrudgill.

1
I use the post-1975 label "Radio New Zealand" for the radio division of the former
New Zealand Broadcasting Corporation. Since 1974, a further private station has
been added, otherwise the Auckland radio scene is not much different today. The
wire system, however, has changed in one major respect. NZPA no longer edits the
AAP-Reuter wire for private radio stations, which now take a Reuters wire direct.
For information on the flow of news in New Zealand and for access to wire copy, I am
I 14 ALLAN BELL

indebted to staff members of Radio New Zealand (especially Ken Gibson), Radio
Hauraki, Radio i, Radio Windy/Wellington (Fraser Foister), N.Z. Press Association,
N. Z. Herald, and Auckland Star.
1
The term •'copy editor" is used rather than the normal New Zealand (and British)
"sub-editor" (or just ..sub"}. The label "sub-editor" is opaque to non-newsworkers
everywhere, who understandably misinterpret it to mean a deputy editor. In many
news organizations, the person who selects stories to pass to the sub-editors is called a
"copy taster". Research on news flows (e.g. White, 1950) labels these people the
•·gatekeepers". Remarkably few people are involved in the editing process: perhaps
only 5-10 even at major ••gates" in the system (Boyd-Barrett, 1980, p. 78).
Editing rules have similarities to rules proposed in two other studies. Bailey (1971)
suggested that the distance of a text from standard Jamaican English and from
Jamaican creole could be measured by counting the number of rules necessary to tum
the text into one or other dialect. Klima {1964} proposed a class of "extension rules"
to be added to the grammar of one style to generate other styles .
.i I would expect the use of video technology to alter the pattern of sentence deletion,

for instance. Video makes certain kinds of actions (e.g. viewing later pages of a wire)
harder to perform, so that a story of several pages cannot be laid out and seen at a
glance. Video editing may encourage even further the rejection of later pages of
copy. Interestingly, Garrison ( 1979) found that many copy editors liked to work from
printed copy as well as the video screen, especially for substantial rewriting.
5
For detail on the rule and its social significance, see Bell {1977, 1982b). As well as
deleting the determiner, the rule subordinates the (first) descriptive NP to the
(second) name NP. Variable syntactic constraints include determiner type, and
presence and complexity of embedding in the descriptive NP.
'' On the precise nature of the copy editor's linguistic knowledge I do not wish to
speculate here. While it is clearly closer to the surface of consciousness than is most
language use, it is by no means as self-aware as it may appear. Copy editors work
largely by unexamined intuition. In my experience, they prove to be frustratingly
unaware of the details of their linguistic function. Copy editing may be a deliberate
process, but editors have in mind only a very general notion of style for their station's
news. Style books are not much more precise. And when a style book does pronounce
on structures (not just lexical items), actual practice tends to ignore it. Some style
books prohibit determiner deletion or negative contraction, but the stations still
delete or contract up to 60%.
7
The Glasgow Group's studies are regrettably flawed. Their linguistic analysis and
sociological methods (cf. Schlesinger, 1978, p. 47) are both rather inept, which leaves
their case vulnerable to counter-attack. This is a pity, since the broad lines of their
critique are sound, but inadequately supported by analysis of their evidence. Kress's
and Trew's work is far more linguistically sophisticated. Again, however, the detail is
less convincing ~han the general impression. They treat ideology as a far more
conscious thing than the unexamined frame of mind which Tuchman {1978) finds in
newsworkers. This leads them to a conspiracy theory of writers' motives for applying
rules like agent deletion in passives and nominalizations. Such a strong view of
linguistic processes seems to me only justified where a rule is subject to a deliberate,
focussed choice, as it is in the editing changes studied here. Nevertheless, Kress and
Trew's work is linguistically able and addresses the socially important questions.
8
Note that sub-condition (a) corresponds to Grice's first sub-maxim {1975, p. 46)
"Do not say what you believe to be false". The second sub-maxim "Do not say that
for which you lack adequate evidence" corresponds to my conditions (b) and (c).
 SYNTAX AND SEMANTICS OF NEWS EDITING 115
9
International time differences can trap a copy editor who attempts to add a little
colour to a story. Rosenblum (1979, p. 112) tells of the US news agency editor who
brightened up a hijack story with description of the aircraft sitting on a "sun-baked
runway" in Dacca. He forgot that while the sun shone in New York, it was night in
Bangladesh.
10
The reading of some as roughly equivalent to a few is, I think, standard in New
Zealand English, and makes the approximation of fourteen by some unacceptable.
However, in some(!) dialects of American English, some has to be at least eight or
ten (I am indebted to Dwight Bolinger for this observation). The fourteen of the
original is in fact already an error, since Mrs Peron was a cabaret dancer in 1956,
eighteen years previously.
11
This was demonstrated again in reporting of the Falklands/Malvinas conflict in
1982. The war produced a textbook crop of irreconcilable statistics from the two
sides, Argentina and Britain. Phillip Knightley (1975) of the London Sunday Times
has chronicled the inglorious history of war reporting from the Crimea to Vietnam in
a book that proves truth is indeed the "First Casualty" of war. Knightley shows that
official war statistics are rarely to be taken at face value. The Falklands/Malvinas war
also threw light on the practice of censorship. Censorship is a form of editing, and its
effects can be assessed by the methodology used here. Normally the pre-censorship
text will not be available. However, during the Falklands/Malvinas war, there was
some assessment of the effects of censorship. The Economist (22 May, 1982)
examined how the censors cut one particular text. Our methodology enables us to
assess (a) whether censorship achieves its aim of withholding information useful to an
enemy, and (b) whether and where it distorts meaning and thus misinforms the
public.
12
(d) astrology follower to astrologer (XA/Sl)
(e) fourteen to some years ago (XA/S 1)
(f) also deleted (PA/S2)
(g) and his wife added (XA/S3)
(h) constant to close companion (XA/S3)
(i) managed Mrs Peron to managed her career (XA/Sl: softens the innuendo)
(j) Mrs Peron referred to as Isabel (XA/$2: unprecedented in the data for a head
of state).
On this particular story, note Rosenblum's anecdote (1979, p. 114) that "one news
agency desk insisted on adding 'former cabaret dancer' to every mention of Isabel
Peron".
13
Most newspaper inaccuracy studies have simply counted errors. However, to treat
a mis-spelt name and a major refocus of a story as equally serious seems inadequate.
"Objective" errors of fact (Berry, 1967) such as age or date may be quite wrong in an
absolute sense, but not do much harm to the content of a news item. "Subjective"
errors of emphasis may be hard to specify as wrong at all, but have significant effects
on the meaning of a story. Ideally errors should be judged by a panel, but in this study
I have been the sole judge. I therefore claim only reasonable internal consistency for
the scores, not an absolute validity.
14
The course of the debate on the New Information Order is hugely complex, and
the literature is multiplying to match. Righter (1978) and Smith (1980) are good
introductory surveys and chronologies. Articles appear in journals such as Gazette,
Inter Media, and Journal of Communication (especially 29, 2, 1979). Boyd-Barrett
(1980) is the basic text on the structure and workings of the "Big Four" international
news agencies: Reuters, Associated Press, United Press International, and Agence
116 ALLAN BELL

France Presse. These agencies provide most of the world's media with most of their
international news, and set the agenda of what other media regard as newsworthy.
Besides selection and editing, an additional obstacle to the flow of accurate inter-
national news is inter-language translation. Especially for non-English speaking
countries, much news must be translated at one of the gates in the flow. Inter-
language translation is even more liable to inaccuracy than intra-lan~age editing.
15
This has been the experience of many media researchers: e.g. Bums (1977), the
Glasgow Group (1980, p. 417). It was also the reaction from some quarters to the
present research, and not just to the obviously threatening inaccuracy findings.
Seemingly harmless observations on radio language style met an equally defensive
response.
Educational Issues
 6
Applications of the social psychology
of language: sociolinguistics and education

John Edwards and Howard Giles

It can be argued that, as a distinctive and robust branch of sociolinguistics,

the social psychology of language has only really emerged in the 1970s as a
coherent entity with its own conceptual, theoretical and methodological
traditions {Giles, et al., 1980), although there have been earlier proponents
(e.g. Brown, 1965; Lambert, 1967). As in other sociolinguistic disciplines,
the topics investigated by social psychologists of language are diverse and,
despite the short history of the area, it has already been seen to have
relevance to a wide range of applied social settings. These include the
clinical-medical (e.g. Pendleton et al., 1982; Scherer, 1980), legal (e.g.
Danet, 1980; Line and O'Barr, 1979), occupational (e.g. Kalin, 1982),
educational (e.g. Gardner, 1982; Robinson, 1978) and language planning
policy (e.g. Ryan, 1981; Taylor and Simard, 1981). There is, however, the
possibility of selecting one area of sociolinguistic interest which in itself may
contain information of wider generality. Thus, in this chapter, we examine
sociolinguistic issues in an educational context, with the intention of dis-
cussing matters of wider importance and of showing how a social psycho-
logical perspective can shed new light on some old problems.
In choosing the educational setting, we have not been completely arbit-
rary; a number of points can be made in support of this choice. First, of all
the areas to which sociolinguistic study is relevant, education is clearly one
of the most intrinsically important. As an institution of the middle class,

APPLIED SOCIOLINGUISTICS Copyright© 1984 by Academic Press, London

ISBN: 0-12-701220-6 All rights of reproduction in any form reserved
120 JOHN EDWARDS AND HOWARD GILES

school is one of the obvious places in which to investigate language and class
contact. Second, home-school and teacher-pupil interactions can be seen
as microcosms of wider social interaction. Here, we might think of the
school as an arena in which minority-majority relations are reflected, in
which general issues of social mobility are first encountered, and in which
social policy of the broadest kind (cultural pluralism or assimilation, for
example) is first brought to bear upon individuals. Third, many general
sociolinguistic matters are particularly susceptible to interpretation in edu-
cational terms. Issues like the validity and acceptance of dialects, the posi-
tion accorded to standard usage, and the relationships between language
and identity can all be profitably explored and, to a certain extent, under-
stood in educational terms.
Obviously, these points do not reflect watertight or mutually exclusive
categories. Rather, they represent variations on the basic theme that socio-
linguistic issues can legitimately be considered in an educational framework.
We shall endeavour to justify this theme, and to demonstrate that attention
given to education-related matters has an importance extending beyond the
school. In this chapter, we follow the child, as it were, from home to school.
We consider, first of all, aspects of pre-school life and language, then
comment upon the home-school transition and, thirdly, treat aspects of
school life itself. Following this, we introduce two social psychological
models which we believe make theoretical sense of some of the important
processes underlying the complex relationships among language, class and
education. Finally, we attempt to indicate how future social pyschological
studies in the educational context may inform sociolinguistics in general,
and advance some modest pragmatic orientations for the classroom.

The sociolinguistics of language and education

Before school
It is not our purpose here to review the psycho- and sociolinguistics of first
language development (see for example Bruner, 1981; Hamers and Blanc,
1982; Robinson, 1981a). It is enough to say that, well before school age,
normal children develop a well-formed system of language. The ease with
which language is developed, in the absence of direct instruction, and in
circumstances in which uniformity of support and reinforcement (especially
from parents) cannot be assumed, has been a continuing source of inspira-
tion for linguists and others for a very long time. It has, in fact, prompted a
revival of Cartesian rationalism in linguistics, which explains the rapidity
and the complexity of language learning through brain readiness or "pre-
 SOCIOLINGUISTICS AND EDUCATION 121
wiring". This is not the only theoretical stance possible, of course. However,
whether one accepts Chomskian or Skinnerian explanation (to put the
matter at its most simplistic level; see Love, 1981), the debate is not over
what the child learns, but how (Wells and Robinson, 1982).
We used the term "normal children" above. What is the normal child in
this context? The linguistic and anthropological viewpoint here is, and has
been for some time, that any child not handicapped by one of the recognized
syndromes in speech pathology learns the language of its speech community
without formal instruction (see, however, Berko-Gleason (1980) on direct
instruction in certain sociolinguistic skills). Thus, although the specifics of
what is learned obviously vary enormously, the process itse.lf is entirely
natural to the human species. Given that it is possible to argue, as Trudgill
(1975) has done, for example, that "all normal adult native speakers know
and therefore use their own dialect of English perfectly" (p. 45), it would
seem that, whatever the specifics may be, virtually every child is, by an early
age, well on the way to becoming a fluent speaker.
While not a logical requirement of the preceding viewpoint, it is none-
theless true that those subscribing to it also generally hold that the language
variety learned by the normal child is itself of complete linguistic validity.
That is, not only will the child become fluent in the prevailing language
variety, but also the variety itself, insofar as it is the regularly-used pattern of
communication of the community, possesses the linguistic range and com-
plexity associated with fully-formed systems. Thus, Sapir (1949) tells us that
"the gift of speech and a well-ordered language are characteristic of every
known group of human beings" (p. 1). The linguistic and anthropological
evidence is most relevant here, in that it rules out the possibility of
"primitive" or "debased" or "illogical" or "substandard" varieties (Trudgill
and Giles, 1977).
With regard to languages, such a view has some longevity. Gleitman and
Gleitman (1970), for example, reject entirely the notion of the primitive
language, incapable of expressing ideas communicable in some other
language (see also Lenneberg, 1967). This does not mean, of course, that all
languages are the same in their complexity, nor exactly equivalent in
expressive power. Atthe least, however, we can take it that languages, while
differing because of different varieties of the human condition, cannot be
seen in terms of "better" or "worse". So far as dialects are concerned,
however, the position is somewhat more involved. Although linguistically
no dialect can be seen as inferior or superior to another, the popular
viewpoint is often at odds with this. While few would want to argue,
perhaps, that French is "better" than German, many can be found who
would claim that standard English is "better" than Cockney, say (Edwards,
1982), or that Quebec French is "worse" than Parisian (Bourhis, 1982), or
122 JOHN EDWARDS AND HOWARD GILES

that Mexican-accented Spanish is "worse" than Castillian (Carranza, 1982).

This is an important point, worth re-emphasizing: linguistic data notwith-
standing, there is abundant evidence of prejudice towards certain dialects,
and this seems to be almost universal among languages in which standard
dialects have emerged (Giles and Edwards, 1983; Ryan and Giles, 1982).
This implies that, in the eyes of many, not all normal children can learn
"proper" language. This perception can arise in two ways: (a) a given dialect
is seen to be somehow not as good as another (e.g. Cockney is viewed as
inferior to standard English); (b) a given speech pattern is not seen as a valid
dialect at all (e.g. the perception of black English as some haphazard,
illogical, substandard approximation to "proper" English). These two
possibilities are not always clearly separable in the arguments of those who
have claimed that some children suffer from "verbal deprivation". The
classic examples here are surely the pre-school programmes of compensatory
education, usually having a central language component. For example,
Bereiter and Engelmann (1966) constructed a programme of intensive
instruction in "correct" English for black American children. These children
were seen as having "immature" and "rudimentary" language which, in
some cases, was actually dispensable. Bereiter and Engelmann attempted,
specifically, to break down what they called the "giant word syndrome";
thus, disadvantaged children were seen as unable to recognize single words
and, instead, amalgamated words (e.g. "I ain't got no juice" became
"Uai-ga-na-ju"; p. 34).
However, as we have already suggested, programmes like this are con-
structed upon very inadequate appreciations of dialect variation - which
includes pronunciation and syntax differences, as well as lexical variation.
Bereiter and Engelmann have been criticized by Labov, who is well-known
for his demonstration of the linguistic validity of Black English Vernacular
(1973, 1976). This rather undercuts the philosophy of intervention and
compensatory language programmes for the pre-school child since there is,
in effect, nothing for which to compensate.
Generally, the points of view noted above may be seen as the deficit and
difference positions on language (see Edwards, 1979a; Hudson, 1980). The
former, fuelled substantially by the writings of Bernstein on elaborated and
restricted codes, is more and more giving way to the latter which, as we have
seen, has the force of linguistic and anthropological evidence behind it. This
does not mean that the deficit viewpoint is entirely without current pro-
ponents. As examples, we might cite the work of Tough (1977) in the United
Kingdom and Ramey and Campbell (1977, 1979) in the United States, who
discuss the developmental (including linguistic) retardation among poor
children supposedly deprived of early stimulation and learning experiences.
We will return to these matters in the following sections, for the difference-
 SOCIOLINGUISTICS AND EDUCATION 123

deficit issue underlies much of the sociolinguistic investigation in education.

We have simply been concerned here to delineate the positions as they bear
upon the pre-school child. It is clear that we agree with Gumperz and
Hernandez-Chavez (1972) when they state that for all children, regardless of
dialect, there exists "control of a fully formed grammatical system" (p. 84)
at the time of school entry (although perhaps "well-formed" might be
slightly more accurate here).

Going to school
The beginning of a child's school life has some noteworthy sociolinguistic
aspects. There is, first of all, the relative degree of continuity/discontinuity
existing between the home and the school. For middle-class children whose
homes are ones in which standard dialect is used, there need be little
linguistic discontinuity; the speech patterns used and encouraged in school
will be essentially the same as those of the home (see Ammon, 1977).
Additionally, there is evidence to suggest differences in language use among
social classes (Robinson, 1978, 1979), differences which will first become
salient at school entry. For example, Hess and Shipman (1965, 1968a,
1968b) have claimed that, whereas middle-class mother-child communica-
tion is broad and rich in detail, lower-class patterns are o(ten ones of
"imperative-normative" control (see also Bernstein and Henderson, 1969;
Cook-Gumperz, 1973; Robinson, 198lb; Robinson and Rackstraw, 1972).
Working-class language use is geared towards things concrete and specific;
middle-class use is more abstract and elaborated. There is, of course, a
relation between all this work and Bernstein's notions of code. While
Bernstein himself has recently been concerned to explain that his codes
would be better interpreted as sociolinguistic than as linguistic, and that his
studies were always aimed at elucidating group performance differences and
not ones of basic competence (e.g. Bernstein, 1972, 1973), his work has
quite often been taken as support for the language deficiency argument.
Hess and Shipman, for example, see the impact of this work as leading to the
necessary "re-socialization" of the child, for whom "the meaning of depriva-
tion would thus seem to be a deprivation of meaning" (1968b, p. 103).
We obviously do not concur with such sentiments, nor do we believe that
there is adequate support for them (see Robinson, 1979). We mention this
work on class differences in language use only to acknowledge that dif-
ferences may in fact exist which may prove a hindrance to children's school
progess; this is an acknowledgement of social deficits rather than substantive
cognitive or linguistic ones. And, indeed, there is reason to think that these
class differences are themselves not as pronounced as the deficit theorists
would have us believe. There is, for example, the general caution that must
124 JOHN EDWARDS AND HOWARD GILES

be attached to any study conducted by middle-class researchers into working-

class speech, communicative skills and life-styles, in which differences may
be found in rather artificial, unfamiliar or actually alien circumstances (see
Edwards, 1979a; Stipek and Nelson, 1980). Also, the work of A. D. Edwards
(1976) among 11-year-old middle- and working-class children suggests that
"large and consistent speech differences" (p. 108) may not always exist. At
the very least, such work indicates how important the context is in any
speech study (see also Labov, 1973). Educators must appreciate that
children's language behaviour in the classroom is largely determined by
their definitions of that context (see Bradac, 1982; Ryan and Giles, 1982;
Smith, Giles and Hewstone, 1972) and that these situational construals may
be quite different depending upon the ethnic and class background of the
children concerned.
To the extent to which dialect differences exist, children may experience a
considerable home-school discontinuity. It has been suggested, in fact, that
for some children this may amount to "culture shock": the phenomenon
many will have experienced upon being immersed in a foreign and un-
familiar context. For immigrant and ethnic minority children particularly,
the discontinuity is frequently severe enough to produce this effect, since
they are entering a foreign-language environment and a quite different
cultural setting (see Abrahams, 1976; Ashworth, 1975; Kochman, 1981). In
short, the transition from home to school may, linguistically, be a difficult
one for children of certain backgrounds. Although children from all sorts of
backgrounds experience anxiety at the home-school transition, working-
class, immigrant and ethnic minority children are of particular concern.
They are entering a new environment which is very different from the one
thay have known, in which the same sorts of behaviour are not valued or
normative (McKirnan and Hamayan, 1980) and in which language may for
the first time become an issue.

School life
Much of what will be discussed here can be considered to pertain especially
to the early years of school life. Given what has already been mentioned, it
seems obvious that the first few years of school are of special importance
when dealing with sociolinguistic matters. Children arrive possessing a
well-formed linguistic system; this much is clear. Difficulties arise, however,
because this system may not be the one encouraged and reinforced at school;
it may in fact be seen as inferior and substandard. Gumperz and Hemandez-
Chavez (1972) have, as noted above, supported the child's basic linguistic
competence. They go on to argue further, however, that language dif-
ferences .. do have a significant influence on a teacher's expectation, and
 SOCIOLINGUISTICS AND EDUCATION 125

hence on the learning environment" (p. 105). This is essentially the crux of
the matter. Dialects which, linguistically, are just as valid systems as any
other may nevertheless be viewed as invalid. Thus, having ruled out sub-
stantive linguistic dialectal deficiency does not, unfortunately, also remove
the social deficiencies associated with certain speech styles. This means that,
for practical purposes, certain children's speech is deficient; we have simply
tried so far to point out the aetiology and nature of this deficiency, and
particularly to stress the power of social convention and prejudice.
The school is important here because, for most children, it represents the
first concentrated evidence of this social deficiency. In addition, of course,
there is the fact that schools have traditionally seen themselves as responsible
for promoting "correctness" in linguistic matters (Trudgill, 1975). This
implies that not only may linguistic discontinuity be noted at school, it may
also provoke some action. Typically, as we shall see, this has been of a
remedial nature, even though (from what we have already discussed) we can
also see that in one sense there is nothing to be remedied. The general issue
here, then, is what (if anything) has been done (or should be done) about
disadvantaged speech at school. At a very general level, one might say that
two broad possibilities exist: either the pupil must accommodate to the ideas
and methods of the school, or the school must adapt to the ideas and
methods of pupils. Historically, the first of these has prevailed, while latterly
we see more evidence of schools changing to deal with different pupil
backgrounds. For convenience, we propose to discuss these matters from
the perspectives of the teachers, parents and children.

Teachers
Teachers' attitudes have typically been built upon an assumed correctness of
certain speech styles, usually those of the middle class. This has led logically
to attempts to teach children "proper" linguistic habits, and to the assump-
tion that their maternal varieties may not always be completely adequate
(see, however, Wiggen's (1978) discussion of the rather different Norwegian
situation). As Trudgill (1975) notes, teachers have not been averse to
labelling non-standard speech as "wrong", "bad", "careless", "sloppy",
"slovenly", "vulgar" or even "gibberish". One can appreciate, of course,
that teachers are not alone in these views; they are, after all, widely held,
and teachers are members of society too. Nevertheless, teachers' views are
of particular importance. Although teachers may express linguistically en-
lightened attitudes, they are still "quite likely to be influenced by what they
perceive as deviant speech" (Gumperz and Hermandez-Chavez, 1972, p.
105). Also, however they may feel before entering their school career,
teachers are soon initiated into the ways of the school. Thus Fuchs (1973, p.
85) has noted that, on the matter of disadvantage generally, the teacher.
126 JOHN EDWARDS AND HOWARD GILES

unarmed with the strength that understanding the social processes involved
might have given her ... (is] socialized by the attitudes of those around her ...
she has learned to behave and think in a way that perpetuates a process by which
disadvantaged children continue to be disadvantaged.

This process, incidentally, tends to undercut the argument that having

children taught by teachers from the same social background, and familiar
with the same non-standard speech patterns, would substantially decrease
linguistic and educational disadvantage. That is, the influence of the school
in linguistic and other matters extends to the teaching staff as well as to the
children.
There is, therefore, some reason to think that teachers' stereotypes and
predispositions may unfairly hinder children's school progress. The work of
Rosenthal and Jacobson (1968) is relevant here. They claimed that, on the
basis of manipulated information provided to teachers, the latter formed
different expectations for what they thought were groups of bright and
not-so-bright children. Rosenthal and Jacobson also claimed that these
expectations were reflected in differential school progress made by the
children. This work has been strenuously criticized (Dusek, 1975, provides a
useful review) and the results have not always proved to be replicable.
However, whatever the verdict on Such manipulated expectations, it is
surely beyond doubt that, in the normal course of classroom life, teachers
routinely hold expectations about their pupils. Such expectations are per-
fectly normal and, in most cases, are probably fairly accurate assessments.
However, in instances in which there are differences between the social
backgrounds of pupil and teacher, we would expect such judgements to be
more fallible. Rist (1970) outlined here a self-fulfilling prophecy originating
from teacher expectations of children. Teachers may feel that certain
children (or groups of children) are less able than others; this may be
communicated in subtle (and, as Trudgill, 1975, noted above, not so subtle)
ways to the children (e.g. by non-verbal reactions; see Rosenfeld, 1978)
who, sensitive to this,. respond in ways consistent with the expectation.
The power of social stereotyping in this regard is illustrated by the work of
Snyder (1980) and his colleagues who have shown, in a series of investiga-
tions, how people can set up situations non-consciously so as to confirm their
behavioural expectations of others. For example, if a teacher believes that
working-class children are inarticulate and aggressive, he or she might,
when encountering a new class, pose irritating or involved questions that
would unduly tax anybody. Nevertheless, the precise behaviours anticipated
of the children might thereby be elicited, confirming the teacher's negative
stereotype. Although many features may figure in the formation of teacher
expectations, Rist noted that children's speech patterns are ofsome import-
ance here. The dangers are obvious: a teacher may unfairly (though not out
 SOCIOLINGUISTICS AND EDUCATION 127
of malice) categorize a child on the basis of linguistic and other features, this
categorization may influence the subsequent school progress of the child,
and disadvantage may thus be perpetuated (Eltis, 1980; Seligman et al.,
1972).
In this vein, recent research has often shown that reactions are not always
based upon objective or portrayed characteristics of the speaker but rather
upon characteristics expected by the listener (Aboud et al., 1974). Williams
et al. ( 1971) also provide data supporting this sort of wish fulfillment. In one
of their experimental conditions, they asked teachers to evaluate the speech
patterns of a biack child whom they saw and heard on a videotape recording.
Despite the fact that another (white) child's middle-class speech patterns
were superimposed on the tape, the black child was nevertheless perceived
as sounding "ethnic-non-standard" and was also rated as low in "confidence-
eagerness" (see also below). In other words, teachers had biased their
perceptions of the speech (and hence their evaluation of the child) in the
direction of their stereotyped vocal expectations. Thus, even when listeners'
speech stereotypes are disconfirmed in the form of speakers' actual vocal
characteristics, discrepancies need not always arise at the listeners' level of
cognitive awareness. Instead, listeners may organize the linguistic input
according to their own predetermined and valued cognitive structures
(Street and Hopper, 1982). Moreover, additional information about a
speaker's class and ethnic background can not only bias one's reconstruc-
tions of their past performances but can also influence the perceptions of
their subsequent behaviour (Ball et al., 1982; Thakerar and Giles, 1981). Of
course, the pragmatic implications of such findings are enormous; even if the
speech of non-standard dialect children is "improved", there is no guarantee
that it will be recognized as such by teachers and others anyway.
Having noted these generalities, let us turn now to some more specific
considerations of teachers' views of disadvantaged speech, bearing in mind
as we do so the assumption made here of the linguistic validity of all dialects.
Some more useful American research has been done by Williams and his
colleagues, working with black, white and Mexican-American children (see,
generally, Williams, 1976). In a number of studies it was found that when
white and black teachers evaluated children on the basis of their speech,
along a number of semantic-differential dimensions, two underlying factors
emerged (see also above). One of these, labelled "confidence-eagerness",
reflected such things as the perceived confidence and social status of the
child. The other factor, "ethnicity-non-standardness", was also associated
with judgements of social status and, as well, with perceptions of ethnicity
and the standardness-non-standardness of the child's speech. These findings
were expanded when Williams et al. (1972) asked white and black teachers
to evaluate low- and middle-status black, white and Mexican-American
128 JOHN EDWARDS AND HOWARD GILES

children in Texas. First, the teachers were requested to provide semantic-

differential evaluations of the three ethnic groups as presented to them via a
simple written label - this presumably being a measure of teachers' general
and overall stereotypes. Second, teachers evaluated videotaped samples o(
children's speech and, third, they were asked to estimate each child's class-
room achievement.
The first important finding was the confirmation of the previously estab-
lished two-factor model of judgement. Second, relationships were found
between teacher ratings and the status and ethnicity of the children; thus,
low-status children were seen as more ethnic/non-standard, and were rated
less favourably on the confidence/eagerness dimension. Within this low-
status group, white children were seen most favourably. For middle-status
children, results were similar with regard to the ethnicity/non-standardness
dimension. However, on the confidence/eagerness dimension, black middle-
class children were rated slightly more favourably than were white children
(see Giles and Bourhis, 1976a), and both black and white children were
rated more positively than were the Mexican-American children. A third
finding was that the teachers' general stereotypes (of the groups overall)
correlated reasonably well with their evaluations of actual group members.
Williams suggests, then, that the tnore general teacher stereotypes may have
affected judgements made of individual children (and this, of course, is
exactly the problem referred to above). Fourth, predictions of children's
scholastic achievement were related to the speech evaluations. It is of
interest here that the ratings made by black and white teachers were
remarkably similar. This work by Williams has been discussed in some detail
here since it represents a series of investigations in which similar results have
emerged, and since it underlines once again the power of speech patterns to
influence teacher ratings specifically.
A study by Edwards (1979b) of Dublin school children provides some
further evidence here. Lower-class and middle-class children were tape-
recorded reading a short passage, and were then evaluated along a large
number of dimensions by teachers-in-training. In this study, the judges were
also asked, for each rating made, to indicate how confident they were about
their assessment. On every scale, the lower-class children were viewed less
favourably than their middle-class counterparts. Factor analysis of all
ratings revealed only one underlying factor of importance; that is, all rating
scales were interrelated. This finding is comparable to those of Williams in
that his "ethnicity" factor is clearly not operative in the Irish context, and it
also rather confirms the notion that teachers' judgements of specific children
along specific traits proceed from some overall elicited stereotype. We
would not wish, of course, to deny that other scales and other situations
might evoke other factors (see Day, 1982; Ryan and Giles, 1982); in the
 SOCIOLINGUISTICS AND EDUCATION 129

school context, however, and on scales relating mainly to language and

school ability, teachers' ratings may well be rather unidimensional. It is also
of some interest here to note that the teachers-in-training employed as
judges were not equally confident about all their ratings. In general, they
were more confident when asked to evaluate aspects of personality more or
less directly related to the speech sample itself (i.e. scales relating to fluency,
reading ability and pronunciation), and less so when asked to judge how
happy the child seemed, or his family socioeconomic status. This indicates
that some dimensions would appear to have greater face validity than
others, and shows judges to have, quite properly, some reservations about
making assessments involving rather far-flung matters. Of course, they do
make such judgements (Choy and Dodd (1976) even found judges willing to
assess, on the basis of brief speech samples, the happiness of the children's
marriages) but, as the confidence ratings indicate, they may not be very
comfortable with them. This is a useful cautionary note for those engaged in
such studies to bear in mind, especially since subjects often seem willing to
make ratings along any lines suggested to them by investigators. At the same
time, however, it is worth emphasizing that, with regard to the difficulties
often involved in generalizing from attitudes to behaviour (Fishbein, 1980), a
number of field studies have shown that attitudinal prejudices towards
non-standard dialect speakers have been reflected behaviourally (e.g. in
lack of co-operation with such speakers; see Bourhis and Giles, 1976; Giles
et al., 1975; Giles and Farrar, 1979).
These studies tend generally to suggest that teachers do hold stereotyped
impressions of certain children, and that these are related at least in part to
children's speech styles. To make the point once again: it is not that teachers
are alone in this, for stereotyped reactions to certain accents and dialects are
widespread throughout society. However, teachers' views in this regard are
rather special since negative evaluations may lead to damaging consequences
for children at school. If we consider that many disadvantaged children from
materially poor backgrounds, especially if they are members of visible
minority groups, have real and continuing difficulties in many aspects of
their interaction with mainstream society, then it seems unfortunate, to say
the least, that teachers may unfairly add to their difficulties through categor-
ization based upon inaccurate and ill-informed stereotypes and judgements.
While we may not be able to hope that schools alone can substantially alter
the lot of certain social groups, we should at least hope that they will not
exacerbate the problems of the disadvantaged.

Parents
What are the views held by parents with regard to variations in dialect and
accent? There are one or two general points to make here first. We must
130 JOHN EDWARDS AND HOWARD GILES

remember that speech styles, however non-prestigious or non-standard they

may be in a wider social sense, are still the maternal and regularly-used
varieties for given groups; furthermore, in the absence of intergroup con-
tact, such varieties would hardly be matters of conscious concern at all.
There is, however, ·evidence to suggest that in modem society where cross-
group linguistic comparisons are inevitable, low-prestige groups are very
much aware of the non-standardness of their speech (Trudgill, 1974). This is
the background against which we should consider the views of parents:
which is to say, the views of adult members of the child's speech com-
munity.
In 1960, Lambert and his colleagues in Montreal studied the reactions of
French- and English-speaking students towards French and English speakers
(Lambert et al., 1960). English-speaking judges, predictably enough, rated
English speakers more favourably than French ones. The French judges'
reactions, however, were more intriguing. They too evaluated the English
speake~s more favourably than they did the French speakers; additionally,
they actually gave less favourable responses to the French speakers than did
the English judges. Lambert interpreted this as evidence of a "minority
group reaction": that is, members of a group seen as inferior in some ways to
another adopt the stereotyped views of the dominant group, downgrading
themselves in the process (see also Ammon, 1977). Moreover, as Ryan
(1981, p. 8) has argued:
members of lower class minority communities (especially those with distinctive
skin color like Blacks and many Hispanics) often feel that their social mobility
may be blocked primarily by factors beyond their control. If the minority group
is politically powerless and economically disadvantaged, learned helplessness (a
low sense of personal control over one's destiny, low expectations for success,
maladaptive causal attributions for success and failures) among the children can
further limit school achievement (see Brody, 1968; Dweck, 1975; Harari and
Covington, 1981; Louden, 1978; Nicholls, 1979).

While there is reason to believe that Lambert's 1960 results would not now
be found in Montreal, because of social changes affecting French-English
relations (Bourhis and Genesee, 1980), the fact remains that low-prestige
groups c~ and do adopt views of the dominant society (see Milner, 1981, for
a critical appraisal of this); that these views are not static only serves to point
out how useful language evaluation studies can be in assessing the tenor of
intergroup relations (Giles and Marsh, 1979; Giles and Powesland, 1975).
At the same time as a "minority group reaction" may exist, there is also
the bonding or solidarity function to be reckoned with (Ryan et al., 1982).
That is, one migJ:it ask why group members would retain a non-prestigious
speech style if they accepted the more general social verdict as to its low
prestige? There are several points to be made here. First, and most simply,
 SOCIOLINGUISTICS AND EDUCATION 131
members of certain visible minorities may find no advantage in changing
speech styles; their disadvantage will not thereby disappear. Second, as
Ryan (1981, p. 8) has pointed out:
minority children have relatively few examples available in their communities of
individuals who have successfully achieved this goal since these people frequently
move away from their speech community· phsyically as well as psychologically ...
Also ... since partial achievement is often downgraded by standard speakers
(Ryan and Carranza, 1977; Sebastian and Ryan, in press), only the very high
goal offull achievement is worth aiming for.
It should of course be pointed out here that Ryan refers only to members of
the group itself; there are many examples of standard dialect speakers
(people in the media, teachers, etc.) regularly presented to minority group
children. Third, any speech style is characteristic of a particular group's
background and life-style, and therefore serves as a bond between group
members (Drake, 1980; Giles, 1977; Giles and Saint-Jacques, 1979). This is
especially so for groups undergoing a revitalization of group pride (e.g.
blacks in the United States, the Chicanos, the Welsh, etc.). But the group
bonding function of language extends beyond those groups experiencing
some renewed internal vigour. It also provides a group identity which is a
known, safe quantity; attempts to alter it may result in marginality: a feeling
of no longer belonging to one group and yet not quite fitting in with a new
one either (Lambert, 1967). Carranza and Ryan (1975), for example, advert
to the Mexican-American notion of the vendido: the sell-out, the defector to
the other side (see also Khlief, 1979; Kochman, 1976). Fourth, there is the
factor of "covert prestige" (Labov, 1966); non-standard varieties, although
low in prestige, may yet possess a toughness and a type of masculinity which
is appealing and which constitutes a prestige of its own. Indeed, there is
some evidence that this prestige is attractive even to those with more
standard styles. Thus, Trudgill (1972) noted that some middle-class males in
Norwich claimed to use non-standard forms more often than, in actuality,
they did.
All of this suggests that, when considering the clash of linguistic forms in
the school context, there are a number of important elements to bear in
mind. These centre upon the fact that a group's speech patterns may be at
once generally non-prestigious and essential for identity (see also Day,
1982). Here, we should recall some of the findings of Giles and his colleagues
(Giles and Powesland, 1975). Drawing upon Lambert's (1967) distinction
among the personality dimensions of competence, personal integrity and
social attractiveness, it was found that although non-standard dialects and
accents typically are evaluated less favourably on scales reflecting the com-
petence dimension, they often evoke higher ratings in terms of the latter two
factors. This may be related to the notion of ingroup solidarity, inasmuch as
132 JOHN EDWARDS AND HOWARD GILES

speakers of non-standard forms may be judged as trustworthy, friendly,

warm, etc., without necessarily being seen as competent. While this pro-
vides some comfort for non-standard speakers, it cannot be denied that the
scales relating to competence (e.g. intelligence, industriousness, ambition
and confidence) are of some particular importance, especially within educa-
tional and other institutional settings (see Kalin, 1982).
In any event, we can note that the speech patterns of non-standard
speakers, and of disadvantaged groups, are not without their positive
features in a social sense. We have already mentioned that, linguistically,
they are valid systems. What, therefore, might we expect to observe in the
school context in which the first sustained language contact occurs, insofar as
the parents are concerned? Schools are often seen as essentially feminine
institutions, and this perception certainly extends to the language patterns
used and encouraged there. In fact, this is related to the work noted above
on covert prestige: if non-standard English is seen as essentially masculine,
then standard English must be relatively more feminine (see also Giles et al.,
1980). Right away then a barrier may exist between the home and the
school. Parents may see the school as a soft, ineffectual place which their
children would do well to leave as soon as possible. This feeling is, of course,
often amplified by the desire to have children enter the work-world as soon
as possible in order to help support the family.
On the other hand, it is also the case that many working-class parents see
in the school an avenue for the upward mobility of their children; far from
wanting their children to leave, they may be most supportive of formal
education in general and their children's scholastic progress in particular.
Labov (1973), Ovando (1978) and others have pointed to the high educa-
tional aspirations of lower-class and minority-group parents for their
children. How would this translate into perceptions of a sociolinguistic sort?
There is some evidence to suggest that parents often adopt a very common-
sense and pragmatic view here. They may reject the idea that schools should
conduct classes in non-standard dialect, or that schools should provide
textbooks written in non-standard forms - suggestions put forward by some
recent linguists and educators (see below)- because they wish their children
to learn and use what they see to be the dialect of upward mobility. At the
same time, however, they do not wish to see a return to the traditional school
intolerance for any non-standard usage in class. Covington (1976), for
example, has referred to the views of a group of black parents in Washington.
They wanted teachers to tolerate their children's use of Black English
Vernacular in the classroom, while at the same time they were strongly
opposed to texts written in this form, and to teachers using it. In this, the
parents were apparantly calling for what many have seen to be the most
appropriate course of action (Edwards, 1979a).
 SOCIOLINGUISTICS AND EDUCATION 133

Children
As with their parents, children are often caught in some no-man's-land
between home speech patterns and the standard forms of the school. Many
features other than purely linguistic or sociolinguistic ones are relevant here.
First, recall the notion that the school may be seen as essentially feminine by
children whose native culture is one stressing action, immediacy, etc.
Second, bear in mind that the school and the teachers are seen as outposts of
authority, and that different social groups react to authority in different
ways. If for. parents the discontintuity between home and school is ex-
perienced somewhat indirectly, for the children it is direct and immediate.
Even at the school itself, the difference between the classroom and the
playground, for example, is marked in terms oflanguage use, and represents
a contrast which has been ignored at some cost by certain writers on
lower-class language (e.g. Bereiter and Engelmann, 1966). Lower-class
children may often be torn between a desire to do well at school (with the
support of their parents) and the desire to maintain out-of-school peer
contacts. In this connection, Labov (1976) has discussed certain Black
English Vernacular-speaking children referred to as "lames". Separated
from the rest of the peer group because of lack of facility in verbal games (see
Kochman, 1981), or because of parental pressure, or because of a percep-
tion of the advantages of mainstream culture, these children are more likely
to accommodate to the norms (including the linguistic ones) of the school.
Their peripheral relation to the non-standard culture opens the way (or
perhaps impels them) to greater success in school. In achieving this, they
distance themselves from their mates and create, for many black speakers,
yet another association between unacceptability and the school. This
American example is perhaps suggestive for other cultural settings as well;
indeed, we might suspect that wherever discontinuity exists between home
culture and that of the school, children may experience difficulties.
However, not all non-standard dialect-speaking children are "lames".
What of the majority of those who experience a home-school difference:
how is the linguistic difference perceived and dealt with? Evidence suggests
that speakers of non-standard forms acquire facility in standard varieties
fairly early. For example, Marwit et al. (1972) demonstrated that black
second-grade children cons,stently used more black vernacular forms than
did white children; that is, the study supported the notion of Black English
Vernacular as a regular and rule-governed dialect. However, when these
same children were looked at again in fourth grade, black children showed
an increased use of standard forms (although they still used more vernacular
ones; Marwit and Marwit, 1976). Marwit (1977) found that this process
extended to seventh-grade children as well. The point of these studies is that
134 JOHN EDWARDS AND HOWARD GILES

children possessing a non-standard dialect, and who are in regular contact

with standard speakers, may very soon come to use standard forms in their
own speech (see also Giles and Bourhis, 1976b).
This does not mean, of course, that children abandon their maternal
varieties; peers and community would militate against that. However,
children from an early age are aware of the differences, substantive and
evaluative, between standard and non-standard (see Day, 1982), and as they
develop they become increasingly able to use either as the situation de-
mands. Gay and Tweney (1976) found that black children's production of
vernacular forms decreased with age although their comprehension of those
forms increased. This rather suggests that the use of standard forms need not
necessarily eradicate the original variety; at the same time, it does indicate
an increasing facility with standard dialect. Gay and Tweney, discussing
their experiments with sixth-grade children, report that
on several occasions when a white adult mistakenly entered a testing room (the
experimenter was black], black subjects immediately switched from black
English responses to standard English responses when presented with black
English stimuli. (1976, p. 266)

Interestingly, Piestrup (1973) reports that when white teachers "punished"

black children for speaking in non-standard ways, the children's use of black
speech forms increased during the course of their time at school; that is, they
accentuated their ethnolinguistic styles. The opposite was true in classrooms
in which the teacher did not punish the use of non-standard speech forms
and thereby threaten the pupils' positively-valued distinctiveness.
These sorts of findings demonstrate that for non-standard speaking
children much the same process operates as for all non-standard speakers;
the standard is understood well enough, and it may be produced if the
situation is seen to warrant it. We should also point out (with Labov,
Williams, and others) that non-standard forms are usually not completely
distinct from standard ones: much is shared in common. As Labov (1976,
p. 64) put it, "the gears and axles of English grammatical machinery are
available to speakers of all dialects". This is a useful point to bear in mind in
discussions of this sort; that is, we are considering dialect differences and not
language differences. Emphasis upon the features of non-standard dialects
which do differ from standard varieties may be overdone. In any event, this
apparent difference between children's comprehension of standard forms
and their production of them is important. It applies to almost all dialects
although there may be some few which differ so markedly from the standard
that problems are greater (see, e.g., the work of V. K. Edwards, 1979, on
West Indian speech patterns in Britain). The pragmatic consequences of
these matters are taken up in the last section of this chapter.
 SOCIOLINGUISTICS AND EDUCATION 135

Theoretical and empirical considerations

Despite the large literature on language and social class, it can be argued
that the area is distinctly atheoretical (at least social psychologically) in-
asmuch as it gives more attention to the details of the presumed structural
and functional differences between working- and middle-class linguistic
skills than to systematizing the processes underlying them. In fact, little of
value for formulating a substantial theoretical perspective has emerged in
the last ten years apart from Robinson's (1979, 1980) and Ryan's (1979)
proposals advocating the consideration of ingroup-outgroup dichotomies
and identities, respectively. In this section we intend to elaborate upon these
proposals by recourse, in the first part, to "ethnolinguistic identity theory"
(Ball et al., in press; Giles and Johnson, 1981) which, as the label implies,
aims to clarify the relationships between language and ethnic group. This is
an issue of considerable importance given our concerns in this chapter with
the education of minority group children. The approach is seen to be highly
relevant to the role of language in any intergroup context, and aspects of it
have been found useful in, for example, the analysis of language in male-
female relations (Kramarae, 1981). Hence we shall devote ourselves speci-
fically to extending ethnolinguistic identity theory (ELIT) to any situation
involving intergroup discontinuities between the school and the home,
whether these be ethnically- or class-based. As well, we shall introduce the
possibility that the "intergroup" model of second language acquisition (Ball
et al., in press; Giles and Byrne, 1982), itself a natural progression from
ELIT, will be useful in understanding the conditions likely to impede or
promote proficiency in the standard variety of the language among non-
standard-speaking children. It is our belief that until we conduct more basic
research from theoretical bases such as these, we cannot hope to develop
successful educational policies. Robinson (1979) has underlined here the
past difficulties resulting from educational programmes founded on in-
adequate methods and data.

Ethnolinguistic identity theory

Central to our orientation is a definition of social group membership as those
individuals who identify themselves as belonging to the same class or ethnic
category (Turner, 1981; see also Barth, 1969; Harris, 1979). This ingroup
identification may either ·be based upon a shared set of ancestral or com-
munity traditions, or it may stimulate the creation of traditions (Fishman,
1977). In both cases, the main concern is with the establishment and main-
tenance of distinct group boundaries. It involves a subjective definition of
136 JOHN EDWARDS AND HOWARD GILES

social group membership, and has the advantage of avoiding categorization

of individuals solely on the basis of objective criteria (e.g. income and
~ducational level, area of residence, skin colour, etc.). It implicitly allows
for the possibility that an individual may act in terms of group membership in
some, but not necessarily all, situations. For instance, Tajfel and his
associates have distinguished between two kinds of encounters (Tajfel, 1978;
Tajfel and Turner, 1979). At one extreme is the inter-individual situation,
occurring between two or more individuals, fully determined by inter-
personal relationships and individual temperaments, and not at all affected
by the individuals' social memberships. At the other extreme are inter-group
situations, involving two or more individuals, completely determined by the
individuals' group memberships, and not at all affected by the interpersonal
relationships and personalities of those involved. Stephenson (1981} has
pointed to the fact that these two extremes might be better conceived of as
two separate dimensions, thereby allowing for the possibility that a situation
may be defined simultaneously in inter-individual and inter-group terms.
Therefore, a classroom discussion between a white, middle-class teacher
and a black, working-class adolescent may be defined by each of them as
occupying, at different times, different positions in two-dimensional space.
The more they conceive of the situation in inter-group terms, for example,
the more they will tend to treat each other simply as members of an
outgroup, rather than in terms of individual characteristics possessed.
Tajfel and his colleagues have developed a theory for understanding
behaviour at this inter-group extreme, and this forms the core of ELIT
(Giles and Johnson, 1981). It runs as follows: We categorize the social
world, and perceive ourselves as members of various groups (class, ethnic,
peer, etc.). This knowledge of ourselves as group members is defined as our
social identity and it may be positive or negative, according to how ingroups
fare in social comparisons with relevant outgroups. It is argued that we strive
to achieve a positive identity by seeking dimensions which afford favourable
comparison with outgroups (i.e. we strive for positive psychological dis-
tinctiveness). Language enters the picture when a group regards its own
language, dialect or slang as a dimension of comparison with outgroups (e.g.
as something worth having, which the outgroups lack}. In this case, the
group would be said to have achieved a positive psycholinguistic distinctive-
ness (Giles et al., 1977). Naturally, and as we have already outlined, there
can be cases of dialect being salient for the group but affording them only
unfavourable inter-group comparisons (e.g. "we speak sloppily, badly,
etc.").
What happens when an individual experiences negative ethnic or class
identity? Three basic responses are suggested here, in order of likely
priority. First, he or she may apply a strategy of individual mobility, trying to
 SOCIOLINGUISTICS AND EDUCATION 137
pass into the more privileged social category by acquiring its characteristics,
and quitting the ingroup so as to attain a new, positive identity (see "lames",
above). This is likely if the inter-group boundaries are perceived as soft and
permeable (Giles, 1979). Linguistically, the individual will converge on
areas where the outgroup's linguistic habits are believed to lie (see the notion
of "hypercorrection"; Labov, 1966; Trudgill, 1974), and will discard in-
group speech markers (see Thakerar et al., 1982). If individual mobility
through language is widespread the ingroup language variety will not, of
course, be maintained (see Denison, 1977 on "dialect death").
A s~cond strategy, applicable if individual mobility fails or is thwarted by
the existence of impermeable inter-group boundaries, is social creativity.
This involves a redefinition of the ingroup-outgroup comparison in various
ways. These include: referring to other, low-status, groups, or fellow
members of the ingroup for social comparison, rather than the outgroup;
simply reversing the previous evaluative polarity of existing inter-group
comparisons (e.g. "Black is beautiful" in the United States); inventing new,
more favourable dimensions of inter-group comparison (see Drake, 1980;
Kochman, 1981). The last two sub-strategies occur in linguistic form when
individuals flaunt, in a divergent way, their supposedly "debased" or "sub-
standard" speech forms in inter-group contexts (e.g. in certain classroom
situations).
The third strategy, outright social competition, is likely to be adopted
when the previous strategies have failed. Here, members still identify with
their subordinate ingroup, and make unfavourable inter-group compari-
sons, but perceive alternatives to the status quo by viewing their subordinate
status as illegitimate and unstable. Numerous situations in ·which language
has been the focus of civil unrest can be seen as linguistic illustrations of this
(e.g. Quebec, Soweto, Washington).
What does the dominant group do in response to a subordinate group's
adoption of any of these sociolinguistic strategies? Assimilation of the
subordinate group through individual mobility would threaten the dominant
group's psychological distinctiveness, so it would be likely to reassert this by
not recognizing the assimilation that had occurred (as, for example, in the
Williams et al. videotape experiment, 1971), or through creatively divergent
counter-strategies (e.g. inventing new ingroup linguistic markers, or even a
new dialect; see Elias, 1978; Ullrich, 1971). Subordinate group creativity or
competition (strategies 2 and 3) would be likely to elicit reciprocal competi-
tion which might include verbal derogation, abrasive verbal humour or
overtly political action (for a more detailed discussion, see Giles, 1978, 1979;
Giles and Johnson, 1981).
138 JOHN EDWARDS AND HOWARD GILES

Factors affecting the salience ofsocial group membership

Although a person's membership groups are part of his_ or her social identity
and can be remarkably potent (even by seven years of age; Vaughan, 1978),
they will not all be equally salient at any one time. It is only when groups with
distinctive linguistic characteristics (often ethnic and social class collectivi-
ties) are salient that we would expect accentuation or attenuation of ingroup
speech markers as inter-group strategies. Four variables may be suggested
as relevant in this regard: perceived ethnolinguistic vitality, perceived group
boundaries, multiple group membership and social attributions {cf. Ball et
al., in press; Giles and Johnson, 1981). Each of these can be briefly dealt
with here.
Giles et al. (1977) have suggested that ethnic groups be compared in terms
of ethnolinguistic vitality, to which three main groups of factors contribute:
status factors (e.g. economic, political and linguistic prestige), demographic
factors (e.g. absolute numbers, birth rate and geographical concentration)
and institutional support factors (e.g. recognition in mass media, validity of
language/dialect in education, government and industry). They proposed
that groups with high vitality are most likely to thrive and remain distinct, as
well as to act collectively in accord with their group memberships. We will
further suggest, here, as the 1977 study implied, that vitality is just as
important for other social groups as well, including class collectivities.
However, the actions of individual group members are likely to be governed
less by actual vitality of their groups than by their perception of the relevant
factors (Bourhis et al., 1981). Giles and Johnson (1981) suggest that high
perceived vitality increases the salience of group identity for members and,
therefore, their inclination to accentuate group speech markers to establish
a favourable psycholinguistic distinctiveness.
Sociological and anthropological scholars have emphasized the importance
of social boundaries in inter-group activity (e.g. Lyman and Douglass,
1973). It is suggested that group members try to maintain a high level of
perceived inter-group boundary hardness. When successful, as is the case
with many low-prestige dialects (Ryan, 1979), this clarifies social categoriza-
tion and the norms for conducting inter-group encounters, and thereby
increases the salience of group membership. Thus, perceiving an authentic
standard dialect as extremely difficult to achieve (hard linguistic boundary)
would result in the consequences just noted. Moreover, those who see
themselves as belonging to many different, overlapping groups (see
Salamone and Swanson, 1979), should possess a more diffuse social identity
than persons who view themselves as members of only one or two. That is,
class or ethnic attachments should be stronger for those who can identify
with few other social categories. Moreover, the favourability of inter-group
 SOCIOLINGUISTICS AND EDUCATION 139
comparisons offered by these other social membership groups will likely
affect individuals' reliance upon their class or ethnic identities, as will their
status within each group.
The role of the attribution process at the inter-group level (Deschamps,
1977; Hamiton, 1979; Hewstone and Jaspars, 1982a; Pettigrew, 1979) lies in
the analysis of how members of subordinate power groups explain their
inferior social positions and relative states of deprivation (low wages, poor
employment prospects, etc.). This application of attribution theory was
envisaged by Guimond and Simard (1979), who examined explanations
offered by English and French Canadians for their social positions. It was
found that English Canadians (the dominant ethnolinguistic group) tended
to blame French Canadians themselves for the latter's lower social position.
That is, they preferred individualistic causes for economic inequality. In
contrast, French Canadians (the subordinate collectivity) offered structural,
more social explanations for their inferior standing. Indeed, one regularly
witnesses such differential attributions in British and American allocation of
responsibility for the social positions of the middle and working classes, by
members of those categories. Given the salience of one's own dialect, and
the importance of ethnic and class membership, it seems very likely that
individual members of subordinate minorities may ascribe their failings to
their possession of a certain dialect, or to discrimination by others because of
this dialect. The process of ascribing one's failure to membership in the
"wrong" social group has been demonstrated by Dion and associates (e.g.
Dion et al., 1978).
Seeing the status of one's language and dialect as a possible cause of social
inferiority may be strengthened by the operation of the covariation principle
(Kelley, 1967). This suggests that an effect is seen to be caused by the factor
with which, over time, it covaries. It is relatively easy to see how children
from lower-status backgrounds might provide themselves with such covaria-
tion data. For example, if an individual perceives that almost everyone who
uses his or her non-standard dialect occupies a relatively low social position,
while almost everyone using the standard dialect occupies a higher one, then
dialect might easily be seen as the cause of the differential status. In
addition, it is easy to see how this covariation rule would be strengthened by
ingroup members who had passed upwards, and were now occupying posi-
tions of higher status. At the same time, it is possible to consider other causal
explanations for the lack of economic and social success of non-standard
dialect speakers; for example, there may exist negative outgroup charac-
teristics. This kind of explanation derives support from a growing literature
on intergroup attributions (see Hewstone and Jaspars, 1982a). From this
perspective, the individual attributes the behaviour of another to charac-
teristics and intentions associated with the group to which the other belongs,
140 JOHN EDWARDS AND HOWARD GILES

and not simply to individual ·characteristics. Thus, if the individual is the

recipient of negative behaviour by an outgroup member (e.g. being refused
employment), then the preferred explanation is more likely to be in terms of
negative outgroup attributes (e.g. "they are discriminatory") than in terms
of negative ingroup characteristics (e.g. "we are lazy"). This pattern of
attributions has been demonstrated by Hewstone and Jaspars (1982b).
Against this background, one might predict that members of lower-status
groups might expect unfair practices on the part of middle-class standard
dialect speakers, and hence they may not bother to acquire the outgroup's
dialect at all. To deal with this possibility, we might hypothesize that when
the inferior position of subordinate class or ethnic group members is
attributed to intransigent and unchangeable discriminatory practices of the
outgroup, then their group identification will become more salient (Ball et
al., in press). Therefore, the threshold for acting in terms of their social
group membership wiJI be lowered, and the probability of accentuating
non-standard dialect patterns thus increased.
To conclude this discussion on the theoretical state of ELIT, we advance
five propositions (Ball et al., in press; Giles and Johnson, 1981). People will
see themselves in either class or ethnic terms, and will strive for positive
psycholinguistic differentiation from the standard dialect outgroup, when
they:
( 1) see themselves strongly as members of an ethnic/class group, with
dialect/language an important dimension of this identity;
(2) regard their group's relative status as changeable and attribute the
cause of their relatively low social status to advantages taken unfairly
by the outgroup.
(3) perceive their ingroup's ethnolinguistic vitality as high;
(4) perceive their ingroup boundaries as hard;
(5) identify with few other social groups - ones which offer unfavourable
social comparisons.
Conversely, the opposites of these conditions should lead to erosion of the
ingroup dialect, since attempts will be made to become assimilated by the
middle class.

The inter-group model of second language/dialect acquisition

Giles and Byrne (19820 have formulated an inter-group model of second
language acquisition, which is largely based on foundations just discussed. It
is an attempt to place second language acquisition in an inter-group context,
while maintaining and integrating some of the components of earlier social
psychological models (see Clement, 1980; Gardner, 1979). It is our belief
 SOCIOLINGUISTICS AND EDUCATION 141
that the same processes are involved in the acquisition of proficiency in the
standard dialect by non-standard speaking children.
In order to illustrate this, let us designate as subgroup "A" those members
of a lower-status group (class or ethnic) for whom the preceding five proposi-
tions apply. Those from backgrounds in which the propositions do not apply
are designated subgroup "B". Subgroup "A" would, we propose, be un-
likely to achieve proficiency in the standard dialect. For them, acquiring
such a dialect would be "subtractive" to their sense of class or ethnic identity
and, even if the acquisition was seen as economically and politically profit-
able (Ryan, 1981), subgroup "A" members would still experience a fear of
assimilation by the outgroup and the likely rejection by their valued ingroup
peers (Clement, 1980). Subgroup "A" members would not seek informal
acquisition contexts (Fig. 1), and would become proficient only in classroom
aspects of the standard dialect, such as vocabulary and grammar. Individual
differences in attaining this limited proficiency would reflect intelligence as
well, perhaps, as differences in the efficiency of pedagogical techniques.
Subgroup "B", on the other hand, are those lower-status group members
most likely to achieve proficiency in the standard dialect. These individuals
do not strongly see themselves as members of an ethnic or class group,
regard their group's status as unchangeable, attribute their inferior status to
inferior group qualities, perceive their ingroup's vitality as low and its
boundaries as permeable, and identify with many other social groups offering
favourable social comparisons. For such children, acquisition of the standard
variety should be "additive" (Lambert, 1974), and their motivation should
be "integrative" (Gardner, 1979). Subgroup "B" members would take
advantage of available informal acquisition contexts to further their standard
dialect skills and would, in addition to formal knowledge, gain high oral
competence. As well, positive attitudes towards the middle class and its
norms and values would be expected. Anxiety experienced in situations of
standard dialect would, for subgroup "B", relate more strongly than intel-
ligence to standard dialect proficiency. Clement (1980) proposed that the
kinds of inter-group experiences the second-language learner has with
members of the other language community will affect his or her motivation
to learn their language. More specifically, he proposed that the more
frequent and the better the quality of inter-group contact, the higher will be
the learner's confidence in using the second language, and in taking risks in
so doing (Beebe, in press). Particularly in this context, however, we would
obviously prefer to rely more on the notion that it is the learner's perceptions
of inter-group contacts that will facilitate or hinder the acquisition process
(see Brown and Turner, 1981). In fact, we can argue that the children's
attributions about the quality of these inter-group experiences are the critical
determinants (Jaspars and Hewstone, in press). If features of external
142 JOHN EDWARDS AND HOWARD GILES

circumstances provide plausible causes for pleasant or unpleasant encounters,

children's motivation will be less affected than if no such causes can be found
and if, therefore, they are forced to consider that actions of the standard
dialect outgroup create unpleasant encounters (see Worchel and Norvell,
1980). The presence or creation of poor environmental conditions is hy-
pothesized to obviate the need for scapegoating (or blaming outgroup
members); this in turn prevents the resurrection of previous social group
boundaries.

L2 1eornin9 Feedback flow to

outcome - ~-,'A') s;bg~p
High proficiency Feed1>ock of 'a' outcome
ottamed only on ..,. inie;p;:;i18doscii'11iiroT--c 1 Feedback of 'e'
Informal leom1n9
'classroom' skill, 'betrayal' by 'A' members proficiency to
contexts are outgro11p likely
vocabulary, CVOlded. to prec1pitote
grammar, etc. upward divergence .

1
Intelligence and
•
TRUE

oplitude determine
proficiency.
AnKiety levels
in situations of
L2 use determine
praf1ciency.

I
t ..
Feedback of A
non-proficiency L2 leorning
likely lo confirm outcome
prlljudices of Feedback of 'A' outcome High oral and
out group ~int-;rp;;ie"d-;;;·"failu~' sociolinguistic
by 'e' members ".!~~oc.!,_!:~~­ proficiency and
own ('B')subgroup fovouroble non-verbal
oulcomes

Fig. 1. The intergroup model of second language learning (as schematized in Ball
and Giles, 1982).

What would be the collective consequences of the contrasting individual

outcomes portrayed above and in Fig. 1? These would depend partly on the
relative proportions of "A" and "B" subgroup members in the class or
ethnic group. However, both subgroups would, if large enough, feed back
information to themselves, to each other and to the standard dialect
outgroup. The maintenance of non-standard dialect features by subgroup
"A" would probably be interpreted by its members as successful retention of
class or ethnic language distinctiveness in the face of cultural or subcultural
threat. Subgroup "B" members, on the other hand, would more likely see
this as a set of individual failures, due perhaps to socioeconomic or educa-
 SOCIOLINGUISTICS AND EDUCATION 143

tional inadequacies. And, the standard-dialect middle class (including

school staff) would see their prejudices confirmed about subgroup "A".
Subgroup "B"'s mastery of the standard variety should fortify its members'
integrative motivational tendencies (see Fig. 1). Subgroup "A" members,
however, might regard this as class or ethnic betrayal, and become yet more
fearful of assimilation and even less motivated to become standard dialect
speakers. The outgroup might feel its distinctive identity threatened and
might respond by differentiating itself further, perhaps creating a new or
altered standard variety.
To incorporate constructs and processes from the second-language acquisi-
tion literature into the standard/non-standard educational context, it might
also be useful to consider subgroups "A" and "B" in relation to Dulay and
Burt's (1978) "socio-affective filter". They suggest that the motives, atti-
tudes and emotional states of learners filter the language input that is
processed by them, and affect the rate and quality of acquisition (of standard
dialect, in the present discussion). They propose (p. 556) that the socio-
affective filter contributes to at least three aspects of selective learning:
(1) the preferences for certain input models over other; (2) the acquisition of
certain aspects of language before others; (3) the (subconscious) determination
of the point at which language acquisition efforts ~hould cease.

Here we could propose that subgroup "A" members are likely to operate
the filter maximally such that processing of incoming standard dialect data is
inhibited, whereas subgroup "B" members are likely to filter minimally,
thereby facilitating standard dialect input. However, in the theoretical
framework presented above, "A" and "B" individuals have been con-
sidered as archetypes: least and most likely, respectively, to achieve
standard dialect proficiency. However, it would not be unreasonable to
suppose that, in most groups, the greatest proportion of individuals would
fall somewhere between these polarities. This body of "intermediates" can
be thought of as comprising those individuals who cannot easily answer
"true" or "false" (see Fig. 1) to Propositions 1 and 2. In other words, they
cannot subscribe wholeheartedly to a strong or a weak sense of ethnic or
class membership, are unsure about the possibility of changes in group
status, and feel that responsibility for their group's inadequate conditions is
borne partly by the ingroup and partly by the middle class and its institutions
(including the school). While Propositions 3 to 5 may be thought to provide
little more than situational and personal support for making class or ethnic
identification salient for group members, they may also have more direct
influence upon the linguistic actions of the "intermediates". These indi-
viduals' perceptions of ethnolinguistic vitalities, group boundaries and
multiple group memberships may well determine the extent and manner of
144 JOHN EDWARDS AND HOWARD GILES

the operation of the socio-effective filter and, therefore, the degree of

proficiency attained in the standard dialect. Ball et al. (in press) have
incorporated such '"intermediates" into the inter-group model of second
language acquisition, through the application of the mathematical "catas-
trophe" theory (Thom, 1975; Zeeman, 1976). Space precludes a discussion
of this elaboration of the inter-group model, but those who are interested in
the theoretical underpinnings discussed in the present chapter will be able to
translate this new perspective into the standard-non-standard context quite
easily. However, before allowing ourselves the further luxury of translations
from the second-language literature (see Beebe and Giles, in press), there
are more immediate priorities. We should develop adequate methods for
testing the viability of the constructs already introduced and for assessing
the predictive value of the processes posited. From the previous discussion
arise many empirical issues regarding the language of children from lower
socioeconomic-status backgrounds. These include the following:
(a) To what extent do these children subscribe to a sense of local, ethnic
or class belongingness, and what contributions do these and other
group memberships make to their personal and social identities at
different ages?
(b) To what extent are dialect, and the specific features of it, salient and
valued dimensions of class, local and ethnic identity, in relation to
other, non-linguistic, characteristics of group membership?
(c) Under what conditions do children define classroom situations as
essentially inter-group in nature, and what are the subjective com-
ponents of such construals? How, if at all, do teachers' language,
actions and pedagogical strategies contribute to the perceived inter-
group situation?
( d) What are children's attributions of their socioeconomic status posi-
tion, how do they see the vitality of it, and what cognitive alternatives
do they see open to them educationally?
(e) To what extent can children, in reality, be adequately represented in
terms of subgroups "A" and "B"; are these predictive of the types of
motivation proposed by the inter-group model?
It is obvious from the tenor of this section that we believe traditional
research into the relationships among language, class and education has only
scratched the surface. Moreover, we require the type of conceptual frame-
work proposed by Ryan et al. (1982) to make comparative statements about
different educational problems in different speech communities and so-
cieties. Hence, we are extremely reluctant to list priorities now, on the basis
of what are insufficient and sometimes non-existent data. This is, of course,
of little comfort to children who come from lower-class backgrounds, and
 SOCIOLINGUISTICS AND EDUCATION 145
who speak a non-standard dialect (or to their teachers). What we can
propose must therefore be very modest. However, we make no apologies for
this especially since, in the current economic/political climate, many re-
searchers are induced to become prematurely applied and policy-orientated.

Pragmatic considerations

As mentioned in the previous sections, the traditional policy of the school

was most often to require adaptation on the part of the pupils. Armed with
an unshakeable sense of the correctness of their own (i.e. middle-class)
speech norms, teachers typically had little hesitation in correcting pupils
who spoke "slovenly" or "illogical" English. While this policy is under-
standable, it doubtless created many difficulties for children to whom it was
applied. For example, many teachers of non-standard-speaking children in
many parts of the world have noted the sullen and non-verbal nature of these
children in class. This is sometimes due to cultural differences with regard to
such matters as behaviour in front of authority figures (Abrahams, 1976), or
public behaviour in front of peers (Philips, 1972). As well, however, there is
reason to believe that teachers have, in large measure, created non-verbal
children through early and excessive "correction" policies; the children,
unwilling to undergo the inevitable, simply cease to participate in classroom
activites. Labov, in a well-known article (1973), demonstrates how reticent
and non-verbal black children can be suddenly transformed into the verbally
adept individuals we now associate with the orally rich black culture, by the
removal of white, middle-class examiners. In so doing, Labov counters the
claims of Bereiter and Engelmann (1966) who had maintained that, for
black children, language may not be of vital importance and may, in some
cases, even be dispensable. Similarly, Shuy (1971) has noted that no non-
verbal child that he has encountered has remained so in a non-intimidating
context (see also Bernstein, 1973).
On this basis alone we might ask if something could be done at school to
remove the possibility of children becoming uncommunicative in the class-
room - children who are clearly not at all so in the playground, or at home. It
is evident that a minimum requirement here is for teachers, and the school
system generally, to be tolerant (and not begrudgingly so) of dialectal
variation, and not to brand it as anything inferior. Given current social
psychological and linguistic data, we can now provide teachers with the
evidence that this, in fact, is the most reasonable thing to do. Dialects and
accents are possessions one shares with those who are psychologically
similar (see Bishop, 1979; Delia, 1972; Sebastian, Ryand and Corso, in
press) and with whom one identifies. They are, as we outlined in the
146 JOHN EDWARDS AND HOWARD GILES

previous section, clearly linked to a child's conception of group membership

and self-respect. Children will therefore not wish to converge towards
another speech variety unless they want to alter their identity in some way. If
teachers reject children's dialects they are also rejecting them. Teachers
should be made more aware that negative evaluations of non-standard
dialects are culture-bound, and that the unfavourable stereotypes are un-
doubtedly more group-biased than language-biased (St. Clair, 1982; Trudgill
and Giles, 1977). To change a child's accent consistently and consciously
means that the teacher must pay more attention than normal to how some-
thing is said, and less to what is said. The psychological costs here may be too
great, and may even hinder educational progress. It is true that how one says
something can sometimes be just as important as what is said, but one should
not be pressed for one aspect at the expense of the other. Given this proposed
tolerance towards variation in children's oral language, we might expect
(based upon what we have discussed here) that children would come
towards bidialectalism of their own accord. This seems to some a rather
passive approach; Ammon (1977) has said that those who advocate educa-
tional tolerance of non-standard dialects, and who further imply that
children should gradually learn a standard variety, have offered no practical
advice as to how the transition from non-standard to standard can be
achieved. Passive approaches have the affect, Ammon claims, of restricting
meaningful social mobility. These are, of course, useful points. However,
since more active approaches often have the effect of stigmatizing non-
standard dialects in ways which are now considered by linguists to be based
upon faulty ideas about language varieties, an apparently more passive
stance may not be as undesirable as Ammon suggests. For we should
remember here that the passivity is really on the part of the teachers, not the
pupils: i.e. the children may adopt another variety, and this is an active
process. The question is really whether teachers, by adopting a more active
strategy for promoting the standard variety, will help children or hinder
them. On ba)ance, it would seem that children do not respond well to
approaches (however benevolently inspired) which, in essence, suggest to
them that their maternal dialect is in some way flawed (see Edwards, 1979a
for fuller discussion here). The general practice advocated here, then, could
be termed one of linguistic addition rather than one of replacement, where
any addition proceeds from a desire of the non-standard speaker and not
from direct suggestion by teachers or others. This practice would lead to the
expansion of a child's linguistic repertoire, thereby creating bidialectalism.
It is a further question as to whether or not bidialectalism can be maintained
permanently. Some have suggested that bidialectalism is one step on the
road to eradication of the maternal variety (e.g. Moses et al., 1976) but much
depends here upon the context. Just as it is possible for bilingualism to be
 SOCIOLINGUISTICS AND EDUCATION 147
stable or transitory, so bidialectalism can exist in either mode. In any event,
the extent to which facility in a second language or dialect leads to eventual
linguistic or dialectal assimilation depends largely upon societal variables
which extend far beyond the school.
The approach advocated in the previous paragraph has, in particular,
seemed too laissez-faire to those concerned with reading, a central element
in any discussion of educational practice. Some have called for texts written
in non-standard dialect, for example (see Baratz, 1970), to ease the lot of the
non-standard-speaking child. There are several reasons, however, why this
policy seems ill-advised. First, as we have noted, dialect differences are not
usually so marked as to make standard English comprehension a real
problem for children. Since reading involves decoding to meaning and not to
sound (Edwards, 198la, 1981b), we might agree with Torrey (1973, p. 68)
who points out that, with regard to Black English:

The difference in phonology between standard English and black English is not
directly relevant to reading. All children who learn to read English have to
break a fairly complex code of sound-spelling relationships. The fact that the
correspondences are different for speakers of Afro-American does not in itself
prove that they are more difficult than for standard speakers.

We could generalize to all dialects where there is reason to think that the
"gears and axles" are essentially similar. Other varieties, including perhaps
West Indian English in Britain, may require other approaches; it is inter-
esting to note here, however, that some (e.g. Bailey, 1966) have argued that
West Indian English is sufficiently divergent from the standard to be deemed
a separate language.
Second, the effort and expense entailed in production of many different
non-standard texts, and the difficulties involved in effecting a transition to
standard English texts (something which supporters of non-standard texts
agree is necessary), rather indicate that this is an inappropriate course to
take without much further evidence of real and substantial dialect inter-
ference. Third, we would recall here the desires of many parents who would
not wish to see non-standard texts used in schools, for the reason that they
would perpetuate their children's disadvantage. While linguists might argue
that this view is correct, but for the wrong reason, it is clear that parental
attitudes must not be ignored. Thus, with reading as with oral language, the
reasonable policy seems to be one of enlightened tolerance. Teachers could
well allow children, when reading out loud, to render the meaning of
standard English texts in their own dialect. This would coincide with the
view of reading as decoding in meaning; it is only after the meaning has been
assimilated that the child then produces something (i.e. this part of reading
aloud is an encoding process).
148 JOHN EDWARDS AND HOWARD GILES

With regard to written language, instruction in the standard form seems

less controversial since all children, regardless of dialect, must learn to
translate ideas into writing. It has been argued elsewhere (Edwards, 1979a)
that a more or less formal equation of written English with standard English
would not likely be destructive of psychological identity (cf. Scollon and
Scollon, 1979).
General tolerance of dialectal variation does not seem, perhaps, much to
come to. However, any more formal approach (language drills in standard
English, formal guidance on "correctness" by the teacher) is not likely to be
successful and may in fact widen the gap between the non-standard speaker
and the school. We must allow children who have competence (passive, at
least) in the standard to choose for themselves linguistically (with an
abundance of models around them, of course, chief of which is the standard
English-speaking teacher).
This in turn implies that changes in schools may involve teachers more
than children: that is, with reference to the two broad possibilities noted
earlier (in the section on school life), we are clearly coming out in favour of
the one which requires schools to adapt to pupils, rather than vice versa. At
the same time, however, we are also saying that this adaptation does not,
and should not, inevitably involve formal changes in curriculum, or in texts,
or in language forms used by teachers. Rather, it involves a more thorough-
going tolerance, which in itself is no small matter. The general point here is
that, given a tolerant atmosphere which is accepting of dialectal variations
and which understands them for what they are (non-standard rather than
substandard), children will define themselves linguistically vis-a-vis the
mainstream varieties as seems most appropriate. In addition (with the help
of parents perhaps), educators may consider ways in which the variety of
linguistic experiences children from different backgrounds bring with them
into the classroom may be used as a resource when planning curricula.
Finally, classroom interaction between teacher and pupil is of great
potential importance for inter-group definition. Both teacher and pupil will
likely use the other's speech style in making group categorizations which
lead to expectations concerning the other's abilities, temperaments, etc.
The more that pupils strongly identify with some social category, the more
likely it is that they will differentiate themselves linguistically from others
who represent different group values. This may be particularly so when
these others are perceived as representatives of a group which, to the
children, symbolizes an illegitimate and oppressive force which has deprived
them of socioeconomic and political rights. Hence, some children find
teachers' speech styles alien, and come to resent the social gulf which these
styles reflect. Thus, they may find teachers' voices "snobbish", "posh" etc.,
and they may react accordingly. Indeed, "downward convergence" (Giles
 SOCIOLINGUISTICS AND EDUCATION 149
and Powesland, 1975) on the part of the teacher trying to bridge this gulf may
be seen as false or even insulting if it is seen to be patronizing (Edwards,
1979a; Giles and Smith, 1979; Platt and Weber, in press; Shuy, 1977). The
implication here is that barriers between teachers and pupils must be
reduced if successful education is to occur.
It is crucial, here, that children see the wisdom of identifying more with
their classmates as a group, than with only a small faction, selected because
they belong to the same ethnic or class category. At the same time, this
should not mean the dissolution or devaluation of such categories in the
classroom; rather, the sense of the class as an important collectivity which
can be identified with in its own right should become more salient. As well, a
useful strategy would be to emphasize that teacher and pupils belong to the
same social group, for educational purposes, in the school context (see
Brown and Turner, 1981; Sherif, 1967; Turner, 1981). Such a collectivity
could then take its place among other relevant groups and could assist,
through inter-group comparison, in the development of more accurate,
favourable, and therefore healthy social identities.
We hope that our use of the educational context has been informative for
important aspects of a more general sociolinguistics. Certainly, there are
issues with which we have not been able to deal here: bilingual education,
for example, suggests itself as an important instance of schools' adaptation
to pupils of non-mainstream linguistic backgrouds (see Edwards 1977, 1980,
1981c, in press). However, much of what has been treated here has a wider
relevance. The issues of language and dialect contact, the prescriptivist ideas
of middle-class institutions, the linguistic evidence bearing upon dialect
validity, and the persistence of non-standard varieties: these are all matters
which can be usefully examined, from a social psychological perspective,
within the educational setting, while also having a much broader significance.

References
Aboud, F. E., Clement, R. and Taylor, D. M. (1974). Evaluational reactions to
discrepancies between social classes and language. Sociometry, 37, 239-250.
Abrahams, R. D. (1976). "Talking Black". Newbury House, Rowley, Mass.
Ammon, U. (1977). School problems of regional dialect speakers: Ideology and
reality. Results and methods of empirical investigations in Southern Germany.
Journal of Pragmatics, 1, 47-68.
Ashworth, M. ( 1975). "Immigrant Children and Canadian Schools", McClelland and
Stewart, Toronto.
Bailey, B. L. (1966). "Jamaican Creole Syntax", Cambridge University Press,
London.
Ball, P., Byrne, J., Giles, H., Berechree, P., Griffiths, J., Macdonald, H. and
McKendrick, I. (1982). The retroactive speech halo effect: Some Australian
data. Language and Communication, 2, in press.
150 JOHN EDWARDS AND HOWARD GILES

Ball, P. and Giles, H. ( 1982). Do I choose to master your language? Polycom, 30, (in
press).
Ball, P., Giles, H. and Hewstone, M. (in press). Second language acquisition: The
intergroup theory with catastrophic dimensions. In "The Social Dimension:
European Developments in Social Psychology" (H. Tajfel, ed.), Cambridge
University Press, Cambridge.
Baratz, J. (1970). Teaching reading in an urban Negro school system. In "Lanugage
and Poverty", (F. Williams, ed.), Markham, Chicago.
Barth, F. (1969). "Ethnic Groups and Boundaries: The Social Organization of
Culture Difference", Little, Brown and Company, Boston.
Beebe, L. M. (in press). Risk-taking and the language learner. In "Classroom
Language Acquisition and Use: New Perspectives", (H. Seliger and Michael
Long, eds.), Newbury House, Rowley, Mass.
Beebe, L. M. and Giles, H. (in press). Speech Accommodation Theories: A
discussion in terms of second language acquisition. International Journal of the
Sociology of Language.
Bereiter, C. and Engelmann, S. (1966). "Teaching Disadvantaged Children in the
Pre-school", Prentice-Hall, Englewood Cliffs, N .J.
Berke-Gleason, J. (1980). The acquisition of social speech: Routines and politeness
formulas. In "Language: Social Psychological Perspectives", (H. Giles, W. P.
Robinson and P. M. Smith, eds), Pergamon, Oxford.
Bernstein, B. (1972). Social class, linguistic and socialization. In, "Language and
Social Context'', (P. Giglioli, ed.), Penguin, Harmondsworth.
Bernstein, B. (1973). A brief account of the theory of codes. In "Social Relationships
and Language", (V. Lee, ed.), Open University Press, Milton Keynes.
Bernstein, B. and Henderson, D. (1969). Social class differences in the relevance of
language to socialization. Sociology, 3, 1-20.
Bishops, G. D. (1979). Perceived similarity in interracial attitudes and behaviours:
The effects of belief and dialect style. Journal of Applied Social Psychology, 9,
446-465.
Bourhis, R. Y. (1982). Language policies and language attitudes: Le monde de la
Francophonie. In "Attitudes Towards Language Variation: Social and Applied
Contexts", (E. B. Ryan and H. Giles, eds), Edward Arnold, London.
Bourhis, R. Y. and Genesee, F. (1980). Evaluative reactions to code switching
strategies in Montreal. In "Language: Social Psychological Perspectives", (H.
Giles, W. P. Robinson and P. M. Smith, eds), Pergamon, Oxford.
Bourhis, R. Y. and Giles, H. (1976). The language of cooperation in Wales: A field
study. Language Sciences, 42, 13-16.
Bourhis, R. Y., Giles, H. and Rosenthal, D. (1981). Notes on the construction ofa
"subjective vitality questionnaire". Journal of Multilingual and Multicultural
Development, 2, 145-155. ·
Bradac, J. J. (1982). A rose by another name: Attitudinal conseuqnces of lexical
variation. In "Attitudes Towards Language Variation: Social and Applied
Contexts", (E. B. Ryan and H. Giles, eds), Edward Arnold, London.
Brody, E. B. (1968). "Minority Group Adolescents in the United States", Williams
and Witkins, Baltimore.
Brown, R. (1965). "Social Psychology", Collier Macmillan, London.
Brown, R. J. and Turner, J. (1981). Interpersonal and intergroup behaviour, In
"Intergroup Behaviour", (J.C. Turner and H. Giles, eds), Blackwell, Oxford.
Bruner, J. (1981). The social context of language acquisition. Language and Com-
munication, 1, 155-178.
 SOCIOLINGUISTICS AND EDUCATION 151
Carranza, M. A. (1982). Attitudinal research on Hispanic language varieties. In
"Attitudes Towards Language Variation: Social and Applied Contexts", (E. B.
Ryand and H. Giles, eds), Edward Arnold, London.
Carranza, M.A. and Ryan, E. B. (1975). Evaluative reactions of bilingual Anglo and
Mexican American adolescents toward speakers of English and Spanish. Inter-
national Journal of the Sociology of Language, 6, 83-104.
Choy, S. and Dodd, D. (1976). Standard-English-speaking and non-standard
Hawaiian-English-speaking children: Comprehension of both dialects and
teachers' evaluations. Journal of Educational Psychology, 68, 184-193.
Clement, R. (1980). Ethnicity, contact and communicative competence in a second
language. In "'Language: Social Psychological Perspectives", (H. Giles, W. P.
Robinson and P. M. Smith, eds), Pergamon, Oxford.
Cook-Gumperz, J. (1973). "Social Control and Socialization: A Study of Class
Differences in the Language of Maternal Control", Routledge & Kegan Paul,
London.
Covington, A. (1976). Black people and Black English: Attitudes and deeducation in
a biased macroculture. In "Black English: A Seminar", (D. Harrison and T.
Trabasso, eds), Erlbaum, Hillsdale, N.J.
Danet, B. (1980). Language in the courtroom. In "Language: Social Psychological
Perspectives", (H. Giles, W. P. Robinson and P. M. Smith, eds), Pergamon,
Oxford.
Day, R. R. (1982). Children's attitudes towards language. In "Attitudes Toward
Language Variation: Social and Applied Contexts", (E. B. Ryan and H. Giles,
eds), Edward Arnold, London.
Delia, J. G. (1972). Dialects and the effects of stereotypes on interpersonal attraction
and cognitive processes in impression formation. Quarterly Journal of Speech,
58, 285-297.
Denison, N. (1977). Language death or language suicide? lnternationallournal ofthe
Sociology of Language, 12, 13-22.
Deschamps, J-C. (1977). "L'Attribution et la Categorisation Sociale", Peter Lang,
Berne.
Dion, K. L., Earn, B. M. and Yee, P.H. N. (1978). Theexperienceofbeingavictim
of prejudice: An experimental approach. International Journal of Psychology,
13, 197-214.
Drake, G. (1980). The social role of slang. In "Language: Social Psychological
Perspectives", (H. Giles, W. P. Robinson and P. M. Smith, eds), Pergamon,
Oxford.
Dulay, H. and Burt, M. (1978). From research to method in bilingual education. In
"International Dimensions of Bilingual Education: Georgetown University
Roundtable on Languages and Linguistics, 1978", (J.E. Alatis, ed.), Georgetown
University Press, Washington, DC.
Dusek, J. (1975). Do teachers bias children's learning? Review 'of Educational
Research, 45, 661-684.
Dweck, C. S. (1975). The role of expectation and attribution in the alleviation of
learned helplessness. Journal of Personality and Social Psychology, 31, 674-685.
Edwards, A. D. (1976). Social class and linguistic choice. Sociology, 10, 101-110.
Edwards, J. R. (1977). Ethnic identity and bilingual education. In "Language,
Ethnicity and Intergroup Relations", (H. Giles, ed.), Academic Press, London
and New York.
Edwards, J. R. (1979a). "Language and Disadvantage", Edward Arnold, London.
152 JOHN EDWARDS AND HOWARD GILES

Edwards, J. R. (1979b). Judgements and confidence in reactions to disadvantaged

speech. In "Language and Social Psychology'', (H. Giles and R. N. St. Clair,
eds), Blackwell, Oxford.
Edwards, J. R. (1980). Critics and cr,iticisms of bilingual education. Modem Language
Journal, 64, 409-415.
Edwards, J. R. ( 1981a). Psychological and linguistic aspects of minority education. Jn
"World Yearbook of Education 1981: Education of Minorities", (J. Megarry, S.
Nisbet and E. Hoyle, eds), Kogan Page, London.
Edwards, J. R. (1981b). "The Social Psychology of Reading", Institute of Modem
Languages, Silver Spring, Maryland.
Edwards, J. R. (1981c). The context of bilingual education. Journal of Multilingual
and Multicultural Development, 2, 25-44.
Edwards, J. R. (1982). Language attitudes and their implications among English
speakers. In "Attitudes Towards Language Variation: Social and Applied
Contexts", (E. B. Ryan and H. Giles, eds), Edward Arnold, London.
Edwards, J. R. (in press). Bilingual education revisited: A reply to Donahue.
Journal of Multilingual and Multicultural Development.
Edwards, V. K. (1979). "The West Indian Language Issue in British Schools'',
Routledge & Kegan Paul, London.
Elias, N. (1978). "The Civilizing Process", Urizen, New York.
Eltis, K. (1980). Pupils' speech style and teacher reaction: Implications from some
Australian data. EngliSh in Australia, 51, 27-35.
Fishbein, M. (1980). A theory of reasoned action: Some applications and implica-
tions. Nebraska Symposium on Motivation, 28, 1-25.
Fishman, J. A. (1977). Language and ethnicity. In "Language, Ethnicity and
Intergroup Relations", (H. Giles, ed.), Academic Press, London and New
York.
Fuchs, E. (1973). How teachers learn to help children fail. In "Tinker, Tailor ... The
Myth of Cultural Deprivation", (N. Keddie, ed.), Penguin, Harmondsworth.
Gardner, R. C. (1979). Social psychological aspects of second language acquisition.
Jn "Language and Social Psychology", {H. Giles and R. N. St. Clair, eds),
Blackwell, Oxford.
Gardner, R. C. {1982). Language attitudes and language learning. In "Attitudes
Toward Language Variation: Social and Applied Contexts", (E. B. Ryan and
H. Giles, eds), Edward Arnold, London.
Giles, H. (1977). "Language, Ethnicity and Intergroup Relations", Academic Press,
London and New York.
Giles, H. (1978). Linguistic differentiation between ethnic groups. In "Differentiation
Between Social Groups", (H. Tajfel, ed.), Academic Press, London and New
York.
Giles, H. (1979). Ethnicity markers in speech. In "Social Markers in Speech", (K. R.
Scherer and H. Giles, eds), Cambridge University Press, Cambridge.
Giles, H., Baker, S. and Fielding, G. (1975). Communication length as a behavioural
index of accent prejudice. International Journal ofthe Sociology of Language, 6,
73-81.
Giles, H. and Bourhis, R. Y. (1976a). Black speakers with white speech - a real
problem? In "Proceedings of the Fourth AILA Congress, Volume l", (G.
Nickel, ed.), Hochschul Verlag, Stuttgart.
Giles, H. and Bourhis, R. Y. (1976b). Methodological issues in dialect perception:
Some social psychological perspectives. Anthropological Linguistics, 18, 294-
304.
 SOCIOLINGUISTICS AND EDUCATION 153
Giles, H., Bourhis, R. Y. and Taylor, D. M. (1977). Towards a theoryoflanguagein
ethnic group relations. Jn "Language, Ethnicity and Intergroup Relations". (H.
Giles, ed.), Academic Press, London and New York.
Giles, H. and Byrne, J. L. (1982). An intergroup approach to second language
acquisition. Journal of Multilingual and Multicultural Development, 3, 17-40.
Giles, H. and Edwards, J. R. (1983). Language attitudes in multicultural settings.
Special Issue of the Journal of Multilingual and Multicultural Development, 4.
Giles, H. and Farrar, K. (1979). Some behavioural consequences of speech and dress
styles. British Journal of Social and Clinical Psychology, 18, 209-210.
Giles, H. and Johnson, P. (1981). The role of language in ethnic group relations. Jn
"Intergroup Behaviour", (J.C. Turner and H. Giles, eds), Blackwell, Oxford.
Giles, H. and Marsh, P. (1979). Perceived masculinity and accented speech. Language
Sciences, l, 301-315.
Giles, H. and Powesland, P. F. (1975), "Speech Style and Social Evaluation",
Academic Press, London and New York.
Giles H., Robinson, W. P. and Smith, P. M. (1980) ...Language: Social Psychological
Perspectives", Pergamon, Oxford.
Giles, H. and Saint-Jacques, B. (1979). "Language and Ethnic Relations", Pergamon,
Oxford.
Giles, H. and Smith, P. M. (1979). Accommodation theory: Optimal levels of
convergence. In "Language and Social Psychology", (H. Giles and R. N. St
Clair, eds), Blackwell, Oxford.
Giles, H., Smith, P. M., Ford, B., Condor, S. and Thakerar, J. N. (1980). Speech
styles and the fluctuating salience of sex. Language Sciences, 2, 260-282.
Gleitman, L. and Gleitman, H. (1970). "Phrase and Paraphrase", Norton, New
York.
Guimond, S. and Simard, L. M. (1979). Perception et interpretation des inegalites
economiques entre Francophones et Anglophones au Quebec. Paper presented
to the 40th. Congress of the Canadian Psychological Society, Quebec City, June.
Gumperz, J. J. and Hernandez-Chavez, E. (1972). Bilingualism, bidialectalism and
classroom interaction. Jn "Functions of Language in the Classroom", (C. Cazden,
V. John and D. Hymes, eds). Teachers College Press, New York.
Hamers, J. F. and Blanc, M. (1982). Towards a social psychological model of
bilingual development. Journal of Language and Social Psychology, 1, (in
press).
Hamilton, D. L. (1979). A cognitive-attributional analysis of stereotyping. Jn
"Advances in Experimental Social Psychology, Volume 12", (L. Berkowitz,
ed.), Academic Press, New York and London.
Harari, 0. and Covington, M. V. ( 1981). Reactions to achievement behaviour from a
teacher and student perspective: A developmental analysis. American Educa-
tional Reserach Journal, (in press).
Harris, R. McL. (1979). Fever of ethnicity: The sociological and educational sig-
nificance of the concept. In "Mosaic or Melting Pot", (P. R. de Lacy and M. E.
Poole, eds), Harcourt Brace, Jovanovich, Sydney.
Hess, R. and Shipman, V. (1965). Early experience and the socialization of cognitive
modes in children. Child Development, 36, 869-886.
Hess, R. and Shipman, V. (1968a). Maternal attitudes towards the school and the
role of the pupil: Some social class comparisons. Jn "Developing Programs for
the Educationally Disadvantaged", (A. Passow, ed.), Teachers College Press,
New York.
154 JOHN EDWARDS AND HOW ARD GILES

Hess, R. and Shipman, V. (1968b). Maternal influences upon early learning: The
cognitive environments of urban pre-school children. In "Early Education,,,
(R. Hess and R. Bear, eds), Aldine, Chicago.
Hewstone, M. and Jaspars, J. (1982a). Intergroup relations and attribution processes.
In "Social Identity and Intergroup Behaviour", (H. Tajfel, ed.), Cambridge
University Press, Cambridge.
Hewstone, M. and Jaspars, J. (1982b). Explanations for racial discrimination: The
effect of group discussion on intergroup attributions. European Journal ofSocial
Psychology, 12, 1-16.
Hudson, R. (1980). "Sociolinguistics", Cambridge University Press, Cambridge.
Jaspars, J. and Hewstone, M. (in press), Cross-cultural interaction, social attribution
and intergroup relations. In "Cultures in contact: Studies in Cross-Cultural
Interaction", (S. Bochner, ed.), Pergamon, Oxford.
Kalin, R. (1982). The social significance of speech in medical, legal and occupational
settings. Jn "Attitudes Toward Language Variation: Social and Applied Con-
texts", (E. B. Ryan and H. Giles, eds), Edward Arnold, London.
Kelley, H. H. (1967). Attribution theory in social psychology. Nebraska Symposium
on Motivation, 15, 192-238.
Khlief, B. (1979). Insiders, outsiders and renegades: Towards a classification of
ethnolinguistic labels. In .. Language and Ethnic Relations", (H. Giles and B.
Saint-Jacques, eds), Pergamon, Oxford.
Kochman, T. (1976). Perceptions along the power axis; A cognitive residue of
interracial encounters. Anthropological Linguistics, 18, 261-273.
Kochman, T. (1981). "Black and White Styles in Conflict", Chicago University
Press, Chicago.
Kramarae, C. (1981). "Women and men speaking", Newbury House, Rowley, Mass.
Labov, W. (1966). "The Social Stratification of English in New York City", Center
for Applied Linguistics, Washington, DC.
Labov, W. (1973). The logical of nonstandard English. Jn "Tinker, Tailor ... The
Myth of Cultural Deprivation", (N. Keddie, ed.), Penguin, Harmondsworth.
Labov, W. (1976). "Lahguage in the Inner City", University of Pennsylvania Press,
Philadelphia.
Lambert, W. E. (1967). A social psychology of bilingualism. Journal of Social Issues,
23, 91-109.
Lambert, W. E. (1974). Culture and language as factors in learning and education. In
"Cultural Factors in Leaming'', (F. Aboud and R. D. Meade, eds), Western
Washington State College Press, Bellingham.
Lambert, W. E., Hodgson, R., Gardner, R. C. and Fillenbaum, S. (1960). Evalua-
tional reactions to spoken languages. Journal of Abnormal and Social Psy-
chology, 60, 44-51.
Lenneberg, E. (1967). "Biological Foundations of Language", Wiley, New York.
Lind, E. A. and O'Barr, W. M. (1979). The social significance of speech in the
courtroom. In "Language and Social Psychology", (H. Giles and R. N. St. Clair,
eds), Blackwell, Oxford.
Louden, D. M. (1978). Self-esteem and locus of control in minority group adolescents.
Ethnic and Racial Studies, l, 196-217.
Love, N. ( 1981). Making sense of Chomsky's revolution. Language and Communica-
tion, 1, 275-287.
Lyman, S. M. and Douglass, W. A. (1973). Ethnicity: Strategies of collective and
individual impression management. Social Research, 40, 344-365.
 SOCIOLINGUISTICS AND EDUCATION 155
Marwit, S. (1977). Black and white children's use of standard English at 7, 9 and 12
years of age. Developmental Psychology, 13, 81-82.
Marwit, S.- and Marwit, K. (1976). Black children's use of nonstandard grammar:
Two years later. Developmental Psychology, 12, 33-38.
Marwit, S., Marwit, K. and Boswell, J. (1972). Negro children's use of nonstandard
grammar. Journal of Educational Psychology, 63, 218-223.
McKirnan, K. J. and Hamayan E. V. (1980). Language norms and perceptions of
ethnolinguistic group diversity. In "Language: Social Psychological Perspectives",
{H. Giles, W. P. Robinson and P. M. Smith, eds), Pergamon, Oxford.
Milner, D. (1981). Racial prejudice. In "Intergroup Behaviour", (J.C. Turner and
H. Giles, eds), Blackwell, Oxford.
Moses, R. A., Daniels, H. A. and Grundlach, R. A. (1976). Teachers' language
attidues and bidialectalism. International Journal of the Sociology of Language,
8, 77-91.
Nicholls, J. G. (1979). The development of perception of own attainment and causal
attributions for success and faiJure in reading. Joumal of Educational Psychology,
74, 94-99.
Ovando, C. (1978). Male and female Latino college aspirations: The implications for
pluralistic education. Educational Research Quarterly, 2, 106-122.
Pendleton, D., Brouwer, H. and Jaspars, J. (1982). Communication difficulties: The
doctor's perspective. Journal of Language and Social Psychology, Vol.?
Pettigrew, T. F. {1979). The ultimate attribution error: Extending Allport's cognitive
analysis of prejudice. Personality and Social Psychology Bulletin, S, 461-476.
Philips, S. (1972). Acquisition of roles for appropriate speech usage. In "Language
and cultural diversity in American education", (R. Abrahams and R. Troike,
eds), Prentice-Hall, Englewood Cliffs, N.J.
Piestrup, A. M. (1973). Black dialect interference and accommodation of reading
instruction in first grade, Monographs of the Language Behaviour Research
Laboratory, University of California, Berkeley.
Platt, J. and Weber, H. (in press). Speech convergence miscarried: An investigation
into inappropriate accommodation strategies. International Journal of the
Sociology of Language.
Ramey, C. and Campbell, F. (1977). Prevention of developmental retardation in
high risk children. Jn "Research to practice in mental retardation: Care and
intervention, Volume l" (P. Mittler, ed.), University Park Press, Baltimore.
Ramey, C. and Campbell, F. (1979). Compensatory education for disad-..:antaged
children. School Review, 87, 171-189.
Rist, R. (1970). Student social class and teacher expectations: The self-fulfilling
prophecy in ghetto education. Harvard Educational Review, 40, 411-451.
Robinson, W. P. (1978). "Language Management in Education: The Australian
Context", Allen and Unwin, Sydney.
Robinson, W. P. (1979). Speech markers and socioeconomic status. In "Social
Markers in Speech", (K. R. Scherer and H. Giles, eds), Cambridge University
Press, Cambridge.
Robinson, W. P. {1980). Language management, socio-economic status and educa-
tional progress. In "Language and Language Disorders in Childhood", {L. A.
Hersov and M. Berger, eds), Pergamon, Oxford.
Robinson, W. P. {1981a). "Communication in Development", Academic Press,
London and New York.
Robinson, W. P. {1981b). Mothers' answers to children's questions: From socio-
156 JOHN EDWARDS AND HOWARD GILES

economic status to individual differences. In "Communication in Development",

(W. P. Robinson, ed.}, Academic Press, London and New York.
Robinson, W. P. and Rackstraw, S. J. (1972). "A Question of Answers, Two
Volumes", Routledge & Kegan Paul, London.
Rosenfeld, H. (1978). Conversational control function of nonverbal behaviour. In
"Nonverbal Behaviour and Communication", (A. W. Siegman and S. Feldstein,
eds), Erlbaum, Hillsdale, N.J.
Rothenthal, R. and Jacobson, L. (1968). "Pygmalion in the Classroom", Holt,
Rinehart and Winston, New York.
Ryan, E. B. (1979). Why do low-prestige language varieties persist? In "Language
and Social Psychology'', (H. Giles and R. N. St. Clair, eds), Blackwell, Oxford.
Ryan, E. B. (1981). Language planning from an attitudinal perspective. Paper
presented at the International Conference on Language Problems and Public
Policy, Cancun, Mexico, December.
Ryan, E. B. and Carranza, M.A. {1977). IngroupandoutgroupreactionstoMexican
American language varieties. In "Language, Ethnicity and Intergroup Relations",
(H. Giles, ed.), Academic Press, London and New York.
Ryan, E. B. and Giles, H. {1982). "Attitudes Toward Language Variation: Social
and Applied Contexts", Edward Arnold, London.
Ryan. E. B., Giles, H. and Sebastian, R. J. (1982). An integrative perspective for the
study of attitudes toward language variation. In "Attitudes Toward Language
Variation: Social and Applied Contexts'', (E. B. Ryan and H. Giles, eds),
Edward Arnold, London.
Salamone, F. A. and Swanson, C. H. (1979). Identity and ethnicity: Ethnic groups
and interactions in a multi-ethnic society. Ethnic Groups, 2, 167-183.
Sapir, E. (1949). "Culture, Language and Personality", University of California
Press, Berkeley.
Scherer. K. R. (1980). Personality.emotion, psychopathology and speech. /n"Social
Markers in Speech", {K. R. Scherer and H. Giles, eds), Cambridge University
Press, Cambridge.
Scallon, R. and Scallon, S. B. K. (1979). Literacy as inter-ethnic communication: An
Athabaskan case. Sociolinguistic Working Papers, 59.
Sebastian, R. J. and Ryan, E. B. (in press). The effects of speech cues on social
evaluation and behaviour. In "Recent Advances in Language, Communication
and Social Psychology", (H. Giles and R. N. St. Clair, eds), Erlbaum, Hillsdale,
N.J.
Sebastian, R. J., Ryan, E. B. and Corso, L. (in press). Social judgement of speakers
with differing degrees of accent. Social Behaviour and Personality.
Seligman, C., Tucker, G. R. and Lambert, W. E. (1972). Theeffectsofspeechstyle
and other attributes on teachers' attitudes toward pupils. Language in Society, 1,
131-142.
Sherif. M. (1967). "Group Conflict and Co-operation", Routledge & Kegan Paul,
London.
Shuy, R. (1971). Sociolinguistic strategies for studying urban speech. Bulletin of the
School of Education (Indiana University), 41, 1-25.
Shuy, R. ( 1977). Problems of communication in the cross-cultural medical interview.
ITL, 35.
Smith, P. M., Giles, H. and Hewstone, M. (1982). Speech variation and social
situations. Jn "Sociogenesis of Language and Human Conduct", (B. Bain, ed.),
Plenum, New York.
 SOCIOLINGUISTICS AND EDUCATION 157
Snyder, M. (1980). Seek, and ye shal1 find: Testing hypotheses about other people. In
"Social Cognition: The Ontario Symposium", (E. T. Higgins et al., eds),
Erlbaum, Hillsdale, N.J.
St. Clair, R. N. (1982). From social history to language attitudes. In "Attitudes
Towards Language Variation: Social and Applied Contexts", (E. B. Ryan and
H. Giles, eds), Edward Arnold, London.
Stephenson, G. M. (1982). Intergroup bargaining and negotiation. In "Intergroup
Behaviour", (J. C. Turner and H. Giles, eds), Blackwell, Oxford.
Stipeck, D. and Nelson, K. (1980). Communication efficiency of middle-and lower-
SES dyads. Human Communications Research, 6, 168-177.
Street, R. L. and Hopper, R. (1982). A model of speech style evaluation. In
"Attitudes Toward Language Variation: Social and Applied Contexts", (E. B.
Ryan and H. Giles, eds), Edward Arnold, London.
Tajfel, H. (1978). "Differentiation Between Social Groups", Academic Press,
London and New York.
Tajfel, H. and Turner, J.C. (1979). An integrative theory of intergroup conflict. In
"The Social Psychology of Intergroup Relations", (W. G. Austin and S.
Worchel, eds), Brooks/Cole, Monterey.
Taylor, D. M. and Simard, L. (1981). "Les Relations lntergroupes au Quebec et la
Loi 101 ". Gouvernement du Quebec: Office de la langue fran9aise.
Thakerar, J. N. and Giles, H. (1981). They are-so they speak: Noncontent speech
stereotypes. Language and Communication, l, 251-256.
Thakerar, J. N., Giles, H. and Cheshire, J. (1982). Psychological and linguistic
parameters of speech accommodation theory. Jn "Advances in the Social
Psychology of Language", (C. Fraser and K. R. Scherer, eds), Cambridge
University Press, Cambridge.
Thom, R. (1975). "Structural Stability and Morphogenesis", Benjamin, Reading,
Mass.
Torrey, J. (1973). Illiteracy in the ghetto. Jn "Tinker, Tailor ... The Myth of Cultural
Deprivation", (N. Keddie, ed.}, Penguin, Harmondsworth.
Tough, J. (1977). "Talking and Learning", Ward Lock, London.
Trudgill, P. (1972). Sex, covert prestige and linguistic change in the urban British
English of Norwich. Language in Society, 1, 179-195.
Trudgill, P. (1974). "The Social Differentiation of English in Norwich", Cambridge
University Press, London.
Trudgill, P. (1975). ••Accent, Dialect and the School", Edward Arnold, London.
Trudgill, P. and Giles, H. (1977). Sociolinguistics and linguistic value judgements:
correctness, adequacy and aesthetics. In "The Functions of Language and
Literature Studies", (F. Coppierters and D. Goyvaerts, eds), Story-Scientia,
Ghent.
Turner, J. C. ( 1981). The experimental social psychology of intergroup behaviour. Jn
"Intergroup Behaviour", (J. C. Turner and H. Giles, eds}, Blackwell, Oxford.
Ullrich, H. E. (1971). Linguistic aspects of antiquity: A dialect study. Anthropological
Linguistics, 13, 106-113.
Vaughan, G. M. (1978). Social categorization and inter-group behaviour in children.
Jn "Differentiation Between Social Groups", (H. Tajfel, ed.), Acadmic Press,
London and New York.
Wells, G. (1981). "'Learning through Interaction: The Study of Language Develop-
ment", Cambridge University Press, Cambridge.
Wells, G. and Robinson, W. P. (1982). The role of adult speech in language
158 JOHN EDWARDS AND HOWARD GILES

development. In ••Advances in the Social Psychology of Language", (ed.?),

Cambri~ge Univeristy Press, Cambridge.
Wiggen. G. ( 1978). The use of dialects in the initial teaching of the written language:
The Norwegian case. Paper presented at A/LA conference, Montreal.
Williams, F. (1976) ...Explorations of the Linguistic Attitudes of Teachers", Newbury
House, Rowley, Mass.
Williams, F., Whitehead. J. L. and Miller, L. M. (1971). Ethnic stereotyping and
judgements of children's speech. Speech Monographs, 38, 166-170.
Williams, F., Whitehead, J. L. and Miller. L. M. (1972). Relations between attitudes
and teacher expectancy. American Educational Research Journal, 9, 263-277.
Worchel, S. and Norvell, N. (1980). Effect of previous environmental conditions
during cooperation on intergroup attraction. Journal of Personality and Social
Psychology, 38, 764-772.
Zeeman, E. C. ( 1976). Catastrophe theory. Scientific American, 234, 65-83.

Note
We are most grateful to W. Peter Robinson and Miles Hewstone for their comments
and contributions to an earlier draft of this chapter.
 7
Applied sociology of language:
vernacular languages and education

A. G. H. Walker

The average English person in the average English town will grow up with
English as the language of their home and of their education. Only rarely
will he or she come into serious contact. with a foreign language. In other
words most English people spend their lives in a state of monolingualism.
However, such monolingualism is a minority phenomenon in terms of the
total world population; bi- or multilingualism is found in most countries in
the world. There are an estimated 394 languages in Nigeria (Hansford, 1976)
an~ 760 languages in Papua New Guinea (Wurm, 1979). This does not mean
that all speakers of these languages are multilingual, but it is probable that
such a concentrated variety of languages in a relatively small area will lead to
multilingualism, especially with increased mobility. Similarly, Fishman and
Hofmann (1966) estimate that in America in 1960 there were some 18 352 351
speakers of the 23 major non-English mother-tongue groups and Spolsky
(1972) considers there were a futher 40 or 50 smaller languages or language
groups. In England it is estimated that in 1978 in London alone there were
over 100 languages spoken by school-age children.' .
In the monolingual setting, the concept of "vernacular language" is not too
difficult to define, nor is there such a wide variety of sociolinguistic problems
connected with education. In the bi- or multilingual setting, however,
"ve~acular language" is a more nebulous concept, and the sociolinguistic
problems connected with its representation in education more numerous.

APPLIED SOCIOLINGUISTICS Copyright © 1984 by Academic Press, London

ISBN: 0-12-701220-6 All rights of reproduction in any fonn reserved
160 VERNACULAR LANGUAGES AND EDUCATION

Thus, in the wide field of education in bi- or multilingual societies socio-

linguistics has an important role to play, first of all in the identification and
comprehension of problems, and secondly in the solving of these problems.
In this chapter I shall briefly discuss the concept of "vernacular language",
describe various reasons why vernacular language is necessary in education
and then illustrate some of the problems and possible solutions to be found
in this context. As my own particular area of research is North Frisia (NF) in
West Germany, where at the time of writing Frisian is being taught in 12
schools, many examples are taken from here.

What is a ''vernacular language''

A vernacular language might be defined as the indigenous language used as
the primary means of socialization within the family or tribal unit within a
certain area. 2 In a monolingual setting such a definition is not problematical,
whereas it can pose some problems in a multilingual context. What, for
example, is the vernacular language in the quinquelingual villages of
Rodenas and Neukirchen in NF where Frisian, Jutish and Low German can
all be considered indigenous in contrast to the High German and Danish
standard languages also spoken here (Larsen, 1983; Spenter, 1976; Walker,
1980a)? All these languages can be used within the family unit, as each
language can be used for addressing specific members of the family group.
Thus, for example, a boy may address his father in Jutish, his mother in
Frisian and his brother-in-law, who comes from the same village, in Low
German. As all languages are in daily use in the family and in the sur-
rounding neighbourhood, and as it is not possible to ascribe each language
an exclusive set of domains generally valid for the two villages, they must
presumably all be considered vernacular languages of more or less equal
standing. However, although the Frisian case may be of some theoretical
interest, I should like to adhere to the above definition of vernacular
language and contrast this concept with those of the (1) regional language,
(2) official language, (3) national language, and (4) international language.

Regional language
UNESCO (1968, p. 689) gives the following definition for a regional
language: "A language which is used as a medium of communication
between peoples living within a certain area who have different mother
tongues". Thus, for example, NF had Low German as its regional language
as speakers of mutually unintelligible Frisian dialects had to revert to it for
communication. Today High German has largely taken over this role.
 A. G. H. WALKER 161
A regional language is practically synonymous with a lingua franca, the
main difference being perhaps that the latter is not so geographically
restricted as the former. The four languages Swahili, Hausa, Arabic and
Mandingo are, for example, all African linguae francae, all of which
encompass at least six different African states.

Official language
An official language is one designated by governmental decree to be the
official means of communication of the given state in government, admin-
istration, law, education and general public life. Kloss (1968) distinguishes
between endoglossic and exoglossic official languages, whereby an endo-
glossic official language is one indigenous to the state, e.g. English in
England, French in France, whereas an exoglossic one is imported, e.g.
English in Kenya, Ghana, Sierra Leone, or French in Zaire, and Portuguese
in Mozambique. Tanzania has both an endoglossic (Swahili) as well as an
exoglossic (English) official language.

National language
Several writers have discussed the concept of national language (Fishman,
1968a; Le Page, 1964; Nida and Wonderly, 1971) but I should like here to
follow Heine (1979) where he differentiates three types of national
language:
(1) The de jure national language. In this case the national language has
been officially chosen by governmental decree. Thus, for example,
Tanzania elected Swahili its national language in 1961. Namibia has
three national languages: Afrikaans, English and German, and Nigeria
has nine. It is interesting that of the 46 African states only 22 have an
"official" national language.
(2) The de facto national language. This "unofficial" national language
must fulfill two of the following criteria:
(a) it must be used as a spoken medium throughout the nation and be
spoken by more than half the population;
(b) the language must symbolize national unity or identity;
(c) the language must be considered as a means of expressing national
culture and the national way of life.
An example of the de facto national language would be the Wolof
language in Senegal.
(3) The de jure and de facto national language. Examples of an "official"
national language with a firm numerical base are Swahili in Tanzania,
Somali in Somalia and Arabic in Algeria.
162 VERNACULAR LANGUAGES AND EDUCATION

International language
An international language is one used over wide parts of the world for
inter-territorial communication.
In each state of the world one or more languages are spoken, each of
which corresponds to one or more of the categories outlined above. In
Britain, for example, English is a vernacular (together with Scots Gaelic and
Welsh) and a lingua franca, as well as the official and national language of
Great Britain. In West Germany, East and North Frisian are purely ver-
naculars, Low German both a vernacular and a regional language, and High
German a vernacular, lingua franca and the official national language of the
Federal Republic. In Nigeria English is the official language, there are nine
national languages: Hausa, Yoruba, lgbo, Ful, Kanuri, Efik, Edo, ldoma
and Ijo. Hausa, Yoruba and Igbo are linguae francae and there are an
estimated 394 vernaculars. Cameroon, Liberia and Mozambique are
examples of countries with no national language.
Having briefly discussed different types of language, I should now like to
differentiate three language groups, i.e. groups of speakers all using the
vernacular, from which I shall be taking my examples. These are: (1)
indigenous linguistic minorities, (2) speakers in developing countries, and
(3) immigrants.

Indigenous linguistic minorities

In Western Europe alone there are some 50 minority languages {Stephens,
1978) and in North America there was an estimated 147 American Indian
languages in 1941 (Ohannessian, 1972). A linguistic minority can perhaps be
defined as a group of people who, within the family or tribal unit and in their
everyday lives, predominantly speak a language which differs from the
language spoken by the majority of the population in that particular
country. Linguistic minorities can range in size from, for example, the
estimated 2000 speakers of East Frisian (Saterlandic) (Fort, 1980; Sjolin.
1969) to the 10-12 million speakers of Occitan (Haarmann, 1975).

Developing countries
A developing country is one which is generally considered to be not as highly
developed technologically or economically as for example the United States
or the countries of Western Europe. In a great number of developing
countries an international language, often the language of the colonialists,
has been used as the official language and language of education. However,
 A. G. H. WALKER 163

following independence, there have often been attempts to introduce the

vernacular into more societal domains including education.

Immigrants
An immigrant is a person who has moved into a different country from the
one he was born in. Immigrants are a somewhat different case to the two
language groups already mentioned as they are not indigenous and have
usually entered the country they now live in of their own volition. However,
they often form a linguistic minority, using a language which is indigenous
elsewhere and which is not spoken by. the majority of the population in their
present country of abode. As a lot of the problems encountered by immi-
grants are similar to those of other linguistic minorities, I have included
them here.

Why vernacular language in education?

Fallowing the short description of the types of language and language groups
relevant to this paper, I shall now consider the reasons why vernacular
education is desirable.

The social factor

When a child goes to school for the first time, it is usually the first major
change in his life. The relative freedom of home life where he can run
around, make noise and where he is usually one of a few children under his
mother's eye, is exchanged for the relatively disciplined atmosphere of
school where he has to sit still, is but one of a larger group in the classroom
and where he has to do as he is told (UNESCO, 1968). It is an even greater
change if the child completely leaves home to go to boarding school as often
happened with the Apaches in America (Liebe-Harkort, 1980). In order to
facilitate the structural change in the child's daily life, the language used at
the new school should be the same as the one used at home so that he is not
confronted with a completely new linguistic as well as social situation.
A further aspect is the communication between the child's two worlds of
home and school. If the same language is used for both, the child can relate
his school experiences at home and vice versa, and comprehension is assured.
If, however, the child uses a second language at school which is incompre-
hensible to the parents, they will experience difficulties in understanding
their child's school life and in helping his education (UNESCO, 1968), and
this could lead to estrangement.
164 A. G. H. WALKER

In order to avoid the possible conflicts arising from primary socialization

at home in Ll and secondary socialization at school in L2, parents some-
times retain their mother tongue for communication between themselves
but prefer to teach their children L2. Similarly, when parents use L1 with
their first child but then notice that the child is having difficulties with L2 at
school, they sometimes change to L2 for their second and later children.
These two examples are often the cause for language shift in minority
languages. However, as the parents in such a situation often only have
imperfect knowledge of the dominant language L2, the children inherit the
mistakes made by their parents. If these mistakes become firmly imbedded
in the child's language performance, the teacher has the greatest difficulties
in ironing them out again. One classic example is the distinction in High
German between mich and mir. This is not known in Frisian or Low German
and consequently children of Low German and/or Frisian parents may well
confuse the usage of the two forms. Thus one often hears such constructions
as: Was hast du fur mir? As a result teachers often maintain that parents
should teach their children their LI and leave the teaching of the dominant
language L2 to those trained for it.

Ethnicity
One possible definition of ethnicity is "a sense of group identity deriving
from real or perceived common bonds such as language, race or religion"
(Isajiw, 1974, quoted after Edwards, 1977, p. 254). Nationalism can be
considered an extension of ethnicity, as the belief that nations possess
certain characteristics which differentiate them from other nations is as-
sociated with the desire to promote the strength and ambitions of the nation
one belongs to. Language is an important part of group identity, as it
provides a link with the past, and a distinctive feature differentiating one
group from those around it. Thus the Republic of Ireland states in Article 8
of its Constitution that "The Irish language as the national language is the
first official language". This is because, as stated in paragraph 6 of the white
paper The Restoration of the Irish Language (1965, p. 6).

The Irish language is the most distinctive sign of our nationality. Our present
position as an independent State derives in large measure from the idealism
evoked by the Irish language movement.

The Irish language is then considered a major unifying and consolidating

factor in the Irish nation. As a consequence it became national policy to
introduce Irish into the curriculum of all primary schools in Eire. Irish is,
however, an example of a language being used in schools to promote the
 VERNACULAR LANGUAGES AND EDUCATION 165

feeling of national identity even though the language is not native to the
majority of the population.
A more genuine mother-tongue education can be found in the teaching of
Eskimo children in Alaska. Trifonovitch discovered (1976) that one import-
ant by-product of the bilingual education programme in Alaska, which had
introduced the tuition of Eskimo children in their mother tongue, was an
increasing awareness of their cultural and ethnic identity. In contrast to this,
the monolingual English education of Apache children in North America
was leading to suicides among the children (Paulston, 1980) or a complete
rejection of American culture (Liebe-Harkort, 1980) as their own ethnic
identity was being negated.

The child's cognitive development

Wells (1981) argues that the nature of interaction between a child and his
parents and teachers is important for the literacy and scholastic development
of a child. If a child is unable to communicate adequately with his teachers
because of being taught in L2, this will stultify his linguistic development.
Furthermore, if the area of intellectual activity at school is in a language
which is not readily comprehensible to the child's parents, this will also
restrict the possibilities of intellectual stimulus in the interaction between
the child and his parents. Intellectual stimulation is, however, necessary as it
develops the child's ability to make meaning explicit in a context-reduced
situation, which in turn is a prerequisite for scholastic achievement.
Swain (1983) points out that the skills most basic to academic progress and
achievement, such as the ability to master speech as a symbolic system, to
generalize and abstract, are most easily learnt in Ll. As these skills are
cross-lingual, they can easily be applied to L2 as well. Thus it is easier to
learn to read in L1 and then to apply this skill to L2 than to learn to read and
learn L2 simultaneously. Once the reading skill is automated through Ll,
more attention can be paid to the acquisition of L2.
Macnamara (1967) considers that education through a weaker language
has negative consequences for a child's academic progress, which Cummins
(1976) in turn interprets in terms of a threshhold of linguistic competence
which has to be attained if a child is to have any benefit from education in a
certain language. For a child to be able to cope with two languages, he needs
a proper education in Ll, as experiments, for example in the French immer-
sion programmes in Canada, show that children having difficulty with L2
leap forward once Ll is properly taught. H education is to be partly in L2,
this can only be in a bilingual situation where the child has achieved the
necessary threshholds in both languages. Bilingualism itself is also con-
sidered positive as it seems to accelerate the development of a child's verbal
166 A. G. H. WALKER

and non-verbal abilities and there is a positive association between bi-

Ii ngualism, cognitive flexibility and divergent thinking.

The preservation of a language

One aspect of linguistic imperialism has been the systematic attempt to
eradicate minority languages within the sphere of a nation's or state's power.
Thus Gregor (1980, p. 340) can quote the French MinisterofEducation who
stated in 1925 that "For the unity of France the Breton language must die".
The Times can also be quoted for its disparaging attitude towards Welsh
when it says in 1866 that "the Welsh language is ... the curse of Wales ... its
prevelance and the ignorance of the English language have excluded the
Welsh people from the civilisation of their English neighbours" (quoted
after Stephens, 1978, p. 159). Part of this policy of eradication was the
exclusive use of the dominant language in education, although this was often
coupled with the genuine belief that education in a minority language was
detrimental to a child's development. Thus Billigmeier (1979, p. 338) quotes
a teacher from Trin (Switzerland) who writes:

The school inspector, . . . , would not let an inspection conclude without

emphasizing to the teachers, to the school board, and to the students that
Romansh was an impediment to instruction and that it must be eliminated as
quickly as possible.

This policy, often reinforced by punishing children caught speaking the

minority language at school, unfortunately proved quite successful with a
number of minority languages, and one can conclude that the lack of
education in a minority language is one major factor for its demise. For this
reason vernacular education is deemed a major goal in most language
movements.
The reasons for the importance of vernacular education within a language
movement are: (1) vernacular education will instill children with a sense of
pride in their native tongue and culture which counteract the sense of
inferiority long connected with minority languages (cf. Ethnicity above); the
vernacular language can then be raised to the level of a prestige language
such as for example Manx on the Isle of Man (Gregor, 1980); (2) vernacular
education serves to develop a child's command of the language in an attempt
to make it a viable medium in the modem world; (3) vernacular education
will enable those children living in an area with a linguistic minority, but who
themselves have no command of the minority language, to learn it. This
would facilitate integration into the linguistic community and hopefully
prevent ostracization or language shift.
 VERNACULAR LANGUAGES AND EDUCATION 167

The official view

On the question of vernacular education, it is pleasing to note that politicians
and linguists are now obviously in agreement. The Recommendation 928
(1981) on the educational and cultural problems of minority languages and
dialects in Europe from the thirty-third ordinary session of the Parliamentary
Assembly of the Council of Europe gives a good example:

The Assembly, . . . Considering that the following principles should form the
basis for the scientific, human and cultural treatment of each language: ... - the
right of children to their own language and culture, ... Recommends that the
Committee of Ministers consider whether it would be possible for governments
of member states to implement the following measures in whatever manner most
appropriate: ... b. With regard to the human aspect, the gradual adoption of
children's mother tongues for their education ... c. With regard to the cultural
aspect, respect and official support for the local use of standardised minority
languages, and for their current use in higher education . . . in so far as this
approach is favoured by the communities which speak them; ...

The implementation of vernacular education

The problems arising in the context of the implementation of vernacular
education fall into three main fields of enquiry: (1) the society in which the
language is found (societal problems); (2) the language itself (linguistic
problems); and (3) the school in which the language is taught (problems
pertaining to the school).

Societal problems
One prerequisite for the implementation of a policy of vernacular education
is the analysis of the linguistic community concerned, in order to avoid
unfavourable reactions on the part of the populace. Stewart (1968, p. 532)
points out that:

Where reactions of this [unexpected] type have caught language planners un-
awares, it has not necessarily been because they were totally unpredictable, but
rather because not enough information was sought in advance about the ways in
which languages may interact with other aspects of society.

Thus, although it may be possible to enumerate a number of factors common

to different sorts of linguistic community (Ferguson, 1966; Gumperz, 1968;
Kloss, 1966, 1968; Rustow, 1968; Stewart, 1962), nevertheless each linguistic
community will have a configuration of factors unique to itseli, and it is this
168 A. G. H. WALKER

configuration which the linguist has to discover. (The validity of this state-
ment does admittedly depend upon the degree of abstraction one is working
with. It will, however, hold for the field-worker we are concerned with here
who is daily confronted with tangible problems on the micro-level).
A second prerequisite is to decide what the aims of the vernacular
education policy are in each particular instance (cf. Karam, 1974). There
may be a policy of language revival where a language which had become
extinct has been reintroduced as a spoken medium such as Hebrew (Fellman,
1973), Manx or Comish (Gregor, 1980; Stephens, 1978); a policy of language
maintenance as in the case of linguistic minorities where the language is in
decline, such as Welsh (Khleif, 1980) and Frisian (Boelens, 1976; Walker,
1980a); a policy confronted with divided language loyalties as in the case of
two parallel language systems in a country like Norway (Haugen, 1966); a
policy connected with large-scale language reform as in Turkey (Gallagher,
1971); a policy of language expansion introducing vernacular education into
an area where previously a colonial language had been the medium of
instruction. Here one would further differentiate between a vernacular
which is supposed to be developed simultaneously as the national language
as in Malaysia (Omar, 1976), and a vernacular which is purely a regional
language as in Zambia (Ohannessian, 1978a).
In the analysis of a language community the linguist would have to
consider the following factors (Spolsky et al., 1976):

The social structure of the area

In many areas where one finds a declining minority language, one of the
principal reasons for its decline is a change in the social structure. This can
have direct consequences for the schools in the area. As linguistic minority
areas are often rural and attractive holiday spots, e.g. Wales, Scotland,
Brittany, North and West Frisia, this leads to tourism. As the rural areas are
often economically underdeveloped the tourism brings a welcome influx of
jobs and wealth and many inhabitants become dependent upon tourism for
their living. However, the tourists introduce the dominant language into the
community and strengthen its position there. In addition, a larger labour
force is often needed for the "season" than can be culled from the area itself,
which again means introducing speakers of the dominant language into the
community. In both instances mixed marriages are often the result whereby
the' member of the linguistic minority usually gives up his or her mother
tongue in favour of the dominant language.·
Tourism also has the effect of raising the price of houses and land so that
members of the indigenous population can no longer afford to live there and
are forced to move. On the island of Sylt in NF, the schools have great
difficulty in finding teachers, as potential candidates cannot afford to live on
 VERNACULAR LANGUAGES AND EDUCATION 169
the island in the manner to which they are accustomed. As a result of the
very high cost of living on Sylt, a large part of the labour force has to
commute from the mainland each day. In a similar case in Wales some Welsh
people have taken to burning down holiday homes belonging to English
people as a protest against the influence of the English language in the
Welsh-speaking area and the consequences of English tourists on the
property market.
Another consequence of the economic underdevelopment of the minority
language areas is the shortage of jobs available for young people, especially
qualified young people, so that many are forced to emigrate. However, in an
attempt to solve this problem, governments sometimes try to develop areas
that are economically weak by introducing industry. This may bring more
prosperity into the area but tends to have a disastrous effect on the minority
language. Kramer (1978) describes the beginning of the new era some 28
years ago, when the East Frisian language really went into decline, as
coming with the advent of the 'Emsland plan' which included the building of
roads and the settling of industry in the area. In Eire the government tried to
settle industry in the Gaeltacht, but discovered that the local populace had
little interest in working in factories and that the trained and qualified staff
needed to run a factory had largely to be imported. This meant another
English incursion into the Irish-speaking area (Gregor, 1980). Similarly,
in NF, there are large units of the German Air Force and of the
"Bundesgrenzschutz" {border police) which on the one hand bring em-
ployment for the local populace but on the other hand change the ethnic
composition of the area so that many children going to school in NF are now
neither linguistically nor ethnically Frisian.

Ethnicity and attitudes

It was noted earlier that language is an important part of a person's or a
linguistic community's ethnic identity. However, as Edwards points out
{1977), one has to differentiate between the communicative and the non-
communicative aspects of language. Although many in Eire see "the value
of Irish as a symbol of national or ethnic identity, or as a symbol of cultural
distinctiveness" (Committee on Irish Language Attitudes Research, 1975),
and although 76 per cent questioned by the Irish Marketing Surveys approved
of Irish being taught in national primary schools (Macnamara, 1971), there
was nevertheless a large amount of hostility towards the Irish Language
Education Programme set up by the Government. In other words, people in
Eire were prepared to accept the symbolic value of Irish but not the
attempted coercion by the Government in order to reinstate Irish as the
main medium of instruction in schools and as the main medium of com-
munication in the country. As a consequence, certain measures taken by the
170 A. G. H. WALKER

Government, such as the obligatory examination in Irish for candidates for

the Civil Service, have since had to be revised.
In contrast to Eire, Wales and West Frisia provide examples where the
language plays an important role in a person's identity. Here, the proof of
ethnicity is not only the language as such but also the ability to speak the
language. This, however, applies mainly to members of the language move-
ments, and in Wales the members of the Welsh Language Society had to
convince the public that giving Welsh a place in the school curriculum would
not lead to a decline in standards attained in English before they could
introduce a programme of Welsh education (Williams, 1973). According to
Chapman et al. (1977) there are now three groups of people in Wales who
consider themselves Welsh: (1) fluent speakers of Welsh, (2) learners of
Welsh, and (3) those who consider Welsh a waste of time but who still show
their "Welshness" by the linguistic distinctive feature of a Welsh accent in
English. In fact all three groups reject standard RP English. Thus, in the
third case, the Welsh accent can be considered the equivalent of the Welsh
vernacular. It is interesting to note that in all three examples there is -a
reaction against one form or other of government coercion. In Eire the
Government tried to impose Irish on the populace and met with a hostile
reception, whereas in Wales and West Frisia the central Government
(London and Den Haag) tried in one form or other to suppress the minority
languages.
Ethnicity is one of the greatest problems facing plurilingual and pluri-
ethnic societies. If a pluriethnic nation decides to give preeminence to one
particular language (and therefore also to the ethnic group it represents) by
making it the national language, this will probably cause bitter resentment
amongst the other ethnic groups. In Nigeria, for example, the naming of one
of the indigenous languages as the national language to the exclusion of the
others would probably lead to civil war. Similarly the name of the language
can have the same effect, as witnessed in Malaysia. Following independence
in 1957 the national language was called "Bahasa Melayu". This, however,
caused much unrest among the Tamil and Chinese sections of the population
as they interpreted the name as meaning the language of the Malays.
Following a severe racial clash in 1969 the Government changed the name to
Bahasa Malaysia, which meant "the language of Malaysia". This name
seems to be acceptable to all ethnic groups and problems of ethnicity should
not now disrupt the education programme the Government has embarked
upon, which envisages "one type of education for all, which uses Bahasa
Malaysia as medium of instruction right up to the university level" (Omar,
1976, p. 5).
An integral part of ethnicity is the question of attitudes. Each group or
community has certain attitudes towards itself and towards the other groups
 VERNACULAR LANGUAGES AND EDUCATION 171

around it. These attitudes usually offset one group from the others and are
evaluative, e.g. our group A is better than group B which in turn is better
than group C. The attitude can apply to the group as such, but also to
certain features of the group such as language. A negative attitude towards a
group will probably include the group's language, but a negative attitude
towards a language need not include the group which speaks it. In Schleswig-
Holstein, for example, many people consider Danish an "ugly" language,
principally because of its glottal stop (st(f)d), but are very fond of Denmark
and the Danes. The attitude one has towards one's own group can be
influenced by the attitude another group has towards it, especially in the
asymmetric relationship of a minority group versus a majority group. Thus,
if the majority group considers the minority group inferior, it is possible that
the members of the minority group will accept this opinion and take it as
their own. The attitude a person has towards his own group is, however,
important, as it may influence his desire to retain or reject those features
which determine his membership of the group (Giles et al., 1977), which
in turn will be of relevance for education, as it may affect a parent's willing-
ness to allow his child to participate in vemaculai education and/or a child's
willingness to study (in) the vernacular.
The black population in North America, the Commonwealth immigrants
;n Britain and the immigrants from Turkey, Yugoslavia etc. (Gastarbeiter)
in West Germany tend to be considered socially inferior. Thus, in Germany
Turks are often addressed with the intimate but in this context socially
degrading du. As these groups are considered socially inferior, so are their
languages and their accents. In England, for example, this often results in
making the Indian and Pakistani English accents the butt of jokes (cf.
Chapman et al., 1977 for a discussion of this phenomenon with reference to
Welsh). We find a similar case with indigenous linguistic minorities. As
minority languages are often found in economically underdeveloped regions,
the group speaking the language is often relatively poor and holds a low
social standing. Gregor {1980, p. 294f) quotes for example An t-Ultach {The
Ulsterman) of June 1951 which had stated:

For some, Irish is synonymous with poverty and social inferiority. . .. Young
people think it is a sign of social inferiority if their parents speak Irish, especially
in the presence of a stranger.

Such attitudes to language are also found in pluriethnic and plurilingual

societies. Heine (1979) reports for Africa that vernacular languages are
often considered symbols of the somewhat limited horizon of the tradi-
tional tribal way of life and as such do not have much prestige. The linguae
francae such as Swahili are preferred as they represent the free modem way
172 A. G. H. WALKER

of life and offer better job opportunities and social recognition. Similarly, in
Ghana many people see English (the official language) as the key to the new
material civilization. In this they find active support from the politicians who
consider the vernacular languages to have a negative influence on the
nation's future, and who prefer English to promote national unity, the
suppression of tribalism, rapid industrialization and accelerated economic
development (Ohannessian, 1978b). Another reason why politicians and
officialdom in developing countries tend to have a negative attitude towards
vernaculars is their desire to retain the status quo for socio-economic reasons.
Many senior officials have attained their posts through, among other things,
a good command of the official language when their country was a colony of a
foreign power. Because of their expertise in this language they can com-
mand a certain elitist position in society. If, however, the former colonial
language is reduced in status to that of the vernacular, it is possible that the
officials wiIJ Jose at least part of their own status and priviliges as they would
then practically be on a par with aIJ other vernacular speakers. In this
instance the officials wish their additional linguistic competence to be
rewarded, although it does not necessarily mean that they simultaneoulsy
reject the vernacular.
In another example, the lack of material benefits the vernacular brings
makes the vernacular unpopular so that children or students see little point
in learning or using it. In Andhara Pradesh, a state in Southern India,
Telugu was introduced as the official language in 1966. Telugu proved very
unpopular as a medium with the students and parents as there was no policy
for preferential treatment for Telugu-medium students and because em-
ployers such as banks and businesses still carried out their work in English
and preferred English-medium students (Krishnamurti, 1978). This is in
contrast to NF, where it seems that a command of Frisian may soon prove
economicalJy useful. There is at the time of writing a greater supply of
teachers in Germany than the schools can employ, and it is probable that
teachers speaking Frisian will be given preferential treatment in NF. Such
measures should enhance the language's prestige.
Socio-economic considerations may persuade a person to abandon his
ethnic group and its language. Edwards (1977) mentions two examples
where members of an ethnic group have considered it to their advantage to
identify more strongly with a different ethnic group.

In Montreal, in which the majority of the citizens are French-Canadian, big

business has traditionally been an English-speaking domain ... French speakers
clearly had to become bilingual in order to ascend the economic ladder. Such a
situation has given rise to the phenomenon of French-speaking Montrcalers
attempting to identify more strongly with English Canadians than with members
oftheirown group. (1977, p. 257).
 VERNACULAR LANGUAGES AND EDUCATION 173

The second example relates to the immigrants in the "melting pot" of the
United States where there was a

process of Americanization by which ethnic groups of every origin submerged

the characteristics of these origins, were assimilated into the mainstream of
American society, and reappeared after a generation or so as more or less
homogenous Americans . . . (Thus] the evident rewards that full participation
in American life offered, have been the causes of the apparently rapid 'de-
ethnicizing' of immigrants. (1977, p. 261).

Both these examples show a group wishing to assimilate into another group
which sometimes results in parents refusing to send their children to schools
set up for their ethnic language group. In NF we have a similar example, but
there a further factor must be taken into consideration. On the mainland of
NF the Frisians can be roughly divided into two groups: the Frisians oriented
mainly towards Germany and the Frisians oriented mainly towards Denmark.
The majority of the Frisians are not very politically minded but, if forced to
make a decision, would consider themselves more members of the first
group because they are not members of the second group. Thus we have
here the phenomenon of the "minority within the minority". The Frisians as
such are a minority within the German majority, but they in turn can be
sub-divided into the German Frisian majority and the Danish Frisian
minority. It is, however, the Danish Frisian minority which is most active in
the promotion of the Frisian tongue. Because in this area to be called a
"Dane" has certain negative connotations, this has the effect of inhibiting a
number of potential activities, as people would often rather acquiesce to
German than support Frisian for fear of being considered a Dane.
The negation of one's ethnicity is often the result of the desire to be
assimilated into the mainstream of the society one is living in. However,
once one has been assimilated to a certain degree and has achieved a certain
standing in society, one can then afford to return to one's true ethnic
heritage. This explains why leaders of language and "cultural revival"
groups are often politicians or academics. The question, however, is
whether these leaders actually do articulate the true needs and desires of the
group they claim to represent, and also to what extent the average member
of the ethnic group can identify with them. To what extent, for example, will
the average working man accept the credibility of the person who advocates
that everyone should speak the minority language with their children and
that it should be taught in schools, when he himself speaks the dominant
language with his family as a result of the previous assimilation process?
Similarly, as Tholund (1980) asks, is the resurgence of ladies' native cos-
tumes in NF really a sign of the awakening of the Frisian identity or rather an
excuse for politicians to show their "love" for the country and to have a
174 A. G. H. WALKER

couple of beautifully clad fair maidens to accompany them on festive occa-

sions? It is difficult to assess the true effect of such people and such
manifestations of "culture" in an ethnic group. On the one hand they tend
not to represent the true needs and desires of the average member of the
group, but on the other hand they do lend the language and culture more
prestige by their personal dedication to it.
In a discussion of attitudes one can differentiate not only various types of
society but also the sexes. Labov (1966) notes that women in a minority
group tend to be more language conscious and prefer the dominant language.
Girls may then be more resistant to being taught the vernacular at school
than boys. However, in a different situation girls may be the principal
bearers of the vernacular. On the island of Fohr, for example, where the
menfolk used to go to sea for a good part of the year, they became multi-
lingual. The women who stayed at home retained their Frisian and ensured
the continuity of the Frisian community and language. It is possible that the
boys here preferred to learn High and Low German in expectation of their
own travels in contrast to the girls who saw their futures in the narrow
confines of the village. Similarly Liebe-Harkort (1980) reports the dif-
ferences in attitudes between Navajo boys and girls. As Navajo men tend to
do traditional outdoor jobs and Navajo women do subordinate work in
offices and hospitals, girls are more proud of their English and use it more
whereas boys tend to regard a compliment on their English as an insult.

Demographic factors
The number of speakers of a language can be of importance for educational
planning, as economic factors may forbid the inclusion of too small a
language in the school curriculum, and speakers of this language would then
be forced to use a second language. However, establishing the true number
of speakers of a language can be problematical as (1) the concept "speaking
a language" would need an exact definition, and (2) a method would have to
be evolved for eliciting accurate information corresponding to the work-
ing definition. Censuses are not always reliable. In Ireland and Wales,
for example, they do not differentiate between those who have Irish or
Welsh as their mother tongue and those who have acquired the language
later in life, e.g. at school or in evening classes {Price, 1973). Similarly, the
Lapps living in the coastal districts of Northern Norway refused to admit
their true identity in a census, as Lapp is considered a negative identity in
Norway (Aarseth, 1969). However, even a linguist conducting a personal
interview is not free from error, as the interviewees will often give the
answer expected or hoped from them rather than the true answer. In the
case of speakers of a language with low prestige there will be a tendency to
negate one's language competence, and in the case of a minority language
 VERNACULAR LANGUAGES AND EDUCATION 175

where the interviewer is obviously a supporter of the language, there will be

a tendency to overestimate one's linguistic ability. One possible method of
checking the true linguistic competence of an interviewee is to converse in
the given language. If the interviewer switches codes, however, it is possible
that the interviewee will not be able to follow suit if he has the linguistic rule
'one person one language'. {I have experienced this in my field work in NF.)

Linguistic relationship
In a bilingual community (in the sense that two languages are used in the
community, while not all members of the community are necessarily pro-
ficient in both) with two completely unrelated languages, such as the Eskimos
in Alaska (Yup'ik and English) or the Apaches in North America (Apache
and English) it will probably be more difficult for the children to com-
prehend and acquire the second language than in a community like West
Frisia, where West Frisian and Dutch are closely related West Germanic
tongues. On the other hand, the incidence of interference will probably be
more marked between two related and similar languages than between two
unrelated ones.
Similar to the linguistic relationship of the languages in question is the
relationship between cultural concepts embodied in the two languages.
Finnish and Swedish, although genetically completely unrelated, will have
more European concepts common to both than English and Eskimo (Mackey,
1978).

The functions of each language

An analysis of the functions each language has in the community, whereby a
function can perhaps be defined as the set of speech acts ascribed to a
language in the community (Johnson, 1977), is important, as a language
which has hitherto been restricted primarily to the domains of the home and
the hearth will face different problems when introduced into education than
an officially recognized language already in wide use. (The term "language
function" has been used in a variety of ways; see for example (after Johnson,
1977); Barker, 1947; Fillmore, 1973; Fishman, 1971; Le Page, 1964; Nida
and Wonderly, 1971; Okonwo, 1975; Stewart, 1968; Weinreich, 1953).
Whiteley (1969) gives a brief description of the language functions in the
multilingual state Tanzania. Swahili is the national language and is used in the
National Assembly, the Party, the Trade Unions, the lower courts, regional
administration, primary education and in certain areas of the Civil Service.
English is used in inter-territorial affairs, higher education, the High Court
and certain technical fields of government, e.g. medicine. The home language
is probably the vernacular if both parents are from the same language
community. If the parents come from different language communities, then·
 176 A. G. H. WALKER

probably Swahili will dominate. However, speakers of Haya, Nyakyusa,

Chaga and Goga have a certain sense of language loyalty and would probably
rather retain the vernacular than use the lingua franca Swahili. If the parents
have enjoyed a certain education and have been to secondary school, they
will probably include English in the languages used at home.

Linguistic problems
Two main problems can be differentiated with reference to the vernacular
language itself. These are the language's range offunctions and the linguistic
norm.

The language's range offunctions

It was mentioned earlier that an analysis of the functions of the languages
used in the community was important. If, as is often the case, the vernacular
is restricted in its functions, the number of styles will also be restricted,
perhaps to the extent of mono-stylism as witnessed in Brittany, where young
Bretons only have a command of the casual style. Formal situations are
usually coped with in French (Dressler and Wodak-Leodolter, 1977). How-
ever, as the spectrum of education must go beyond the purely casual level,
the problem arises how further levels of style can be introduced. There is a
similar problem with respect to the vocabulary if the language is used only in
certain domains. If, as in the case of Welsh, the language is spoken primarily
in the home in a rural area, the vocabulary will be rich in terms for the
domestic and agricultural domains, but lacking in technical, commercial and
other such "sophisticated" teminology (Jones, 1973). Here the vocabulary
must be extended (cf. Modernization).

The linguistic norm

For languages lacking a standardized norm, Kloss (1969) has described one
[ aspect of language planning as "corpus planning". This can, in agreement
with Ferguson (1968), be subdivided into three sections: (1) the creation of a
written form (graphization), (2) the choice of a standard form (standardiza-
tion), and (3) the modernization of the language (modernization).

Graphization. The question of the making of alphabets has long been the
subject of debate, and Berry (1968) posits certain scientific principles and
social factors which must be taken into account in order to create a successful
alphabet. He states that "an alphabet is successful in so far and only in so far
as it is scientifically and socially acceptable." (1968, p. 737). In a later article
 VERNACULAR LANGUAGES AND EDUCATION 177

(1977) he shifts the emphasis slightly from the scientific aspect and stresses
the social aspect more. Fishman (1977) points out that a language community
in a bi- or multilingual situation may be guided by two opposing forces when }
forming their attitudes to a possible orthography, as they may desire their e~~~
orthography to be similar or dissimilar to those of the languages around lbv-vt
them.
It is said of the West Frisians that, if anyone attacks their language, they
will man the barricades, but even on the barricades they will continue their
squabbles about the orthography (Jorgensen, 1979). The same applies to the
North Frisians. The present North Frisian orthographic system states that all
long vowels shall be written with a double vowel, e.g. naame [na:m~] "to
take", and short vowels with ~ single vowel, e.g. ham (ham] "him". No
consonants can be doubled. This is in contrast to the German orthographic
system where short vowels are indicated by a doubling of the following
consonant, e.g. Butter, Holle, and long vowels are followed by a single
consonant or a length symbol such as h in Mehl or e as in Sieg. Arhammar .,- 1
(1976) considers that this radical deviation from the dominant language's _J
norm is to emphasize Frisian's independent identity. This orthographic
principle has been widely accepted although one symbol has caused some
trouble. As the mainland dialects are characterized by an abundance of
vowel phonemes, the symbol ii was introduced in the phonemes /A/ and /Ml
to differentiate them from the phonemes /oo/ = [o:) and /oo/ = {~:),the
latter two phonemes being examples of the desired principle of one phoneme
one grapheme not being fulfilled as each phoneme is written oo. As, however,
Danish also uses the symbol a as in Arhus, some Frisians rejected the
orthography as they considered it too Danish. A further more serious
argument, however, evolved around the discussion whether or not nouns
should be written with a capital letter as in German. One argument forwarded
was that all children are used to the German orthography and that it must be
pedagogically sensible to use the same principles in Frisian i.e. the ortho-
graphic principles a child is used to in his literacy L1 should be transferred to
his literacy L2. The counter argument was that all other languages in Europe
use small letters for nouns, such as English, which the children also learn at
school. This "pedagogical" argument was in fact a political one, as the
German-oriented Frisians saw the noun with the capital letter as their last
bastion of German identity in the Frisian orthography. Thus dissimilarity
from German is sought by that group striving for a degree of Frisian
autonomy and similarity is desired by that group which wishes to see Frisian
firmly clutched to the bosom of the German state. At present a certain form
of anarchy reigns as each group publishes its works according to its own
principles.
 178 A. G. H. WALKER

Standardization. Feguson {1968, p. 31) sees language standardization as

the process of one variety of a language becoming widely accepted throughout

the speech community as a supradialectal norm - the 'best' form of the language
- rated above regional and social dialects, although these may be felt appropriate
in some domains.

r Standardization is necessary if a language is to be included in the school

I
'--- curriculum as a subject or a medium. If, however, the language is an as yet
!_..-- uncoded or insufficiently coded vernacular, there are various factors which
may influence the choice of one variant as a norm. Fishman (1968c, p. lOf),
for example, lists the following considerations:

number of speakers, . . . past association with a Great Tradition, current

association with major social trends ... , greater purity in the sense of fewer
influences from varieties or languages considered undesirable or, conversely,
greater similarity to other highly regarded varieties or languages, and, finally, a
middle-ground position vis-a-vis overly pure and overly indistinguishable
varieties.

I should, however, like to discuss the question of the norm by considering

three axes. These are the temporal, the areal and the societal axes.
(1) The temporal axis. Language is permanently changing, which perhaps
helps explain why each generation complains that the following generation
no longer speaks the language properly. However, change can have dif-
ferent consequences for different languages. English has, for example,
relatively recently gained the words streaker and punk and is in the process
of losing the pronouns thee and thou. These changes will, however, have
little effect on the English language community. English, as result of these
changes, is in no way endangered. It is a different case, however, for a
linguistic minority, where changes may be interpreted as symptoms of
language decline. Haugen (1980) has coined the concept of the cycle of
language shift whereby a monolingual community gradually introduces L2,
the dominant language. For a period of time there is a lot of interference in
L2 from LI. After a while there comes a diglossic situation. Following this
the interference stems from the dominant language L2 and is found in the
minority language Ll. Finally monolingualism is reintroduced following the
extinction of L 1.
In the declining East Sutherland Scots Gaelic dialect showing interference
from the dominant L2 English, Dorian (1977, p. 23)

discovered considerable differences in the Gaelic of the oldest available fluent

speakers and the youngest, the Gaelic of the latter showing reduction and loss in
certain areas in comparison with the former.
 VERNACULAR LANGUAGES AND EDUCATION 179

Walker ( 1977) found in three generations of trilingual Frisian-Low German-

High German speakers (where both Frisian and Low German are L1 in
Haugen's cycle, 1980), that the forms of address had changed in Frisian and
Low German from the binary system of second person singular (familiar)
and third person singular (respect) to the equivalent of the High German
system with du and Se in Low German, and to an insecure system in Frisian
where the High German system had been partly adopted, but where also
part of the old Frisian system with the third person singular had been
retained. These two examples indicate that a declining language shows ( 1) a
reduction in forms, and (2) a change in various sub-systems which then
adopt the dominant language form. As a result of this it is difficult to
establish exactly what the norm is. In traditional dialectology the field
workers have always preferred the oldest inhabitants of a village as in-
formants as they "reflect the speech of a bygone era" (Chambers and
Trudgill, 1980, p. 35) and are generally considered to speak the "best"
dialect. This could also be used as the criterion for selecting the norm,
especially if the older generation grew up monolingual. However, a grammar
based on the dialect variant represented by the older generation may well be
severely criticized by the younger generation of speakers and school children
who would not wish to accept that variant as a norm. In NF we then hear
such comments as: "That's alright for old grannies and philologists, but
no-one speaks like that now." If, however, the norm is based on the speech
of the younger generation, in all probability various forms unique to the
minority language and which helped differentiate it from the dominant
language would be lost. Thus the question arises whether one wishes to -1

retain or reintroduce the old pre-language-contact grammar, or work with a

grammar which can be interpreted as showing all the symptoms of a declining ~
language. As it were, puristic prescriptivism versus pure descriptivism.

(2) The areal axis. Most languages of the world display a variety of dialects.
With those languages with a standardized norm, there often arises a case of
bidialectal diglossia whereby the "local" dialect is, for example, used at-r
home and the "standard" dialect i.e. the standardized form of the language,
1

is spoken at school. There are problems here (cf. Trudgill, 1975) but the ',
teacher does at least know what standard he or she is supposed to be....;
teaching. With a non-standardized language showing great dialectal varia-
tion, the fundamental problem of the norm is most acute. One may find that
the speakers of a non-standardized language consider the language inferior
to the standardized dominant language, but that within the non-standardized
language each speaker considers his dialect to be the only correct norm,
making him reject the choice of any other norm. For NF, for example, Wilts
(1979, p. 198) writes:
180 A. G. H. WALKER

Der einzelne J.!Ordfriesische Dialekt darf sich auf Grund seiner subjektiv
empfundenen Uberlegenheit iiber die friesischen Nacbbardialekte bereits als .
Vertreter des Nordfriesischen par excellence fiihlen.

As a consequence it is possible that a dialect speaker will reject books,

dictionaries etc. which are not written in his norm. A multidialectal dic-
tionary as a compromise would probably only result in confusing the reader.
If, however, one dialect is chosen as the norm this may lead to the speakers
of the other dialects considering their dialect incorrect and inferior which, in
the case of a declining language, might make them refuse to speak it and
revert to the dominant language.
When constructing a norm for a language consisting of a variety of
dialects, a factor which has to be taken into consideration is the mutal
intelligibility of the dialects. Statements on this subject can, however, be
misleading as they may be subjective and subject to a certain ideology. Thus,
an objective linguist considers some North Frisian dialects to be so dissimilar
that mutal comprehension is practically impossible (Hofmann, 1956),
whereas certain eminent members of the language movement see no reason
why intra-dialectal comprehension should be difficult. It is, however,
difficult to find objective criteria for comprehensibility. Linguistic similarity
does not suffice, as socio-psychological factors such as prestige may render
two objectively similar languages incomprehensible. In Nigeria, speakers of
the Kalahari dialect of the Ijo language are relatively prosperous and
consider the other dialect groups such as Nembe to be backward. The latter
have no difficulty in understanding Kalahari, whereas the Kalahari maintain
they are unable to understand the Nembe. This is an example of non-
reciprocal comprehension. Similarly, socio-political factors may be at work,
as exemplified in the Urhobo language in Southern Nigeria, where speakers
of the Isoko dialect group maintain they cannot understand the other
dialects of the language. This is accredited to the Isoko group's desire for .
autonomy (Heine, 1979).
There are also more objective factors which can render dialects incom-
prehensible. The Lapp language is spoken over an area which encompasses
four different countries. There are seven main dialects, some of which are
mutually incomprehensible, so that a Lapp from Finnmark cannot under-
stand one from Tr0ndelag (Norway). One of the reasons for this is that
different languages are used as sources of loan words. As Finnish and
Norwegian belong to two completely different language families (Finno-
Ugrian and Germanic), any loan words taken from them can only widen the
already existent gap between the two Lapp dialects (Aarseth, 1969).
In NF a certain amount of tolerance is exercised in the orthography
designed for the mainland dialect of the Bokingharde in NF around the
 VERNACULAR LANGUAGES AND EDUCATION 181

Risum Moor, as words are written and accepted using forms from both the
main sub-dialects (East and West Mooring). Admittedly there are only
minor differences between the two dialects, but nevertheless people do
recognize a given form as belonging to a certain village. If, however, one
goes seven kilometers to the south or ten kilometers to the north, one
encounters two completely different dialects which adhere to the same
orthographic principles outlined above (p. 177). but which demand that the
lexicographical realization of the individual words should conform to their
dialect and not to the norm established for the Mooring dialects. This is quite
understandable, as various features with a high frequency rate differ quite
strongly as, for example (after Walker, 1980c):

East Mooring Fahretoft (south) Emmelsbull (north)

Old 0 > 0 (fOtj) OU (fout) oi (foitj) foot
Frisian e >a (bradj) ai (braid) ai (braid) bride
a > uii (iiiilj) u(ill) ii (ill) old

For these dialects a different orthography has to be used.

The diversity of dialects has sometimes led to a compromise whereby
books etc. are published in a number of dialects. In Eire, three dialects are
used (Connacht, Munster and Ulster) and in NF sometimes six. In Eire,
speakers from each dialect take it in turn to read the news bulletin in the
radio (Stephens, 1978).

(3) The societal axis. In a language consisting of a multiplicity of dialects, -- 1

history would appear to show that the dialect of the elitist group in the \
society usually develops as the standard language. Present-day standard """
English, for example, originates from the language spoken at the Royal
Court in London in the fourteenth century. This may cause a conflict
between various criteria, as an elitist dialect is not necessarily numerically
the strongest, the one most free of external influence, or the most developed.
The choice of the norm depends upon the question of priorities and the·- !
degree to which the elitist group is capable of influencing the relevant .,;
authorities.

Language planning can suffer set-backs if language planners are insuf- 1 rt> \ji.
ficiently trained or if inadequate field-work or sociolinguistic survey-work is 1, "'
carried out beforehand. Ohannessian (1978a), for example, complains that_.......... :
the members of the committees set up to standardize the seven officially t
taught languages in Zambia had, with few exceptions, no training in Bantu
linguistics, Zambian languages or language pedagogy. As there is often an
insufficient number of qualified indigenous linguists in a developing country
to establish a system of language planning, experts are frequently called in
 182 A. G. H. WALKER

from outside. These, however, usually Jack the intimate k~owledge neces-
sary to appreciate the subtleties of the society they are working in, and as a
result produce a language form which is considered by the indigenous . .
population to be lifeless and stilted. as in the case of Swahili in 1934
(Whiteley, 1969). Furthermore, certain parts of the population may feel
offended, as in the case of Malawi, where a group of impartial linguistic
experts created the language "Union Nyanja" as a synthesis of two main
dialects. Despite the neutrality of the language planners, members of each
dialect group rejected the new creation as each group saw too many features
of the other language in it and not enough of their own (Heine, 1979).
I. As a result of various problems that have arisen in language planning,
I Miilhausler et al. (1979) suggest that the best method would be to observe
\. how communication works between various language varieties, where the
L difficulties arise which need standardization, and to note what degree of
variation is tolerated without impairing communication. As an example of
tolerated variation in an otherwise very norm-conscious society, Tauli
(1968, p. 154) quotes a count published in 1936 which established that

the first 300 pages of the German Duden dictionary comprised over 1.100
substantives with variant inflectional forms, as des Aales/Aals, die Karnevale/
Karnevals, over 600 orthographic variants, e.g. Grafik/Graphik, over 100 words
with gender variants, e.g. der/die/das Klafter, hundreds of verb form variants.
e.g. du fragstlfriigst, ich angele/angle.

-· Modernization. Ferguson (1968, p. 32) considers the modernization of a

language as being "the process of its becoming the equal of other developed
languages as a medium of communication; ... " and sees "the expansion of
the lexicon" as an important aspect in this process. With respect to the
expansion of the lexicon it would appear that one positive aspect of language
decline is that it sometimes encourages laymen to write dictionaries in order
to preserve the diminishing vocabulary and in a valiant effort to produce
i- new terms for the modern world. There is perhaps the belief that a dictionary
~- is a language's redemption. Bendsen ( 1860), for example writes with respect
to North Frisian:

Da nun diese, gewiss uralte. Sprache mit jeder Generation von alien Seiten
durch die danische und plattdeutsche immer mehr beschrankt wird, und ihrem
allmahligen Aussterben entgegengeht, so habe ich mich der, wahrlich nicht
Ieichten, Muhe unterzogen, sie durch gegenwartige Arbeit von ihrem ganzlichen
Untergange bei der Mit- und Nachwelt zu retten.

However, despite the trouble that laymen put themselves to, their lexical
creations are often not accepted by the language community. This could also
 VERNACULAR LANGUAGES AND EDUCATION 183
happen to an official lexicographical institution, though in this case there
may well be a tendency to feel that a dictionary produced by trained
academics carries an aura of infallibility, although this will depend on the J
general esteem accorded academics in the society concerned. In the case of
the North Frisian dictionary for which a university department was especiallyl
created in 1950, the informants are very reluctant to question the lemmata, 1
even though in some instances they may be contrary to their native-speaker_,
intuition. Admittedly most of the words listed are the result of field work, or
have been culled from books written over the past three centuries, but many
. words are included which must have been attempts by earlier lexicographers
to modernize their language and which in all probability were never fully
accepted by the language community. However, a word which can be .
proved to have been recorded a long time ago tends to be rather uncritically •
1
regarded as correct, although this in no way means that the word will then be J
used.
In many countries, institutions have been set up to standardize and
modernize a language, such as the Icelandic Language Commission founded
J
in 1964. This institution tries to modernize Icelandic by reviving old Icelandic -~.
words which appeared in the sagas rather than accepting loan words. They (
have two five-minute broadcasts a week on the radio and have succeeded in - 1
making the population take an interest in their language. Examples of
revived words with a new semantic content are: simi (telephone) which is the
old word for "wire" or "thread", pjalfa (to train) which meant "to work
hard". A lot of new creations are composita using tw9 Icelandic words, such
as ftug-maOur (pil~t) coming from flug (flight) and matiur (person, or gervi-
tung/ (satellite) coming from gervi (artificial) and tungl (moon) (Jonsson,
1979). Such a policy is possible in such a small and compact community as ~1
1
Iceland, especially as the language here is an integral part of the Icelander's •

identity. In other communities a better policy might, according to Wilts .;

( 1983), be to observe the forms created by the speakers themselves rather .J
than try to impose forms upon them.

Problems pertaining to the school

The problems affecting vernacular education and pertaining to the school
can also be discussed under various headings.

The school in the community

As there is considerable interaction between the school and the community,
the role of the school in the community will play a decisive part in the success
or failure of education in general and vernacular education in particular.
Burns (1965) considers a school's responsibilities to be social, moral and
184 A. G. H. WALKER

pedagogic, and the success of an educational programme to be dependent

upon the school's ability to enter into the life of the community and identify
itself with the community's needs. The Eskimo programme in Alaska is a
positive example of this, where the children are highly motivated because of
high community involvement in the project (Trifonovitch, 1976). When,
however, the community is dissatisfied with the school, a programme can be
jeopardized as witnessed in NF where one reason for the parents in Risum-
Lindholm rejecting Frisian tuition was that they thought the school was not
educating their children properly (Walker, 1980b).
Various factors may determine the relationship between the school and
the community. Spolsky et al. (1976, p. 239), for example, consider the
authorities responsible for the school policy and the framework they work in
to be important. They suggest three types of authorities:

the superintendent of a school district, his specialist advisors and the principals
of the schools in the district; ... the top educational bureaucrats in a nation, ...
[and] group representatives of the community that a particular school serves.

A further factor may be the proximity of the school to the community. If

there are village schools or similar, these will probably be integrated into the
village community and the teachers will play a part in community life. If, on
the other hand, the children are obliged to go to large centralized schools or
boarding schools, there is little connection with the community, and one
possible consequence is that the parents take little interest in their children's
education. Burns (1965) reports that one factor leading to outbreaks of
indiscipline at some boarding schools (which generally constituted the
secondary schools) in Rhodesia (1962), Swaziland (1960), Uganda {1959)
and Zambia (1960) was the lack of personal ties between the school and the
community. He also considers it unfortunate that the parents were not
unduly perturbed when their children were expelled as a result.
The centralized school is also affected by the problem of logistics. If a
central school has a large catchment area, this can involve a lot of travelling
for pupils, which in turn may impair their ability to concentrate during the
lessons and also renders the school time-table less flexible for additional
tuition. In the Chinle Public School District (a Navajo Reservation), for
example, there are nine schools for 4559 children within an area of 7200
square miles, which results in four-hour bus runs for some of the children
(Liebe-Harkort, 1980). Mackay (1969) similarly reports that the children in
the Scottish Highlands often have to leave home at 8 am and do not return
until 6 pm.
Perhaps one of the main problems facing the centralized school, though, is
the plurilingual intake of a large catchment area. This can have a negative
 VERNACULAR LANGUAGES AND EDUCATION 185

effect on a minority language such as Scots Gaelic, which was mostly used in
village schools in the Highlands. When, however, these were closed and the
children sent to a large central school, they were found to adopt the
dominant English language (Mackay, 1969). Gaelic could no longer be
fostered here as it was in the small village school. There is sometimes,
however, the possibility for a minority language community to open its own
school, although this is very dependent upon the good will of the majority.
In Finland, for example, 18 children are needed to open an elementary
school for a language community (Ahlskog, 1969). Once a small school like
this has been opened, it is possible that it will gain more pupils than
originally planned, as other people decide to send their children there for a
variety of reasons, e.g. less stress, better teacher-pupil ratio, different
methods and principles of education. This could prove beneficial for the
vernacular concerned, though on the other hand parents might sometimes
be tempted to send their lesser gifted children there for the above mentioned
reasons, thus giving the school more quantity but less quality.
A school may enhance its standing in the community by itself becoming an
employer in a possibly underdeveloped area. Spolsky et al. (1976) reported
that as a result of the introduction of Navajo into the school curriculum for
the 53 OOO Navajo pupils, a further 1000 Navajo speaking teachers were
required, which naturally brought more funds and employment into the
Navajo community.

The educational programme

If vernacular education is to be given in school, it will usually be part of a bi-
or multilingual programme unless, of course, the vernacular is English or
similar (cf. p. 162). The actual design of the programme will depend upon the
educational policy being pursued (cf. p. 168). In a developing country in a
plurilingual setting, a decision will first have to be made as to which language
is to be considered the vernacular for educational purposes, and then as to
which languages shall also be included in the school curriculum. Ferguson
states (Ferguson et al., 1977, quoted after Paulston, 1980, p. If) two of the
implicit goals of bilingual education are: "To unify a multilingual community"
and "To enable people to communicate with the outside world." Thus a
regional and/or national language could be introduced in a developing
country as an attempt at unification, and at a later date the official language
(usually a world language) to facilitate communication with the outside
world. Another goal is seen in the desire "To assimilate individuals or groups
into the mainstream of society." This is usually again in the interests of unity
and is often practised with respect to immigrants and linguistic
minorities. Edwards (1977) argues that many of the bilingual projects in
America are designed to facilitate the transition to the dominant language
186 A. G. H. WALKER

for the immigrants. Two further goals are "To preserve ethnic or religious
ties" and "To give equal status to languages of unequal prominence in the
society". These goals are relevant for linguistic minorities, firstly as bilingual
education could re-educate children in their parents' language once they
have completed the language shift away from their ancestral tongue to the
dominant language. This is the case on the mainland of NF, where Frisian
has to be taught to a certain degree as a foreign language. The second reason
is that if a minority language is given equal prominence, it will be less
stigmatized, so that language shift can perhaps be halted, if not reversed.
This, it might be argued, is the case on the island of Fohr in NF. The
particular goal will determine whether a policy of transfer is adopted,
whereby the vernacular is later abandoned for the dominant language as in
the process of acculturation, or a policy of maintenance, whereby the
vernacular is maintained next to the second language (Mackey, 1978).
There are a number of other factors influencing the design of a bilingual
programme which can be discussed individually:

The linguistic competence of the children involved. The monolingual

Navajo or Eskimo child would need a different programme from his bi-
lingual cousin who, perhaps due to the proximity of an English-speaking
town or television station or because of some English-speaking neighbours
in the village already has a certain competence in English (Trifonovitch,
1976). However, even though in the latter instance a certain knowledge of
the dominant language might be present, would it be sufficient for use as a
medium in education? Paulston (1980, quoting Gaardner n.d.) refers to the
concept of "folk bilingualism" which refers to bilinguals who are members
of an underprivileged section of society and whose bilingualism arises
through confrontation with the dominant language. This confrontation does
not necessarily lead to a sufficient command of the dominant language for
use as a medium of instruction.

The status of bilingual education. Bilingual education must take cognisance

of its special status and not just teach two languages parallel to but inde-
pendent of each other. The minority child with some competence in the
dominant language can by no means be taught the dominant language in the
same way as a native-speaker. One unfortunate phenomenon observed in
Schleswig-Holstein is that there are no text-books describing High German
grammar as contrasted with Low German and/or Frisian. Thus points of
interference are not rectified and the child is punished without knowing why.
There are, however, contrastive grammars in other parts of Low German
speaking Germany, e.g. Niebaum (1977).
 VERNACULAR LANGUAGES AND EDUCATION 187

The applicability of immersion education. One very successful bilingual

programme is the immersion education practised in Canada, and Swain
( 1980) suggests that it might be applicable in other situations. In this
programme, a group of children homogenous in its complete lack of
knowledge of L2 is taught L2 as a medium by a bilingual teacher. The
teacher understands the children's utterances in Ll but replies in L2. Thus
the children first of all learn to comprehend L2. Ll is, however, later also
incorporated into the program:1J1e, and both languages thereafter serve as
languages of instruction throughout schooling. Swain does, however, point
out that certain cri~eria have to be present if such a programme is to be
successful. In the Canadian programme the parents belonged to the lin-
guistic majority and wanted their children to learn French, on the condition
that their ability in their native English was not impaired. In the case of a
vernacular low in prestige, such a programme would probably lead to rapid
assimilation, or, in Lamberfs words (1975), to "subtractive bilingualism"
where competence is gained in L2 to the detriment of L1. Thus Swain
concludes that such a programme could be successful in some instances, but
not in all.

When should which language be introduced? There are two possible ways
of answering this question, firstly with respect to simple expedience, and
secondly with respect to a child's learning ability.
In some developing countries the observation has been made that there
has been a shift since independence "in the direction of the Europeanization
of the media of instruction with a concommittant neglect of the teaching of
African languages" (Bokamba and Tlou, 1977, p. 38). In the independent
francophone and lusophone African states, education is carried out purely
in the earlier colonial languages, and in the one-time British and Belgian
colonies there is a trend back to the colonial language. Where vernacular
languages are taught, the colonial language is usually introduced in the third
or fourth grade. The arguments for this are practical, e.g. national unity,
national progress, and efficiency. The latter case encompasses such con-
siderations as the low degree of standardization and modernization of
African languages, the lack of teaching materials, the lack of teachers
trained in or competent in the vernacular, the lack of syllabuses drawn up in
the vernacular. In some countries like Ghana, the vernacular is taught only
in the first years of primary school, until the children have sufficient
knowledge of English to be able to continue their schooling in this language.
At secondary school level in most countries, pupils need a command of a
world language, as all advanced technical literature is written in them, and in
Zambia the Zambian languages do not qualify for university entrance
(Ohannessian, 1978c). Tanzania is an interesting exception, where it is
188 A. G. H. WALKER

hoped to introduce Swahili as the medium into the complete educational

span (Bokamba and Tlou, 1977).
The question of the theoretically optimal age for starting second language
acquisition has not as yet been satisfactorally answered. McLaughlin (1978,
p. 131) summarizes research findings with the conclusion that the "optimal
time to introduce a child to a second language is at birth". However, if this is
not possible, then research would seem to indicate that "a later, rather than
an earlier, start in the second language is better". There is one exception to
this as young children were found "to be superior in acquiring the sound
system of the second language". The main argument for the later introduc-
tion of second language teaching is the greater cognitive maturity of the
children. Weighed against this, one has the uninhibitedness and spontaneity
of the younger child. Genesee (1978) agrees in principle with these argu-
ments, stating that late instruction has the advantage of the learner's greater
learning efficiency. However, because of the additional factor of length of
exposure to a language he argues that early instruction gives a child
extended opportunities for language learning in and outside school. If then
the language instruction is continued through the higher grades, which
includes the advantage of greater learning efficiency, then early instruction
is probably "more conducive to attaining the higher levels of second
language proficiency.," (1978, p. 153) However, one must not forget that
age alone is not necessarily the decisive factor in determining a child's ability
to acquire a second language, as other factors such as motivation and
attitudes may play an important role.

Should languages be divided according to subjects or time?

Very often the use of the vernacular is restricted in a bilingual school to such
subjects as the vernacular, local history, etc. -essentially subjects belonging
to the humanities - whereas science and technical subjects are taught in the
medium of a world language, as this has the vocabulary and the literature
needed. To give lessons in these subjects in the vernacular presupposes a
programme of modernization. If, however,.it is possible to teach all subjects
in the vernacular, then a division of time may be preferred, e.g. morning
tuition in L 1 and afternoon tuition in L2, to give the children maximum
exposure to both languages.

How many languages can be taught?

In a multilingual setting, children may be forced to learn various languages.
In NF, for example, at the Danish minority school in Risum, children are
taught Frisian, High German, Danish and English, and they learn Low
German in the vilJage. In India children may also be forced to learn up to five
languages. Two questions arise here: (1) can the children cope with so many
 VERNACULAR LANGUAGES AND EDUCATION 189

languages? and (2) does the language element take up too much time in the
curriculum? Possibly one would have to differentiate the degree of com-
petence that should be attained in each language, whether listening com-
prehension, oral, reading or writing competence. This might depend on the
function each language is to have in the pupils' lives.

Should vernacular education be obligatory or voluntary?

This applies principally to minority languages being re-introduced into a
community. Experience has proved voluntary classes in Frisian in NF to be
unsuccessful, because children soon stopped attending the classes after an
initial bout of enthusiasm. In one experiment on Sylt, Frisian was used as a
medium for teaching voluntary classes of wood-work, chess etc. which
proved very successful, as the children could at least understand Frisian, but
they too left as soon as they had acquired the manual skill they desired.
There is as yet no back-up here. A further point is.connected with logistics. If
the lesson is voluntary it must be placed in one part of the time-table where
non-participants could then go home. However, if the child has a long
distance to travel, he must take the school bus along with the non-participants,
which effectively bars him from the tuition. Obligatory tuition ensures full
participation but can cause a negative reaction on the part of some parents.

The teachers
Two of the main problems facing vernacular education are the lack of
suitable teachers and the lack of training. Bums (1965) reports that African
countries which had introduced compulsory primary education had to
employ large numbers of untrained teachers. Thus in 1961, of the 40 OOO
primary school teachers in West Nigeria, 26 OOO (65 per cent) were un-
trained. This was due (1) to a lot of qualified teachers going to work in
secondary schools, and (2) to the large number of extra pupils following the
introduction of compulsory education without a prior teacher-training
scheme to provide sufficient teachers. As a result of the lack of training,
standards dropped, vernacular languages and the teaching profession lost in
status, and wages were reduced. This in tum caused a lot of teachers to leave
for more remunerative or less arduous jobs, so that the remaining teachers
were basically of two groups: those with a sense of vocation, and those who
could not pass exams.
In higher education many teachers are expatriates. In Zambia, for
example, about 90 per cent of the teaching force at secondary schools in 1970
were expatriates who all taught in English (Ohannessian, 1978b). Burns
(1965) thought this figure applied also to the African universities in 1962,
although it was hoped to increase the number of African members of staff.
190 A. G. H. WALKER

However, Brann points out-1 that it is practically impossible to gain qualified

lecturers for Hausa in Nigeria, as the qualified potential candidates seek
better jobs such as in the diplomatic corps. Only when the saturation level
has been reached for such positions will there be candidates for university
posts.
Teacher-training colleges also have their difficulties. Ohannessian (1978a),
for example, sees the following problems in Zambia: the low calibre of the
students due to the low status accorded the teaching profession; the lack of
training of the lecturers themselves; the small amount of time allotted to the
teaching of Zambian languages as such and to the teaching of the pedagogical
and didactic principles of vernacular education; the preponderance of
English in the courses; the confusion arising because of the seven official
languages, as (1) the students have mixed linguistic backgrounds, (2) the
lecturers are not competent in all the languages, and (3) the students do not
know in which language area they will eventually be employed; the dearth of
materials in and on Zambian languages; and the lack of facilities for teaching
practice so that sometimes children have to be brought by lorry to the college
for the students to practice. This is symptomatic of the problems facing
colleges too small to be able to give their students proper training. In
interesting contrast to these problems, Omar (1976) emphasizes that in
Malaysia the curriculum for the teacher trainee has two parts: language
mastery, and the methodology of language teaching. This includes phonetics
and phonology, so that the trainee can cope with the great dialectal diversity
in Malaysia, and a training in the analysis of grammar leading to a better
understanding of the standard language and of the written and oral mistakes
made by the pupils. This reflects recognition of the fact that a teacher's
competence in a language is not sufficient but that he must also know about a
language if he is to teach it.
One reason why many young people do not wish to study a vernacular
language with a view to teaching is the lack of motivation. Ohannessian
( 1978a} mentions the lack of advancement opportunities in teaching Zambian
vernaculars, the lack of advancement opportunities in other fields because
of a knowledge of the languages, the minor role they play in higher educa-
tion so that no premium is set on excellence, and the small amount of
attention given them at the University of Zambia. Perhaps one major reason
is the lack of status afforded the teaching profession and/or the vernacular in
many countries. Linguistic minorities suffer from similar problems. Minority
schools tend to be small, limiting the possibilities of a personal career, and
there is often conflict with the majority schools (Ahlskog, 1969). Also, due
to the smallness of a minority group, a person may refuse to return to his
village as a teacher, as he is afraid of the animosities that might arise if he
failed a nephew or niece in an exam. This is one reason why there is still no
 VERNACULAR LANGUAGES AND EDUCATION 191

teacher of Frisian on the island of Amrum. However, motivation is not

always lacking, as teachers in West Frisia with a Frisian Teaching Certificate
receive an extra bonus in most municipalities, and in the Eskimo programme
in Alaska the teachers of Yup'ik are mostly well trained even though they
are not officially qualified. They hope, however, to gain a certificate through
their work which would endow them with more prestige and probably more
money.
Vernacular languages are sometimes considered a "soft option" at school
and university because of a lack of status and because there is perhaps not
such a literary or philological tradition as in major subjects, meaning that in
theory there is less for a student to learn. Therefore the students are often
not of the best quality or ones who cannot come to terms with a large
department. Conversely though, some students are idealists of very
high calibre. However, one wonders what use some of the students will be to
the language community as, especially in a linguistic minority, much ideal-
ism and personal initiative is required of a teacher. Similarly, if students are
studying a language without prior knowledge, it is most unlikely that they
will be able to learn the language sufficiently to be able to teach it after four
years at university or teacher-training college. However, if these are the only
sources to be culled for teachers, little can be done, unless the authorities are
understanding enough to allow amateur (i.e. unqualified but pedagogically
gifted) native-speakers to teach. In NF the problem of teachers has been
partially solved by persuading Frisian speakers already in the profession to
teach Frisian as well as their own subjects. The school authorities have also
agreed in principle to laymen being allowed to give lessons in their native
tongue (Walker and Wilts, 1979), and the field-workers try to gather in-
formation in the villages about Frisian children scattered through the
various universities of Germany who might be persuaded to return to NF as
teachers. Unfortunately, many move to a different part of Germany fol-
lowing marriage. In all three instances, though, the individuals will have had
no formal training in Frisian and will require assistance from outside. In
recognition of this fact, a teacher's working group is now planned. Formal
tuition in Frisian is given at the University of Kiel and the Teacher Training
College in Flensburg, and it is hoped that in the near future some students
may become qualified.
The teacher actually in school also deserves some discussion. The com-
munity's attitude towards a teacher will influence his ability to teach. If, for
example, the teacher has only limited knowledge of the vernacular so that
the children or the parents can see him making mistakes, or if he speaks a
different dialect or is considered a poor pedagogue, perhaps due to a lack of
talent or laziness, he will enjoy little respect in class. The teacher's own
attitude is important. If he is dedicated, he has more chance of success. If on
192 A. G. H. WALKER

the other hand he considers the job a burden, allowing other people to
prepare his teaching materials for him, his lack of enthusiasm will not escape
the children's notice. The teacher's attitude may be influenced by other
members of staff, and it is important that they be in agreement, or else the
vernacular teacher may give up rather than suffer an unfriendly staff room.
Also, if it is known in the community that some members of staff are against
vernacular education, this may spark off latent hostility.

The class room

In a plurilingual community one basic problem in the classroom can be the
number of languages found there. Ohannessian (1978a) found in a survey in
Zambia that the highest number was 17, and only 82 of the 254 classes
examined had a "unilingual" class where more than 80 per cent of the
children spoke one language. He distinguishes three classroom types: (1) a
heavily multilingual class where only a handful are being taught in their
mother tongue, (2) a homogenous group where the language being taught is
not that of the majority group, and (3) a homogenous group where the
majority speaks the approved language as the mother tongue. The problems
resulting from the multilingualism can be heightened by the mutual un-
intelligibility of the various languages, so that a teacher may be forced to
interpret the official approved vernacular to the pupils who do not under-
stand it. Luckily, teachers themselves tend to be multilingual in Zambia,
with an average of three languages per teacher. However, despite the
teacher's own multilingualism, Ohannessian found that only 39 per cent of
the 254 teachers interviewed were teaching their mother tongue, 55 per cent
the language they were most fluent in, and that 42 per cent were teaching
neither. As a result, the teachers were sometimes not teaching the ver-
nacular in the prescribed lessons, but doing something else like mathematics
or letting the children read. Furthermore, the teachers were sometimes not
familiar with the culture of the children they were teaching, and had to ask
for help with riddles, etc.
If, as a result of such indigenous plurilingualism or as a result of social
c}langes in a linguistic minority community, there is bi- or multilingualism in
the class, how should the children be taught? In Switzerland an experiment
was tried using two schools, one for Romansch children and one for
German. This failed, as the parents sent their children to the prestige
(German) school rather than the one corresponding to their true linguistic
background (Gregor, 1980). If instead of two different schools one has two
different classes, this can have the effect of one class feeling socially inferior,
as happened in Malaysia where tuition was divided between the native
speakers who worked with literature and classical texts, and learners who
concentrated on acquiring the language (Omar, 1976). However, experience
 VERNACULAR LANGUAGES AND EDUCATION 193

in NF would seem to indicate that separate classes are best, as otherwise

frustration soon overcomes the children competent in the language.
The method of teaching used is also important. If children are to learn a
vernacular such as Irish or Frisian, then it must be taught as a living
language, and not as a dead language concentrating on verb paradigms etc.
otherwise the children will probably leave school without any oral com-
petence. Similarly, tuition should not be purely oral, perhaps because of a
fear of the orthography, as reading and writipg help imprint the contents of a
lesson on a child's brain.

Materials
Burns (1965) notes that education in Africa has often had the effect of
alienating the children from their own society, as education was not geared
to Africa. He considers that education should bring a synthesis of the old
values of the tribe and the new ones of the wage-earning society, and as such
must refer to the realities of life in an African community. Thus any
materials used in education should be culturally relevant and not be as
foreign as circuses, railway-stations and cricket bats to the Scots Gaels on
Lewis (Stephens, 1978) or the European and early American settler accounts
to the Eskimos in Alaska (Trifonovitch, 1976). The contents and style
should also be adapted to the age group. Omar (1976) reports that the book
Sejarah Melayu was recommended for primary schools as it was supposed to
represent the best Malay language. However, it proved unsuitable, as the
texts were too heavy and the language dated from the seventeenth century.
One of the characteristics of most vernaculars which are not world
languages is a dearth of materials. To remedy this, books etc. are often
translated from another language. There is, however, some doubt as to the
advisability of this policy. On the one hand children may find it a help to read
a book in the vernacular they already know in the original language, e.g. in
the case of linguistic minorities; but on the other hand it may undermine
their confidence in their own language if they discover that most of the
literature is merely translated from other languages. The authenticity of the
vernacular must also be lost in translation, as the original was conceived in a
different linguistic and cultural background. Perhaps though this is not so
important at school level, and again practical considerations may determine
the choice, as it seems that translation is easier than the creation of new
works. Translation does, however, also pose its problems, as one version
may not suffice. One book was adapted, for example, for primary school use
on Sylt featuring dunes, beaches, the Sylt flag and the Sylt style of archi-
tecture. The book is excellent for Sylt Frisian but has to be completely
re-adapted for the mainland as here there are only dikes, a different flag and
different architecture. Furthermore, the Sylt Frisian grammar differs some-
194 A. G. H. WALKER

what from the mainland Frisian grammar, e.g. two as opposed to three
genders, one as opposed to two forms of the definite article. This neces-
sitates a different division of the pages, a re-ordering of the pictures and the
drawing of new ones. Considering the amount of work involved with this
small project, it might prove simpler to write a completely new text-book.
Who should write the text-books? Few people are prepared to write in a
vernacular as {1) there is probably no financial reward, (2) there is the
possibility of unfavourable reviews, which in a small community could prove
embarrassing, (3) readership is limited, and (4) the editions are smaJI. The
ideal writer should have linguistic and educational qualifications, a good
knowledge of the area and its educational problems, and some experience in
teaching his mother tongue. In the case of a language being reintroduced to
children, he would also have to know how to present his mother tongue as a
foreign language. This would involve a contrastive analysis of the languages
involved (Boon, 1969). The sort of materials required are: (1) primers, (2)
readers, (3) informative reading material, and (4) non-informative reading
material, such as comics, the latter being important as children learn best to
read by reading for entertainment.
One way of solving the problem of materials is for the children to write
them themselves in collaboration with the teachers. This is being actively
encouraged in the Rough Rock Demonstration School (Liebe-Harkort,
1980) and in the Eskimo programme (Trifonovitch, 1976), as it fosters pupil
involvement in and a sense of identity with the school curriculum.

The economics of vernacular education

The theoretical needs of vernacular education are often compromised by the
simple expedience of economics. Many developing countries are obliged to
introduce a world language as a medium of instruction and use the cor-
responding text-books, because their own resources are so slender that they
cannot afford a programme of text-book translation or teacher-training in
the vernacular (Whiteley, 1969). The same applies for linguistic minorities.
Each time a book is written for Frisian tuition, for example, the author has to
find some generous Frisian association or authority who will supply the
funds to have it published. These funds are, however, limited. There is also
usually relatively little return on sales, because books in such small editions
(e.g. 500 copies) are expensive to produce but cannot be sold at too
expensive a price as otherwise the public aimed at will not be able to afford
the book. Thus a large percentage of each grant is automatically lost.
Because of the cost, printing as such has largely been abandoned in NF, at
least for school materials, and the off-set production technique is preferred,
or materials are simply produced, xeroxed and thrown away after use. This
 VERNACULAR LANGUAGES AND EDUCATION 195
does, however, have the disadvantage that no permanent text-book
appears.
Despite these problems, Tadadjeu (1977) argues that an efficient cost-
benefit analysis of the language programme could minimize costs and
maximize benefits, for example by improving cooperation between various
countries sharing a common vernacular, thus rendering education possible
in most African languages.
A further economic problem is the question of when minority schools are
economically viable. This problem tends to become more acute the higher
the school level. In a grammar school there will be fewer pupils than in a
prima.ry school, whereas there will be a demand for a larger number of
teachers covering a variety of subjects if an adequate training is to be
guaranteed. Furthermore, if an absolute number is set, how can fluctuations
in school numbers be catered for? Ahlskog (1969) reports that in Finland the
minimum number for a grammar school was set at 270 but that of the 40
Swedish-language schools only eight had more than 150 pupils.

Tests
If tests are carried out, there is a need to differentiate between why the test is
initiated, what is being tested and who is being tested. In the Eskimo
programme, tests are carried out monthly to evaluate progress in both
English and Yup'ik, to isolate difficult concepts and to identify students'
problems and weaknesses (Trifonovitch , 1976). Tadadjeu (1977) lists the
skills to be tested as (1) listening comprehension, (2) oral competence, (3)
reading, and (4) writing; and Williams (1973) suggests three groups of
pupils in Wales each requiring his own type of exam: (1) pupils with Welsh as
a foreign tongue, (2) pupils with Welsh in the environment but not in the
home, and (3) pupils with Welsh as their mother tongue.

Legislation
Legislation regarding vernacular education usually gives vernacular speakers
the right to education in their own language and is certainly commendable.
However, legislation alone is not enough, as a method of enforcement is also
necessary. Williams (1973, p. 99) reports thatthe Welsh Board of Education
stated that "where Welsh is the mother-tongue of the infants, that language
shall be the medium of instruction in the classes". The local education
authorities and teachers, however, responded only very sluggisly. Similarly,
in West Frisia, where Frisian was introduced as a compulsory subject on first
August 1980, the board of each individual school can decide how much
attention it pays to the law: whether there is one lesson a week for Frisian or
a bilingual programme is introduced (Zondag, 1982).
196 A. G. H. WALKER

Institutions
Institutions working with vernacular languages can prove very beneficial,
and Omar (1976), for example, considers a central language centre neces-
sary at the University of Malaya to coordinate the courses and the different
needs of the various faculties where Bahasa Malaysia is being taught. The
Algemiene Fryske Underrjocht Kommisje in Leeuwarden (Netherlands)
produces excellent materials for Frisian classes in school. However, institu-
tions can also have a negative effect and stultify private enterprise, as
laymen and teachers may refuse to show any initiative, arguing that the staff
of the institutions are paid to do the work and not the laymen. This has the
danger in a minority language movement of reducing the active members to
the "professionals" (i.e. those paid to work with the language), and taking
away the grass roots essential to a successful movement. Similarly, if the
institution is not recognised by the language community, it might do some
harm. In NF, for example, there seems to be the feeling that a person wins
respect by being proficient in his own profession, e.g. as a teacher, farmer or
business man. If he then also works with Frisian, this is applauded and others
will follow. A person, however, who earns his money by working with
Frisian alone seems not to be considered to be making a useful contribution
to society, and thus will not rank very high in the community's esteem.
In this chapter I have tried to illustrate some of the problems facing
vernacular education in the context of developing countries, minority
languages and immigrants. Sociolinguistic knowledge and expertise are pre-
requisites for successful programmes of vernacular education. Without
them, it is impossible to comprehend and, hopefully, solve the problems
involved, ranging as they do in complexity from a simple lack of finance to
the intricate social and linguistic patterns found in a pluriethnic and pluri-
lingual society.

Notes
1
Further research with respect to the number of languages spoken is currently being
conducted in selected areas of England by the Linguistic Minorities Project, based
at the University of London Institute of Education. Even this project does not,
however. expect to arrive at comprehensive figures for the whole country as nothing
short of a census questionnaire could produce such statistics. (My thanks to Euan
Reid of the Linguistic Minorities Project for this information.)
2
I diverge from the UNESCO definition of vernacular {1968, p. 689) which states:
••A language which is the mother tongue of a group which is socially or politically
dominated by another group speaking a different language." A vernacular may be a
 VERNACULAR LANGUAGES AND EDUCATION 197

minority language, but this is not a prerequisite. The Englishman's vernacular is after
all a world language.

C. M. B~ Brann - in the discussion following his lecture at the Sixth AILA

Conference in 1981 in Lund.

REFERENCES
Aarseth, B. (1969). The Situation of the Lapps, especially in Norway. In .. Lingual
Minorities in Europe." (E. Holmestad and A. Lade, eds), pp. 73-85. Det Norske
Samlaget, Oslo.
Ahlskog, S. H. (1969). Educational Problems in Finland. In "Lingual Minorities in
Europe." (E. Holmestad and A. Lade, eds), pp. 73-85. Det Norske Samlaget.
Oslo.
Arhammar, N. ( 1976). Historisch-soziolinguistische Aspekte der nordfriesischen.
Mehrsprachigkeit. Nordfriesisches Jahrbuch, N F 12, 55- 76.
Barker, G. C. (1947). Social Functions of Language in a Mexican-American Com-
munity. Acta Americana 5, 185-202.
Bendsen, B. ( 1860). '"Die Nordfriesische Sprache nach der Moringer rviundart". E. J.
Brill, Leiden. Reprint Sandig, Walluf, 1973.
Berry, J. (1968). The Making of Alphabets. In "Readings in the Sociology of
Language". (J. Fishman, ed.), pp. 737-753. Mouton, The Hague.
Berry, J. (1977). 'The Making of Alphabets' Revisited. In "Advances in the Creation
and Revision of Writing Systems". (J. Fishman, ed.), pp. 3-16. (Contributions
to the Sociology of Language 8). Mouton, The Hague.
Billigmeier, R. H. (1979). "A Crisis in Swiss Pluralism". (Contributions to the
Sociology of Language 26). Mouton, The Hague.
Boelens, Kr. (1976). "Friesisch und Niederlandisch. Zweisprachiger Grundschul-
unterricht in den Niederlanden." Ministerium fiir Unterricht und Wissenschaft,
Den Haag.
Bokamba, E. G. and Tlou, J. S. (1977). The Consequences of the Language Policies
of African States vis-a-vis Education. In "Language and Linguistic Problems in
Africa". (P. Kotey and H. Der-Houssikian, eds), pp. 35-53. Hornbeam Press,
South Carolina.
Boon, I. (1969). Problems of Text-books for lingual minorities. In "Lingual Minorities
in Europe." (E. Holmestad and A. Lade, eds). pp. 169-175. Det Norske
Samlaget, Oslo.
Burns, D. G. (1965). "African Education: an Introductory Survey of Education in
Commonwealth Countries." Oxford University Press, London.
Chambers, J. K. and Trudgill, P. (1980) ...Dialectology". Cambridge University
Press, Cambridge.
Chapman, A. J., Smith, J. R. and Foot, H. C. (1977). Language, Humour and
Intergroup Relations. In '"Language, Ethnicity and Intergroup Relations". (H.
Giles, ed.), pp. 137-169. Academic Press, London.
Committee on Irish Language Attitudes Research (1975). "Report submitted to the
Minister for the Gaeltacht". Government Stationery Office, Dublin.
Cummins, J. (1976). The Influence of Bilingualism on Cognitive Growth: A Synthesis
of Research Findings and Explanatory Hypotheses. Working Papers on Bi-
lingualism 9, 1-43.
198 A. G. H. WALKER

Dorian, N. (1973). Grammatical Change in a dying Dialect. Language 49, 413-438.

Dorian, N. {1977). The Problem of the Semi-Speaker in Language Death. In
"Language Death" (W. Dressler and R. Wodak-Leodolter, eds), pp. 23-32.
{International Journal of the Sociology of Language 12). Mouton, The Hague.
Dressler, W. and Wodak-Leodolter, R. (1977). Language Preservation and
Language Death in Brittany. In "Language Death". (W. Dressler and R.
Wodak-Leodolter, eds), pp. 33-44, (International Journal of the Sociology of
Language, 12) Mouton, The Hague.
Edwards, J. R. (1977). Ethnic Identity and Bilingual Education. In "Language,
Ethnicity and Intergroup Relations" (H. Giles, ed.), pp. 253-282. Academic
Press, London.
Fellman, J. (1973). "The Revival of a Classical Tongue" (Contributions to the
Sociology of Language 6). Mouton, The Hague.
Ferguson, C. A. (1968). Language Development. In "Language Problems of
Developing Nations" (J. Fishman, C. Ferguson and J. Das Gupta, eds), pp.
27-35. John Wiley, New York.
Ferguson, C. A., Houghton, C. and Wells, M. (1977). Bilingual Education: An
International Perspective. In "Frontiers of Bilingual Education" (B. Spolsky
and R. Cooper, eds). Newbury House, Rowley, Massachusetts.
Fillmore, C. A. {1973). A Grammarian looks to Linguistics. Georgetown University
Monograph Series on Languages and Linguistics 25, 273-287.
Fishman, J. A. {ed.) (1966). "Language Loyalty in the United States". Mouton, The
Hague.
Fishman, J. A. (1968a). Nationality-Nationalism and Nation-Nationalism. In
"Language Problems of Developing Nations" (J. Fishman, C. Ferguson and J.
Das Gupta, eds), pp. 39-51. John Wiley, New York.
Fishman, J. A. (ed.) (1968b). "Readings in the Sociology of Language". Mouton,
The Hague.
Fishman, J. A. {1968c). Sociolinguistics and the Language Problems of the
Developing Countries. In "Language Problems of Developing Nations" (J.
Fisman, C. Ferguson andJ. Das Gupta, eds), pp. 3-16. John Wiley, New York.
Fishman, J. A. (1971). The Sociology of Language. Jn "Advances in the Sociology of
Language I" (J. Fishman, ed.), pp. 217-404. Mouton, The Hague.
Fishman, J. A. (ed.) (1974). "Advances in Language Planning" (Contributions to the
Sociology of Language 5). Mouton, The Hague.
Fishman, J. A. {ed.) (1977). "Advances in the Creation and Revision of Writing
Systems'' (Contributions to the Sociology of Lanuage 8). Mouton, The Hague.
Fishman, J. A. and Hofman, J. E. (1966). Mother Tongue and Nativity in the
American Population. In "Language Loyalty in the United States" (J. Fishman,
ed.), pp. 34-50. Mouton, The Hague.
Fishman, J. A., Ferguson, C. A. and Das Gupta, J. (eds) (1968). "Language
Problems of Developing Nations". John Wiley, New York. ..
Fort, M. C. (1980). "Saterfriesisches Worterbuch rnit einer grammatischen Ubersicht".
Helmut Buske, Hamburg.
Gaarder, B. (n.d.) "Political Perspectives on Bilingual Education". Ms.
Gallagher, C. F. (1971). Language Reform and Social Modernization in Turkey. Jn
"Can Language be Planned?" (J. Rubin and B. Jernudd, eds), pp. 159-178.
East-West Centre Press, Honolulu.
Genesee, F. (1978). Is there an Optimal Age for Starting Second Language Instruc-
tion? McGill Journal of Education, XIl, 2, 145-154.
 VERNACULAR LANGUAGES AND EDUCATION 199

Giles, H., Bourhis, R. Y. and Taylor, D. M. (19n). Towards a Theory of Language

in Ethnic Group Relations. In "Language, Ethnicity and Intergroup Relations"
(H. Giles, ed.), pp. 307-348. Academic Press, London.
Gregor, D. B. (1980). "Celtic: A Comparative Study". Oleander Press, Cambridge/
New York.
Groustra, G. R. (1976). Was geschieht in Seelterlound? Nordfriesland, 35/36, 180-
182.
Gumperz, J. J. (1968). Types of Linguistic Communities. In "Readings in the
Sociology of Language" (J. Fishman, ed.), pp. 460-472. Mouton, The Hague.
Haarmann, H. (1975). "Soziologie und Politik der Sprachen Europas". Deutscher
Taschenbuch Verlag, Miinchen.
Hansford, K., Bendor-Samuel, J. and Stanford, R. (1976). "An Index of Nigerian
Languages" (Studies in Nigerian Languages 5). Summer Institute of Linguistics,
Accra.
Haugen, E. (1966). "Language Conflict and Language Planning: The Case of
Modern Norwegian". Harvard University Press, Cambridge, Massachusetts.
Haugen, E. (1980). Language Problems and Language Planning: The Scandinavian
Model. Jn ••Languages in Contact and Conflict" (P. Nelde, ed), pp. 151-157.
Zeitschrift fiir Dialektologie und Linguistik, Beihefte 32. Franz Steiner,
Wiesbaden.
Heine, B. (1979) ... Sprache, Gesellschaft und Kommunikation in Afrika" (Afrika
Studien 103). Weltforum, Miinchen.
Hofmann, D. (1956). Probleme der nordfriesischen Dialektforschung. Zeitschrift fiir
Mundartforschung XXIV, 78-112.
Isajiw, W.W. (1974). Definitions of Ethnicity. Ethnicity l, 111-124.
Johnson, B. C. (1977). Language Functions in Africa: A Typological View. In
••Language and Linguistic Problems in Africa" (P. Kotey and H. Der-Houssikian,
eds), pp. 54-67. Hornbeam Press, South Carolina.
Jones, M. (1973). The Present Condition of the Welsh Language. In "The Welsh
Language Today" (M. Stephens, ed.), pp. 110-126. Gomer Press, Llandysul.
Jonsson, J. H. {1979). Zur Sprachpflege im heutigen Island. In "Friesisch heute"
(A. Walker and 0. Wilts, eds), pp. 43-45. Schriftenreihe der Akademie
Sankelmark NF 45/46, Sankelmark.
Jorgensen, V. T. (1979). Tu di nordfrasche Schriwwise. In ••Friesisch heute" {A.
Walker and 0. Wilts, eds), pp. 102-115. Schriftenreihe der Akademie
Sankelmark NF 45/46, Sankelmark.
Karam, F. X. (1974). Toward a Definition of Language Planning. Jn .. Advances in
Language Planning" (J. Fishman, ed.), pp. 103-112. (Contributions to the
Sociology of Language 5) Mouton, The Hague.
Khleif, B. B. {1980). "Language, Ethnicity and Education in Wales" (Contributions
to the Sociology of Language 28). Mouton, The Hague.
Kloss, H. (1966). Types of Multilingual Communities: A Discussion of Ten Variables.
Sociological Inquiry 36, 135- 145.
Kloss, H. (1968). Notes concerning a Language-Nation Typology. Jn '"Language
Problems of Developing Nations" (J. Fishman, C. Ferguson and J. Das Gupta,
eds), pp. 69-85. John Wiley, New York.
Kramer, P. {1978). W at int Seelterlound geschjut. Nordfriesland 45/46, 45-47.
Krishnamurti, Bh. ( 1978). Language Planning and Development: The Case of
Telugu. Contributions to Asian Studies XI, 37-56.
Labov, W. {1966). "The Social Stratification of English in New York City". Center
for Applied Linguistics, Washington DC.
200 A. G. H. WALKER

Lambert, W. E. (1975). CultureandLanguageasFactorsinLearningandEducation.

In "Education of Immigrant Students" (W. Aaron, ed.), pp. 55-83. Ontario
Institute for Studies in Education, Toronto.
Larsen, N.-E. (1983). "Statistical Investigations of Language Death in a North
Frisian Community". (In press).
LePage, R. B. (1964). "The National Language Question". Oxford Univserity Press,
London.
Liebe-Harkort, M.-L. (1980). Factors Influencing the Survival of Minority Languages:
A Catalogue and its Application to the Situation of the Apachean Languages.
SAIS. Arbeitsberichte aus dem Seminar fur Al/gemeine und Indogermanische
Sprachwissenschaft, Christian-Albrechts-Universitiit Kiel 2, 67-125.
Mackay, D. J. (1969). Educational Problems in Scotland. Jn "Lingual Minorities in
Europe" (E. Holmestad and A. Lade, eds), pp. 142-145. Det Norske Samlaget,
Oslo.
Mackey, W. F. (1978). A Typology of Bilingual Education. Jn "Bilingual Education"
(H. LaFontaine, B. Persky and L. Golubchick, eds), pp. 192-203. Avery
Publishing Group, New Jersey.
Macnamara, J. (1966). "Bilingualism and Primary Education". University Press,
Edinburgh.
Macnamara, J. (1967). The Effects of Instruction in a Weaker Language. J. Soc.
Issues XXIII, 2, 121-135.
Macnamara, J. (1971). Successes and Failures in the Movement for the Restoration
of Irish. In "Can Language be Planned?" (J. Rubin and B. Jernudd, eds), pp.
65-94. East-West Centre Press, Honolulu.
McLaughlin, B. (1978). "Second-Language Acquisition in Childhood". Lawrence
Erlbaum Associates, New Jersey.
Miihlhausler, P., Wurm, S. A. and Dutton, T. E. (1979). Language Planning and
New Guinea Pidgin. In "New Guinea and Neighbouring Areas: A Sociolinguistic
Laboratory" (S. Wurm, ed.), pp. 263-276. (Contributions to the Sociology of
Language 24). Mouton, The Hague.
Nida, E. A. and Wonderly, W. M. (1971). Communication Rules of Languages in
Multilingual Societies. In "Language Use and Social Change" (W. Whiteley,
ed.), pp. 57-74. Oxford University Press, London.
Niebaum, H. ( 1977). "Westfalisch" (Dialekt/Hochsprache Kontrastiv, Sprachhefte
ffir den Deutschunterricht 5) Schwann, Diisseldorf.
Ohannessian, S. (1972). The Language Problems of American Indian Children. In
"The Language Education of Minority Children" (B. Spolsky, ed.), pp. 13-24.
Newbury House, Rowley, Massachusetts.
Ohannessian, S. (1978a). The Teaching of Zambian Languages and the Preparation
of Teachers for Language Teaching in Primary Schools. Jn "Language in
Zambia." (S. Ohannessian and M. Kashoki, eds), pp. 292-328. International
African Institute, London.
Ohannessian, S. (l978b). English and Zambian Languages in Secondary Schools. In
"Language in Zambia" (S. Ohannessian and M. Kashoki, eds), pp. 355-375.
International African Institute, London.
Ohannessian, S. (1978c). Zambian Languages in Secondary Schools: Attitudes of
Teachers in Training. In "Language in Zambia" (S. Ohannessian and M.
Kashoki, eds), pp. 376-397. International African Institute, London.
Okonkwo, C. J. E. (1975). A Function-oriented Model of Initial Language Planning
in Sub-Saharan Africa. Ohio State University Working Papers in Linguistics 19,
37-52.
 VERNACULAR LANGUAGES AND EDUCATION 201
Omar, A. H. (1976). "The Teaching of Bahasa Malaysia in the Context of National
Language Planning". Kuala Lumpur, Malaysia.
Paulston, C. B. (1980). "Bilingual Education: Theories and Issues". Newbury House,
Rowley, Massachusetts.
Price, G. (1973). Minority Languages in Western Europe. In "The Welsh Language
Today" (M. Stephens, ed.), pp. 1-17. Gomer Press, Llandysul.
The Restoration of the Irish Language (1965). Stationery Office, Dublin.
Rustow, D. A. (1968). Language, Modernization and Nationhood- an Attempt at
Typology. In "Language Problems of Developing Nations" (J. Fishman, C.
Ferguson and J. Das Gupta, eds), pp. 87-105. John Wiley, New York.
Sjoberg, A. (1966). Socio-Cultural and Linguistic Factors in the Development of
Writing Systems for Preliterate Peoples. In "Sociolinguistics" (W. Bright, ed.),
pp. 260-276. Mouton, The Hague.
Sjolin, B. (1969). "Einfiihrung in das Friesische". Sammlung Metzler, Stuttgart.
Spenter, A. (1977). Zur Mehrsprachigkeit in der Gemeinde Rodenas. Nordfriesisches
Jahrbuch NF 13, 167-177.
Spolsky, B. (1972). The Language Education of Minority Children. In "The
Language Education of Minority Children" (B. Spolsky, ed.), pp. 1-10.
Newbury House, Rowley, Massachusetts.
Spolsky, B., Green, J.B. and Read, J. (1976). A Model for the Description, Analysis
and perhaps Evaluation of Bilingual Education. In "Language in Sociology" (A.
Verdoodt and R. Kjolseth, eds), pp. 233-263. lnstitut de Linguistique de
Louvain, Louvain.
Stephens, M. (1978 2). "Linguistic Minorities in Western Europe." Gomer Press,
Dyfed.
Stewart, W. A. (1962). An Outline of Linguistic Typology for Describing Multi-
Lingualism. In "Study of the Role of Second Languages." (Ed. F. Rice), pp.
15-25. Centre of Applied Linguistics of the Modem Language Association of
America, Washington DC.
Stewart, W. A. (1968). A Sociolinguistic Typology for Describing national Multi-
lingualism. In "Readings in the Sociology of Language." (Ed. J. Fishman), pp.
530-545. Mouton, The Hague.
Swain, M. (1980). "Immersion Education: Applicability for Non-vernacular Teaching
to Vernacular Speakers." Paper presented at the Symposium on Standard
Language/Vernacular Relations and Bilingual Education; Racine, Wisconsin.
Swain, M. (1983). "Bilingual Education for Majority and Minority Language
Children." (In press).
Tadadjeu, M. (1977). Cost-Benefit Analysis and Language Education Planning in
Sub-Saharan Africa. In "Language and Linguistic Problems in Africa"
(P. Kotey and H. Der-Houssikion, eds), pp. 3-34. Hornbeam Press, South
Carolina.
Tauli, V. (1968). "Introduction to a Theory of Language Planning". Acta Universitatis
Upsaliensis 6, Uppsala.
Tholund, J. (1980). Quo vadis, Sylt? Quo vadis, Nordfriesland? Nordfries/and 56,
121-126. .
Trifonovitch, G. J. (1976). "A Brief Evaluation of the Primary Eskimo Program".
Bureau of Indian Affairs, Bethel, Alaska.
Trudgill, P. (1975). "Accent, Dialect and the School". Explorations in Language
Study. Arnold, London.
UNESCO (1968). The use of Vernacular Languages in Education: The Report of the
202 A. G. H. WALKER

UNESCO Meeting of Specialists, 1951. In ••Readings in the Sociology of

Language" (J. Fishman, ed.), pp. 688-716. Mouton, The Hague·.
Walker, A. G. H. (1977). Sprachwandel in Nordfriesland. Jahrbuch des Vereinsfiir
niederdeutsche Sprachforschung 100, 86-107.
Walker, A. G. H. (1980a). North Frisia and Linguistics. Nottingham Linguistic
Circular 9/1, 18-42.
Walker, A. G. H. (1980b). Some Factors Concerning the Decline of the North
Frisian Tongue. Jn "Languages in Contact and Conflict" (P. Nelde, ed.), pp.
453-460. Zeitschrift fiir Dialektologie und Linguistik, Beihefte 32. Franz
Steiner, Wiesbaden.
Walker, A. G. H. (1980c). :'Die nordfriesische Mundart der Bokingharde. Zu einer
strukturell-dialektolog'ischen Definition der Begriffe 'Haupt-', ·unter-' und
'Dorfmundart"'. Zeitschrift fiir Dialektologie und Linguistik, Beihefte 33.
Franz Steiner, Wiesbaden.
Walker, A. G. H. and Wilts, 0. (eds) (1979). "Friesisch heute". Schriftenreihe der
Akademie Sankelmark NF 45/46, Sankelmark.
Weinreich, U. (1953). "Languages in Contact". Publications of the Linguistic Circle
of New York.
Wells, G. (1981). "Learning through Interaction: The Study of Language Develop-
ment". Cambridge University Press, Cambridge.
Whiteley, W. H. (1969). "Swahili, the Rise of a National Language". Methuen,
London.
Williams, J. L. (1973). The Welsh Language in Education. In "The Welsh Language
Today" (M. Stephens, ed.), pp. 92-109. Gomer Press, Llandysul.
Wilts, 0. (1979). Lexikographie zwischen Sprachwissenschaft und Sprachpflege. Us
Wurk, 28, (Scripta Frisica, Tinkbondel foar Arne Spenter), 197-206.
Wilts, 0. (1983). Sprachplanerische Aspekte der nordfriesischen Lexikographie. In
•'Friserstudier III". (N. Danielsen, E. Hansen, H. F. Nielsen, H. Bekker-Nielsen,
eds), pp. 9-24. Odense Universitetsforlag Odense.
Wurm, S. A. (1979). The Language Situation in the New Guinea Area. In "New
Guinea and Neighbouring Areas: A Sociolinguistic Laboratory" (S. Wunn,
ed.), pp. 3-10. (Contributions to the Sociology of Language 24). Mouton, The
Hague.
Zondag, K. (ed.) (1982). "Bilingual Education in Friesland". T. Wever, Franeker.
 8
Applied discourse analysis and
educational linguistics

Michael Stubbs

Introduction
This chapter is about some aspects of educational linguistics, and the most
basic question to be discussed is: what should teachers and students know
about language? I will attempt to give one possible answer to the question
posed in the following quote from Halliday (1978, p. 207):

I am often asked by teachers if it is possible to give a succinct account of the

essential nature of language in terms that are truly relevant to the educational
process. It is not easy to do this, because it means departing very radically from
the images of language that are presented in our schoolbooks and in the
classroom ... We have to build up an image of language which enables us to look
at how people actually do communicate with each other ...

My general argument will be that a model of language relevant to educa-

tional purposes should focus on language in use in connected discourse. I
will review .some general issues concerning the relation between linguistics
and teacher-training, and then discuss two very different in-service teacher-
training courses which I have prepared, discussing the theoretical rationale
behind the courses, which lies in recent work on discourse analysis. Finally, I
will review some applications of discourse analysis in teaching both the
mother tongue and foreign languages.

APPLIED SOCIOLINGUISTICS Copyright © 1984 by Academic Press. London

ISBN: 0-12-701220-6 All rights of reproduction in any form reserved
 204 MICHAEL STUBBS

In considering discourse analysis and education, there are two logically

possible areas of application. One could (1) study the_Wsconrse_of.teaE!!ffig
· jtself, as part of a study of institutional language. Alternatively, one could
(2) teach about discourse and communication, as part of the content of a
foreign language or mother tongue syllabus. There are two main topics for
(1),_a study__gf classroom_di~rs~. One is_~oken class.room inter~t!.on
between teacher and students, and this has been ~-major research area in the
soci3.lsclen~~s sine~ !ne~e l9.60s. The other ls the study of the discourse
organisation of written teaching texts, primarily "textbooks" and related
material. Education in Western schools and higher education is, in these
senses, predominantly verbal and textual (cf. Van Dijk, 1981, p. 2), and a
stu<;ly of teaching itself therefore demands a study of its written and spoken
discourse. Possibility (2), teaching about discourse and communication,
could also have two rather different emphases. The main aim might be to
increase students' communicative competence, by giving them more ability
in a wider range of discourse types, and therefore increasing the functional
potential of students' language (cf. Halliday, 1978, p. 57). Alternatively,
one could teach directly the theory of discourse analysis, as part of a course
on language or linguistics. Teaching about language in the education system
usually means teaching about either grammar or literature, with only un-
systematic discussion of the wide variety of non-literary discourse.
This framework of possible applications of discourse analysis underlies
this chapter, although not all four areas are equally discussed. The frame-
work is similar to one proposed by Van Dijk (1981), in a very useful survey
article on discourse studies and education. Van Dijk admits however that his
article is entirely speculative and programmatic: I will try here .also to give
more ccncrete examples.
I will not attempt to give a comprehensive overview of the applications of
discourse analysis. It would not be practicable to attempt a survey of such a
recent and open-ended field, which draws on several different academic
traditions including linguistics, cognitive psychology, sociology and artificial
intelligence. It is probably more useful to try and tackle a few more specific
questions with reference to particular examples. In educational linguistics,
this means discussing particular syllabus proposals and teaching strategies.
Amongst other possible applications of discourse analysis which I will not be
able to discuss here are: a view of psychotherapy as conversation (e.g. Labov
and Fanshel, 1977); an approach to developing the communicative com-
petence of mentally retarded patients (e.g. Price-Williams and Sabsay,
1979); training doctors, either native or foreign speakers, to interact more
efficiently with patients (e.g. Candlin, Bruton, Leather and Woods, 1977);
and social skills training (e.g. Hargie, Dickson and Saunders, 1981), in-
cluding the vast literature on effective communication in business and
 EDUCATIONAL LINGUISTICS 205
management. The applied study of discourse in these and other areas is,
however, concerned with training professionals of different kinds to under-
stand better the rules and conventions underlying spoken and written
discourse. Again, it may be more useful to avoid bland overviews, and to
develop most of my points specifically with reference to teachers, making
points relevant to other professional groups by implication. I have preferred
therefore to be mildly original rather than to attempt a comprehensive
survey, and my bibliographic references are intended as representative but
in no way exhaustive.
Milroy (in this volume) provides one general alternative approach to a
socially responsible applied discourse analysis, in her discussion of the
theoretical and practical issues raised by communicative problems. In
particular, she discusses the kind of cross-dialectal communicative problems
which are likely to have serious social consequences when professionals and
clients from different social groups and with different dialects meet in school
classrooms, doctors' surgeries, courtrooms and other comparable settings.
For case studies which analyse the discourse in such settings, and discuss
such cross-dialectal communicative problems, see Harris ( 1980) on magi-
strates' courts and Malcolm (1979) on school classrooms.

Definitions and terminology

There is considerable variation in the use of the term discourse Jm_a_ly__sis, so I
had better insert a brief note on what I mean by it. Ffrst, I do not intend to
draw any important distinction between text and discourse. As these terms
are normally used, they often imply only a difference in emphasis, but
nothing of theoretical importance need hang on the distinction. A distinc-
tion is sometimes implied between written text and spoken discourse.
Alternatively, discourse refers to interactive language versus text as non-
interactive monologue, whether spoken or written. For example, one can
talk of the text of a speech. Another distinction is that discourse implies
length whereas a text may be very short. Halliday and Hasan (1976) define a
text as a semantic unit and point out that complete texts include Exit and No
smoking. Some scholars have used the two terms to label theoretically
important distinctions, but since I will not be concerned with those distinc-
tions here, I will ignore them.
Similarly, I will favour the term discourse analysis over other terms for
reasons of convenience rather than theory. The term text analysis would do
equally well except that it usually implies a particular European tradition of
text linguistics. The term conversational analysis might also serve, except
that it almost always implies work which derives from ethnomethodology. It
is also too narrow in that it implies a restriction to conversation, and the
206 MICHAEL STUBBS

exclusion of more formal discourse, although there are problems in the

analysis of discourse which are common to formal and informal, written and
spoken laguage - for example, the analysis of lexical and grammatical
cohesion.
By discourse analysis I mean therefore the linguistic or sociolinguistic
analysis of naturally occurring discourse or text, spoken or written. This
does not deny the validity of other approaches: a full understanding of
discourse is necessarily interdisciplinary. It merely restricts my topic here to
manageable limits. (This definition of discourse analysis is developed at
length by Stubbs, 1983b).

Discourse analysis in teacher education

Linguistics in teacher-training: some general points

The main question to be discussed in thi~ _chap~~r is: what kind of under-
teaching the mother tongue- or Ioie!~
··stancliiig-of language -is relevant in
languages, or in teaching in general?_Which linguistic concepfs~should be
presented to teachers and how? What therefore is the relation between
linguistics and teacher-training? Language is ooinplex, so soirie-siifrplifica=.
tion In the presentation is inevitable; but how can the simplification be
managed without distorting the material or patronising the audience? These
questions are not often discussed explicitly by linguistics, although it is only
professional linguists who have the necessary knowledge and full-time
commitment to linguistics which would allow them to select what is relevant
to teachers, or to any other professionals. Educationalists could not be
expected to do this for the simple reason that there is so much in con-
temporary linguistics which is of no direct relevance to education. It is
arguably a social responsibility of linguists as a profession to present to
non-linguists in an accessible way the aspects of knowledge about language
which are of important practical value.
The complexity, breadth and specialization of contemporary linguistics
has itself meant that educationalists have often rejected linguistics in toto.
Not being able to see the wood for the trees, they have jumped to the
conclusion that linguistics comprises jargon and formalism, and discusses
only a model of language which is of no use to teachers and pupils, since it
avoids real language behaviour. Descriptions such as linguistics is "scientific",
"mathematical" or "abstract" are often used to refer in a slightly confused
way to such feelings. There is no point in ignoring the problem that linguistics
is now as much as a tum-off for many teachers as grammar used to be. This
 EDUCATIONAL LINGUISTICS 207
may be a superficial reaction, but it must be countered: there is no point in
talking if no-one wants to listen. It follows that questions of presentation are
crucial.
There is often a suspicion in teacher education of anything that practising
teachers might not understand immediately, and linguistics has a reputation
for conceptual difficulty. Many teachers have an image of themselves as
practical and down-to-earth folk, having to cope with everyday life at the
chalk-face, in the blackboard jungle, and as taking a sensible standing against
the impractical theorising of linguists, sociologists and psychologists of
education, and all the rest,

those sinister figures in the wings, faintly contemptuous, armed with the para-
phernalia of expertise and tapping ominously their research findings. (Rosen,
1978)

There is doubtless much justification in such a sceptical attitude towards

theory, and it has to be taken into account in preparing training courses. If
the clients are not interested, they will not learn anything. And they are
often justifiably sceptical after past promises and disappointments. Teachers
are under constant pressure from new ideas and information, not only from
linguistics, but also from psychology, sociology and other disciplines, and
under constant pressure to respond to what often turn out to be short-lived
fashions.
However, the view that teacher-training must have the practical short-
term goal of making them better teachers in the classroom tomorrow may be
short-sighted and dangerous. It is short-sighted since there are other,
longer-term, and possibly more interesting and ambitious goals. And it is
dangerous since it implies that teachers have no need for, or are incapable of
continuing their own general education or engaging in their own scholarly
study about their subject or about their own professional behaviour.
Teachers are scholars. However, some recent educational thinking has
retreated into well meaning but ultimately sentimental concern with child-
ren, sometimes disguised as a practical stance against abstract theorising.
This has sometimes led to teacher training courses which flirt with the
peripheral aspects of academic disciplines. It must be admitted, for ex-
ample, that some courses focus almost exclusively on psycholinguistic and
sociolinguistic aspects of language, but avoid any detailed analysis of
language itself. Carter {1980, p. 228) has likened this to the absurdity of a
mathematics course which avoids doing too much mathematics. However, it
has to be demonstrated that training can provide interesting intellectual
challenges, of inherent interest to teachers as educated people. Also it must
be admitted that a real problem with teaching linguistics is that it requires a
208 MICHAEL STUBBS

very considerable initial investment by students before they begin to see the
general value of what they are doing, and can use their understanding to
prepare their own teaching materials. Studying linguistics is like going on a
blind date.
A danger of an overly practical, short-term view is that without prirtciples
teachers are condemned to follow techniques superficially, without being
able to adapt them for their own particular circumstances. They may take
particular examples as orthodoxy, mistaking them for unchangeable pre-
scriptions, and be unable to invent new examples. Finally, of course, an
insistence on a practical approach may conceal a basic misunderstanding, if
it is taken to imply that teaching can be theory-free. All teaching takes place
on the basis of some theory (whether Piagetian, audio-lingual, or whatever,
or simply classroom folklore), and this should be made explicit. As I have
begun to indicate, discourse analysis can itself begin to analyse the assump-
tions underlying classroom dialogue and written texts.
There are further problems related to how abstract and analytic teaching
about language and linguistics should be. It is plausible that a functional
view of language in use will be of more direct relevance to teaching practice
than a purely abstract view of language as system and structure. A view of
language in use also starts from everyone's everyday experience of language.
However, it is difficult to introduce samples of real language in use, without
immediatley decontextualizing and trivializing them. It has often been
pointed out that the search for authenticity in language teaching materials is
an illusory one. If an authentic text (that is, a text originally written for some
real purpose, without the linguist's intervention and not specially prepared
as teaching material) is taken out of its context, and used for something else
(teaching), it is thereby made inauthentic. However, this is simply to note
that all teaching implies some contrivance, which may be more or less
extreme.
A more basic problem may be that, even if it is accepted that a functional
orientation is of more interest to teachers, this functional view may appear
hopelessly vague if there is no formal analytic model to support it. If one
starts with functions, one often never gets (back) to forms. To argue in a
Hallidayan fashion, for example, that language is as it is formally, because of
the functions it serves, assumes a sophisticated prior understanding of
forms: of concepts such as rank-scale, and the mapping of one layer of
structure onto another. More generally, one might argue that the valuable
thing which is to be gained from a study of linguistics is not the details of
particular formal arguments, but the style of argument itself: the nature of
the data used, the attention to evidence of different kinds, the concept of a
counterexample to a clearly formulated statement, and so on. However, this
kind of argumentation can only be properly demonstrated on detailed
 EDUCATIONAL LINGUISTICS 209
examples, and has been fully developed only in descriptions of phonological
and syntactic forms. Again, at least some analytic work seems necessary.
It is important, in summary, to distinguish between language in educa-
tion and linguistics in education: they are potentially very different. (cf. the
papers in Carter 1982, which make this distinction very clearly.)
Although these questions of selection for pedagogic purposes are not
often explicitly discussed in print by linguists, they are probably constantly
discussed in the course of meetings over syllabuses and examinations, and
they have to be faced in one way or another by any linguist who has to select
some topics in his subject in preparing a lecture course or writing a text
book. Probably all teachers have to adopt a policy of diminishing deception:
starting with plausible but strictly speaking unsound and oversimplified
arguments, to get students used to the style of argument and the basic
subject matter. This starting matter is then gradually refined and replaced by
something better. Different teachers will take different decisions on how
much initial distortion is defensible. However, some is inevitable, and if
anyone feels unhappy with the situation, it can be pointed out that the
process of diminishing deception is in any case what happens in anyone's
intellectual development, and what happens at a much slower pace in the
whole history of intellectual progress. As Feyerabend {1979, p. 156) has
pointed out, an argument does not necessarily reveal the true beliefs of the
arguer. One may, for example, adopt and express arguments which one
believes to be false, in order to persuade an audience. This is a common
pedagogic tactic, since false arguments may have to be followed through to
their logical conclusion, before being revealed as false.
Although all teachers are familiar to some extent with such decisions,
their interest here is that they constitute a possible definition of applied
linguistics: the selection or development of theories for different purposes.
However, the view that applied linguists interpret and mediate linguistics for
practitioners is only one possibility. There has been considerable debate
recently over whether applied linguistics is (1) a body of linguistic knowl-
edge which is applied to practical problems (i.e. linguistics applied), or (2) a
quasi-independent body of knowledge and specially developed theories (i.e
applied linguistics). For example, Widdowson (1977, 1980a, 1980b) pro-
vides a series of arguments which represent his own changing views about the
degree to which applied linguistics should develop its own theories inde-
pendent of theoretical linguistics. He argues in later papers (e. g. t 980a) that
models developed by theoretical .and applied linguistics are incompatible,
since there is a radical difference between analysts' and users' models of
language, and applied linguistics has to take account of the latter.
One also has to be careful about the general rationale which is proposed
for linguistics in teacher-training, or indeed more generally. It is often
210 MICHAEL STUBBS

argued that language is peculiarly central in human society, that humans are
homo loquens, that human society would be impossible without language,
that a detailed understanding of language can tell us how the human mind
works, and so on. In studying language therefore, students learn about
essential and defining characteristics of their psychological and social en-
vironments. These arguments may all be true, and they are convincing
rationales for studying linguistics. However, they do not clearly distinguish
linguistics from other traditional academic disciplines, including biology,
physics, geography, history or mathematics. All these disciplines and others
tell us about the relationship between human beings and their environment.
They also have much to teach students about valid forms of argument,
different ways of evaluating data and evidence, and so on. And, in any case,
the relationship between an abstract knowledge of language (or any other
subject) and behaviour is indirect.
There are, then, several general problems concerning the presentation of
linguistics to educationalists, and by implication to other professionals.
These problems are inseparable from the more specific problem of formulat-
ing "a succinct account of the essential nature of language in terms that are
truly relevant to the educational process" (Halliday, 1978).
Other problems do arise, however, from the admittedly special relation-
ship between speakers and their native language. Everyone bas a native
language, and this is a great advantage in such teaching, since this implicit
linguistic competence provides an enormous resource to draw on in teaching
an explicit and conscious knowledge of language. However, it is also a
disadvantage, since language is so notoriously open to misunderstanding
and myth. Language is so central to everyone's life that it is surrounded with
mystique. It is difficult to see any need for explanation at all. It either all
seems natural and is taken for granted without the need for explanation
being seen at all. Or speakers assume that just being a native speaker makes
them experts, especially if they have been further sensitised by some
language study, for example a training in literary criticism or learning a
foreign language. Carter (1981) therefore argues that teacher-training must
explicitly tackle misunderstandings about linguistics as the systematic study
of language, since the attitude of mind required in linguistics is often
believed to be contrary to the sensitivity required in literary study. It is
almost certain, for example, that any course will have to tackle the common
prejudice against formalism, idealization and explicitness, held by many
students with literary training. Linguists are regularly accused of wanting to
have things neater than they are, and of idealizing away just what is inter-
esting about instances of language in use. Since linguists themselves differ in
their view of how much idealization is permissible or useful, this provides a
major source of confusion for students. So any course therefore starts from a
 EDU CA TI ON AL LINGUISTICS 211

certain necessary demystification: some students find it exhilarating to have

their assumptions challenged; others find it uncomfortable and destructive.
To discuss here any further the issues involved in the dissemination of
information about language and linguistics would take me too far from my
main theme, so I will make just a few concluding comments central to any
application of sociolinguistics. Information is never neutral: it is always
transmitted in the face of prevailing expectations, preconceptions and
entrenched professional interests. Any serious discussion of such problems
would have to examine the way in which such prevailing views are supported
and legitimated by institutions. In Britain, this would involve, for example, a
study of the role of NATE (National Association for the Teaching of
English) and its journal English in Education in forming the ideas of
teachers, teacher-advisers and other educationalists. Or see Gordon (1980)
or Stubbs (1980, Chapter 7) for discussions of the way in which Bernstein's
ideas have often been distorted and simplified in their transmission to
teachers and their subsequent application. To take a more general example:
sociolinguistics is the study of language variation. It argues, contrary to
much recent theoretical linguistics, that language is inherently hetero-
geneous. When they have written on educational issues, sociolinguists have
therefore tended to stress the value of diversity, and have seen bidialecal-
ism, bilingualism and biliteracy as positive resources which teachers can use
and encourage. By and large, however, the educational system, and govern-
ment itself, has seen such diversity as a problem. For such reasons and
others, applied sociolinguistics cannot avoid consideration of the practical
sociology of knowledge.

Discourse analysis and foreign language teaching

1
" It has often been argued by both linguists and teachers that theoretical
linguistics has little or nothing to offer the practice of language teaching. For
1
example, one extreme statement is by Sampson (1980, p. 10):

'1 I do not believe that linguistics has any contribution to make to the teaching of
\I English or the standard European languages.

Sampson admits that linguistics may contribute to the teaching of exotic

languages, but only insofar as it provides descriptions which are not other-
wise available of such languages. Sampson's statement clearly recalls, in an
even more extreme fashion, a famous quote from Chomsky (in Lester, ed.,
1970, p. 52). It may be, however, that such beliefs are due to looking for the
contribution of linguistics in the wrong place. Sampson's and Chomsky's
view appears to be that theoretical descriptions of syntax have nothing to
212 MICHAEL STUBBS

offer the teacher, who has his own more appropriate pedagogic descriptions.
However, what I will _!l_ow argue more directly is tf!~l_recent work in dis-
cours~ £~fl_ -be- ~ery-helpful in con~~ructing an._appropr.iate. and coherent
_pec:h1gogic d~_~c;ripiioo of langu~ge. -
On the face of it, the most obvious application of discourse analysis to
foreign language teaching is to help to construct the kind of model dialogue
common to so much language teaching material. Role-playing, drama and
simulated conversations are one established method in foreign language
teaching (e.g. Maley and Duff, 1978). It is therefore plausible that a better
understanding of real dialogue should lead to better dialogues for teaching
purposes. However, it is obvious that a close transcript of a real conversation
is very far from what is normally required for teaching. Any conversation
will contain many characteristics which are relevant only to its original
context of occurrence. The indexicality of everyday conversation has been
the particular study of conversational analysis which derives its theoretical
impetus from ethnomethodology (cf. Atkinson, 1981, for a discussion of
ethnomethodology and applied linguistics, and the argument that situation
and notional syllabuses are both rather crude attempts to apply sociological
ideas.) If real conversations are used as the basis for pedagogic material,
they will have to be carefully adapted to be at the right level of generaliza-
tion. Davies (1978) compares in detail differences between an audio-
recording of a real family breakfast and a foreign language textbook
representation. He discusses the several different kinds of idealization
required if one is to be turned into the other (cf. also Burton, 1980, for a
detailed comparison of real discourse and simulated dialogue in playscripts.)
Another problem is that both teachers and students are, in general,
ignorant of the structure and functions of conversations: discourse has
simply not been studied in the educational system in the way that grammar
has been for hundreds of years. Roulet (1981) therefore argues that foreign
language teaching must be supported by mother tongue teaching, in which
conversational analysis based on authentic documents is taught. Such
suggestions are made within the more general context of suggested rap-
prochements between foreign language and mother tongue teaching which
are currently (in the early 1980s) being debated.
If one is thinking of the direct applications of discourse analysis to
language teaching, then this could mean also several other things. In general
it suggests teaching language as communication (cf. Widdowson, 1978;
Brumfit and Johnson, eds, 1979), and communicative syllabuses rather than
grammatical syllabuses (cf. Munby, 1978). Teaching English for science and
technology (EST) or for other academic purposes (EAP) or more generally
teaching English for special purposes (ESP) imply teaching communicative
competence, since the aim is generally to teach adults a foreign language for
 EDUCATIONAL LINGUISTICS 213

some specific, real, possibly quite restricted purpose, not for the artificial
purpose of passing an examination. Applied linguistics has to recognise that
language is studied for different reasons. Often the study of language is
instrumental: not an end in itself, but a means to an end. The concept of
notional or functional syllabuses is closely related here (Wilkins. 1976; Van
Ek, 1975): that is, the view that the syllabus can be constructed round a list of
speech acts, communicative and semantic categories, rather than the tradi-
tional grammatical organisation of most syllabuses. Whilst a communicative
syllabus would necessarily be partly functional, it has been pointed out
however, that a notional syllabus may be a list of isolated functional
categories, and not take fully into account the sequential organisation of
connected discourse (Widdowson, 1979b). In fact this important criticism
has been levelled against speech act theory itself: that it studies isolated acts,
although often the illocutionary force of an utterance can only be inter-
preted from knowledge of its place in a discourse sequence.
More narrowly still, applied discourse analysis might imply teaching
something which has been neglected in the past, but about which we now
have·information due to recent research. For example, Brazil et al. (1980)
propose teaching discourse intonation. Or one can teach directly other
interactional skills, such as teaching students to interrupt politely. In
general, discourse analysis is beginning to provide information at the level of
contrastive pragmatics. Different speech communities differ in their rules
for tum-taking, expression of politeness, amounts of talking, use of ritual-
istic formulae and the like, and such information is of potential use to the
language learner. Textual conventions similarly vary in different languages:
written Arabic, for example, makes little if any distinction between sen-
tences and paragraphs, and punctuation conventions therefore differ con-
siderably between Arabic and English. Detailed contrastive analyses of
specific speech events have begun to appear: for example, Godard (1977)
compares behaviour on the telephone in France and the USA by analysing
sequential rules for openings. On the other hand, such work clearly has a
long way to go before comprehensive contrastive descriptions are available.
It has frequently been pointed out that much of the work on speech act
theory and conversational maxims is western European in its assumptions.
For example, Ochs Keenan (1976) criticises Grice (1975) on these grounds,
showing that not all of Grice 's conversational maxims hold in Malagasy.
The papers in Sinclair (ed., 1980) provide other views on applied dis-
course analysis and foreign language teaching.
There are however alternatives to these kinds of direct application. One
alternative is to try to convey to teachers a general view of language which
constantly takes into account its use in connected discourse in different
social contexts. This is the topic of the next section.
214 MICHAEL STUBBS

Discourse analysis and training EFL teachers: course description and

discussion
The course material described here was part of an intensive eight-week
course taught in China. 1 The course material filled about90 hours of lectures
and seminars: 8 weeks of 6 days each and roughly one lecture and one seminar
each day. Another one or two hours each day were generally filled with
other related topics in modern English language, language teaching method-
ology, cultural background, films, and so on.
The students were over fifty Chinese lecturers in English at institutes of
higher education, including universities and teacher-training colleges. They
came from all over China, some from high prestige institutes in Peking and
Shanghai, others from small institutes a long distance from main centres.
Their command of English varied from near native speaker competence to
some students who had almost no useful comprehension of normal spoken
English at all. Many had, of course, no opportunity to hear native speakers:
none had e.ver spent any period of time in an English-speaking country; and
some had recently been "turned around" from teaching Russian. The break
in Sino-Soviet relations came in 1960, and from 1966to1972 all universities
in China were closed during the Cultural Revolution, and the study of
foreign languages was stopped. In general, the students' comprehension of
written English was much better, and many of them were in fact teaching
intensive or extensive reading. Their knowledge of descriptive linguistics
was in general restricted to traditional grammar. Quirk et al. 's (1972)
grammar was well known, although in many cases students had clearly not
understood the linguistic principles on which it is based. The students were
also very familiar with the International Phonetic Alphabet and with broad
phonemic representations of words, as this is standardly marked in even
elementary EFL textbooks in China. The 50 students were divided into
three groups on the basis of a cloze passage and a listening comprehension
test, plus subsequent minor adjustments to the groups. The course described
here was given to the "top" group of 16 students. On the oral testing
procedures used by the Foreign Service Institute (Oller, 1979, p. 320) these
students would have been at points 3 or 4 on the 5-point scale:

(3) Able to speak the language with sufficient structural accuracy and
vocabulary to participate effectively in most formal and informal
conversations, on practical, social and professional topics.
(4) Able to use the language fluently and accurately on all levels normally
pertinent to professional needs.

The aim of the course was to teach neither EFL nor TEFL, but to teach
 EDUCATIONAL LINGUISTICS 215
descriptive linguistics with reference to modem English language and to
TEFL. It attempted to provide a coherent approach to describing English
which was particularly appropriate to TEFL, with the underlying theoretical
coherence coming from work in discourse analysis, text analysis, narrative
analysis, cohesion, speech act theory and related areas. In other words, the
course was predominantly theoretical, but was theory explicitly geared to
teaching practice. This would hopefully: (1) improve the student's own
communicative competence in English; and (2) allow them to improve their
own teaching and testing techniques, by (3) teaching them about linguistic
description. Hopefully this would be of both practical value and also be
intellectually interesting. As the course progressed, I realized how import-
ant requirement (2) was. Since we knew next to nothing about actual
teaching conditions in Chinese higher education, we could not impose actual
teaching methods, but only provide the underlying principles in the hope
that our students would then be able to adapt our ideas to their own
circumstances.
There were several constraints on the course, which may seem extreme,
but which doubtless have parallels elsewhere. These must be taken into
account, since there is little point in providing students with impractical
ideas. First, a communicative approach to foreign language teaching re-
quires, to all intents and purposes, native speaker competence in the
teacher. It is worth remembering that the communicative approach was
developed very much through courses in ESP where native English speakers
were developing basically study skills courses for improving reading ability
to handle written academic English (EAP). There are considerable dangers
in having non-native speakers produce texts for teaching purposes. For a
teacher who has less than native speaker competence, the safest method
may well be to base teaching firmly on given texts. Much teaching in China is
very traditional and text-based, for this and other reaons. Many of the ideas
in the course therefore aimed to provide students with ways of manipulating
naturally occurring texts.
At the outset, I had intended a course fairly evenly balanced between
spoken and written discourse. However these various practical constraints
led to a concentration on written texts with some work on listening com-
prehension: the less than native competence of the students; the need to
start from and develop the traditional text-based methods already used by
the students; the lack of books and the need to exploit available texts to the
maximum effect; the difficulty of using native models of spoken language;
and the fact that many of the students were explictly teaching extensive or
intensive reading for EST or EAP. Given the general isolation of the
students (and their students) from native English speakers, there was in any
case no direct motivation for attempting to develop their competence in
216 MICHAEL STUBBS

spoken English. Teaching communicative skills in spoken English w~

therefore not a direct aim, although during the course they heard a lot of
spoken English in lectures, seminars and more informal conversation.
One basic decision was that all materials used should be authentic texts. I
have already admitted above that the search for authenticity is illusory, since
material is taken out of its original context if it is used for teaching.
(Allwright, 1979, proposes one way round this problem, albeit in a very
special situation.) Given the students' need to exploit texts to the maximum
effect in their own teaching, I relaxed the authenticity criterion to allow the
manipulation of texts for teaching purposes. (I give examples below.)
Nevertheless, here authentic texts means material originally produced for
native speakers, and not produced for teaching and designed for learners.
There are several very useful collections of such material from a wide variety
of written styles (e.g. Maley and Duff, eds, 1976; Levine, ed., 1971) and
collections of short stories (e.g. Cochrane, ed., 1969; Dolley, ed., 1968).
This decision leads inevitably to other consequences. It follows that all work
was on analysing and interpreting connected text. It also means that no
linguistic feature or content will ever be introduced for its own sake. This
point applies both to syntax ("Today we learn the passive") and also to
function ("Today we do polite requests"). Texts may be selected because of
some central or recurrent feature, but that feature will always be con-
textualized in other features. The choice of materials also makes it fairly
easy to avoid both linguistic correction and linguistic aid (e.g. supplying
words) if this is thought desirable. All theory and practice was, therefore,
explicitly related to this aim of handling real connected text.
The general model of language underlying the course, and put over more
or less explicitly at different stages, was as follows. Learning a language is
essentially learning to make correct predictions. As soon as something is
said, one can make predictions about what is likely or unlikely to be said
next. Expectations may be broken, predictions may be wrong, and people
say unexpected things. However, unfulfilled predictions show that there
were predictions made, and whatever does occur is interpreted in the light of
what was expected. If something is not expected, this surprisal value is part
of its meaning. These points can be reformulated in terms of redundancy: if
an item is predictable, then this means that it is redundant. Linguists are
fully familiar with this information theory approach to language and mean-
ing. And they are fully familiar with the way in which it is equally applicable
to all levels of language: phonology, graphology, lexis (e.g. collocations),
syntax, semantics and discourse. It is also evident to linguists how this
concept may be reformulated into a concept of structure as constraints on
linear sequence. In one way or another, all modern linguistics is based on the
concept that language is polysystemic: that is, there are always constraints
 EDUCATIONAL LINGUISTICS 217
on linear sequence, and different paradigmatic choices are available at
different points in the sequence.
From a theoretical point of view, these ideas are very powerful and
general, although of course they run into all kinds of difficulties as a
theoretical model, and these are well known. However, what we are con-
cerned with here is their appropriateness as a model for understanding
foreign language learning, and for producing coherent and interesting
language teaching materials. They relate well, for example, to the influential
view of reading as a psycholinguistic guessing game (Smith, 1973), with all its
implications for teaching reading. Leaming a foreign language is seen,
therefore, as acquiring preditive competence. Halliday {1978, p. 200) points
out that there are certain things which are particularly difficult for a speaker
of a foreign language. These include: {1) saying the same thing in different
ways; (2) hesitating and saying nothing much; and (3) predicting what the
other person is going to say. These aspects of linguistic competence are all
closely related; they all have to do with understanding and producing
language in discourse under the constraints of real time. They all also
concern ways of exploiting the redundancy of natural language in use.
The expectancy model is a good one for foreign learners for very practical
reasons. First, when they are listening to spoken language, learners are
often worried when they do not understand every word. They miss a word,
wonder what it was, and miss the next few words. However, native speakers
do not listen to every word: they exploit the redundancy of any piece of
language, make predictions and then check their predictions by sampling. It
can be good for foreign learners' confidence to be made aware of this. And
the principle immediately suggests ways of preparing listening comprehen-
sion materials (cf. Brown, 1978, and below on "helping the listener").
Learners have to listen like a native, as Brown puts it. A comparable point
holds for reading comprehension where it is sometimes difficult to break
learners of the habit of looking up every unfamiliar word in a dictionary.
Again this leads to many obvious reading exercises involving guessing word
meanings from context and the like (e.g. see Clarke and Nation, 1980, for
many suggestions). It is worth also pointing out to students, that in order to
find the meaning of a word in a dictionary, this assumes that part of the
meaning has already been guessed from context. All words are ambiguous in
isolation, and dictionary users have to select the relevant dictionary entry.
Students can therefore gain a more sophisticated theoretical understanding
of both word meaning and of the organisation of dictionaries. The general
aim of the model of listening and reading comprehension is to make students
independent: of dictionaries, teachers and so on. This might be proposed as
the whole aim of education: to make students independent of teachers. This
was certainly an important consideration on the present course, where
218 MICHAEL STUBBS

students were used to very formal teacher-centred classes, and lacked any
confidence in their own ideas.
Taking now one topic on the course in a little more detail, a major
component was classic structural lexical semantics. However, this was
taught as a way of analysing texts, and therefore proposed as one kind of
discourse analysis. At one level it was presented as a way of teaching
directed reading: by forcing students to identify key-words in arguments,
and by identifying hyponyms, antonyms and synonyms, to identify the
outline of the argument. At a more theoretical level, this led immediately to
a discussion of lexical cohesion. This in tum was taught both as a technique
of linguistic description, and also as a further method of intensive reading,
with particular reference to the stylistic analysis of literary texts. This led
further to a redefinition of such lexical relations in terms of relations
between sentences: entailment, paraphrase, contradiction, presupposition,
and so on- and hence to other ways of analysing the organization of texts. It
is also possible to relate structural semantics directly to language teaching
strategies. For example, Blum and Levenston (1978) have proposed that
there are universals of lexical simplification which include the use of
superordinate terms, synonymy, paraphrase and the like, and therefore
principled ways of making do with less words: precisely what a language
learner often has to do. Hudson (1980, pp. 93-4) also puts forward the
suggestion that in hierarchic lexical taxonomies such as sets of terms for
plants or animals, there are maximal information levels. To take a simple
example, a term such as mammal is less useful than animal for most everyday
purposes, and collie will less often be useful than dog.
This aspect of the course involved practical work on intensive reading,
summarizing, note taking, explaining and reformulating. All such activities
involve understanding the semantic structure of texts, both in their local and
global organization. As well as these aspects of lexical cohesion and logico-
linguistic relations, the theory also covered narrative structure (cf. Labov,
1972b) and speech act theory, as well as communicative competence in
general. Such practical activities blur the distinction between EFL and study
skills. This means that language is not being taught in an intellectual
vacuum, but as a tool. It also blurs the distinction between EFL and mother
tongue teaching, since many such analytic activities are also useful with
native speakers.
I have room here to give only a few practical examples of the kind of
classroom activities which were based on such a view of semantic organisa-
tion and discourse predictions.
(1) The technique of cloze passages is well known. Passages are specially
prepared by deleting words; students have to make predictions from
context and complete the gaps. Such exercises are linguistically
 EDUCATIONAL LINGUISTICS 219
principled, but nevertheless involve artificial preparation of texts. A
real alternative which I used was to take a newspaper article in East
African English which contained a large number of Swahili loan
words, incomprehensible out of context to an English speaker. These
loan words provided real lexical gaps for students to translate into
English.
(2) A common situation in which hearers have to predict large parts of a
conversation occurs when they hear one end of a telephone call. It is
usually possible to predict much of what is said at the other end of the
line. It is easy to tape record a telephone call, and to delete one
speaker's contributions from the transcript to form a discourse cloze
passage.
(3) A short story can be divided into sections and fed to students one
section at a time. Their task is to predict what will happen next, and to
write the continuation of the story. Again, this involves some manipu-
lation of a text, but forces students to make explicit their expecta-
tions in a way which is essentially similar to that involved in an
intelligent first reading of a literary text. Any such exercises can
provide material for subsequent more formal analysis of the students'
own predictions. This will inevitably involve comparison between
different students' predictions, and between these predictions and the
original. This will inevitably lead also to an analysis of the gram-
matical and lexical cohesion in the passages, of semantic relations
such as paraphrase and entailment, as well as of the macrostructure of
narratives and other discourse types. (In Stubbs, 1983a, I discuss in
more detail some aspects of the semantic organisation of a literary
text and give other examples of such classroom activities, suitable for
mother tongue teaching in secondary schools.)
It might be argued that the model of language proposed here is not
specifically linguistic, and that the concepts of predictability and redundancy
are applicable to many aspects of psychological activity (e.g. memory) and
social behaviour. Nevertheless they are particularly clear when applied to
language, and the theory has been most explicitly developed with reference
to linguistic examples. Furthermore, as the concern with autonomous
linguistics mellows, it may be useful to start looking for a basis of linguistic
organisation in wider psychological and social competence. It is not entirely
plausible that linguistic competence is as distinct from other cognitive
abilities as some linguists have proposed.

Discourse analysis and analysing classroom language

The view of teaching which has been widely held in the west for centuries is a
220 MICHAEL STUBBS

predominantly verbal one. A teacher does things such as lecturing, explain-

ing, asking questions and telling students to do things. Students have a
largely complementary role of listening, understanding, answering and
basically responsing to the initiative of the teacher. Many people also hold
some version of the view that people learn things by expressing them in their
own words. This is why we distrust students' work if it is copied verbatim
from a book. And it appears reasonable that "talking through" a problem can
often clarify it. A widely held and often taken for granted view of classroom
behaviour is therefore based on some version of teacher-student verbal
dialogue, with a high value placed on the public, explicit, verbal expression
of knowledge. This view of education, with its equation of teaching and
talking, is of course, culture-specific. Not all cultures take it for granted that
the verbal channel is the primary channel for learning, but believe that
learning occurs through silent observation, participation, self-initiated test-
ing, experience, and so on. In addition, the western model of teaching has
often been attacked by educational theorists, but it has proved remarkably
resistant to such criticism. In a culture such as ours which assumes a close
relationship between teaching, learning and talking, an obvious application
of discourse analysis is to analyse the teacher-student classroom dialogue
itself. This is the educational process as it is experienced day by day by most
students. It is important that teachers have systematic ways of analysing
their own daily professional behaviour; and such reflection on the process of
classroom interaction itself is becoming a standard component of teacher-
training courses. For example, micro-teaching is now a common teacher-
training technique. Such training is also becoming increasingly common for
doctors, managers and other professionals.
A very substantial body of work on classroom interation has been
published since the late 1960s. There is no room to review this work here,
and it is in any case fairly well known. The work varies according to how it
draws on linguistic, sociological, anthropological and psychological methods,
but broadly speaking there have been three influential kinds of study of
classroom language. Type 1 could be called i~sightful observation. This
involves detailed study and commentary on recorded lessons. It is valuable
in that it demands close attention to be paid to the details of real language,
but is inevitably limited since it is restricted to impressionistic and selective
commentary. The best known British work is probably that of Barnes et al.
( 1969). There is no doubt that many teachers find Barnes' work very helpful:
it has made them aware of all kinds of things they had never noticed before,
and Barnes is a very sensitive observer. This is precisely one of the problems;
there is no method or guiding principle for those of us who are not as
sensitive and full of insight as Barnes. Such work can be made more
principled and theoretically secure by using fieldwork methods developed in
 EDU CA TI ON AL LINGUISTICS 221
sociology and anthropology, and by relating the observational data to an
explicit theoretical framework such as symbolic interactionism. Type 2
involves the use of coding schemes: that is, sets of categories designed to
code or classify large amounts oflanguage, usually as it happens in real time.
This may be valuable in allowing broad trends to become visible and in
making gross comparisons between different teachers, different school
subjects, even groups of teachers in different countries, and so on.
However, it inevitably means that close attention is no longer paid to the
actual language used. This approach derives from work done in the 1950s by
the American social psychologist Robert Bales. The best known work on
classroom language is by Flanders (1970). Such coding schemes are often
frequently used in micro-teaching when this is used as a teacher-training
technique. Type 3 could simply be called discourse analysis. The aim here is
to describe spoken discourse as a linguistic system in its own right: to
discover what the units of analysis are, and how these units relate into
sequences.
There are various applications of such linguistic analyses of classroom
discourse. Several researchers have studied teacher-pupil interaction to
investigate whether teachers and pupils understand each other. Willes
(1978, 1981, 1983) and Holmes (1983) report on miscommunications be-
tween teachers and pupils in infant classrooms in Britain and New Zealand
respectively. And Malcolm (1979) reports on communicative interference
between teachers and Aboriginal pupils who speak varieties of non-standard
English in Western Australia. He also goes further to propose how such
sociolinguistic study can lead to action research, involving teachers and
Aboriginal classroom aides.
I will not be concerned here further with such direct applications of studies
of classroom discourse, but will discuss whether it is possible to combine an
increased understanding of teacher-pupil interaction with an increased
understanding of language in general. Ideally, a linguistic approach to
classroom discourse would: sensitise teachers to the complex but orderly
nature of classroom dialogue; improve their teaching via this increased
sensitivity; provide them with a firmer theoretical basis for understanding
their own professional behaviour in the classroom; and provide them with a
theoretically interesting account of an important aspect of language. This is
undoubtedly too tall an order to be fulfilled on many teacher-training
courses, and in the next section I will discuss what might be possible with
reference to part of a second in-service training course which I have
prepared.
222 MICHAEL STUBBS

Discourse analysis and classroom language: course description and

discussion
The material which forms the basis of discussion in this section is published as
Stubbs and Robinson {1979). This is part of an Open Univeristy course on
Language Development. Other material on the course covers phonetics and
phonology, lexis, syntax, semantics, communication and context, assessing
children's language and the language curriculum in schools. The course is
designed as a post-experience course for schoolteachers. The course is
reviewed by Carter ( 1980), which is in tum criticized by Czemiewska {1981).
At the beginning of the material on discourse, we express the objectives as
follows {Stubbs and Robinson, 1979, p. 9):

After studying (this part of the course) students should (a) have a broad view of
different approaches to analysing classroom language; (b) have had experience
in the problem of transcribing natural spoken language from classrooms; (c) be
able to describe classroom lessons using one particular system of analysis; (d)
understand some general limitations on all systems for describing language
behaviour; (e) have several ideas for ways of exploring the language of their own
classrooms; (t) have an increased understanding of the study of discourse as a level
of linguistics, as are phonology, syntactic study and semantics; (g) have a set of
criteria to think about the work they read on classroom language; {h) have a way
of talking precisely about their own classroom language, and of studying aspects
of it.

In my discussion here, I will concentrate on the objectives which have to

do with an appropriate approach to analytic techniques and theory and their
relation to practice, which I have commented on above. Note that the
objectives do not hold out the promise of improved teaching, and we add also
the following caveat {Siubbs and Robinson, 1979, p. 13-14):

We are concered at every stage with the details of real language in classrooms,
and for some readers this may lead to an assumption that such work is directly
and obviously relevant to teaching practice. However, while we do see such work
as very relevant, such an assumption needs qualification. There is no reason, for
example, why increased insight into teacher-pupil discourse should in itself lead
to better teaching. It may do or it may not (it depends on the educational
decisions and action that the teacher takes) ... Nor is there any special value in
the analysis of classroom discourse for its own sake, and it would be wrong of us
to promise that there is.

The material summarizes and criticizes different approaches to classroom

language: insightful observation and coding schemes (pp. 16-26); defines
criteria for linguistic descriptions of discourse: descriptive categories should
be finite in number, relatable to data, comprehensive in coverage, and
 EDUCATIONAL LINGUISTICS 223
restricted in their possible sequential combinations (pp. 26-31); discusses
problems of transcription (pp. 32-3); discusses differences between
grammar and discourse (pp. 33-4); discusses the nature of teachers'
questions (pp. 34-9); and then presents a summary of Sinclair and
Coulthard's (1975) analysis of classroom discourse in enough detail to allow
students to analyse their own data (pp. 39-55). This analytic approach is
also the subject of an accompanying television programme by Willes.
One major problem concerns the amount of analytic skill which students
should have at the end of such a course. As I discussed briefly above, this is a
general problem for any course in linguistics. How much should students be
expected to be able to apply the description in the narrow sense of applying
its categories in a replicable way to data? An emphasis on textual analysis
means concentration on one descriptive framework, ignoring or playing
down its limitations, and neglecting alternative descriptions. Whereas a lack
of detailed description may mean vagueness. With reference to the gram-
matical sections of the course, Czerniewska (1981, p. 38), one of the Open
University staff who prepared the course says

Our decision . . . was merely to provide students with an awareness of a

descriptive approach . . .

However, merely to be aware that pupils' language requires to be seen

systematically, for example, is surely not adequate for an in-service course
for teachers. Mere consciousness-raising can lead to a course which is "all
bricks and no foundations" (Carter, 1980 p. 226).
Another problem is how such a course can be evaluated in general.
Czerniewska (1981) accuses Carter (1980) of judging the course by in-
appropriate criteria, those of academic linguistics, and in commenting
(1981, p. 39) on the whole course, says that "the real test is its usefulness to
teachers". In fact, she defines (p. 37) an in-service course in an even
narrower fashion as "one that will lead to improved classroom practice".
The claim that applicability to the classroom teacher in this sense is primary,
is a common type of argument, which I have already discussed. This is only
one test. Others are accuracy, consistency, clarity, interest, academic and
intellectual value. Furthermore, the relation between analysis (e.g. of
discourse) and behaviour (e.g. in classroom interaction) will rarely be
direct.
The main question is one of the practical educational value of analysis. A
major part of the material is concerned with analysing characteristic teacher-
pupil exchanges which have a structure initiation-response-feedback (IRF),
for example:
224 MICHAEL STUBBS

Teacher: Now what can you tell me that all reptiles do, all reptiles do it.
Pupil: Lay their eggs on land.
Teacher: Good, lay their eggs on land, lay their eggs on land.

There is a danger that such descriptions may be taken as prescriptions. They

might either be taken as a model of good, clear teaching, confusing what is a
norm with what ought to be a norm. Or they might be taken as a warning of
what should be avoided as restrictive teaching practice. However, the
question of whether such exchanges are pedagogically good or bad is a
separate question from their analysis. Presumably they are good for some
purposes (e.g. checking factual knowledge), but hopelessly restircted for
others (e.g. discussing literature). What such analyses can do is to provide a
firmer basis for such interpretation and value judgments, by providing a
precise way of talking about recurrent patterns in classroom discourse. This
is analogous to the role of linguistic description in stylistics. Linguistics can
provide more evidence and a firmer basis for a literary interpretation, but
the analysis is not an interpretation.
The fact that people make this kind of interpretative leap is itself inter-
esting, however, and suggests another educational use of discourse analysis.
It is almost impossible to avoid such value considerations in doing such
analyses. Ideological questions are thus opened up by any critical analysis of
institutional discourse, whether between teachers and students, magistrates
and defendants, doctors and patients, and in general between professionals
and their clients. In such areas, discourse analysis can provide ways of
studying the language of social power and control, prestige, status and
deference, manipulation and misunderstanding, and can provide evidence
for discussion of the moral, ethical, social and political questions which
arise. This is a modem study of rhetoric.
Related work which manages a skillful integration of descriptive theory,
methodology and practical educational concerns is by Willes (1978, 1981,
1983). (Willes, 1978, is part of the set reading for the Open University
course.) She has studied the teacher-pupil interaction in reception classes in
British infant schools. Her descriptions are based on Sinclair and Coulthard
(1975), but the descriptive system is not mechanically applied. It is put to
practical exploratory use in an innovative and imaginative way. She uses
Sinclair and Coulthard's description of classroom discourse as a theoretical
statement about communicative competence in classrooms. She argues that
not all classroom discourse fits the IRF pattern which they have identified (a
descriptive point); and she investigates the different rates at which pupils
learn to conform to this pattern (findings based on the description). But she
also discusses the social and educational value of the IRF pattern, and
argues that teachers should try to depart from it (a pedagogical point). She
 EDUCATIONAL LINGUISTICS 225

also uses the descriptive framework to develop other research methods. For
example, some children seldom speak in the classroom, and their com-
municative competence cannot therefore be directly observed. Willes
devised a discourse doze procedure to test their competence: a story about
classrooms with blanks in the teacher-pupil dialogue for the children to fill
in. Such an integration of theory, methods and practice is rare, but provides
a model study in applied discourse analysis. 2

Discourse analysis in classroom practice

Introductory points
It is traditional to consider language under separate headings of reading,
writing, listening and speaking. There is obviously much to recommend
these distinctions, although they are often artificial. For example, there are
processes of linguistic comprehension common to both reading and listen-
ing, despite the often used divisions of reading comprehension and listening
comprehension, and foreign language or mother tongue courses designed to
teach one or the other. In addition, the division into two productive and two
receptive aspects of competence must clearly not be taken to imply that
speaking and writing are active, whereas listening and reading are passive.
Listening and reading comprehension clearly involve active processes of
prediction, for example. Given these caveats, I will use the traditional
distinctions to structure the following sections, although the categories will
sometimes overlap. So far, I have discussed discourse analysis in teacher-
training. The remainder of the chapter discusses the content of syllabuses for
· students, whose aim is either to develop students' communicative com-
petence or to teach them about language and linguistic theory.

Helping the reader: literacy and stylistics

It is now commonplace to assume that the written language used in basal
readers should be adjusted to the spoken language of learner. Otherwise the
beginning reader has to learn a new style or dialect at the same time as
learning to read; and it is assumed best to learn one thing at a time. Thus,
much scorn has been poured on "primerese" of the type Run, Spot, run. See
Spot run, Jane. Such primerese does of course make a plausible attempt to
con trot the language presented to the beginner, but the basis of the selection
may be wrong. Lexical items are chosen for their regular sound-letter
correspondences, shortness, and so on, rather than for their contribution to
meaningful connected prose. More recently, it has been argued that an
226 MICHAEL STUBBS

approximation of reading primers to the spoken language of beginners

should also involve discourse organization. This is important, given the
general failure of attempts to break reading down into discrete subskllls,
such as vocabulary recognition, identifying key ideas, and so on. Both local
and global textual organization appear to contribute to reading comprehen-
sion, and to be inseparable.
A useful summary statement of discourse factors to consider in preparing
literacy materials is by Longacre (1977). He argues that discourse is the
primary unit of linguistic structure, in the sense of being the unit which
people are aware of. Examples include functional units at the level of speech
acts and speech events: story, explanation, request, giving directions, and so
on. If literacy materials are to approximate to the actual usage of spoken
language, the choice of discourse genre is crucial. The general argument is
that materials should be in line with learners' expectancies. This general
approach fits well with Goodman and Smith's model of reading as a psycho-
linguistic guessing game, although Goodman and Smith have not developed
their model at the level of discourse. An active, interpretative search for
meaning takes place at all levels of language. (For more general reviews of
the contribution of sociolinguistics to literacy teaching see Gudschinsky,
1976; Stubbs, 1980) Much discourse oriented work on the preparation of
literacy materials has been carried out by linguists working with the Summer
Institute of Linguistics. As a result, there is much work which is based on
tagmemic theory, or which is predominantly concerned with Bible transla-
tion into exotic languages, or both (e.g. Callow, 1974).
A different academic tradition which has contributed a great deal to the
structural analysis of prose passages is psychological work on the cognitive
processing of discourse. Both the overall macrostructure of narratives,
descriptions, explanations and the like, and also the micropropositional
development of texts from sentence to sentence, are seen as cognitive
schemas which play an important part in the comprehension and production
of texts. Mandler and Johnson (1977) and Van Dijk and Kintsch (1978)
provide useful reviews of this work. And examples of applied educational
studies are provided by Stein and Glenn (1979) who investigate young
children's comprehension of stories, and by Waters (1980) who provides a
case study of a single child's written production over a year. Much of this
work is within the cognitive psychological approach to studying memory,
which is defined as the ability to recall the semantic content of texts. This
derives from the classic work of Bartlett (1932) who showed that remembering
a story is not mere repetition, but an active process of interpretation and
reconstruction based on familiar structures and standard story schemas.
Such work turns out to be very compatible with work which sets out to
specify the discourse structure of academic textbooks and articles. Such
 EDUCATIONAL LINGUISTICS 227
analysis has often studied science texts, both ·because they are highly
structured in some rather obvious ways, and also because of the importance
of such texts to foreign learners of English in EST and EAP courses. The aim
of such work is to identify discourse plans such as: problem-solution;
assertion-justification; exemplification-clarification-c.onclusion. (See Roe,
1977, Hutchins, 1977, on science books; and Tadros, 1980, for similar work
on economics textbooks). Montgomery's (1977) work on the discourse
structure of science lectures also contributes to this approach to EAP.
Academic lecturing is characteristically a mixed mode: partly spontaneous
spoken language, but based on written notes. But Montgomery's work could
also be considered under listening comprehension below.
Such work is of theoretical interest to linguists, since it is concerned with
the semantic organisation of texts. It has obvious applied interests in helping
students to understand academic materials. The basic pedagogical rationale
is clear enough. We try at least to teach students explicitly about the
organization of written language at the level of graphology: although our
teaching may often be inadequate. We do less well at the level of syntax:
since there is as yet inadequate description of the differences in grammar
between spoken and written English. But systematic teaching about semantic
organization (including cohesion and paraphrase) and discourse (e.g. nar-
rative structure) is almost non-existent.
Such work is therefore beginning to have an educational impact. Work is
beginning to be done on the way in which pupils in British schools actually
use textbooks in classrooms, and this involves both observational studies of
school classrooms and also analyses of the textual structure of school text-
books (Lunzer and Gardner, eds, 1978, and subsequent unpublished work
at. the University of Nottingham). In foreign language teaching, it is now
common to have courses designed specifically tp promote reading ability in
relatively well defined academic areas. Typically, for example, students of
chemistry or music might require a reading knowledge of German, students
of art might require a reading knowledge of Italian, or students of history or
law might require French. There are now several ESP textbooks on the
market which aim to develop skills of effective reading in the sense of
understanding textual organization and finding sense relations in texts. (See
Allen and Widdowson, 1974, for a discussion of such material; Widdowson,
1975b, 1979a, for discussion of the underlying discourse theory; and Von
Faber and Heid, eds, 1981, for a description and discussion of several
specific courses.)
It might be thought that such work on written texts is rather far from the
analysis of spoken discourse or conversation, since written texts are not
interactive. However, as Sinclair (1981) has pointed out, a written text could
in principle consist only of strings of propositions with logical connectors.
228 MICHAEL STUBBS

Anything else is interactive, including: predictive structures, discourse

labelling, cross references, and so on. Any such organizational features
serve to present the text interactively, by taking account of the readers'
likely knowledge and reactions at different points in the text. Thus, to take a
rather obvious example, I began this paragraph by writing "It might be
thought that ... ", in order to take into account an objection to my argument
which I predicted at this point from some readers.
As far as an educationally relevant model of language is concerned, it is
also important that this view of reading, as coping with the textual organiza-
tion of books and articles in order to read for meaning, is compatible with
the increasingly accepted view that we learn to read and write by using
language. In real life, discourse always has a reason for being interpreted.
With written discourse, these reasons range from passing a pleasant hour
with a detective novel to retrieving a specific bit of information in a scientific
article or telephone directory. It does not occur to us as fluent adult readers
to confuse such different functions of reading. However, there is increasing
evidence that many children have problems learning to read because they
never understand what reading is for (Stubbs, 1980, p. 98ff.). So, to
emphasize the main theme of this article, such a view of reading contributes
to a coherent overall view of language. It must be admitted that this view is
not yet well defined or very explicit, but it may be explicit and coherent
enough for practical eductional purposes.
At a more analytically sophisticated level again, from the students' own
point of view, stylistics can be regarded as a training in close reading, and
work on discourse has also begun to influence analysis here. Stylistics is
usually taken to mean the linguistic analysis of literary texts: a study of how
literary effects are created and how readers' intuitive reactions to texts can be
explicitly accounted for. In so far as stylistics uses descriptive linguistic
techniques to explain literary effects, it is already applied linguistics. There
are many problems in such a view of stylistics which attempts to take such a
linguistic account of texts as a theoretically adequate account of literary
effects. For example, it is arguable that a systematic linguistic analysis will
necessarily concentrate on superficial features of linguistic form, that it can
only provide comprehensive accounts of s~ort, and possibly minor, texts,
and that any analysis only provides the literary critic with more data, but
does not in itself lead in any rigorous way to an interpretation. These
problems do not directly concern us here. However, as Widdowson (1975a)
argues in detail, stylistics can also be useful a teaching strategy. Techniques
which force students to pay close attention to linguistic features of texts can
provide a way in to understand the organization of complex texts and
therefore a help in the interpretation of literature.
Recently many studies have drawn on work in discourse analysis and
 EDUCATIONAL LINGUISTICS 229
speech act theory. Pratt (1977) proposes a speech act theory of literature.
Searle (1975) discusses the logical status of fictional discourse and what
kinds of speech acts literature is performing. Carter (1979) uses discourse
concepts in an analysis of Auden's poetry. Stubbs (1983a) uses the concepts
of conversational implicature, following Grice (1975) in an interpretation of
a Hemingway story. Stylistics and literary criticism in general have often
been restricted to prose and poetry, and drama has been neglected. It is
dramatic dialogue which provides an obvious area of application of discourse
analysis. Burton (1980) provides a detailed study of short plays by lonesco
and Pinter, and a general discussion of the relationship between theatrical
dialogue and natural conversation, both using discourse analysis to develop
literary theory and also using the insights of dramatists as data for the
description of natural discourse. (See also Short, 1981.) A related body of
work studies narrative structure: this includes work by scholars such as
Propp (1928), Todorov (1969), and Genette (1980). Some of this work is
predominantly literary criticism; other work on oral narrative is more
obviously sociolinguistic (Labov, 1972b). All of this work uses basically
linguistic-structural techniques of description to provide a more explicit
account of the local and global structure of literary texts. Where literary
criticism previously discussed related questions at all, it was often restricted
to rather superficial commentary of, for example, differences between
spoken and written language (e.g. Page, 1973, on the uses of direct speech in
novels).

Helping the writer: written composition

Teachers are often understandably at a loss when require to correct
students' written compositions. First, there is often no clear dividing line
between coherent and incoherent text, as there typically (or at least often) is
between grammatical and ungrammatical sentences. A text in which surface
lexical or grammatical cohesion is faulty may often seem unclear or in bad
style, rather than obviously "wrong". Second, our intuitions about gram-
matical well fonnedness have been sharpened by two thousand years of
explicit syntactic study. Comparably explicit work on discourse organisation
is only now starting to appear. (For a detailed discussion of how far the
concept of well formedness applies to discourse, see Stubbs, 1983b, Chapter
5.)
The tradition of teaching written composition is mainly confined to
schools in Britain, and almost unknown at university level, although there
are occasional courses in report writing for engineers and the like. The
American tradition of teaching rhetoric in the form of freshman composition
tends to be derided (in Britain) as remedial English, and its often vague aims
230 MICHAEL STUBBS

satirised as "a course in existential awareness and the accurate use of the
comma" (Bradbury, 1976, p. 111). This is precisely the kind ofobservation
which has kept linguists away from discourse analysis in the past: the fear
that discourse either involves questions of mere surface style; or that
discourse is impossible to delimit, and that there is no way to prevent
semantics, pragmatics, culture and the world from flooding in. The plot in
several campus novels revolves around the frustrations of teaching freshman
composition in American colleges (e.g. Bradbury, 1965; Lodge, 1975, not to
mention Pirsig, 1974). However, an excellent recent British book on writing
to a directive which could be used on such courses is by Nash (1980). He aims
to strengthen students' intuitions about structures in discourse, and dis-
cusses varieties of rhetorical design and textual cohesion.
There is no doubt that manuals of style such as the famous American
high-school text by Strunk and White ( 1979) discuss important issues of
textual organisation. What such manuals often lack is any systematic
framework within which to discuss such organisation, and often matters of
information structure or cohesion are described vaguely as questions of
emphasis, balance, rhythm, monotony or variety, or simply as good or bad
style. Similarly, work by British educationalists (e.g. Britton et al., 1975) has
usefully discussed the different functions of written language, such as poetic,
expressive and transactional. They have also usefully pointed to the unfair
demands often placed on school pupils who are expected to produce final
draft writing before producing and revising preliminary drafts, arguing for
the value of both exploratory talk and exploratory writing. Again, however,
such work is often inexplicit in its discussion of form-function relations.
Enkvist (1981) provides a useful summary of some rhetorical application
of text linguistics, which is very relevant here, and suggests how the teaching
of both mother tongue and foreign languages could benefit from a more
explicit discussion of text strategies which is now available. In teaching
English as a mother tongue, this could make explicit the ways in which
language is adapted to hearers and readers. In EFL, such work could
provide an explicit basis for explaining the ways in which information may be
concentrated or diluted for different audiences: for example, a discussion of
how much redundancy foreign learners require in a text. One way to
"dilute" a text is to insert existential structures (Enkvist, 1981, p. 199):

There were three books on the table. All fell down.

The three books on the table all fell down.

Such topics have received much explicit discussion in recent work on natural
conversation (e.g. Ochs, 1979, on left-dislocation structures in casual con-
versation versus formal or written language).
 EDUCATIONAL LINGUISTICS 231
At a practical level, Keen (1978) and Gannon and Czerniewska (1980)
demonstrate ways in which teachers can analyse the textual cohesion in
children's writing, and therefore better assess it and correct it.

Helping the listener: listening comprehension

For obvious reasons, listening comprehension has traditionally been a topic
for the foreign language classroom rather than in mother tongue teaching,
although it should become clear that some of the points discussed here are
relevant to both. It is worthwhile also bearing in mind that the strict foreign
language/mother tongue distinction is often much less clear than often
appears at first sight, especially in the increasingly multilingual classrooms in
Britain and the USA, and with the existence of varieties of language such as
creoles which blur the language/dialect distinction. I have also pointed out
above that recent courses on ESP blur the distinction between foreign
language learning and study skills.
Brown (1977, 1978) has discussed the inadequacy of many EFL tests of
listening comprehension, which make demands on learners which are never
made on native speakers. Her main general point is that it is inappropriate to
judge spoken language by criteria only applicable to written language. For
example, a test which requires hearers to extract discrete details of informa-
tion from casual conversational language rather than the overall significance
of the utterance, may be confusing the forms and functions of written and
spoken language. Spontaneous speech is not usually used for transmitting
detailed information. Where it is, it is usually backed up with written or
visual aids, as in much teaching. Or the propositional information occurs in
short bursts, as in giving directions in the street, or giving orders in a shop.
Alternatively, hearers will probably record at least the gist of what is said in
writing, as in giving complex orders to workmen. It is therefore inappropriate
to ask questions about the detailed cognitive content of casual conversa-
tional language. Brown's work is based on an examination of both the
phonological obscurity (including elisions and assimilations) which charac-
terizes most spoken English (Brown, 1977); and also of the differences in
discourse organization between written transactional and spoken inter-
actional language. Adult interactional language is characterized by: slow
tempo; division into short chunks with a lot of pauses; one-place predicates
in which one thing is said about one referent at a time; topic-comment
structures; paratactic structures which rarely make explicit logical relations
between clauses (cf. Ochs, 1979). In general, these features mean that
information is not densely structured. Brown points out that when spoken
language is intended to transmit detailed factual information, then special
232 MICHAEL STUBBS

discourse structures have evolved. For example, in a radio news broadcast, a

typical structure allows information to be repeated three times. Brief
headlines are followed by an expansion of the news items which is followed
in turn by a repetition of the main points.
Crystal and Davy (1975) have published transcripts of unedited audio-
recorded conversations, representing standard, educated colloquial English
usage. They admit, as might be predicted from the arguments which Brown
puts forward, that they are "unclear as to how data of this kind can best be
used in a teaching situation" (p. x). A paradox is whether informal language
can be formally taught or tested. Such conversational English is important as
it is different from the language presented to learners in most textbooks.
Usually students are exposed to formal varieties, although informal con-
versation must provide some kind of baseline for a description of English, if
only because of its massively common occurrence. Crystal and Davy do not
suggest teaching the productive use of such a variety of English; they
propose a policy of exposure to increase the receptive skills of discrimination
and comprehension. Similarly, Brown (1977, p. 156) does not approve of
teaching foreign learners to produce assimilated and elided phonological
forms, but only developing students' listening comprehension of such forms.
One of the main teaching points suggested by such work is that there is
much more variety in English than is often realized. This is often not realized
since spoken language varies much more than written language, but is more
difficult to observe. (cf. Stubbs, 1980: Chapters 5 and 6 for a more detailed
discussion.) This point is relevant to both foreign language and mother
tongue teaching.

Helping the speaker: rhetoric and oracy

Traditionally, rhetoric studies the effect of a text, written or spoken, on its
audience. Classical rhetoric starts from a belief that audiences are open to
persuasion. It holds also that ways of presenting arguments can be taught,
and that the validity of these arguments can be analysed. There is therefore
much debate on questions such as whether eloquence or style of presenta-
tion of an argument can compensate for its faulty logic. The systematic study
of rhetoric and the structure of discourse was founded by scholars such as
Aristotle in his work on narrative and tragedy. Scholars such as the first
century Roman orator Quintillian wrote textbooks on the art of speaking,
discussing the choice of subject matter and the style of delivery appropriate
to different speakers such as politicians, attorneys and preachers. It is
perhaps not too much of an exaggeration to say that little progress was made
between such work and the twentieth century. Indeed, Corbett (1965) uses
the categories of traditional rhetoric to analyse famous public speeches. As
 EDUCATIONAL LINGUISTICS 233
well as work on narrative structure already mentioned, however, several
scholars have recently pointed out ways of developing the traditional con-
cerns of rhetoric in linguistically interesting ways, which draw on con-
temporary work in discourse, semantics and pragmatics. For example,
different approaches are represented by Nystrand (ed., 1983) which has
particular reference to written discourse; Widdowson (1979) which has
particular reference to EFL and EST; and papers by Sperber and Wilson
(e.g. 1983) which have particular reference to semantic and pragmatic
theory.
Rhetoric traditionally has to do also with formal spoken language. Work
by Sophists, two thousand years ago, on the successful pleading of legal
cases is applied discourse analysis of great social relevance. Teaching of
spoken language in the mother tongue often means hints on speech making,
or training in interactional skills such as interviewing. However, formal
spoken language is influenced by written style: spoken legal language
provides an obvious example. Many types are in fact mixed: partly spon-
taneous spoken language but supported by written notes, such as much
lecturing and public speaking. The general topic of the relationship between
spoken and written language is too large to discuss fully here, but the
following points are particularly relevant. Both written and spoken language
show stylistic variation according to the formality of the context of utterance.
However, spoken language varies more in form, between casual and formal,
than written language does. Furthermore, the more formal spoken English
becomes, the closer it moves towards written lexis and syntax. These
generalizations are valid for educated standard English, although not always
for non-standard varieties of spoken English, nor for other languages. It
follows that extending students' functional command of spoken English, by
giving them access to a wider variety of styles, means extending their
competence in the direction of the standard written language.
As linguists have often pointed out, there are paradoxes involved in
correcting or teaching informal language. They would argue that everyone
has competence in the informal conversational varieties of their native
language: this is simply what is meant by being a native speaker of a
language. And I have already mentioned some arguments against teaching
foreign learners productive competence in informal spoken varieties of
language. Much of this section may therefore seem rather negative or to
shade into something else, namely teaching written language. However,
some of the confusions involved are rife in much educational research, and
current work in the forms and functions of written and spoken discourse can
make explicit some of these confusions.
Within the education systems in Britain and the USA, spoken language
has in any case been largely undervalued until recently. Education has
234 MICHAEL STUBBS

usually been based predominantly on written language: indeed education

has often been equated with literacy. These assertions are inevitably broad
and rather crude, but the general point should be clear enough. It is
relatively recently that educationalists such as Barnes and Todd (1977) have
argued for the value of informal small group talk amongst pupils with no
teacher present. Such work usefully draws into question the taken-for-
granted equation between education and formal written language. How-
ever, it may lack both a systematic formal description of the spontaneous
spoken discourse in such teaching situations, and also lack a.very convincing
educational rationale. It is plausible that small group discussions help
children to formulate their ideas, for example, but this is a commonsense
observation, rather than a firmly demonstrated point about the relation of
language and thought.
The term oracy is used to mean the ability in spoken language, either
spoken production or listening comprehension. If we are thinking of teach-
ing or assessing oral production, then this could involve, for example,
getting pupils to tell stories, give explanations or short lectures to the class,
criticise and challenge arguments put forward by other speakers, take part in
or chair small group discussions, and so on. Such language would then be
judged according to its appropriateness to the situation, whether this in-
volves talking in a group of other pupils, or talking to single adult, teacher or
examiner. The term "oracy" is particularly associated with the work
of Wilkinson, which began to appear in 1965, with an influential book
(Wilkinson et al., 1965). {Many other books and articles by Wilkinson and
his colleagues have since appeared: see Stubbs, 1981, for a more detailed
review of part of this work.) The concern with assessing spoken English is,
however, wider than this. In Britain, CSE (Certificate of Secondary Educa-
tion) boards are obliged to set an oral component in their examinations, and
such examinations characteristically attempt to assess such abilities as
fluency, clarity, audibility, liveliness, intelligibility, developing an argument
or sustaining an interesting discussion. Work on oracy is at present influential
amongst British teachers, although it is not entirely satisfactory. The
implication of the term appears to be that oracy is parallel to literacy, but has
been neglected because no term happens to have been available in English.
However, this implied parallelism is very dubious for reasons I have already
discussed. The most general logical problem is that spontaneous behaviour
is not intended to be assessed. In fact, the question arises as to whether such
aspects of a pupil's life should be open to assessment at all: it is in practice
impossible to separate a pupil's personality (e.g. his confidence in the test
situation) from such language ability. There is a consequent danger that it is
the examiner's competence (to elicit effective language) which is assessed,
rather than the pupil's competence (to produce it). There is in any case no
 EDUCATIONAL LINGUISTICS 235
consensus about what constitutes effective speech appropriate to different
purposes. This is because such speech can create and define social situations,
as well as defining an individual's membership of social groups. Further, we
know very little about language development after the age of about five
years, and therefore have very little idea of the conversational competence
to expect of ten- and eleven-year-olds.
At the very least, such attempts to assess children's spoken language
would have to draw on what we do know about children's discourse and on
work on the semantics and pragmatics of natural language. Recent work on
child language has moved away from the "sentence centrism" which charac-
terized Chomskyan work, and studies the acquisition of communicative
competence in social contexts. Bates (1976), Ervin-Tripp and Mitchell-
Keman, eds (1977) and McTear (1981) are representative of what is now a
very extensive literature, although it is largely concerned with the language
of young children of up to about six or seven years. In turn, such approaches
to child discourse have clear implications for educational research into the
differences between linguistic interaction in the home and in school (e.g.
Wells, 1981). In addition, any proposals to test children's oracy would have
to come to terms with problems su~h as: the distinction between sentences,
utterances, propositions and speech acts; overt and covert meanings, in-
cluding concepts such as presupposition, entailment and implicature. The
ability to make such distinctions is precisely what is being tested in the pupils
(in tests proposed, for example, by Wilkinson et al., 1974) and it is only fair
that the testers should be able to make explicit just what distinctions the
pupils are supposed to be able to make.
In summary, there are both educational research studies and also teaching
and assessing techniques which ignore the organization of classroom talk as
sequential discourse (see Stubbs, 1981).

Teaching about discourse

As part of the content of a syllabus, it is possible to teach directly about
different kinds of discourse, such as casual conversation, formal meetings,
diaries, songs, legal contracts, and so on: clearly a very long list of such
discourse types is imaginable. The aim here would be to give a systematic
understanding of discourse organization and structure, the great variety of
discourse types in spoken and written language, and the relation between
discourse types and social contexts. Such teaching may have different
emphases. It may aim to develop students' own communicative competence
by increasing the functional range of their language, productive or recep-
tive. Or it may aim to teach linguistic theory directly {cf. p. 221). With
younger pupils an intuitive approach would be appropriate, moving towards
236 MICHAEL STUBBS

more explicit and theoretical approaches with older students. (cf. Stubbs,
1980, for discussion of how such textual analysis could fit into a more general
English language syllabus; and Tinkel, 1979, for discussion of teaching
linguistics in schools.)
There is one aspect of such teaching that I have so far only briefly
mentioned, however (p. 224). This is the use of such analysis to open up
ideological questions. Recent linguistics has largely ignored the rhetorical,
social and public uses of language which are of central concern to educators,
for example: the language of politics, law and religion; journalism and the
media; technical language; translating and interpreting; and in general the
kinds of socially weighted language used to establish and maintain control in
school classrooms, courtrooms, doctors' surgeries, mental hospitals, or by
"experts" and "science". There are, of course, isolated exceptions to this
neglect (e.g. Bolinger, 1980a, 1980b). However, these areas have largely
been the province of sociologists, literary critics and others. This is un-
fortunate, since linguists could offer a great deal to such topics. As Milroy
argues (this volume), if socially responsible linguists do not do such
analyses, then they will be done, but less well, by others.
A major principle, well studied by linguists, is that people have a very
strong tendency to make sense out of nonsense. This is one way in which
hearers or readers exploit the redundancy of discourse: they assume that
utterances make sense and make predictions about what they think was
meant. This applies at all levels, from typographical errors and slips of the
tongue to the interpretation of political rhetoric, advertising or whatever.
Linguistics has also developed powerful ways of analysing the syntax and
semantics of deceptive language. There are different ways of deceiving
through language, by smuggling in propositions without explicitly stating
them, and this has been a major topic of the current linguistic interest in
semantics, pragmatics and discourse presuppositions. There is very con-
siderable theoretical debate in this area but significant progress has been
made in studying the differences between propositions which are asserted,
or presupposed or entailed by other propositions, or implicated but not
stated in so many words. The theoretical debate centres not so much on the
surface description of such facts, but on how precisely they should be
accounted for within linguistic theory, for example within semantics or in a
pragmatic component. (Kempson, 1977, provides a clear summary of the
basic issues.)
Much of this work therefore gives detailed definitions of what speakers
are committed to in discourse and what they can deny without logical
contradiction. Such analyses are directly applicable to the ways, for example,
in which news is presented in the media. To cite one very brief example, a
recent BBC radio news programme announced:
 EDUCATIONAL LlNGUISTICS 237
Sir Geoffrey Howe explained that the budget measures were necessary, because

The use of the factive verb explain assumes the truth of the following
proposition, in a way that a non-factive such as claim would not. Embedding
propositions in this way can make them more difficult to identify, and more
difficult to challenge. It would therefore be possible to study discussions
between political commentators on radio and politicians or other public
figures. One could study the propositions to which one or both speakers are
committed at a given point in the discourse, whether such propositions have
been asserted or are taken for granted, and which propositions are, con-
versely, under explicit questioning in some way. A detailed study of the
syntax and semantics of factive verbs and related linguistic devices is a
necessary prerequisite for such a study. 3
Semantics has not characteristically been applied to practical issues,
although it is clear that there are many problems in, for example, the
interpretation of legal documents, which are essentially semantic. Linguistics
is, however, beginning to provide the tools which would allow such applica-
tions in a principled way, and which would therefore answer this complaint
from Enoch Powell (1980):

To sit down to write about •the English of politics now' is to be appalled by the
difficulty of finding any objective instruments which would prevent description
from being mere whimsy or subjective guesswork.

See Lerman (1980) for one such extended attempt.

A more adequate account of such features of language in use would
demand a detailed discussion of work such asFoucault's (e.g. 1972) attempts
to define the discourses which constitute such fields as medicine or econ-
omics, and Habermas' (e.g. 1979) work on communicative competence and
universal pragmatics. Such a discussion is well beyond the scope of this
chapter.
However, it should already be clear that recent work in pragmatics has
therefore contributed many concepts which can help to analyse such rhe-
torical strategies. People have many everyday ways of talking about language,
but they do not normally have available ways of talking precisely about such
aspects of meaning. Again, I have had room here only for the briefest
examples of a type which might be developed by a teacher in the classroom.
The basic argument is that language is used for social control, but that the
mechanisms of such control are describable and understandable, and that
some escape is possible. As Bolinger (1980a, p. 387) argues:
238 MICHAEL STUBBS

... people ... are bright enough to learn the language of language -with a bit of
help from linguists who have acquired a sense of their social responsibilities.

Conclusions
There is no well established body of work thai represents the applications of
discourse analysis. Since discourse analysis is itself not a well defined field,
this is hardly surprising. What I have tried to illustrate in this chapter is the
re-emergent interest amongst both linguists and educators in analysing
connected discourse in socially important contexts. And I have argued that
linguistic approaches to discourse are beginning to provide explicit ways of
discussing aspects of language which are very relevant to the educational
process. I have no doubt that current work in discourse is a very rich source
of ideas for educational theory and practice, if it is well selected and
interpreted. Good teachers may justifiably feel that it provides only a
different slant on what they already do. A general problem with much
applied social research is that it tells practitioners, in different words, what
they know already, if only unconsciously. However, making explicit the
principles of good teaching practice is precisely one important aim of applied
discourse analysis. The systematic study of language in use provides many
ideas for teaching, from lesson plans to whole syllabuses. And just as
importantly, it provides a principled and explicit basis for work that is done,
by relating it to a coherent theory. This is what is meant by applied linguistics:
theory which suggests and illuminates good practice.

Notes
I . This course was taught in summer 1980 at the Peking Language Institute (Beijing
Yuyan Xueyuan) under the auspices of the British Council. My colleagues on the
course were Alan Cunningsworth and Cliff Garwood, and I am grateful to them
for many valuable ideas. I have also used similar material in a much reduced
version of the course taught at the University of Sana'a, Yemen Arab Republic, at
Easter l 981. I am grateful to students on both courses for their ideas and reactions.

2. I should also make explicit what is otherwise not clear from my bibliographical
references alone, that many of the studies discussed in this article derive from
work originally done at the University of Birmingham, England, or are develop-
ments of ideas put forward in Sinclair and Coulthard (1975), although in several
cases these studies have moved a long way from their origins. These studies are:
Brazil et al., 1980; Burton, 1980; Carter, 1979; Harris, 1980; Malcolm, 1979;
McTear, 1981; Montgomery, 1977; Roe, 1977; Stubbs, 1983b; Stubbs and
Robinson, 1979; Tadros, 1980; Willes, 1978, 1981, 1983.

3. I am grateful to Andrew Gilling (personal communication) for discussion of such

 EDUCATIONAL LINGUISTICS 239
analysis of argumentative discourse. My comments here summarize some aspects
of his current work in progress.

Acknowledgements
For comments on a previous draft of this article, I am most grateful to
Margaret Berry, Ron Carter and Mike McTear.

References
Adelman, C., ed., (1981). "Uttering, Mutteringu. Grant Mcintyre, London.
Alatis, J. E., ed., (1968). Contrastive Linguistics and its Pedagogical Implications.
Report of the 19th Round Table Meeting on Linguistics and Language Studies.
University of Georgetown Press, Washington.
Alexander, R. (1980). A learning-to-learn perspective on reading in a foreign
language. System 8, 2, 113-19.
Allen, J.P. B. and Widdowson, H. G. (1974). "English in Physical Science". English
in Focus Series. Oxford University Press, London.
Allwright, R. (1979). Language learning through communication practice. In
Brumfit and Johnson, eds, pp. 167-82.
Atkinson, P. (1981). Ethnomethodology and applied linguistics. In Eichheim and
Maley, eds, pp. 64-89.
Barnes, D. and Todd, F. (1977). "Communication and Leaming in Small Groups".
Routledge and Kegan Paul, London.
Barnes, D., Britton, J. and Rosen, H. (1969). "Language, the Learner and the
School". Penguin, Hannondsworth.
Bartlett, F. C. (1932). "Remembering". Cambridge University Press, London.
Bates, E. (1976). "Language and Context: the Acquisition of Pragmatics". Academic
Press, New York.
Blum, S. and Levenston, E. A. (1978). Universals of lexical simplification. Lang.
Learn.28,2,399-415.
Bolinger, D. (1980a). Fire in a wooden stove: on being aware in language. In Michaels
and Ricks, eds, pp. 379-88.
Bolinger, D. (1980b). "Language the Loaded Weapon". Longman, London.
Bradbury, M. ( 1965). "Stepping Westward". Secker and Warburg, London.
Bradbury, M. (1976). "Who Do You Think You Are?" Secker and Warburg,
London.
Brazil, D., Coulthard, M. and Johns, K. (1980). "Discourse Intonation and Language
Teaching". Longman, London.
Britton, J., Burgess, T., Martin, N., McLeod, A. and Rosen, H. (1975). "The
Development of Writing Abilities (11-18)". Macmillan Education, London.
Brown, G. (1977). "Listening to Spoken English". Longman, London.
Brown, G. (1978). Understanding spoken language. TESOL Quart. 12, 3, 271-83.
Brumfit, C. J. and Johnson, K. eds (1979). "The Communicative Approach to
Language Teaching". Oxford University Press, London.
Burton, D. ( 1980). "Dialogue and Discourse". Routledge and Kegan Paul, London.
240 MICHAEL STUBBS

Callow, K. (1974) ... Discourse Considerations in Translating the Word of God".

Zondervan, Grand Rapids, Michigan.
Candlin, C. N., Bruton, C. J., Leather, J. L. and Woods, E. (1977). "Doc.tor-
Patient Communication Skills". Graves Medical Audiovisual Library, Chelms-
ford.
Carter, R. A. (1979). "Towards a Theory of Discourse Stylistics". Unpublished PhD
Thesis, University of Birmingham.
Carter, R. A. (1980). Linguistics, the teacher and language development: a Review
of Open University Course P232: •Language Development'. Educ. Rev. 32, 2,
223-28.
Carter, R. A. (1981). Back to basics: assessment, language and the English teacher.
Times Educational Supplement, 20 February, 1981, 31.
Carter, R. A., ed. (1982). "Linguistics and the Teacher". Routledge and Kegan Paul,
London. ·
Clarke, D. F. and Nation, I. S. P. (1980). Guessing the meaning of words from
context: strategy and techniques. System 8, 3, 211-20.
Cochrane, J. ed. (1969). "The Penguin Book of American Short Stories". Penguin,
Harmondsworth.
Cole, P. and Morgan, J. L. eds (1975). "Syntax and Semantics", Vol. 3, "Speech
Acts". Academic Press, New York.
Corbett, E. P. J. (1965). "Classical Rhetoric for the Modem Student". Oxford
University Press, New York.
Crystal, D. and Davy, D. (1975). "Advanced Conversational English". Longman,
London.
Czerniewska, P. (1981). The teacher, language development and linguistics: a
response to Ronald Carter's Review of the Open University Course PE232,
'Language Development'. Educ. Rev. 31, 1, 37-9.
Darian, S. ( 1979). The role of redundancy in language and language teaching. System
7, 1, 47-59.
Davies, A. (1978). Textbook situations and idealised language. Work in Progress 11,
120-33. Department of Linguistics, University of Edinburgh.
Dolley, C. ed. (1967). "The Penguin Book of English Short Stories". Penguin,
Hannondsworth.
Dressler, W. V. ed. (1978). "Current Trends in Textlinguistics". De Gruyter, Berlin.
Eichheim, H. and Maley, A. eds (1981) ...Fremdsprachenunterricht im Spannungsfeld
zwischen Gesellschaft, Schute und Wissenschaften". Goethe lnstitut, Munich.
Enkvist, N. E. (1981). Some rhetorical aspects of text linguistics. In Eichheim and
Maley, eds, pp. 172-206.
Ervin-Tripp, S. and Mitchell-Kernan, C. eds (1977) ...Child Discourse". Academic
Press, New York.
Feyerabend, P. (1978) ... Science in a Free Society". NLB, London.
Flanders, N. (1970). "Analysing Teacher Behaviour". Addison Wesley, London.
Foucault, M. (1972). "The Archaeology of Knowledge". Tavistock, London.
Gannon, P. and Czerniewska, P. (1980). "Using Linguistics: an Educational Focus".
Edward Arnold, London.
Genette, G. (1980). "Narrative Discourse". Blackwell, London.
Givan, T., ed. (1979). "Syntax and Semantics" vol. 12, "Discourse and Syntax".
Academic Press, New York.
Gleason, H. A. (1968). Contrastive analysis in discourse structure. In Alatis, ed. pp.
39-63.
 EDUCATIONAL LINGUISTICS 241
Godard, D. (1977). Same setting, different norms: phone call beginnings in France
and the United States. Lang. in Soc. 6, 209-19.
Gordon, J. C. B. (1980). A case-study in misinterpretation: a note on some dis-
seminations of Bernstein by educationalists. UEA Papers in Linguistics ( =
University of East Anglia), 12, 45-52
Gorman, T. ed. (1977). "Language and Literacy: Current Issues and Research".
International Institute for Adult Literacy Methods, Tehran.
Grice, H. (1975). Logic and conversation. /nColeandMorgan,eds, 1975,pp. 41-48.
Gudschinsky, S. C. (1976). "Literacy: the Growing Influence of Linguistics".
Mouton. The Hague.
Habermas, J. (1979). "Communication and the Evolution of Society". Heinemann,
London.
Hall, E.T. (1959). "The Silent Language". Doubleday, New York.
Halliday, M. A. K. (1978). "Language as Social Semiotic". Edward Arnold,
London.
Halliday, M. A. K. and Hasan, R. (1976). "Cohesion in English''. Longman,
London.
Hargie, 0., Dickson, D. and Saunders, C. (1981). "Social Skills in Interpersonal
Communication". Croom Helm, London.
Harris, S. (1980). "Language Interaction in Magistrates' Courts". Unpublished PhD
Thesis, University of Nottingham.
Holmes, J. (1983). The structure of teachers' directives: a sociolinguistic analysis. Jn
"Communicative Competence" (J. Richards and R. Schmidt, eds). London,
London.
Hudson, R. A. (1980). "Sociolinguistics". Cambridge University Press, London.
Hutchins, W. J. ( 1977). On the structure of scientific texts. UEA Papers in Linguistics
(=University of East Anglia) 5, 106-28.
Kaplan, R. B. ( 1967). Contrastive rhetoric and the teaching of composition. TESOL
Quart. l, 4, 10-16.
Kaplan, R. B. ed. (1980). "On the Scope of Applied Linguistics". Newbury House,
Rowley, Massachusetts.
Keen, J. ( 1978). "Teaching English: A Linguistic Approach''. Methuen, London.
Kempson, R. M. (1977). "Semantic Theory". Cambridge University Press, London.
Labov, W. (1972a). "Language in the Inner City". University of Pennsylvania Press,
Philadelphia.
Labov, W. (1972b). The transformation of experience in narrative syntax. In Labov,
1972a, pp. 354-96.
Labov, W. and Fanshel, D. (1977). "Therapeutic Discourse". Academic Press, New
York.
Lerman, C. L. (1980). A Sociolinguistic Study of Political Discourse: The Nixon
Whitehouse Conversations. Unpublished PhD Thesis, University of
Cambridge.
Lester, M. ed. (1970). "Readings in Applied Transformational Grammar". Holt,
Rinehart and Winston, New York.
Levine, A. ed. (1971). "Penguin English Reader". Penguin, Harmondsworth.
Lodge, D. (1975). "Changing Places". Secker and Warburg, London.
Longacre, R. E. (1977). Discourse analysis and literacy. In Gorman, ed. pp. 71-88.
Lunzer, E. and Gardner, K. eds (1978). "The Effective Use of Reading".
Heinemann, London.
McTear, M. F. (1981). The Development of Conversation in Pre-School Children.
Unpublished PhD Thesis, Ulster Polytechnic.
242 MICHAEL STUBBS

Malcolm, I. (1979). Classroom Communication and the Aboriginal Child: A Socio-

linguistic Investigation in Western Australian Primary Schools. Unpublished
PhD Thesis, University of Western Australia.
Maley, A. and Duff, A. eds (1976). "Words". Cambridge University Press, London.
Mandler, J. M. and Johnson, N. S. (1977). Remembrance of things parsed: story
structure and recall. Cognitive Psychology 9, 111-51.
Michaels, L. and Ricks, C. eds, (1980). "The State of the Language". University of
California Press, London.
Montgomery, M. (1977) ...The Structure of Lectures". Unpublished MA Thesis,
University of Birmingham.
Munby, J. (1978). "Communicative Syllabus Design". Cambridge University Press,
London.
Nash, W. (1980). "Designs in Prose". Longman, London.
Nystrand, M. ed. (1983). "What Writers Know." Academic Press, New York.
Ochs Keenan, E. (1976). The universality of conversational postulates. Lang. in Soc.
5, 67-80.
Ochs, E. (1979). Planned and unplanned discourse. Jn Giv6n, ed., pp. 51-80.
Oller, J. W. Jr. (1979). "Language Tests at School". Longman, London.
Page, N. (1973). "Speech in the English Novel". Longman, London.
Pirsig, R. {1974). "Zen and the Art of Motorcycle Maintenance". Bodley Head,
London.
Powell, J. E. (1980). The language of politics. Jn Michaels and Ricks, eds, pp.
432-39.
Pratt, M. L. (1977). "Toward a Speech Act Theory of Literary Discourse". Indiana
University Press, Bloomington.
Price-Williams, D. and Sabsay, S. (1979). Communicative competence amongst
severely retarded persons. Semiotica 26, 1/2, 35-63.
Propp, V. (1928). "Morphology of the Folktale". Trans. L. Scott. Indiana University
Press, Bloomington.
Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (1972). "A Grammar of
Contemporary English". Longman, London.
Roe, P. (1977). "The Notion of Difficulty in Scientific Text". Unpublished PhD
Thesis, University of Birmingham.
Rosen, H. (1978). Signing on. The New Review, Feb. 1978. Reprinted in BAAL
Newslett. 7, June 1979, 4-13.
Roulet, E. (1981). L'analyse de conversations authentiques dans une pedagogie
integree de la langue matemelle et des langues secondes. Paper presented to
Colloque de Linguistique Appliquee de Berne, 1981.
Sampson, G. (1980). "Schools of Linguistics". Hutchinson, London.
Searle, J. R. (1975). The logical status of fictional discourse. New Lit. Hist. 6, 2.
Short, M. H. (1981). Discourse analysis and the analysis of drama. Applied Linguistics
2, 2, 180-202.
Sinclair, J. McH. ed. (1980). Applied Discourse Analysis. Thematic Issue of Appl.
Linguistics 1, 3.
Sinclair, J. McH. (1981 ). Planes of discourse in literature. Mimeo, University of
Birmingham.
Sinclair, J. McH. and Coulthard, R. M. {1975). "TowardsanAnalysisofDiscourse".
Oxford University Press, London.
Smith, F. (1973). "Psycholingu~tics and Reading". Holt, Rinehard and Winston,
London.
 EDUCATIONAL LINGUISTICS 243
Sperber, D. and Wilson, D. (1983). Irony and the use-mention distinction. In
"Radical Pragmatics" (P. Cole, ed.). Academic Press, New York.
Stein, N. L. and Glenn, C. G. (1979). An analysis of story comprehension in
elementary school children. In "Discourse Proce~ing (R. Freedle, ed.). Ablex,
Norwood, New Jersey.
Strunk, W. Jr. and White, E. B. (1979). "The Element of Style", 3rd edition.
Macmillan, New York.
Stubbs, M. (1980). "Language and Literacy". Routledge and Kegan Paul, London.
Stubbs, M. (1980). What is English? Modem English language in the curriculum.
English in Australia 51, 3-20. Reprinted in Carter, ed. 1982.
Stubbs, M. (1981). Oracy and educational lingusitics: the quality (of the theory) of
listening. First Lang. 2, 21-30. ·
Stubbs, M. (1983a). Stir until the plot thickens. In "Literary Text and Language
Study" (R. A. Carter and D. Burton, eds), pp. 56--85. Edward Arnold,
London.
Stubbs, M. (1983b). "Discourse Analysis: the Sociolinguistic Analysis of Natural
Language". Blackwell, Oxford.
Stubbs, M. and Robinson, B. (1979). Analysing classroom lanugage. /n "Observing
Classroom Language (M. Stubbs, B. Robinson and S. Twite, eds), Block 5,
PE232. Open University Press, Milton Keynes.
Tadros, A. (1980). Prediction in economics text. ELR Journal ( = English Language
Research) 1, 42-59. Mimeo, University ofBirmingham.
Tinkel, T. (1979). A proposal for the teaching of linguistics at the secondary school
level, MALS Journal(= Midland Association for Linguistic Studies) 4, Spring
1979, 79-100.
Todorov, T. (1969). "Grammaire du Decameron". Mouton, The Hague.
Van Dijk, T. A. (1981). Discourse studies and education. Appl. Linguistics 2, 1,
1-26.
Van Dijk, T. A. and Kintsch, W. (1978). Cognitive psychology and discourse:
recalling and summarising stories. In Dressler, ed., pp. 61-80.
Van Ek, J. A. (1975). "The Threshhold Level". Strasbourg, Council of Europe.
Von Faber, H. and Heid, M. eds (1981). "Lesen in der Fremdsprache". Goethe
lnstitut, Munich.
Waters, H. S. (1980). 'Class news': A single-subject longitudinal study of prose
production and schema formation during childhood. J. Verbal Leaming and
Verbal Behaviour 19, 152-67.
Wells, G. (1981 ). Describing children's linguistic development at home and at school.
In Adelman, ed., pp. 134-62.
Widowson, H. G. (1975a). "Stylistics and the Teaching of Literature". Longman,
London.
Widdowson, H. G. {1975b). "An Applied Linguistic Approach to Discourse
Analysis". Unpublished PhD Thesis, University of Edinburgh.
Widdowson, H. G. (1977). The partiality and relevance of linguistic descriptions. In
Widdowson, 1979a, pp. 234-45
Widdowson, H. G. (1978). "Teaching Language as Communication". Oxford
University Press, London.
Widdowson, H. G. (1979a). "Explorations in Applied Linguistics". Oxford
University Press, London.
Widdowson, H. G. (1979b). Notionalsyllabuses. /n Widdowson, 1979a, pp. 247-50.
Widdowson, H. G. (1980a). Applied linguistics: the pursuit ofrelevance. In Kaplan,
ed., pp. 74-87.
244 MICHAEL STUBBS

Widdowson, H. G. (1980b). Models and fictions. Appl. Linguistics l, 2, 165-70.

Wilkins, D. A. (1976). "Notional Syllabuses". Oxford University Press, London.
Wilkinson, A. M., Davies, A. and Atkinson, D. (1965). Spoken English. Educational
Review Occasional Publications, 2, Supplement to 17, 2.
Wilkinson, A. M., Stratta, L. and Dudley, P. (1974). "The Quality of Listening".
Schools Council Research Studies. Macmillan, London.
Willes, M. (1978). Early lessons learned too well. Supplementary reading to Language
Development. PE232. Open University Press, Milton Keynes.
Willes, M. (1981). Learning to take part in classroom interaction. In "Adult-Child
Conversation" (P. French and M. Maclure, eds). Croom Helm, London.
Willes, M. (1983). "Children into Pupils". Routledge and Kegan Paul, London.
 9
Variation theory and
language learning

Ralph W. Fasold

Introduction
In spite of the use of a special term, "variation theory" should not be thought
of as a new linguistic model in competition with existing ones, but as an
attempt to add a dimension to linguistic theory. It is self-evident that
variation and change are pervasive throughout language. All one has to do is
listen to the various geographical and social dialects of English (or any other
language) or listen carefully to one's own language use in less and more
formal styles to become aware of linguistic variation. And of course anyone
who has tried to read Chaucer in the original or even the King James Version
of the Bible is immediately made aware of language change over time.
Variation theory is the attempt on the part of a number of linguists to
develop linguistic theory in such a way that it can account for variation and
change as an ongoing and observable phenomenon.
The very attempt to do this is somewhat controversial because it has been
the usual practice in this century to think of language as a set of photographs
taken with a still camera arranged on a page in an album (the various styles
and dialect that exist at a given time). Historical linguistics has usually been
thought of rather as a series of these pages arranged in (temporal) order.
The kind of linguist I have just referred to would like to develop a theory that
treats language more like a motion picture film. Most of the linguists
246 RALPH W. FASOLD

working on variation theory, at least in the US, have used a version of

transformational-generative syntax and phonology. This theoretical choice
has been more a matter of convenience than conviction, and not all studies
of linguistic variation have made that choice (cf. Trudgill, 1974).
The attempt to develop a "motion picture" theory oflanguage has forced
linguists to face two crucial issues that have a bearing on language learning.
The first issue is the systematic nature of variation and change. The second,
which in a way follows from the first, has to do with the nature of the
linguistic knowledge or ability of people who know more than one language
or dialect.

Variation and change

The serious investigation of linguistic variation and change that was begun by
Labov (1966, 1969) revealed that the linguistic environment around an
element undergoing variation systematically influenced the frequency with
which each variant could be expected to appear. The term "linguistic
environment" is being used here in a technical sense to mean the kinds of
vowels, consonants and grammatical categories that occur close to the
varying element. For example, it is a fact about English that the voiceless
stops/pt k/ are always aspirated (i.e. pronounced [ph), [tti) and [kh) when
they occur at the beginning of a word. but are unaspirated (pronounced [p I,
It J and Ik J after /s/. The environment (no sound at all in the same word
before the /p/ in pill, /sf before the /p/ in spill) has a categorical effect on
aspiration - categorical in the sense that there is always aspiration in word-
initial position and never after /s/. The careful study of linguistic variation
has revealed that the same kinds of environmental considerations.regularly
influence the frequency of one variant or the other. That is, linguistic
environments have not only an always-never effect but an often-seldom
effect as well.
To see how this might work, consider the rule of contraction in English.
There is no doubt that the rule applies variably. A speaker can equally well
say He's a good man or He is a good man without having violated the rule of
English. The effect of contraction, of course, is to remove the first vowel (in
some cases also the first consonant) in a particular class of words. The
traditional way to express this in linguistic grammars is to invoke the concept
of optionality. Contraction would be formally designated as an optional
process and nothing more included in the theory about it. A grammar writer
might point out in passing that contraction is rather more colloquial and the
full form more formal, but this kind of information would not appear in
theoretical statements.
 VARIATION THEORY AND LANGUAGE LEARNING 247

The close study of contraction carried out by Labov (1969) revealed that,
quite apart from stylistic considerations, particular aspects of the linguistic
environment had a systematic influence. For example, if the word preceding
a contractable word ends with a vowel, contraction is more likely. All else
being equal, a speaker who says Mary is going is more likely to contract is
that a speaker who says Marion is going. The grammatical category of the
following construction also has a predictable effect. If a verb follows, con-
tracton is most likely; if a noun phrase follows, contraction is less likely; and
if the following construction is a locative expression or an adjective, con-
traction is least likely. If one had to bet, one would be well advised to bet that
a speaker who said The play is beginning would contract rather than if they
were to say The play is a musical or The play is at the Folger Theatre. And of
the latter two sentences, the first is a better bet for contraction than the
second.
It is not only possible to identify environmental influences that favour or
disfavour contraction, but to determine that some of them have a greater
effect than others. In the case of contraction, the various environments
would be ordered as in Table 9.1. The percentages are from Labov's New
York City study, but the relative ordering can be expected to be the same in
any variety of English.

Table 9.1 Percent frequency of contraction by weighted environment

(data from Labov, 1969) ·
Percent
Environment Example deleted
Vowel precedes, verb follows Mary is going 86
Vowel precedes, noun phrase follows Mary is a girl 80
Vowel precedes, adjective or locative follows Mary is at home 70
Consonant precedes, verb follows Marion is going 65
Consonant precedes, noun phrase follows Marion is a girl 37
Consonant precedes, adjective or locative follows Marion is at home 25

A set of environments arranged in this way is ordered according to what

Bailey (1973) calls environment weight. The environment that includes the
most and strongest influencing elements is called the heaviest environment;
the one with the fewest and weakest is the lightest. The contraction environ-
ments have been ordered from heaviest to lightest in Table 9.1. This
arrangement makes it clear that the effect of the preceding sound is greater
than the effect of the following construction. It also makes it clear what is
meant by an uoften-s~ldom" effect. Unlike the aspiration case, even the
248 RALPH W. FASOLD

heaviest environment does not always produce contraction. In Labov's data,

full forms were observed 14 per cent of the time in the environment most
favourable to contraction. Similarly, the lightest environment did not
prevent contraction, since speakers contracted once out of every four times
in this environment. Except for the often-seldom rather than always-never
nature of the case, the analysis of contraction parallels the analysis of stop
aspiration. The environments are made up purely of linguistic elements in
both cases. For this reason, a linguistic theory that takes variation into
account will account for both.
The view of language variation I have just presented is related to a theory
of language change through time (Bailey, 1973). Basically, this theory
suggests changes begin in heavy environments and proceed through inter-
mediate environments to the lightest ones. As this process continues, the
newer forms become more frequent in the heavier and earlier environments
until they may exclude the older form entirely. The replacement of the older
forms by newer ones continues until it reaches the lightest environment, at
which time the change is said to have gone to completion. Of course, not aU
variable rules in languages are on their way to completion. There is no
reason to believe that contraction, for example, will go to completion, and
the word is will disappear from the English language! Rules that are variable
but are not on their way to completion are called stagnant rules.
Some of the more recent research on second language acquisition indi-
cates that the same process may be at work when an individual learns
another language. In an excellently conceived and executed study of bi-
lingualism, Gatbonton-Segalowitz (1976) found evidence that a second
language learner acquires features in the same way that new features move
into a language over time. Gatbonton-Segalowitz examined the acquisition
of English by native French speakers in Quebec. She studied in greatest
detail the pronunciation of English /l'J/ and /9/ and word-intial /h/.
Canadian French speakers are inclined to substitute (d} for [cl), (t) for
[9} and to drop initial /h/. For each of these pronunciations, Gatbonton-
Segalowitz found that the English pronunciation would be mastered in some.
environments more readily than in others. She was able to arrange five
environments from heavy to light, and discovered that the speakers who
were learning English fell into a pattern similar in principle to the one for
English contraction in Table 9· 1. The results appear in Table 9· 2. The
differences between the two tables are, firstly, that Table 9· 1 is a summary of
the treatment of contraction by an entire sample of speakers, whereas Table
9·2 displays the results for each speaker individually. Secondly, Table 9· 1
gives percentages of contraction in each environment, whereas Table 9· 2
includes only three levels of variation; (1) indicates the exclusive use of [C)J,
(2) indicates the exclusive use of (d), and (1,2) indicates variation between
 VARIATION THEORY AND LANGUAGE LEARNING 249

(~) and [d J. Thirdly (and trivially), weighted environments from a column in

Table 9· land a row in Table 9·2.
A statistically significant majority of the subjects fit the above pattern. It
would appear that the speakers towards the top of Table 9· 2 have learned
the pronunciation of English [5] better than those towards the bottom of the
table. Gatbonton-Segalowitz obtained independent evidence that this is the
case. A group of native speakers of English were asked to rate the same
subjects subjectively on their ability to speak English well. There was a
statistically significant positive correlation of the rank order of their judg-
ments and the rank order of the speakers as determined by Table 9· 2.

Table 9.2 Variation between [OI and Id I in the speech of Quebec French
speakers learning English (from Gatbonton-Segalowitz, 1976)
Environments
Speakers Heaviest Lightest
I 2 3 4 5
l 1 1 1 1 I
2 1 1 I 1 12
3 1 I I 12 12
4 1 1 12 12 12
5 1 12 12 12 12
6 12 12 12 12 12
7 12 12 12 12 2
8 12 12 12 2 2
9 12 12 2 2 2
10 12 2 2 2 2
11 2 2 2 2 2

The Quebec French speakers are not the only ones who acquire foreign
language pronunciations by the same mechanism that is found in language
variation by native speakers. In a study of Japanese speakers learning
English, Dickerson (1975), and Dickerson and Dickerson (1977), found the
same patterns. One of the English consonants that is difficult for Japanese
speakers is /z/. Dickerson found that the difficulty in pronouncing an English
(zJ depends very much on the phonological environment in which it occurred.
The lightest environment for the production of an acceptable (zJ (i.e the
most difficult one) was immediately preceding /0/, /()/, It/, /di, /cl and /JI.
Slightly less difficult were instances in which the target /z/ occurred immedi-
ately before silence. The next heavier (easier) environment was when lz/
preceded any consonant except the ones that define the lightest environ-
ment, and the heaviest was when lz/ occurred directly before a vowel. The
percentage of acceptable pronunciation for a typical speaker when reading
dialogues appears as Fig. 1. In the figure, "V" symbolizes vowels, "T", the
250 RALPH W. FASOLD

100

90 ·----.""'
·~.
80

~ 70
u

~ 60
50

00'----~~~~~~~~~~~
_v _C _## _T
Environment

Fig. l. Percentage acceptable pronunciation of English /z/ in four environments by a

Japanese speaker. (Adapted from Dickerson, 1975, p. 403.)

consonants that make up the lightest environment, "'C", all other consonants
and ··~ f', silence.
The English /r/ is difficult for learners from many language backgrounds,
including Japanese speakers. Dickerson and Dickerson (1977) present data
that demonstrates that some environment inhibit the English pronunciation
of /r/ more than others, as in the case of /z/. In particular, for prevocalic /r/,
the Dickersons found that following low vowels were the heaviest environ-
ment, mid vowels next and high vowels the lightest. The data for one typical
speaker appear as Fig. 2.
Gabonton-Segalowitz' work suggests that French speakers acquire more

90

-~
70
60
+-
c
50
QI
u
Qi
a.
40 •
30

20

10

OOL--~~~~~~~~~~~~-

_vlo -Vmld

Fig. 2. Percentage of acceptable pronunciations of /r/ in three environments by a

Japanese speaker. (Adapted from Dickerson and Dickerson, 1977, p. 20.)
 VARIATION THEORY AND LANGUAGE LEARNING 251

English-like pronunciations in order by environment in the same manner as

sound change through time. In other words, it seems quite natural to assume
that Table 9·2 could have been produced with data from one speaker over
time as well as with data from several speakers of varying proficiency levels
at one time. The Dickersons were able to provide evidence of this through
longitudinal study of their subjects. Data from the Japanese speakers
learning English were collected three times over nine months. The pattern
of progress was clear: by and large speakers increased the proportions of
English pronunciations of /z/ and /r/ as time went on. At all three test times,
the proportions of acceptable pronunciations was greatest in the heavier
environments and less in the lighter environments.
All of this suggests the possibility of a higher level of sophistication in
teaching the pronunciations of a second language than is usual in language
classrooms. If it is generally the case, as it seems to be, that the difficult
sounds of a second language are easier to pronounce in some phonological
contexts than they are in others, it should be possible to structure the
language teaching programme so that a difficult new sound is introduced and
practised at first in the heaviest environments only. When the learners have
gained proficiency in these contexts, the next harder environments could be
drilled and so on. In the long run, the phenomena discovered by Gatbonton-
Segalowitz and the Dickersons might well lead to really elegant tests of the
mastery of pronunciation proficiency. The learners' speech could be tape-
recorded and key target language sounds tabulated in various phonological
environments. The score would be derived from the frequency of high
quality pronunciations of the sound relative to environment weight. While a
test of this sort is conceivable at present, it would no doubt be impratical to
score such a test as long as a human evaluator has to listen to each sound in
each environment for each speaker. Future advances in automatic speech
recognition technology, however, may evenutally make it possible to score a
test like this using a computer.
The research on second language learning from the variation theory
perspective leads to an emphasis on the continuous nature of language
learning that contrasts with the discrete stage approach often used in this
kind of research. There is a continuous increase in the use of acceptable
pronunciations across environments. The longitudinal studies reported by
the Dickersons shows that there is another continuum in which learners
produce ever higher proportions of target language pronunciations as time
passes. A second look at Fig. 2 shows that there is yet another continuum, in
the environments themselves. Figure 2 suggests, a bit misleadingly, that
acceptable English /r/ is produced ·by Japanese learners least often before
0

high vowels, somewhat more frequently before mid vowels and most often
before low vowels, as if these were three distinct environments. Apparently
252 RALPH W. FASOLD

the real generality is that /r/ is increasingly easy to pronounce the lower the
tongue height of the following vowel. Some reflection on the articulatory
facts of this suggested generality will show how plausible it is.
But there is more. Not only does the frequency of quality pronunciations
increase along an environment continuum (where the environments them-
selves may be continuous in nature) and also through time, but there is a
style continuum as well (Dickerson and Dickerson, 1977, p. 21, cf. Beebe,
1980). As the style becomes more formal, acceptable pronunciations of the
target language sound increase, even when environment is controlled.
Perhaps most surprisingly, it is not only the quantity of acceptable pro-
nunciations that systematically varies, but the quality as well. So far, by
simply presenting percentages of acceptable pronunciations, we have given
the impression that either a learner can pronounce a target language sound,
or not. But as the Dickersons point out, there is a range of increasingly close
approximations of the target language sound. For example, a Japanese
speaker may use any of the following sound segments as an approximation
for English /z/: [t] (a shortened [z]), (dz], [s] or nothing at all (L.
Dickerson,
v
1975, p. 401). In learning /r/, Japanese speakers use [r], [f], [I],
(1] and [OJ (Dickerson and Dickerson, 1977, p. 19). In each case, the variants
that are observed can be understood as a more or less successful attempt by
the speaker to master the foreign sound.
The examples we have used so far all have to do with pronunciation. Does
the same apply to syntax? In an attempt to answer this question, Adamson
(1980) reviewed quantitative data on second language acquisition and
conducted his own experiment on the acquisition of English main verb
negation by native speakers of Spanish. He concluded that the acquisition of
this structure involves the variation between a series of less English-like and
more English-like approximations of main verb negation, as the Dickersons
found in phonology. However, learning syntactic rules was less likely to be
influenced by environmental factors than learning phonology. While there
was some evidence of environmental influence, much of the variation was
better accounted for by combinations of grammar rules, some of them
neither English nor Spanish, to which probabilities could be assigned. These
probabilities were associated with the rule as a whole and not determined by
the linguistic environment. Furthermore, as learning progresses, use of the
more English-like rules increased in likelihood at the expense of the less
English-like rules. Adamson found that much of what he observed was best
analysed using the model developed by the Heidelberg Forschungsprojekt
"Pidgin-Deutsch" (1978), a study of the acquisition of German by "guest-
workers" in Germany from other countries.
According to Dickerson (1975, p. 406-407), all of this implies a new
outlook on the part of language teachers.
 VARIATION THEORY AND LANGUAGE LEARNING 253

The language teacher should not despair when he encounters variability; rather,
he should adopt a point of view which expects variability . ... The teacher is urged
to look at degrees of attainment, not just at a right/wrong dichotomy of English
versus non-English. Even if the student never reaches the target sound, credit
should be given when, over time, he modifies his pronunciation from 'very
wrong' to 'not so wrong' or from 'not so wrong' to 'ahnost right'. In addition,
even though the student may not reach the target 100 per cent of the time, credit
should be given when, over time, he increases the use of variants closer to the
target.

Code-switching

Once linguists decided to construct theories of language that included

change and variation, they were forced to revise the prevalent notion of
what it meant to speak a certain dialect or style. The prevalent notion was
that the object of study in linguistics was ultimately a single style of an
individual speaker. This entity, called an idiolect, could then be collected
with a set of sufficently similar idiolects to make up a dialect. A language
would then be made up of a set of dialects. Strictly speaking, a separate
technical grammar would have to be constructed for each individual idiolect,
which meant that a language would be made up of hundreds of thousands,
perhaps millions of grammars. This was not seen as a very great problem,
sinc.e these myriads of grammars would be very similar to each other,
differing in the occasional minor detail. Nonetheless, sometimes a linguist
would become uncomfortable with this view of language and attempt to
modify linguistic theory accordingly. None of these attempts was very
successful.
The implications of the one grammar for each style and dialect view of
language became particularly clear with the study of Vernacular Black
English (VBE) in the United States. Early studies of VBE assumed a
separate grammar with code-switching (e.g. Dillard, 1970). When a speaker
of VBE was observed using the more exotic features of his dialect, such as
the absence ofthe present tense forms of to be, he was assumed to be using
the grammar of VBE. H he sometimes used the to be forms as called for by
the grammar of standard English, it was assumed either than he had switched
to his standard English grammar or was experiencing interference from one
dialect while using the grammar of the other. The inadeqtiacy of such a
model becomes quite clear in passages such as the following one, spoken by
a preadolescent black youngster in Detroit (Wolfram, 1969, p. 170):

Raymond, I thing he_ thirteen. Oh, and I got another one, he live back over
that way- his name is Robert. I think he_ eleven. And I'm eleven and Lonnie
Joe, he_ twelve and Little Man is fourteen and Richard is twelve.
254 RALPH W. FASOLD

The speaker switches between constructions in which is is absent and

those in which it is present rapidly in the same discourse. It would be difficult
to discover a credible motivation for the claim that he has switched between
standard English and VBE within this discourse. Furthermore, closer
inspection would reveal that copula absence and presence is governed by the
environmental conditions discussed earlier. In particular, the copula is
absent whenever the subject is a pronoun and it is present whenever the
subject is a noun phrase. Although the presence of a pronoun versus a full
noun phrase determines treatment of the copula in this short passage,
investigations of larger amounts of data show that subject type is really an
often-seldom effect, not an always-never one. Similar variable constraints
can readily be found for other features of this dialect and others. The
discovery of variable constraints makes it clear that separate grammars with
code-switching is not the correct way to account for this kind of variation.
Further evidence leading to the same conclusion was presented by Berdan
(1975), who showed that the alternation between non-standard and standard
features of English did not always co-occur in narratives by speakers of
VBE. If speakers are assumed to switch dialects by switching from one
grammar to another like a bilingual speaker switching from English to
French would, it is reasonable to expect that all the features of VBE would
be dropped in favour of the corresponding features of standard English at
about the same time. Berdan's tabulations revealed that this was not the
case; speakers would adopt some features of standard English while main-
taining VBE for others. These latter features would alternate with standard
English features later in the narrative, but some of the earlier standard
features would have been replaced by VBE. There was very little evidence
indeed for a cohesive .. code".
Even if there is doubt about how appropriate the separate grammar
approach is for explaining dialects and styles, it would seem to work much
better as a model for what a bilingual speaker knows. When bilinguals use
one language rather than another, they are .assumed to switch from one
grammar to another. When they speak a second language with a foreign
accent, it is generally assumed that they are using grammar and pronuncia-
tion rules from the grammar of their first language while speaking the second
(interference). A few scholars who have studied bilingualism have suspected
that the notion of separate grammars for separate languages might be too
simple even in some cases of bilingualism. Discussing this issue, Haugen
(1972, p. 317) comes to the following conclusion:

In the world of the bilingual anything is possible, from virtual separation of the
two codes to their virtual coalescence. The reasons for this are clearly rooted in
the possibilities for variable competence in the human brain.
 VARIATION THEORY AND LANGUAGE LEARNING 255

Haugen sees quite clearly that it is not necessary to assume that every
bilingual has two separate and distinct grammars. This is one possibility, but
it is also possible that they have a single grammar which contains elements of
two languages, but which is different from the grammars of each that a
monolingual speaker would have. Degrees of distinctness and coalescence
between these two extremes are also possible.
One scholar who agrees with Haugen is Poplack (1980). Poplack has
studied switching between English and Spanish by people of Puerto Rican
ancestry in New York City. A great deal of the code-switching she observed
consisted of switches back and forth between English more than once in the
same sentence (Poplack, 1980, p. 589):

Why make Carol sentarse atras pa' que (sit in the back so) everybody has to move
pa' que se salga (for her to get out)?

Her study led her to hypothesize that "bilingual speakers might have
expanding grammars of the type depicted in Fig. 9.3." (Poplack, 1980,
p. 615). I have reproduced her figure in slightly modified form as in Fig. 3.
The shaded areas represent the relative amount of overlap as evidenced by
code-switching.

( a ) Inter-sentential switching ( b ) "tag" switching ( c l lntr11 -sentential switching

Fig. 3. Representation of bilingual code-switching grammars. (From Pop lack, 1980,

p. 615.)

A dramatic example of just how extensively grammars of different

languages can be merged is provided by Gumperz and Wilson ( 1971). They
investigated the multilingual case of Kupwar, a village in the state of
Maharashtra, India. The major linguistic and social groups are the two
land-holding groups, the Jains, speakers of Kannada, and the Moslems,
speakers of Urdu; and the landless labourers who speak Marathi. The three
languages have been in contact for hundreds of years without any of them
replacing any of the others. There are cultural and social reasons for such
stable multilingualism. Not only do the three languages exist in the village,
but almost all the local men use two or all three of them. Therefore, not only
the village, but the individual villagers are multilingual. The duration and
intensity of the contact among the three languages have had a profound
effect on their grammars.
The three languages are not particularly closely related historically. Urdu
256 RALPH W. FASOLD

and Marathi are both lndo-Aryan languages, but not especially closely
related members of that family. The degree of relatedness might be com-
pared to the relationship between French and Spanish. Kannada is a
Dravidian language from an entirley different language family with strikingly
different grammatical features. Over the centuries, the three languages as
spoken in Kupwar have developed a converged grammar to the extent that
they all share the same grammar differing only in lexicon and morpho-
phonemics. As Gumperz and Wilson (1971, p. 155) put it: "We may say,
therefore, that the codes used in code-switching situations in Kupwar have a
single syntactic surface structure".
An example of how it works is to be found in the subordinate construction
of indirect statements and questions. In Urdu and in Marathi, indirect
quotations are introduced by the conjunction ki meaning "that", in the
following manner:
SI ki 5.?
He said that I'm going now.
In Kannada, the indirect quotation comes first and the conjunction is a form
of the verb meaning "say", as follows:
s2 conj s,
I'm going now so saying he said
In Kupwar, all three languages use the construction found in standard Urdu
and Marathi. The examples below compare standard Urdu (U), Kupwar
Urdu (KuU), Kupwar Marathi (KuM), and Kupwar Kannada (KuK).
u bol-o ki k~ha g;)y-a tha k~I
KuU bol-o kikhfi g~e te k;,I
KuM sang-a ki kutt gel;) ho ta kal
Kuk kel ri ki y31Ji hog idni ninni
tell that where went yesterday

The sentence means "Tell (me) where you went yesterday." Kupwar
Kannada, although it retains Kannada words, has altered the syntax
radically to fit the coalesced local grammar. The case of Kupwar bilingual-
ism seems strong evidence that Haugen is right and that even whole
language bilingualism need not entail a separate grammar for each language
a bilingual knows.
It may appear that Haugen's insight has simply turned up a new horror for
language teachers. Not only do language teachers have to get across new
syntactic patterns, pronunciations and vocabulary but, it would seem, have
to do their best to prevent students from acquiring a merged grammar that
does not conform to the standard of either the old language or the new one.
There are two reasons why this is not something to worry about. First of all,
 VARIATION THEORY AND LANGUAGE LEARNING 257

it seems that merged grammars are most likely to develop in communities

where two languages are used in everyday life, such as among Puerto Ricans
in New York or the residents of Kupwar. Secondly, and more important, it is
not clear that merged grammars are always and everywhere to be avoided.
Poplack concluded that the most intimate "intra-sentential" switching
represented both the highest degree of merger and the greatest degree of
bilingual acquisition. To see why this is so, it is necessary to understand the
two constraints Poplack found on code-switching: (1) the free morpheme
constraint and (2) the equivalence constraint.
The free morpheme constraint predicts that a switch from one language to
another cannot occur if the last item spoken in one language or the first in the
other is a bound morpheme. For example -iendo is the approximate Spanish
equivalent of English -ing. The free morpheme constraint predicts that a
switch producing an utterance like "eatiendo" will not occur. In fact, Poplack
knows of no attestations of such a switch unless one of the morphemes has
been phonologically integrated into the language of the other (Poplack,
1980, p. 586). The equivalence constraint means that a switch will not occur
within a constituent that does not map onto the structure of the other
language in a straightforward manner. For example, consider the following
sentence in the two languages:

English: He : would bring it : fast

Spanish: (el) : la trajera : ligero
he it would bring fast

By the equivalence constraint, switches can only occur at the points marked
by colons. It is not possible to switch within the phrase meaning "would
bring it" because the mapping between the two languages is not direct. In
English, the pronominal object follows the main verb, in Spanish it precedes
it. The verbal semantics that English handles with the modal auxiliary
"would" are handled in Spanish in the verb morphology. As a result, while it
is possible, for example, for a bilingual to say:

He would bring it ligero

it would not be possible to say either:

*He it trajera ligero

or

*He would bring la ligero

258 RALPH W. FASOLD

Given the equivalence constraint, Poplack's hypothesis is quite credible.

A speaker who is insecure with the syntax of one language or the other will
"play it safe" and switch between sentences most of the time to avoid the risk
of an equivalence constraint violation. At the sentence level, the speaker is
dealing with a constituent large enough so that direct mapping is usually or
always possible. A very skilled bilingual speaker, who is familiar with the
morphology and syntax of both languages, is free to switch within sentences
since his high level of competence in the syntax of both languages will
prevent him from running foul of the equivalence constraint.
Poplack was able to support her hypothesis with quantitative empirical
data. The Puerto Rican subjects whose language use was observed were all
residents of a neighbourhood in New York City. In addition to and in
conjunction with the gathering of tape recorded speech samples, the com-
munity was studied by means of the anthropologist's technique of participant-
observation, much of it by Pedro Pedraza, himself a community member.
Furthermore, the people whose language samples were analysed were asked
to complete a language-attitude questionnaire. One of the questions on the
questionnaire required respondents to self-report themselves as "mainly
Spanish speakers, English speakers, or bilingual." Over half said they were
"mainly Spanish speakers"; the rest claimed to be bilingual. No one claimed
to be mainly an English speaker. Participant observation led the investigators
to conclude that the self-reports were accurate in the main.
Of the 1293 instances of switching by those who reported themselves to be
bilingual, 53 per cent were intrasentential. Of the 542 switches by Spanish-
dominant speakers, only 31 per cent were intrasentential. This different was
significant by a Chi-square test at the .001 level of confidence (Poplack,
1980, p. 609). This result is the more impressive when we examine the sort
of switching the speakers engaged in that was not intrasentential. The
Spanish-dominant speakers used mostly tag-type switches, including inter-
jections, fillers, tags and idiomatic expressions. This kind of switch demands,
if anything, even less proficiency in the second language than inter-sentential
switching, since these elements have a great deal of freedom of occurrence in
the sentence. The bilingual speakers, when their switches were not of the
intra-sentential type, usually switched between sentences.
A danger in an analysis of this sort, where the switch type of Spanish-
dominant speakers versus bilingual speakers is tabulated, is that some other
variable that correlates with language ability might be the real reason for the
results. For example, if the Spanish-dominant speakers were also those who
made frequent trips to Puerto Rico, or had less education than bilinguals,
then it might be these factors that account for the distribution of switch types
and not the language factor directly. A multi-variate statistical technique
could be used to control for this kind of interaction. The usual statistical test
 VARIATION THEORY AND LANGUAGE LEARNING 259

used in such an application would be analysis of variance, but due to the

uneven distribution of switches among groups of speakers representing
various putative independent variables, analysis of variance would have been
unreliable. Instead, Poplack used an adaption of the VARBRUL 2 analytic
procedure developed by David Sankoff and others (e.g. Rousseau and
Sankoff, 1978). This procedure has the capability of providing meaningful
results even with poorly distributed data. The result showed that reported
language ability not only was statistically significant in its association with
switch type independently of other possible factors, but that it was the most
important single factor influencing switch types (Poplack, 1980, pp. 612-613).
We are left with a different impression of code-switching than the usual
one. Rather than being evidence of poor language learning, it may be an
indication of very thorough learning. Does this mean that students in
language classes should be taught to switch continually between their native
language and the one they are learning? The answer is no doubt "no",
largely because rapid and frequent switching is a natural development in
bilingual communities where two languages are used for everday tasks. It is
probably usually the case that students learning a second language in school
are learning it not so much so that they can join in a bilingual community, but
for personal enrichment, to learn to read in a second language, or to aid
them in joining a more or less monolingual community they have immigrated
to. If they do find themselves in a bilingual community where switching is
used, they will acquire the ability soon enough in that context and would not
need to be taught it. On the other hand, it now appears that the ability to
switch frequently and rapidly deserves rather more respect than contempt.

Conclusion
It no longer seems wise to view the variation and interaction of a learner's
old and new languages as a tangled wasteland to be crossed as soon as
possible. On the contrary, on close inspection this kind of variation is found
to be rather orderly, and holds the potential for more accurate, fine-grained
language instruction and testing than we now have. Even frequent switching
between languages within the same discourse by functioning bilinguals now
seems more a measure of success in second linguistic acquisition than of
failure, and, in any case, seems to be required for normal interaction in some
bilingual communities. Basically, language teachers and students of second
language acquisition must learn not to be embarrassed by inter-language
variation, but to approach it with respect and the expectation that the early
and intermediate stages of language learning have their own order and
system. As Bickerton (1981, p. 205) describes a language learner's progress:
260 RALPH W. FASOLD

... as that knowledge [of the target language] increases, he will progressively
reorganize that imperfect system, and the meanings that he attaches to the
morphemes (and, consequently, the ways in which he uses them) will change
accordingly. But those meanings can only be determined within the system he
has at any given time. They cannot be determined from the system of the target
language. Only the acquisition analyst, with his target-oriented dichotomy
between "acquired" and "not acquired", and its artificial, totally arbitrary
criterion points, can make that kind of mistake.

References
Adamson, H. (1980). A Study of Variable Syntactic Rules the Interlanguage of
Spanish-speaking Adults Acquiring English as a Second Language. Unpublished
doctoral dissertation, Georgetown University Press, Washington DC.
Bailey, C.-J. (1973). "Variation and Linguistic Theory". Center for Applied
Linguistics, Washington DC.
Beebe L. (1980). Sociolinguistic variation and style shifting in second-language
acquisition. Language Learning 30, 2, 433-447.
Berdan, R. (1975). The necessity of variable rules. In "Analysing Variation in
Language" (R. Fasold and R. Shuy, eds), pp. 11-26. Georgetown University
Press, Washington DC.
Bickerton, (1981). Discussion in "New Dimensions in Second Language Acquisition
Research" (R. Anderson, ed.), pp. 202-206. Newbury House, Rowley,
Massachusetts.
Dickerson, L. (1975). The learner's interlanguage as a system of variable rules.
TESOL Quart. 9, 401-407.
Dickerson, L. and Dickerson, W. (1977). Interlanguage phonology: Current research
and future directions. In "Interlanguages and Pidgins and Their Relation to
Second Language Pedagogy" (S. P. Corder and E. Ronlet, eds), pp. 18-29.
Librairie Droz, Neufchdtel: Faculte des Lettres and Geneve.
Dillard, J. L. (1972). "Black and English: Its History and Usage in the United
States". Randon House, New York.
Gatbonton-Segalowitz, E. {1976). Systematic Variations in Second Language Speech:
A Sociolinguistic Study. Unpublished doctoral dissertation, McGill University.
Gumperz, J. and Wilson, R. (1971). Convergence and creolization: A case from the
Indo-Aryan/Dravidian border. In "Pidginization and Creolization of Language"
(Dell Hymes, ed.), pp. 151-168. Cambridge University Press, Cambridge, UK.
Haugen, (1972). "The Ecology of Language". Stanford University Press, Stanford.
Heidelberg Forschungsprojekt "Pidgin-Deutsch" (1978). The acquisition of German
syntax by foreign migrant workers. "Linguistic Variation Models and Methods"
(D. Sankoff, ed.), pp. 1-22. Acadmic Press, New York.
Labov, W. (1966). The Social Stratification of English in New York City. Center for
Applied Linguistics, Washington DC.
Labov, W. (1969). Contraction, deletion and inherent variability of the English
copula. Language45 (4), 715-752.
Poplack, S. (1980). Sometimes I'll start a sentence in Spanish y rermino en espano/:
Toward a typology of code-switching. Linguistics 18 (7/8), 581-618.
 VARIATION THEORY AND LANGUAGE LEARNING 261

Rousseau, P. and Sankoff, D. (1978) Advances in variable rule rule methodology.

"Linguistic Variation: Models and Methods" (D. Sankoff, ed.), pp. 57-69.
Academic Press, New York.
Trudgill, P. (1974). The Social Differentiation of English in Norwich. Cambridge
University Press, Cambridge.
Wolfram, W. (1969). A Sociolinguistic Study of Detroit Negro Speech. Center for
Applied Linguistics, Washington DC.
 Index

A Assembly of the Council of Europe, 167

Aboriginal pupils, in Australia, 221 Aspiration, of voiceless stops, in English,
Accent differences, between speakers, 246
57 Assimilation process, the, 173
Accuracy, sub-conditions for, 92, 93, Associated Press, 105-107
108 Attribution process, the, 139
Addition, in news editing, 93, 103-104, Attributions, in news stories, 99, 100,
110 104, 109
Advertizing, sexual stereotypes in, 41, Australian Associated Press, 75-77, 81,
42 88, 100
Africa Australasian languages, 33
education in, 193 Authorship, linguistic analysis for, 74
linguae francae in, 171
news stories from, 106
African languages, 187, 188, 195
see also names of individual languages 8
Afrikaans, 161 Bahasa Malaysia, 170
Alaska, 165, 175, 184, 191, 193 Bantu, 181
Algeria, 161 BBC, the, 90, 91
Alterations, made during the news editing BBC radio news, 236, 237
process, 77-79 Belfast English, 64, 65
America, bilingual projects in, 185 variables in, 15, 16, 61, 65-70
American Anthropological Association, Bernstein's notions of code, 123
36 Bias, news, 90, 91, 105, 11 I
American children, teachers' evaluation Bible,
of, 127-129 King Ja mes Version of, 245
American Heritage School Dictionary, translation of, 226
avoidance of stereotyping by, 39, 40 Bidialectalism, 146, 147
American Indian languages, 162, 165 Bilingual education, 185-188
American Library Association, 36 Bilingualism, 159, 165, 166, 185, 188,
American speakers, 26 248,254,255-259
Amerindian languages, 33 Black English trial, the, 3, 69
Apache, the, 163, 165, 175 Black English Vernacular, 11, 253, 254
Arabic, 161, 216 copula deletion in, 23, 133, 134
Argentina, news stories about, 90, status of, 122, 132, 147, 171
93-95, 102, 103, 105 Brazilian floods, editing of stories about,
Aristotle, 232 100, 101, 105
264 INDEX

Briton, 166, 176 Copy editors, 74, 75

Britain, education in, 233, 234 function of, 77, 78, 81-85, 87-89, 91,
British speakers. 26 92,94-99, 102, 104, 105, 108-110
Copy tasters, 77, 92, ll4
Cornish, 168
c Country of origin, relationship with
Cameroon, 162 accuracy of news story, 105-107
Canada, French immersion programmes Covariation principle, the, 139-140
in, 165, 187 Cross-dialectal communication, 10, 11,
native French speakers in, 248 19-22, 28, 205
Cartesian rationalism, in linguistics, 120 CSE oral examinations, 234
"Catastrophe··, mathematical, theory of, Culture shock, 124
144
Centre for Urban Educational Studies.
41
D
Chaucer, 245
Childcare books, use of pronouns in, 36 Danish, 160, 171, 177, 188
Children's books, sex stereotyping in, 40 Debt incurred, principle of, 4
China, course for English teachers in, De facto national language, 161
214-219 Deficit viewpoint, on language, 122, 123
Chomsky, 9, 19, 121, 211, 235 De jure national language, 161
Class, social, relationship with language, Deletions of information, from news
123-135, 138-141, 143, 144, 149 stories, 82-85
Classroom language, analysis of, 219-235 Demographic factors, affecting vernacu-
Cloze passages, technique of, 218, 219, lar languages, 174, 175
225 Denmark, 171, 173
Cockney, status of, 121, 122 Descriptive linguistics, 214-219
Code-switching, bilingual, 255-259 Descriptive methods, in linguistics,
Cognitive development. of a child, 165, 223-225
166 Developing countries, 106, 162, 163, 181,
Communication 185, 189, 190, 196
cross-cultural, 24 Dialect differences, between speakers,
cross-dialectal, 10, 11, 19 9-12, 14-18, 54
teaching of language as, 212 Dialectal diversity, within the British
Communicative breakdown, 7-11. 15, Isles, 54, 62
21,22,24,26,27,205 Dialects
Communicative competence, 14 local and standard, 179-181
Comprehension, 7-28, 225, 226 standard, model for acquisition of, 135,
listening, 231, 232 140-145
"Confidence-eagerness", 127, 128 status of, 121-123, 125, 131, 139
Consonance, of a new story, with pre- validity of, 121-149
conceptions, 105 see also Vernacular languages
Context, importance of in communica- Dialogue, see Discourse analysis
tion, 8, 13, 14, 22-24, 26 Dictionaries, role of, 182, 183, 217
Continuity, linguistic, between home and Difference position, on language, 122
school, 123 Disadvantaged children, 125-129
Contraction, rule of, in English, 246, 247 Discontinuity, linguistic, between home
Conversations, female and male roles in, and school, 123
44 Discourse analysis, 3
Copy editing, see News editing definition of, 205, 206
 INDEX 265
educational, 203-238 ESP (English for Special Purposes), 212,
teaching about, 235-238 215,231
Dublin children, evaluation of, 128 EST (English for Science and Technology,
Dutch, 175 212,215,227,233
Ethiopian revolution, editing of stories
aboutthe,95-97, 105
Ethnic groups, 24-26, 124, 127, 128, 130,
E 131, 138-141, 143, 149, 170, 171
EAP (Engling for Academic Purposes), see also Ethnicity
212,215,227 Ethnicity, 164, 165, 169-174
East Frisian, 162, 169 "Ethnicity - non-standardness", 127, 128
Economic considerations, among Ethnography of speaking, 3
speakers·of minority languages, 172, Ethnolinguistic identity theory, 135, 136
173 Evaluation, of children on their speech,
Editing 127-129
news, 73-111 Evaluation of speech, sex differences in,
quality of, 105, 107 44,45
rules, 82, 87-89, 114 Exoglossic official languages, 161
Education Expert witnesses, in the Mullan case, 56,
in Africa, 193 57
sociolinguistics and, 119-149
vernacular languages in, 159-197
Educational linguistics, 203-238 F
EFL, 230, 231, 233
training of teachers of, 214-219 Factive verbs, 237
Eire, 164, 169, 170, 181 Falklands war, news coverage of, 90
ELIT (Ethnolinguistic identity theory), Falsification, in news editing, 92-96, 108
135, 136 Father, to, meaning of, 39
"Emsland plan", the, 169 Feminist English Dictionary, The, 40
Endoglossic official languages, 161 Finland, 185, 195
English Canadians, social position of, Finnish, 175, 180
139 Foreign language teaching, see language
English, 159, 161, 169, 172, 175, 176, teaching
178, 187-189,195,211,230-234,236, Free morpheme constraint, the, 257
246-249,252,253,258 French, 256
course for Chinese teachers of, 214-219 for history or law students, 227
relationship between language and sex immersion programmes in, 165, 187
in, 33-46 Quebec and Parisian, 121
see also Black English vernacular, EAP; Speakers of in Quebec, 172, 248
EFL; ESP; Non-standard English; French and English-speaking students,
Standard English attitudes of, 130
Environment weight, 247-253 French Canadians, social position of, 139
Equal Opportunities Commission, The, French language system, the, 14
40,41 Freshman composition, in American
Equivalence constraint, the, 257, 258 colleges,229,230
Error correction, principle of, 4 Frisian, 160, 162, 164, 169, 172-175, 177,
Errors of fact, in news editing, 93 179-184, 186, 188, 189, 191, 193-196
Eskimo children, teaching of in mother orthography, 177
tongue, 165, 175, 184, 186, 191, 194, Function, of different languages in a
195 community, 175, 176
266 INDEX

G ldiolects, 253
Ijo language, the, 180
Gaelic, see Irish; Scots Gaelic Immersion education in Canada 187
Gate_keeping, performance, of copy Immigrants, as linguistic minoriti~s, 163,
editors at, 91, 92, 96, 114 171, 173, 185, 186, 196
quality of, 105 Inaccuracy, resulting from news editing
Generic nouns, of masculine gender 35 80, 81 '
36 ' '
study of, 91-107
German, 177, 192, 227, 252
types of, 92, 93
variations, in, 183 India, multilingualism in 188
see also High German; Low German Individual mobility, strategy of, 136
Germany, West, attitude to Gastarbeiter
Inda-Aryan languages, 256
in, 171 Inferences, conversational, 24-27
Frisian speakers in, 172 rules of, 25, 28
Ghana, 172, 187 Information deletions 82-85
Giant word syndrome, 122 lnitiation-response-fe~dback, 223, 224
Glasgow University Media Group, 90, Inputs
91, 105, 114, 116 accuracy of, 80, 92, 93
Glottal stop, incidence of in Belfast identification of, 78, 79
dialect, 65-68 scope of, 100, 101, 110
God, male personification of, 37, 38 Intensifiers, insertion of, 97, 110
Grammar Internal grammar, 7-10
~ilingual, code-switching, 255-259
"Interg.r?~P" model of second language,
mternal, 7-10 acqu1s1t10n, 135, 140-145
low German, 186 Inter-group situations, 135-141
non-standard, 10-12, 16-24, 27 Inter-individual situations, 136
ofVBE, 253, 254 International Phonetic Alphabet 214
of vernacular languages, 179-181 Interpretation, patterns of 26 '
Irish, 1~, 164, 169, 170, 11i, 174
Grammatic rules, relationship with com-
prehension, 23, 24 lsoko dialect group, the 180
Graphization, 176, 177 Italian, 227 '
Group bonding function, of language
131 '
J
H Japanese speakers, difficulty found by in
Hausa, 161, 190 learning English, 249, 251
Jo~ descriptions, sex-role stereotyping
Hebrew, 168
Hedges, linguistic, 97, 98, 109, 110 m,42
Hibemo-English, 12, 18-22, 27, 28 Jutish, in North Frisia, 160
see also Irish
High German, 160, 162, 164, 174 179 K
186, 188 ' ,
Home-school difference, linguistic, 133 Kalahari dialect, the, 180
House style, 78 Kannada, 255, 256
How long, use of in Hibemo-English 21 Kupwar, multilingualism at, 255, 257
22 ' '
L
I Labov, William, 1-3, 9, 10, 14, 18, 61,
Icelandic, modernization of, 183 121, 133, 134, 145,174,246-~48
Identification of voices, 51-70 "Lames", 133
 INDEX 267
Language Linguistic competence, of children
colonial, 187 receiving vernacular education, 186
international, 160, 162 Linguistic environment, 246-260
legal, 233 Linguistic minorities, 162, 164, 168-173,
minority, 166-196 185, 194
national, 160, 161, 170 Linguistic norm, for vernacular
of news reports, 91 languages,176-183
official, 160, 161 Linguistic relationships, 175
prestige 166 Linguistics
regional, 160, 161 descriptive, 214-219
second, 164,165, 187,188 educational, 201-238
spoken, 233-235 historical, 245
vernacular,159,197 in teacher training, 206-219
written, 232, 235 Linguistics, role of in teacher training,
Language change, 42, 43 206-211,216,236,245,246
Language decline, 178, 179 Listening comprehension, 231, 232
Language deficiency, 122, 123 Literacy materials, preparation of, 226
"Language deficit", 2 Literature, speech act theory of, 229
Language development, in young child- Low German, 160, 162, 164, 174, 179,
ren, 120, 121 186, 188
Open University course on, 222-225 Lexical substitutions, 82, 85, 86, 87
Language function, 175, 176 Linguistic analysis
Language learning, and variation theory, of editing, 73-111
2 of tape-recorded evidence, 60-67
Language planning, 3, 174, 176-183 Linguistic archives, 8, 13
Language revival, 168 Linguistic differences, in male and female
Language shift, in minority languages, speech, 43, 44
164, 166,178,179, 186 Linguistic evidence, status of in cases of
Language teaching, 203-238 voice identification, 52-70
communicative approach to, 215-219 Linguistic hedges, deletion of, 97, 98
courses designed for specific purposes, Linguistic knowledge, relationship with
227 language use, 9
Language use, 9
Language variation, 211
Lapps, 174 M
dialects of, 180 MacBride Report, 106
Larynx, variations in posture of, 58 Main verb negation, 252
Latin America, accuracy of news stories Malawi, 182
from,90,94,95, 100,106 Malaysia, 168, 170, 190, 192
Law, relevance oflinguistics to, 51-53 Mandingo, 161
Leaming, of a foreign language, 216, 217 Manly, definition of, 39
Legislation, on vernacular education, Manx, 166, 168
195 Marathi,255,256
Lewis, 193 Masculine imagery, use of in religion, 37
Lexicon, expansion of the, 182 Memory, limitation of in voice identifica-
Liberia, 162 tion, 58
Linguaefrancae, 161, 162 Men, social attitudes towards, 33-46
African, 161, 171, 176 Mexican-American children, teachers'
Linguistic communities, analysis of, evaluation of, 128
167-176 Merged grammars, 256,257
268 INDEX

Mid-Ulster dialects, 63, 64, 66, 67 Nigeria, variety of languages in, 159,
Minorities, see Linguistic minorities 161, 162,170, 180,189,190
"Minority group reaction", 130 Non-standard dialects, 2, 69
Minority languages Non-standard English
differences in the series in attitudes to, relationship with standard English,
174 10-12,16-19,27,28, 134
preservation of, 166, 167, 171, 174, 175, sex differences in use of, 43, 44
185-196 teacher's attitudes to, 125-129, 130,
Minority schools, 190, 192, 195 134, 145-149,221
Miscommunications, between people Non-verbal children, 145
withdifferentintemalgrammars, 7-11, Normal child, definition of, 121
15,20-28 North Frisia, 160, 168, 169, 172-175,
Misunderstanding, 15, 19 177, 179, 180-184, 186, 188,191, 193,
Modernization, of vernacular langauges, 196
176, 182, 183, 187, 188 see also Frisian
Monolingualism, 159, 178 Norway, 168, 174, 180
Montreal, language evaluation studies Number of speakers, of a language, 174,
in, 130 175
Mother, to, meaning of, 39
Mother tongue teaching, 212, 218, 230-
0
232
"Motion picture theory", of linguistics, Occitan, 162
245-246 Open University course, on language
Mozambique, 162 development, 222-225
Mullan case, the, 53-70 Oracy, 234, 235
Multilingualism, 159, 185, 188, 189 Orthography, of Frisian, 177, 180, 181,
in the classroom, 192, 193, 231 193
Output copy, in congruence with input,
80,92,93, 100, 101
N
Over-assertion, 92, 96-100, 108
Namibia, national languages of, 161 Over-scope, in news editing, 93,
N ation.Jl Association for the Teaching of 100-102, 110
English, 211
National Council for Civil Liberties, 40 p
National Union of Teachers, the, 41
Navajo, 185, 186 Papua New Guinea, variety of languages
Negativity, ofa news story, 104 in, 158
"Neutral" pronouns, 34, 36, 38 Parents, views held by with regard to
News, accuracy of, 90 variations in dialect and accent, 129-
editing of, 74-111 132
News agencies, 75-77, 79, 81, 88, 96-101, Patterned variability in editing, 87-90
105-111 . Patterns of interpretation, 26
News bias, 90, 91, 105 Perfect tense, in Hiberno-English, 21, 27
News media, dissemination of standard Peron, Mrs, new stories about, 93-95,
English by, 11 102' 103' 105
News-readers, 98 Phoneticians, attitude to voice prints,
News reporting, language of, 3 52,53
New Zealand, analysis of news editing Place, expressions of, as a source of
in, 75-100, 106, 107, 109 inaccuracy,95,96, 108
New Zealand Press Association, 81 scope of, 101
 INDEX 269
Plurilingualism, see Multilingualism s
Polylectal grammar, 9
"Powerless' variety of speech, 45 School
Pragmatic intelligence, 13 bilingual, 187-189
Pragmatic locus of miscommunication, centralized, 184
24,25 communication with home, 163, 164
Predictability, concept of, 216-219 feminine nature of, 132, 133
Pre-school children, language develop- minority, 190, 192, 195
ment in, 120-123 role in vernacular education, 183-196
Present perfect tense, the, 94, 95 sociolinguistic aspects of, 123-149
Prestige, associated with use of non-- see also Education
standard English, 131, 132 Schoolbooks,sexismin,38,40,41
"Primerese", 225 Scope, redefinition of, 93, 100-102
Professions, language attitudes of the, Scots Gaelic, 162, 178, 185
51-53 Second language acquisition, 164, 165,
Pronouns, "neutral", 34-36 187, 168
Pronunciation "intergroup" model of, 135, 140-145
foreign language, 249 "Secular linguistics", 1, 2
of English by native, French speakers, Semantic butchery, 90-111
248,249 Semantic condition, for editing rules, 80
Puerto Ricans, code-switching by, 255, Semantic effects of inaccuracies in news
257,258 stories, 104
Pupils, interaction with teachers, 221-223 Semantics, structural, in language teach-
ing, 218
Q Senegal, Woloflanguage in, 161
"Sentence centrism", 235
Quantitative methodology, use of, 61, 62 Sex Distrimination Act, the, 42
Quebec French, 121 Sex,language,and,3,33-46
Quintillian, 232 Sex:ism, in language, 3, 33, 38
Quotation marks, function of, 98, 99 Sexist language, avoidance of, 36-38
Sexual stereotypes, linguistic mainten-
ance of, 34, 38, 42
R
Simple past tense, the, 94, 95
r, pronunciation of in English, 250, 252 Social change, relationship with language
Radio news broadcasts, 75-79, 83-85, change,42
88-90,96,99-101,107,109,113,114 Social class, classification of speech
Reading, by non-standard English according to, 61
speakers, 147 Social class, relationship with language,
Reading, function of, 228 123-135,138-141, 143,144
Recency, in news stories, 104 Social competition, 137
Redundancy, concept of, 216-219, 230, Social creativity, 137
236 Social psychology, of language, 2,
Refocus, in news editing, 93, 102, 103, 119-149
110 Societal axis, the, 181-183
Religious bodies, attitude of to linguistic "Socio-affective filter", 143, 144
sexism, 37, 38 Sociolinguistic distribution, of "youse"
Reuters, 75, 77, 81, 99, 100, 107, 109 in Belfast, 16-18
Rhetoric, study of, 229, 232-235 Socioilinguist methodology, 51-70
Romansch, 166, 192 Sociology, of language, 3
Russian, 214 Somali, 161
270 INDEX

Source, of news stories, 99, 100, 108 Syntax

South, editing quality of news from the, and bilingual code switching, 258
105-107, 111 theories of, 211, 216, 227, 236
South America, see Latin America
Spanish, 122, 255-257 T
Speakers of minority languages, number Tanzania, national language of, 161,
of, 174, 175 175,187, 188 .
Spectograms, see Voiceprints . Tape-recordings, identification of voices
Speech, sex differences in the evaluation on,52-54
of,44,45 Teacher-pupil interaction, 221-223
Speech act theory of literature, 229 Teachers
Speech stereotypes, 126-128 attitudes to dialects, 125
Stagnant rules, 248 discourse analysis for, 219-238
Standard English involvement in vernacular education
feminine nature of, 132, 133 programmes, 189-193
grammar of, 253, 254 speech styles of, 148-149
origin, 181 . status of, 189, 190
relationship with non-standard English, training of, 189-192, 203-219
10-12,16-19,27,28,134 Teaching materials, 193, 194, 214-219
sex differences in the use of, 43, 44 TEFL, descriptive linguistics applied to,
status of, 121-123 214-219
use of perfect tense in, 21 Telephone callers, identification of,
Standardization, of vernacular languages, 55-70
178-182 Telugu, 172
Status, of teachers and of vernacular Temporal axis, the, 178, 179
languages, 190, 191 . Terminological immunization, 13
Stereotyped images of the sexes, mam- Tests, in vernacular education pro-
tenance of by language, 34, 38-42 grammes, 195
Stereoptyes,speech,126-130 Text, definition of, 205, 206
Stories Text-books, for vernacular education,
inaccuracy in, 93 193-195
length of, 82-85 Texts for EFL in China, 214-219
Style, copy editing for, 89, 90 Third,World news, gatekeeping of, 105,
manuals of, 230 106
Style books, 78 114 . Time adverbials, as a source of editing
Sub-editors, see Copy editors errors,94, 101, 104, 108
Subtractive bilingualism, 187 To be, absence of in Vernacular Black
Summer Institute of Linguistics, 226
English, 253 ..
Swahili, 161, 171, 175, 176, 182, 188,219 Tourists, effect of on commumt1es, 168,
Swedish, 175, 195 169
Switzerland, 166, 192 Training, of teachers, 189-193, 203-219
Syllabuses, for language teaching, 212, Turkey, 168
213 Turks, 171
Sylt, 168 169, 189, 193, 194 .
Syntactic and lexical changes, made m u
news editing, 75, 79-87 . .
Syntactic condition, on the apphcatmn Ulster Anglo-Irish dialect area, 64
of editing rules, 80 Ulster dialects, 62-68
Syntactic editing rules, 82, 86, 87 incidence of particular variables in, 64
 INDEX 271

Ulster Scots dialects, 64-68 Voice-prints 52, 53, 62

Unambiguity, in news stories, 104 Voice quality, 58, 59
United States, immigrants in the, 173 Voiceless stops, aspiration of, 246
Upward mobility, 132 Vowels, in Ulster dialects, 65-68
Urhobolanguage,the, 180
Urdu, 255, 256 w
USA, education in, 233
Wales, 169, 170, 195
Welsh, 162, 166, 168-171, 174, 176, 195
v West Frisia, 170, 175, 177, 191, 195
Variability, 253 West Indian creole-speakers, 11
Variables, incidence of, 64, 90 West Indian English, 147
Variants, of particular variables, 61-67 Whenever, in non-standard English,
Variation, toleration of, 182 18-20, 31
Variation theory, 2, 245-260 While, use of in Northern England, 28
"Verbal deprivation", 2, 122 Will/shall distinction, the, 18
Verbs, transitivity of, 108 Wolof language, in Senegal, 161
Vernacular Black English, see Black Womanly, definition of, 39
English Vernacular Women, social attitude towards, 33-46
Vernacular, definition of, 196, 197 Written composition, 229-231
Vernacular education, 166-196
economics for, 194, 195 y
legislation on, 195
materials for, 193, 194 Yorkshire Ripper case, 54
Vernacular languages Youse, distribution of in Belfast, 16-18,
concept of, 159-197 21, 26
graphization, 176, 177 Yup'ik, 191, 195
institutions for, 196
linguistic norm for, 176-183
range of functions, 176 z
standardization of, 178-182 Z, pronunciation of, in English, 249,
see also Black English Vernacular 251-252
Video editing, 84, 114 Zambia, 168, 181, 187, 189, 190, 192
Voice identification, 51- 70