FEMINISM AND LINGUISTIC THEORY

FEMINISM AND
LINGUISTIC THEORY

Deborah Cameron

M
MACMIllAN
 © Deborah Cameron 1985
Softcover reprint of the hardcover 1st edition 1985 978-0-333-37077-3
All rights reserved. No part of this publication
may be reproduced or transmitted, in any form
or by any means, without permission

First published 1985 by

THE MACMILLAN PRESS LTD
London and Basingstoke
Companies and representatives
throughout the world

British Library Cataloguing in Publication Data

Cameron, Deborah
Feminism and linguistic theory
1. Woman and language
I. Title
400 Pl20.W6

ISBN 978-0-333-37078-0 ISBN 978-1-349-17727-1 (eBook)

DOI 10.1007/978-1-349-17727-1
Contents
Preface: On Demystift"catt"on Vll

1. Introduction: Language and Feminism 1

2. Linguistic Theory: Frameworks and Approaches 9
3. The Politics of Variation: Sex Differences in
Language and Linguistics 28
4. False Dichotomies: Grammar and Sexual Polarity 57
5. Making Changes: Can We Decontaminate Sexist
Language? 72
6. Silence, Alienation and Oppression: Feminist Models
of Language (I) 91
7. Feminist Models of Language (II): Semiology and the
Gendered Subject 114
8. Beyond Alienation: An Integrational Approach to
Women and Language 134
9. Conclusion: Feminism and Linguistic Theory:
Problems and Practices 162

Notes 174
Bz·blt"ography 182
Glossary 186
Index 191

v
Preface: On Demystification
As a feminist academic, I am aware of certain obligations and
responsibilities. I am conscious, for instance, that many women
have been denied the privilege of higher education. By this I
certainly don't mean the chance to imbibe 'wisdom' from 'great
minds', nor even the opportunity to develop and discuss ideas,
which feminists do anyway. I mean the right to financial support,
the right to organise your own time, and the right to use the
informational, technical, social and recreational facilities of a
college or university. In our society these are privileges indeed,
and women get less of a share in them than men.
I also realise that many women consider higher education the
very reverse of a privilege. They are only too delighted never to
have sold their souls and brains to the repressive patriarchal
values of academic institutions.
So my responsibilities are these. First, I must be responsive to
the needs and concerns of women outside academic life; and
secondly I must challenge the practices and values that keep
women outside.
Because of these obligations, I have tried to write this book in a
particular way, one that I feel embodies feminist principles. That
in itself is a challenge to the status quo; and the essence of the
challenge lies in my concern to demystify language and
linguistics.
Intellectual mystification occurs when a writer, to put herself
in a position of authority, denies the reader sufficient resources to
understand and dispute what she says. It can be done in a
number of ways.
For example, the writer may leave unexplained and taken for
granted the conceptual framework she is working in, or may
present it as a given rather than something open to question. Or
she may depersonalise herself, hiding behind the spurious
authority of an 'objective commentator' by not making it clear

Vll
Vlll Preface

where she stands, politically and intellectually, in relation to the

ideas she discusses.
In this book, therefore, I have tried to spell out even the most
basic assumptions behind the theories I deal with, and to provide
enough background to suggest how they may be called into
question themselves. I have been at pains to make clear what my
own opinions are, and to present the opinions of others
scrupulously. To do this I have used a lot of quotation - which
allows my subjects to some extent to speak for themselves - and it
is important that the reader scrutinise that quotation carefully.
Another important source of mystification in academic
writings is the language used: indeed, it could be said that
mystification BY language and mystification OF language are the
joint subjects of this book. Writers may prevent readers from
dealing with their ideas as anything more than gibberish, or as
anything less than received truth, by writing in a way that is
incomprehensible. Alternatively, they may be so vague that no
clear line of thought emerges. Then, if they are criticised, it is
easy for them to claim they have been misinterpreted.
In this book I have attempted a relatively simple style. An
important addition to the text is the glossary of linguistic and
other technical terms, which the reader should refer to whenever
necessary for a concise account of what I mean by using various
unfamiliar words.
I have avoided language that conceals the presence of the
writer and the process of writing. The word I appears frequently,
and at many points 1 indicate exactly what argument I am trying
to put forward. The aim here is to give the reader every
opportunity of saying to herself, 'hold on a minute, that doesn't
follow', or 'but what about x?' or 'I can't accept that'. In other
words, the reader is encouraged to be an active maker of her own
ideas in relation to this book, and not simply a passive consumer
of other people's.
I have also avoided offensive and sexist language, replacing it
either with 'neutral' terminology or, more often, with terms that
draw attention to the existence of women. Most sex-indefinite
and generic referents in this book will be she and her. If there are
any men reading who feel uneasy about being excluded, or not
addressed, they may care to consider that women get this feeling
within minutes of opening the vast majority of books, and to
reflect on the effect it has.
 Preface IX

Finally, I acknowledge that I did not write this book unaided:

many groups and individuals contributed to it in different ways.
Some of them participated in discussions of language and sex;
some showed me their work, or shared information and experi-
ences they thought might be useful; some read and commented
on the typescript; some gave me encouragement and support
while I was writing it. One particular group, my students, helped
me by obliging me to concentrate on the basics of linguistic theory
and to work out the best ways- of explaining them.
I would like to thank the following in particular: participants
in the first W A V A W conference workshop on language and
violence; members of Balliol College Women's Group, Pembroke
College Women's Group, Oxford University Women in Politics
seminar and Oxford Rape Crisis Group; Kate Cameron, Tony
Crowley, Liz Frazer, Ian Griffiths, Roy Harris, Caroline Henton,
Rebecca Hiscock, Radhika Holmstrom, Bob Hoyle, Helen
Lawrence, Toril Moi, Peter Miihlhausler, M. Nawaz, Elizabeth
Powell-Jones and Marni Stanley.
D.C.
1 Introduction: Language
and Feminism
We have inherited a
contaminated language ....
Mary Daly
... how can we conceive of a
revolutionary struggle which
does not involve a revolution
in discourse?
Julia Kristeva

The question of language and its political implications has

exercised writers, philosophers and social theorists throughout the
intellectual history of western civilisation. It is noticeable, too,
that the subject has inspired extreme pessimism: from ancient
Greece to Orwell's Nineteen Eighty-Four, speech and writing
have been credited with a malign power to regulate human social
relations in ways we are not aware of, and to disguise abstract
truths in a cloud of misleading rhetoric.
Today's speakers inherit the view that language is a weapon,
used by the powerful to oppress their subordinates. But why
should language, and knowledge about the workings of language,
be a resource for the powerful alone? Why shouldn't the weapons
of reaction be appropriated by the other side?
Mary Daly and Julia Kristeva are among the women who have
argued in their writings for a radical theory oflanguage, not as an
intellectual luxury, but as an essential part of the struggle for
women's liberation. In the last few years it seems their voices have
been heard, and problems of language and linguistic analysis
have entered the arena of feminist debate. Women have begun to
talk about words, and to change them. In this book, then, I am
making a contribution not only to an ancient tradition but also to
feminist work in progress.
But why add another book to the pile? A number of books and

I
2 Feminism and Linguistic Theory

a host of articles are available to feminists already. What do I

hope to contribute? It is true that feminist awareness of language
as an issue has never been greater, and that a good deal has been
published on the subject. Yet in this accumulating literature
there is a diversity of approach and viewpoint which I find
something of a problem. What common ground is there between,
say, the sociolinguist's statistical analyses of sex difference (e.g.
Philip Smith),l the reformists' prescriptions for eliminating
sexism in everyday usage (e.g. Casey Miller and Kate Swift),2 and
the radical call for a revolution in language which will liberate us
all? Are they addressing the same question? What questions
should they be addressing? If their conclusions differ, are they all
equally valid? What, in short, is the state of the art? This is what I
set out to assess.
I have other reasons for writing. First, to supplement the
available material by bringing the different approaches I have
mentioned together in a spirit of critical examination. My aim is
to explain ideas about language clearly, and with enough
background theoretical discussion for readers to decide for
themselves what is correct and what is open to question.
My second motive is that I have never encountered opinions
similar to mine in feminist writings on language. As I am an
active feminist who also happens to be a linguist, it has been
exciting for me to see the growth of interest in language that has
recently occurred in the women's movement: but I am sorry the
movement seems to be adopting an orthodoxy on the subject
which is rarely challenged, and with which I disagree.
Finally, I want to attempt in this book a critique of academic
linguistics. This is not just a matter of pointing out sexist
assumptions and practices, important though that may be: more
radically, it involves questioning the whole scientific/objective
basis of linguistics, and showing how the practices of linguists are
implicated in patriarchal ideology and oppression.
Before I explore any of these points further, however, I want to
consider why feminists should take an interest in language at all,
and to talk about the forms their interest has taken in the past. It
is often said that the most distinctively human quality we possess
is the ability to communicate with each other by means of
languages, and that linguistic communication is crucial to the
organisation of human societies. So people with an interest in the
workings of any society must also concern themselves with its
language - how it is structured and used, what its users believe
 Introduction 3

about it and so on. These are, broadly speaking, the questions

linguistic theory is supposed to deal with.
Feminists are deeply interested in the workings of their
societies, since in order to fight their oppression they must first
understand it. Much feminist effort is directed, therefore, to
reanalysing society as a patriarchy, a system in which men have
power over women. Language is part of patriarchy. If it plays a
crucial part in social organisation it is instrumental in
maintaining male power, and feminists must study its workings
carefully.
So it is not entirely astonishing that the last few years have seen
a noticeable upsurge of interest in language, affecting not only
academics but women throughout the women's movement. You
can see this upsurge, for instance, in the fact that many women's
conferences on subjects ranging from violence to trades unionism
now include a workshop or a paper on language. A few years ago,
this would have seemed a very esoteric subject for grass-roots
feminism: yet at the first WAVAW conference in 1981, I was
struck by the intensity with which language was discussed. The
women present had evidently given it a lot of thought, and were
emphatic on the need to include language in their political
analyses. They were very aware of their own usage, and many
were making conscious efforts to change it.
In the last two years, too, I have heard of women and language
groups being set up and producing material; noted the
proliferation of sections on language in women's studies texts;
watched the subject of language and sex appear in lecture
courses, exam questions and research proposals at Oxford
University; and read with interest a number of new books on
language, produced by feminists for a feminist readership. The
best known of these, Dale Spender'S Man Made Language,
received enough media coverage to put the subject of women and
language on the map for many people outside the women's
movement.
All this indicates that in language, feminists have pinpointed a
major theoretical and practical issue. But how and why has it
emerged? Although it has recently taken new forms and acquired
a new importance, interest in language among feminists is not a
new phenomenon. Language used to be discussed, in the early
years of 'second wave' feminist activity, under the general
heading of 'representation'.
This term embraced not only the language used about women,
4 Feminism and Linguistic Theory

but also the way they were in general depicted by the mass media.
Particular attention was paid to genres thought especially likely to
inculcate or reinforce sexism, such as school textbooks, children's
fiction and advertising. On the simple assumption that people's
attitudes will be affected by something they see repeatedly,
feminists exposed, and tried to eliminate, negative and stereo-
typical portrayals of women.
Where language was concerned, the representation lobby
a.dvocated 'non-sexist writing'. A usage was in need of reform if it
was either blatantly offensive (,Blonde in fatal car-crash'; 'Bitches
wear furs') or androcentric, implying that the norm of humanity
is male ('Man', 'Mankind', 'Man in the street', etc.) Reform
consisted of recasting expressions to make them neutral (so
'mankind' becomes 'humanity', 'man in the street' becomes
'average person' and so on). This was supposed magically to bring
women into our mental landscapes. As early as 1973 a major
publisher, McGraw-Hill, circulated its authors with guidelines on
non-sexist writing.
Recently, however, a new kind of conern with language has
become more common among feminists, and the new theorists
warn that non-sexist language is an illusion. They stress that
language is pervaded by sexism, and that women are alienated
from it because it is controlled by men.
Some linguistic scholars would like to dismiss this development
as mere academic fashion. Weare currently in a period of great
concern among progressive intellectuals about language. These
thinkers, working in areas like literary and film criticism,
psychoanalysis and cultural theory, are often called 'structuralists'
or 'semiologists' because of their debt to the linguist Saussure.
They centre their analyses on language, believing that human
culture is characterised by the creation of languages (discourses)
which structure our view of the world.
Feminism is an important current in this sort of work.
However, it is only recently that British feminists outside the
academic world have become interested in it. Although Dale
Spender, for example, has read semiological writers like Luce
Irigaray and Cora Kaplan, both she and her audience are more at
home in the Anglo-American tradition of linguistics and
anthropology. The resurgence of interest we are discussing may
have brought semiology to the notice of a wider audience, but it is
not the consequence of the current vogue for structuralism.
 Introduction 5

Nor can we attribute it simply to the writings of Dale Spender

and her colleagues (many of whom wrote in comparative
obscurity for some time before her). The publicity fanfare with
which Man Made Language was greeted showed this was an idea
whose time had come. Suddenly, it seemed, feminists wanted to
think and read about language. Why? It could be argued that
forces within feminism itself make women aware of language and
its problems. This is because the use of language is at the heart of
feminist political practice.
Nowhere is this centrality clearer than in the practice of
consciousness-raising (CR), where women uncover the roots and
the precise nature of their oppression by talking to each other
about their experience. Often, women who have spent time in CR
groups emphasise how liberating it is to be able to put into words
experiences which had seemed nebulous, so private and personal
as to be unmentionable; and to find these experiences are
understood and shared by other women.
Many feminist writers have referred to this communication of
shared experience as naming, but in fact a name, a label, is not
always necessary. Sometimes one is coined - sexism, for instance,
which handily encapsulates a whole area of women's experience -
but at other times it is enough to define an experience by
describing it and getting others to acknowledge it. There is still no
name for Betty Friedan's 'problem without a name', but we all
now know what it is. 3 Language has dispelled its air of non-
existence not by naming it but by communicating it.
The need to communicate, to bridge the gap between women,
is a constant theme of feminist writing, reflected very often in the
titles of books and poems (Silences; Lies, Secrets and Silence;
Dream of a Common Language; Finding a Voice; Unlearning To
Not SPeak). Silence is a symbol of oppression, while liberation is
speaking out, making contact. The contact is what matters; a
woman who lies or who is silent may not lack a language, but she
does not communicate.
Women struggling to reinterpret their experience have noticed
again and again that language itself does not guarantee
communication, and many feel actually inhibited by the
inadequacy of words. A woman writes:

Sometimes when I am talking to people I really feel at a loss for

words. I have this idea in my head and a feeling I want to
6 Feminism and Linguistic Theory

express, and I just can't get it out. I have felt like this for years
and I have never been able to understand why .... A vast
number of the words I use all the time to describe my
experience are not really describing it at all. 4

What this woman is describing has been called women's

'alienation' from language. It is an uneasy feeling that your words
are not yours at all - they have been somehow co-opted, or taken
away and turned against you. The feminist view of language is
reminiscent of the feminist view of sexuality: it is a powerful
resource which the oppressor has appropriated, giving back only
the shadow which women need to function in patriarchal society.
From this point of view, reclaiming women's language is indeed
crucial for women's liberation.
Some writers try to reclaim language by making its inad-
equacies visible in wordplay. Many semiologists favour this
approach, aDd it is used by Mary Daly in her classic text
Gyn/Ecology. Daly wants us to look behind the taken-for-granted
senses of words, so she restructures them; re-member, crone-
logical, the/rapist, a-maze. 5 But while this strategy exposes the
deceptiveness of words by confronting the reader with the extent
to which she normally fails to question them, it does not explain
why language presents this problem, nor suggest any solution.
This is one more incentive for women to look towards a
comprehensive feminist theory of language.
Impelled, then, not only by outside influences but more
importantly by feminist practice itself, feminists have begun to
develop a linguistic theory. They have started to explore the
disciplines which look useful, particularly sociolinguistics (the
study of language and society, how social differences interact with
linguistic ones and what the role of language in culture is) and
semiology, which among other things tries to explain how
children develop 'gendered subjectivity' (i.e. a sense of themselves
as men or women in their society) in terms of a sex-differentiated
relation to language.
Three areas of investigation have been identified by
proponents of feminist language theory. First, there is the study
of sex-difference; do men and women use language differently,
and if so, what does this mean? Secondly, there is sexism in
language, its effects, and how to eliminate it. And thirdly, there
 Introduction 7

is alienation: is this 'the oppressor's language', within which we

cannot articulate our experience as women?
In all of these areas, work has been done: in all of them, much
remains to be done. One of the most important outstanding
problems is to clarify our underlying theory of the relation
between language and world view. Like a wolf-whistle, a sexist
remark has a significance above and beyond the immediate
offence it gives: it is the outward manifestation of an unaccept-
able misogyny. But is it also, as many feminists believe, the very
mechanism by which misogyny is constructed and transmitted?
Can we think outside the confines of a woman-hating language?
It is here, and at similar points where general theoretical
matters are at issue, that feminists have much to gain by turning
to linguistic theory, provided they understand, and have some
way of assessing, conflicting views. An additional hazard they face
is the sexism which affects linguistics as it does other academic
disciplines; again, it is necessary to understand the nature of
argument, methodology, etc. in linguistics before you can tell
how much of it is pure moonshine.
Many women feel incapable of making this sort of judgement. I
have often heard feminists complain that the theories underlying
much of the work on language mystify them, and they do not
know whether to accept what they read. Often they are interested
in the broad questions feminist work on language raises, and
would like more background information; but they do not know
where to find it.
This book is addressed to those women. In it they will find an
introduction to the most influential relevant trends in linguistic
theory (Chapters 1 and 2); a critique of sex difference studies
(Chapter 3); an account of sexism in grammar and in language,
with a discussion of possible reform strategies (Chapters 4 and 5);
a detailed consideration of the language-and-world view question,
with an account of several radical linguistic theories (Chapters 6
and 7); an alternative theory based on a communicational
approach (Chapter 8) and a conclusion (Chapter 9) which looks
at the prospects for an adequate feminist theory of language, as
well as for a feminist linguistic practice.
It is inevitable that the feminist debate about language - what
it means and what we should be doing about it - will become ever
more complex and difficult for the non-specialist to evaluate: the
8 Feminism and Linguistic Theory

major challenge is to keep the discussion accessible without

blinking difficult theoretical problems. In this book I have tried
to meet that challenge.
2 Linguistic Theory:
Frameworks and
Approaches
... the study of language has opened
the route to an understanding of
mankind [sicl, social history and
the laws of how a society functions.
Rosalind Coward and John Ellis

I have already mentioned the three areas which a feminist theory

of language must deal with: sex differences, sexism and the
question of alienation. In saying this I am not saying anything
new. Although as far as I know, no one has ever put out a
manifesto for feminist language study aiming to define the sub-
ject matter comprehensively, history and consensus dictate that
these things should be the proper focus for our efforts.
As well as charting the terrain, albeit informally, feminists
have begun to explore it. In their researches they have drawn
mainly on two disciplines: modern linguistics and contemporary
semiology. The most popular texts, and those read in depart-
ments of linguistics, tend to the approach known as socio-
linguistics, a branch of (predominantly Anglo-American)
linguistic science; but there is also a considerable body of work by
feminists whose allegiance is to semiology (represented in Britain
by journals such as mlJ and Screen, and in America by Signs and
Diacritics: in France they are more numerous, and a good
introductory anthology is that by E. Marks and I. de
Courtivron).!
In this book I intend to develop and to criticise both of these
approaches, in themselves and as they relate to feminism. It is
therefore important that the reader should have some basic
understanding of linguistics and semiology from the start; and
this chapter is devoted to explaining the basic principles of
linguistic and semiological theory.

9
10 Feminism and Linguistic Theory

WHAT IS LINGUISTICS?

The usual definition of linguistics simply calls it 'the scientific

study of language'. Scientific is the important word here: for
although language has been studied for at least 2500 years in the
guise of grammar (how to use language correctly), rhetoric (how
to use it persuasively), poetics (literary criticism) and philology
(historical analysis and comparison of different languages), the
scientific study of language is a specifically modern development.
Most introductory textbooks date it from the publication in 1916
of Ferdinand de Saussure's COUTS de Linguistique Generale
(Course in General Linguistics).
Before we turn to Saussure, or consider what it means in
general to say that the study of language can be scientific, it is
important to examine what it means to call any kind of enquiry
'scientific' .
The dictionary definition of science is 'a branch of knowledge
dealing with objects, forces and phenomena of the natural
universe, based on systematic observation of facts and seeking to
formulate general explanatory laws'. Evidently this description in
its entirety will only do for the 'natural' sciences like physics and
chemistry. Other subjects usually thought of as scientific, such as
computing science and psychology, do not deal with 'objects,
forces and phenomena of the natural universe' (and neither does
linguistics.) But we can make these less immediately concrete
areas of enquiry as much like the prestigious natural sciences as
possible by extending to them the most important connotations of
sczence.
According to these connotations, science is factual - it deals
with fact, not wild guesses or opinions. Thus the scientist has to
be objective and methodical, rather than biased, casual,
haphazard or anything else that might interfere with the facts.
Finally, scientists are obliged to provide not only facts but
explanations, by discovering the laws of how things work: which
implies that anything a scientist studies must have laws for her to
find, rather than operating at random.
The other, very important, connotation of science is its high
value and prestige. Science, in the imagination of scientists and
non-scientists alike, is good. The qualities it is thought to have
may in fact be complete myths (objectivity, for instance; and
 Linguistic Theory 11

many non-scientific qualities like intuition and guesswork

actually play a significant part in all investigations) but the
mythology of science is one that our culture worships. We believe
that science, distinguished from the arts by its objectivity and
from scholarship (e.g. history) by its power to explain things, will
lead us to the truth.
So when we examine the development of a scientific linguistics
we must bear in mind that there is a great deal to be gained, in
terms of credibility and respect, from the use of the label
scientific. In linguistics, as in other disciplines, deliberate
attempts have been made to appear as objective, rigorous, etc. as
possible, not because the resultant practices are necessarily more
appropriate than any others, but because they bring with them
an enhanced position in the academic community. Linguists have
had to pay attention to methodology, formulate all observations
as rules, eschew (at least in theory) value judgement. The
scientific ness of linguistics is enshrined in three principles, which
need to be introduced at the very beginning of our discussion.

I Descriptive v. prescriptive: eliminating subjectivity

Linguistics is often described as 'the construction of grammars'

but this is not the same sense of grammar as the one I mentioned
earlier (rules for correct usage). Whilst most of us recall
grammatical rules from English lessons at school - 'don't split
infinitives' and 'never end a sentence with a preposition' for
instance - these are exactly the sort of rules a linguist'S grammar
does not include. They are prescriptive rules telling people what
they should say when they would naturally tend to say something
else. The linguist is interested in descTZptive rules, formulae
which capture the regularities of what people really do say - or
more accurately, because of a sophisticated theoretical sleight of
hand which we will shortly come to, what they unconsciously
know about their language. In other words, linguistic rules are
laws of nature rather than rules of the game, to be discovered
rather than invented. This is, of course, very nice for linguists
who want to think of themselves as scientists.
At the bottom of the belief in description rather than pre-
scription is the scientific requirement of objective investigation.
Correctness in grammar is a social norm, arbitrary and based on
12 Feminism and Linguistic Theory

value-judgement: therefore it cannot be the province of the

objective scientist. To be objective, the linguist must include in
her grammar not what ought to be said (e.g. it is I) but what is
said (e.g. it's me), excluding only what cannot be said in the
language under consideration (e.g. it I i5). Many people have
pointed out how erratically the distinction has been applied in
practice, but it remains a basic principle. Comfortable in their
reiterated assertion that prescription is a cardinal sin, most
analysts do not bother with their own transgressions, still less
ponder how far transgression may be unavoidable.

2 Synchronic v. diachronic: eliminating history

Before Saussure, the dominant approach to language study was

comparative philology, which deals with relations between
languages and traces their descent. Saussure thought that this
historical approach stopped linguistics making a systematic
analysis of language, since it dealt only with change through
time, in his opinion an unsystematic phenomenon. Rather than
watching the organism's development, he wanted to interrupt its
life history and place it under the microscope in order to examine
its internal structure. Therefore he proposed to cut through the
time dimension at a particular point and study everything about
the language as it existed in that frozen moment. He called this
'synchronic' linguistics, opposing it to 'diachronic' (historical)
study.
Saussure obviously idealised the facts of language when he
imagined cutting through a single moment in time, since
languages are in a constant state of change. Moreover, in any
speech community there exist forms from every point along the
time dimension, because speakers vary in age. However, the idea
that language could be studied ahistorically was an important
one, and synchronic linguistics nowadays is not only distinguished
from diachronic linguistics, it usually takes precedence over it.

3 Knowledge v. use: ordering chaos

The third important principle of linguistic science is the

dichotomy between what the language user knows, thl!! system,
 Linguistic Theory 13

and the use she makes of that system as demonstrated by her

actual linguistic behaviour. The data of speech are very
complicated, heterogeneous to the point of chaos: so linguists
propose that they must rest on something much more elegant and
unified, a set of rules or relations which cannot be observed
directly, but which may be inferred by the skilful scientist.
Saussure's version of this dichotomy is the langue/parole
distinction. Langue is the abstract system of relations which make
individual behaviour possible. It is a 'social contract' rather than
the property of anyone speaker, and it is the linguist's primary
object of interest. The individual behaviour regulated by langue,
particular instances of speech, is called parole. Chomsky too
makes a knowledge/use distinction: in his case it is between
competence, the set of rules known by native speakers of a
language, and performance, the actual and imperfect language
these speakers produce on the basis of the rules.
Both Chomsky and Saussure consider it more revealing to study
langue, and to do so is now 'scientific' practice within linguistics.
Saussure's was the first theory to use fully the scientific
principles of descriptiveness, synchronicity and knowledge
orientation, and they have remained influential ever since. This is
not to say, however, that Saussurean theory has dominated
linguistics. The preoccupations of a science change according to
the needs of the moment and the idea of science currently
popular among intellectuals, and linguistics is no exception.
Three approaches to language study have arisen in linguistics
which are relevant to our concerns, since all of them, and the
contradictions between them, have implications for feminist
theory. They are the structural approach, as exemplified by
Saussure and later by the semiologists; the psychological
approach, in which language is above all a property of the mind;
and the sociocultural approach, which views language primarily
as a product and a shaper of human culture.

THE STRUCTURAL APPROACH

Saussure

Many people regard Saussure as a 'thinker' rather than a linguist:

14 Feminism and Linguistic Theory

he is famous for having invented structuralism and founded the

discipline of semiology. However, the major elements of his
thought were developed expressly to deal with language, or more
precisely, with his own abstraction, langue.
The defining characteristic of a structural approach is its
insistence that language must be studied as a self-contained
system, rather than, say, a historical phenomenon, a philo-
'Sophical problem, a social or pedagogical tool. As Saussure put it,
'Language must ... be studied in itself; heretofore language has
almost always been studied in connection with something else,
from other viewpoints.'2 But if language is not just a tool of the
mind, society, education or philosophy, what is left for it to be?
Saussure's solution, from which all his theoretical insights sprang,
was to view language as a sign system. He placed linguistics at the
centre of a new science, the science of signs. 'A science that
studied the life of signs within society is conceivable; it would be
part of social psychology .... I shall call it semiology. Semiology
would show what constitutes signs, what laws govern them. 'S In
other words, Saussure believed that within the field of social
psychology lies a sub-field worthy of separate consideration: the
tendency of human societies to construct symbolic systems such as
languages. And indeed, this 'science' does now exist, having been
brought to fruition by the neo-Saussurians, the semiologists of our
own time.

What is a sign?

A sign is difficult to define adequately. It is what is produced by

linking two theoretical entities, the signifier and the signified. In
the case of language, Saussure exemplified the sign at the level of
the word: so that the sign CAT, for instance, consists of a signifier
which is the sound image /k a t/ and a signified which is the
concept cat, a four-legged feline animal. In other words, we may
say rather crudely that a signified is an idea, and the signifier
what formally expresses it.
One important point about signs is that they separate language
from things, from reality. A signified is not a thing, but a
concept, and the sign is arbitrary in two ways. Firstly, signifiers
are arbitrary. For there is nothing in nature which obliges a cat to
be called a cat: it could just as easily have been called a BLERG, so
 Linguistic Theory 15

long as the appropriate signified had been attached to that

sequence of sounds. Equally, however, the signified is arbitrary;
for prelinguistic reality (which exists for infants and those who
cannot learn to speak, presumably) is an undifferentiated chaos
out of which nothing naturally compels us to pick a class of CATS.
By means of signs we make sense of a world that does not
necessarily fall into neat conceptual classes; and in this process
there is, according to Saussure, a great deal that is arbitrary.

Relations between signs

If signs are arbitrary, their substance does not matter much.

What matters is that they should be distinguished from each
other. It is the difference between signs that defines them.
This point may be grasped by thinking of military ranks such as
private, corporal, sergeant, sergeant-major, etc. What do these
mean without each other? If someone asks you what a private is,
you are bound to reply that it is the lowest rank in the army,
which means you are defining it in relation to other assumed
ranks. Similarly, the signifiers of rank, stripes on a uniform,
would be meaningless if they were not to be compared with each
other.
Saussure talked about two sorts of relationship signs can have
with each other. Either they can combine with each other
(syntagmatic relations) like the sounds /k a t/ in CAT, or they can
replace each other (associative or paradigmatic relations), as p, r,
f, v, b, etc. could replace k in CAT.

American structuralism

An approach similar to Saussure's dominated American

linguistics until the late 1950s. The American structuralist
linguists were concerned with the actual description of native
American languages on the verge of extinction, and their
emphasis was on practical techniques for linguistic analysis.
The method they evolved was to describe a corpus, a sample of
speech from a native speaker. Their grammars were inventories of
the elements of a language (its sounds, words and grammatical
forms) together with rules for the distribution of each element
16 Feminism and Linguistic Theory

(i.e. where it could occur). Although not truly Saussurean, these

linguists were structuralists in their approach and their analytic
method.

SEMIOLOGY

Unlike American structuralism, semiology (a blanket term, as will

become clear, for a number of rather different enterprises) is
Saussurean: it is the 'science of signs' he suggested, and can only
be defined as a single movement on the basis of its debt to the
COUTS. However, it is not thought of as part of the linguistic
tradition, either by linguists or by its own practitioners, and to
discuss the reasons why it is so different is to acknowledge a
problem in the context of this book.
Theoretical linguistics as it exists now belongs to the
predominantly Anglo-American scientific and social-scientific
tradition. This tradition is bourgeois, empiricist and positivistic.
In other words, for all that it has modified its objectives and
methods since Saussure, linguistic theory remains locked within
the view of 'science' that I have already outlined.
Semiology, on the other hand, exists within another tradition,
not Anglo-American and scientific but French and literary. It is
anti-bourgeois, anti-empiricist and sceptical about the mythology
of science. Its own mythology is marxist and Freudian. In its
aims, its methods and its general tone, it is unlike linguistics.
A somewhat similar gulf exists between feminism in France and
in Britain or America. (This is an oversimplification, since
dissenting voices are often heard, particularly among British
intellectuals affected by the French tradition, but it is useful up to
a point.) Marxism has affected both kinds of feminism, but
whereas the English regard Freud as reactionary, the French are
profoundly influenced in their feminism by psychoanalytic
theory. They dislike our bourgeois empiricism, while we find their
ideas over-theoretical and frequently biologistic (i.e. they attach
more importance to the d,ifferences between women and men,
often implying that these are the consequence of biology).
All this means that it is difficult, and perhaps illegitimate, to
examine theories developed in the Anglo-American feminist and
linguistic tradition alongside those developed in the French
 Linguistic Theory 17
semiological context. These theories do not attempt to do the
same things, either in terms of language or in terms of feminism:
why should they be discussed in the same terms? How can I
evaluate them when I belong irredeemably to the bourgeois
empiricist camp?
In fact, it is far beyond the scope of this book to deal with
semiology fully, and in particular I am neither concerned nor
qualified to discuss the psychoanalytic basis of Lacanian theory.
What interests me is the view of language semiologists hold, and
specificially their analysis of the relation between language and
women's oppression.
It is especially interesting that many Anglo-American writers
who are not semiologists hold very similar views to the ones I am
talking about. Some of them have been influenced by semiology,
but others have derived the same ideas from quite different
sources. This suggests two things: first, that something about
women's experience of language makes certain ideas appealing,
and secondly, that linguists and semiologists have something in
common in the way they conceive of languages.
It follows that what I want to bring out when I discuss
semiological theories is their similarities to linguistic theory,
rather than the ways in which the two traditions differ. My
discussion will be partial in both senses of the word; it will be
directed to the three problems of sex-difference, sexism and
alienation, asking whether semiology is useful to feminist theory
rather than considering it entirely on its own terms.

SEMIOLOGY AND STRUCTURALISM

Although the terms semiology and structuralism are sometimes

used as if they were interchangeable, in this book they are not.
Semiology is the field of study which Saussure envisaged and
defined in the Cours as a science of signs - in other words, the
investigation of symbolic systems such as languages. Thinkers like
Levi-Strauss, Barthes and Lacan (respectively an anthropologist,
a literary critic and a psychoanalyst) are semiologists because they
regard their objects of study (kinship, myth, the literary text, the
unconscious) as sign-systems, organised as Saussure maintained
languages are organised.
18 Feminism and Linguistic Theory

Structuralism, on the other hand, is a method for analysing

phenomena, particularly sign-systems. It depends on the idea
that differences and relationships are important, and although
semiologists employ this method, it has also been employed
without any concern for the theoretical framework of semiology.
In brief, then, semiology is a discipline and structuralism a
method. When I mean the contemporary movement associated
with the theories of Lacan, Derrida, Kristeva et al. I shall say
semiology.

CULTURE AS SIGN -SYSTEM

It has already been explained how Saussure defined the system of

signs; the system he was interested in was, of course, language.
But clearly, it is not only language that consists of signs. The
contemporary semiologists take Saussure's linguistic insights and
apply them to other phenomena, treating diverse cultural objects
and practices as sign-systems and studying them in the same way
that Saussure studied language.
A short example of what this means in practice is provided by
Roland Barthes' analysis of a magazine cover.4 The photograph
on this cover (in Saussurean terms the signifier) showed a black
soldier saluting the French flag (its immediate meaning or
signified). In Barthes' analysis, however, this sign (pairing of
image and concept) was itself the signifier of a higher-order sign,
the myth, whose signified was the idea of French imperialism as
acceptable.
This kind of cultural analysis improves somewhat on the usual
notions of stereotyping and representation by giving an account
of why some images have a particular significance. Once an
image becomes a sign, with its place in the network of
relationships which is the system of signs, the union of signifier
and signified which it embodies (out of many other possibilities)
comes to be seen as natural and indissoluble.
The possible applications of this in feminist thinking are clear.
It might be a possible explanation, for instance, of why one
cannot normally use the signifier image of a naked woman
without invoking in those who see it ideas of sexual availability
and degradation. A naked woman does not signal those things
 Linguistic Theory 19

inherently; in principle the image could mean many other things.

Yet because it has become a sign, with the image as signifier and
availability / degradation as signified, one particular meaning, a
misogynist one, has more power than the alternatives, and its
unpleasant connotations cannot simply be discarded.
We must be careful, however, not to argue that this kind of
meaning is acquired in an arbitrary way. Certainly it is
conventional, but it is mediated by the political structures of a
misogynist society. In Barthes' example too, the sign can only be
mythologised if we pay attention to a particular bit of it, the race
of the soldier: and it is no accident that we do this, but a
reflection of our culture's racism. In other words, signs are
historically specific and signification does not occur in a vaccum.
Although some semiologists acknowledge this point, others are
equivocal. This is because they have adopted a deterministic
position, arguing that language creates all meanings within a
society rather than reflecting, or interacting with, anything else.
The kind of semiology feminists have been most influenced by,
Lacanian psychoanalysis, is very deterministic; yet surely the
question of how some images acquire political significance, and
the nature of that significance, also deserves attention.

SEMIOLOGY AND PSYCHOANALYSIS

French feminism is strongly marked by the influence of psycho-

analysis. In particular, many writers have developed their ideas in
response to the psychoanalyst Jacques Lacan. The important
thing about this is that Lacan's is a linguistic theory of the
unconscious and of the subject's development: hence the critiques
and reformulations which feminists have produced also centre on
language, and have interesting things to say about language.

LACAN

In this introductory section my main objective is to explain how

Lacan's psychoanalysic theory is 'linguistic,' and I shall not deal
with its substance in detail.
Lacan is a semiologist because he regards the psychoanalyst's
20 Feminism and Linguistic Theory

object of study, the unconscious, as a sign-system. He is deeply

indebted to Saussure for many of his central concepts. Most
important of all, Lacan's theory of the unconscious gives pride of
place to language, thus producing a radical re-reading of Freud.
Lacan observes the stress which Freud put on language and
linguistic evidence in his accounts of dreams and jokes. In case-
histories too language is inevitably of the utmost importance, for
the analyst has nothing else to go on when probing the
analysand's unconscious. You cannot observe the unconscious: it
is always mediated by language.
Following Saussure, Lacan believes that without language,
everything must remain not only unobservable but undifferent-
iated, without structure: and since (as Freud demonstrates) the
unconscious is highly structured, it cannot exist before language
does. Lacan reasons that the unconscious must in fact be
constructed through language, as language develops in the child.
This process governs and differentiates what is, before language,
a mass of instinctual drives, an 'hommelette' (little man (sz'c)/
omelette) spreading in all directions. For Lacan, then, i.t is
learning language that makes us what we are; and since our
sexuality and gender identity is an important component of what
we are, his theory has important implications for feminists trying
to understand how femininity is constructed in patriarchal
societies.

THE PSYCHOLOGICAL APPROACH

Saussure believed that semiology was part of social psychology,

and indeed no linguist has ever denied that language was
somehow 'in the mind'. American structuralism, however,
subscribed to the behaviouristic kind of psychology in which no
attempt is made to study internal mental processes; and therefore
the 'mental' aspects of language were given little emphasis in the
major texts produced by American structuralist linguists.
The behaviourist position on language was attacked by Chomsky
in his review of B.F. Skinner's Verbal Behavior.5 Chomsky argued
that the structures of language are too complicated for it to be
learned by a stimulus-response mechanism motivated by positive
reinforcement.
 LZ'nguistic Theory 21

Chomsky also attacked the methods of American structuralism,

because he believed linguistics should be not a set of procedures
for analysing and classifying different languages, but a branch of
cognitive psychology interested in the general properties of human
language. These, in Chomsky's opinion, would be found nowhere
but in the structure of the human mind itself. He objected that
the traditional corpus was not only a violation oflinguistic principle
(being parole rather than langue) but an irrationality. For how
could the description of a corpus provide an exhaustive account
of the language? In any language, the number of sentences
we can produce or understand is in principle infinite.
But how could anyone describe an infinite number of
sentences? Chomsky answered this question with one of his own:
how can anyone accomplish the child's task of learning an infinite
number of sentences? Yet all normal children master a language.
Chomsky's belief that this impressive achievement could not be
a matter of mere stimulus and response led him to three
conclusions. First, he concluded that children do not learn
sentences, but rules for producing sentences. These would be
finite in number, though the sentences themselves are infinite.
Secondly, he concluded that the rules must be innate in the
child's mind: the structures of language are too complicated for
children to pick up without some predisposition towards them.
Thirdly, if language is somehow innate in humans, all languages
must all some level be very like each other. Children learn to
speak whatever they hear around them; whatever structures are
innate, therefore, must apply to any human language.
There are several reasons to believe that linguistic abilities
depend at least to some extent on an innate predisposition. Psy-
chologically oriented linguists point to the features which·
differentiate human and animal languages, both structural (e.g.
human languages are hierarchical) and communicational (it is
possible to invent new messages, to lie, to talk about events
removed in time and place, only in human languages).
It is also pointed out that human children show signs of being
'programmed' to learn language at a particular stage of
development and in a particular way. 'Feral' children, who grow
up away from society and are therefore not exposed to language
at the critical period, find it very difficult to learn to speak
subsequently regardless of their general intelligence.
The psychological approach in linguistics dictates that the
22 Feminism and Linguistic Theory

linguist should look not just for any set of rules which accurately
describe a language, but for the set the child has in her brain. She
must also pay attention to isolating linguistic universals, the
features which, if Chomsky is right, all languages must share.

THE SOCIOCULTURAL APPROACH

The sociocultural approach to language views speaking as a

mediator of social relations, a way of expressing social role, a
reflection of, and an influence on, culture.

ANTHROPOLOGICAL LINGUISTICS

The American structuralists always had a particularly close

connection with anthropology, because they necessarily com-
bined the analysis of unfamiliar languages with the observation of
unfamiliar cultures. In addition, many anthropologists have
always been sensitive to the importance of language in regulating
society: they have used linguistic data to pinpoint aspects of social
organisation, and as a guide to what the people being studied
regard as important. So a large cross-cultural literature has long
been available on the social aspects of language.
One question which arose within the anthropological tradition
has proved controversial: the problem of relativity and linguistic
determinism. (Relativity expresses the idea that 'objective reality'
is in fact perceived differently by different cultures or in different
circumstances: determinism means that one particular agency is
responsible for the variation.)
Some anthropological linguists believed that the extreme
variations they observed in different peoples' perceptions of
reality were directly attributable to language. For instance, a
tenseless language like Hopi would make it impossible for its
speakers to conceptualise time as speakers of English, a tense
language, conceptualise it.
This theory is often called the Sapir- Whorf hypothesis after
the two linguists most associated with it, Edward SaPir and
Benjamin Lee Whorf. It is both relativistic (asserts that reality is
 Linguistic Theory 23

perceived differently) and deterministic (asserts that language is

responsible), and several recent feminist theorists acknowledge it
has influenced them (e.g. Spender,6 and Kramarae 6) Determinist
theories of language will be discussed in detail in Chapter 6.

SOCIOLINGUISTICS

The other major sociocultural trend in linguistics is sociolinguis-

tics which emerged as a discipline in its own right during the
1960s. In its genesis several factors were at work, many of them
very typical of the period.
First, there was a desire to study variation within languages in
an up-to-date way. Dialectology, a historical and mainly rural
subject, was not appropriate to study variation in the urbanised
and geographically mobile society most linguists lived in, and a
new sociological discipline was needed to analyse class and ethnic
differences.
Secondly, many academics felt an impulse toward a more
socially relevant linguistics, concerned with the problems of
disadvantaged groups. (Needless to say, they usually ignored the
largest such group - women_) A common motif in sociolinguistics
is the description and defence of non-standard speech, and
explicit opposition to educational theories blaming it for black
and working class underachievement.
Thirdiy, many sociolinguists believed that a different approach
might make good worrying defects in the orthodox model of the
time (i.e. Chomskyan transformational grammar). They wanted
to show that a theory which excluded history and parole was
artificial and unsatisfactory.
The technique of sociolinguistics is usually to correlate
linguistic features in people's speech (different pronunciations
and grammatical forms, for instance) with social indices like race,
sex, age and class or with situational characteristics (how formal
the situation is, what kind of speech event is taking place, etc.) In
this, sociolinguistics might appear to be the linguistics of use and
community rather than knowledge and the individual: yet par-
adoxically it claims to demonstrate, by showing that variations
are structured and systematic rather than occurring at random,
that these social aspects of language are as much part of our com-
24 Feminism and Linguistic Theory

petence as the ability to construct grammatical sentences. A

person who does not know how to vary her speech in different sit-
uations, or who cannot tell an old from a young speaker, is not a
competent user of her language.
As well as subverting the knowledge/use distinction, socio-
linguistics challenges the separation of synchronic and fliachronic
study. Many sociolinguists believe that they are interrelated,
because it is the social significance of variation in the present
which determines how language will change in the future.

SUMMARY

I do not want to imply that the three approaches discussed here

are mutually incompatible in the sense that, for instance,
structuralists think language is not a social phenomenon and
sociolinguists see no cognitive/psychological element to be
explained in language. It is quite possible to see all three
approaches as valid and even useful. However, there are
important inconsistencies, whose implications for feminism
matter a good deal. For instance, the psychological linguist is
committed to universalism and mentalism (belief in innate ideas)
while to the sociolinguist it is more significant that languages and
cultures vary. Whorfian determinism, which holds that the mind
is deeply affected by the language it learns in the course of its
development, cannot be reconciled with the Chomskyan view that
all languages ultimately share one structure, dictated by
properties of the human mind.
There are also tensions between the structual and sociocultural
approaches. Practitioners of the latter believe that classical
structuralism demands too much idealisation, the assumption of
a homogeneity which no investigator will ever find in the real
world of differences and, the sociolinguist reminds us,
inequalities.
Lacan is interesting in that his theories synthesise all the
approaches: structural in method, they are concerned both with
psychological development and with the acculturation of
individual subjects. But this synthesis is not without its problems,
and Lacan would not be comfortable either with Chomsky or
with Whorf. Although he shares the psycholinguist's belief that
 Linguistic Theory 25

he is uncovering universal properties of the human mind, he does

not believe in innate ideas; although his position is deterministic
he seems not to be a relativist (which begs the question of how
very different languages like English and Hopi can generate
identical unconscious structures in their users).
These inconsistencies are relevant for feminist language study
because to date its greatest debts have been to sociolinguistics and
semiology. Both, in their feminist versions, are very deterministic,
a position which is problematic and should not be accepted
without question. On the other hand, where sociolinguistics
stresses variation and heterogeneity, semiology pulls in the
opposite direction toward a rigid and idealised systematicity
which is far from anything we can observe. These are points to
bear in mind, and they will be taken up again in the context of
particular problems.

CONCLUSION

I want to conclude this chapter by indicating the contributions

which the above approaches can make/have made to feminist
language study.

Feminism and semiology

The contribution of semiology to feminist language study is a

double one. First, there is its literary criticism, which has had a
number of things to say about women's use of language in
imaginative literature, female creativity, etc. I cannot give this
aspect of semiology the attention it deserves, partly because I am
not a literary critic or stylistician but also because to deal with
literary writing would detract from my emphasis in his book.
The second contribution, however, is of more general interest
to the linguist: it is the theory of 'gendered subjectivity' derived
from Lacanian psychoanalysis. This amounts to a claim that
women are alienated in society because they have to learn a male
language, and it will be discussed in Chapter 7.
26 Feminism and Linguistic Theory

Feminism and sociolinguistics

Because it deals with language use in social context, sociolinguis-

tics is by far the most fruitful source of data and work on sex-
differences. Whatever the shortcomings of prefeminist surveys
(see Chapter 3), they did sample the speech of a large number of
women. In recent years, a great deal of research on sex-difference
has actually been done by feminists, who are attracted to
sociolinguistics precisely because within it they can work on the
questions which relate to women and language.
Another contribution which sociolinguistics can make to our
understanding as feminists concerns the relation between
language and disadvantage. The last decade has produced a
considerable literature on speech, social evaluation and
deprivation/underachievement. Almost all of it concerns either
racial min on tIes or the working class (groups which,
astonishingly, successive researchers have chosen to represent
with all-male samples) but the findings are suggestive and could
well be helpful to students of female oppression.

Feminism and anthropological linguistics

Anthropology's most significant contributions to feminist think-

ing about language have been theoretical, since it has given
us both the Sapir-Whorf hypothesis and the 'dominant/muted'
theory of Edwin and Shirley Ardener (both extensively used by
feminists). It demonstrates a very helpful concern with the role of
language in thought and culture. Anthropological studies allow
us to see how far linguistic means are used to maintain the social
order in other cultures, and this suggests interesting parallels with
our own (for instance sexist language as verbal violence;
restrictions on women speaking in public as a form of social
control; judgements on speech style as reinforcements of sexist
personality theories).
In addition, anthropology provides cross-cultural data on sex
differences which is of interest both linguistically and as an
insight into other societies' sexual politics.
 Linguistic Theory 27

Feminist linguistics?

Finally, it is difficult to separate the contribution these

approaches can make to feminist theory from the things feminism
can teach linguistics. In so far as linguists have dealt with the
language of! about women, feminists must scrutinise their work
and present a careful critique. I shall begin on this in the next
chapter, which is about sex difference.
28 Feminism and Linguistic Theory

3 The Politics of Variation:

Sex Differences in
Language and Linguistics
In a society where women are
devalued it is not surprising
that their language should be
devalued .. ,.
Dale Spender

Although this chapter is about sex differences in language, it is

not intended to serve as a catalogue of research findings. Rather,
it has two main concerns: the first, which it shares with Chapter
4, is the sexism of linguistic science, as expressed by a mass of
assumptions and practices reflecting the status quo; while the
second is the political significance of sex difference itself. Hence
the chapter heading is deliberately ambiguous, referring not only
to the politics of variation but also to the politics of studying it.
Feminists have frequently pointed out that the study of sex
differences in any subject has a political dimension, and that
many sex difference studies are simply elaborate justifications of
female subordination (girls don't become engineers because they
are inferior to boys on tests of spatial ability) or else overt anti-
feminist propaganda (as in the recent rise of sociobiology, which
sets out to prove that the domestication of women is a biological
imperative, determined by our genes). Sex difference research
inevitably comes out of sexist ideology, and its findings,
unsurprisingly, are interpreted in ways that reinforce that
ideology. There is usually a hidden presumption (if indeed the
experimenters trouble to hide it) that men are the norm from
which women deviate, that the male norm is superior to the
female deviation, and that the difference is ultimately reducible
to biology, that is, 'natural'.
Hence an honourable feminist tradition has arisen of directing
 The Politics of Variation 29

our scholarly attention to the academic explanation of sex

difference rather than to difference itself. The aim is to expose,
and then to destroy, academic legitimations of women's
oppression, which are a powerful force simply because so many
people accept the factuality and objectivity of science.
In linguistics, however, this has not been such a major feminist
concern. Sneering at long dead researchers like Jespersen and
Sapir has become a commonplace, but a thorough feminist
critique of modern sociolinguistics has been much slower to
emerge. (There are some useful comments in this direction by
Jenkins and Kramarae, a few of which will be developed here.)1
In general, feminists within linguistics have been more interested
in furthering the study of sex difference than in criticising it. It is
interesting to ask why this should be the case. What have
feminists to gain from studying sex differences in language?
A survey of the studies carried out by feminist linguists suggests
there are two main motives (which are not mutually incompatible
and may occasionally co-occur). The first is the quest for an
'authentic' female speech or writing, whether this is taken to
reflect some deep-seated cognitive or unconscious difference, or
the existence in specific societies of distinctive female subcultures.
The political point of this approach is to isolate and validate a
female mode of language use, and to relate such a variety
(supposing it exists) to the general notion of female culture.
The second motive has more to do with identifying the sexual
power dynamic in language use. Here, sex differences in
language are related to the power of men and the powerlessness of
women, and the political point being made is that even our
speech behaviour reflects and perpetuates patriarchal norms.
Feminist sex difference researchers in linguistics have therefore
addressed themselves to a number of questions. How great are the
verbal differences between men and women? Are they great
enough, for instance, to justify us in talking about separate
languages or dialects (,genderlects') for the sexes? Are the
differences innate, reflecting anatomical or cognitive factors, or
are they learned? At what stage of development do they arise?
Finally, what can we learn from the substance of the differences?
Are they best understood as expressing a feminine identity, as
functions of the particular social/occupational roles taken by
women in particular societies, or simply as the indices of
oppression and powerlessness? What is their over-all significance?
30 Feminism and Linguistic Theory

These are the very same questions that any account of linguistic
variation would have to deal with, feminist or not. Although the
feminist and the anti-feminist favour different explanatory
strategies, they share a belief that in language use we have a key
to the nature and relative status of the sexes.
In this chapter then, I propose to look at how sex differences in
language reflect, or are said by researchers to reflect, the nature
(or role) and status of women and men. As well as pointing out
the obvious biases and blind spots in current models, I shall also
be trying to demonstrate the problems raised by this idea that
language use is a direct reflection of social identity. The
explanation of linguistic behaviour and linguistic variation is
trickier either than feminists suppose, or than linguists admit.

SEX DIFFERENCE STUDIES: A CRITIQUE

Linguistic accounts of sex difference are open to criticism on three

fronts. First, we may be critical of the way the field has been set
up, so that women's speech is viewed only as a collection of
deviations from the (male) norm. Secondly, we may be highly
suspicious of most studies' actual findings (if that is the right term
for anecdotes, stereotypes and survey results distorted by
methodological bias). Thirdly, we must take up the matter of
interpretation and explanation in some detail. For once again,
sex difference studies do not describe male and female attributes
without comment. Undertaken by people who share the
prejudices of their culture, they end up rationalising and
perpetuating those prejudices.

DELIMITING THE FIELD: WHAT IS SEX DIFFERENCE?

It might seem at first glance that every question about language

and sex is a question of sex difference, but for the linguist the
concept is more limited. It goes back to the important distinction
in linguistics between system and use, langue and parole:
sociolinguistics, the subdiscipline that deals with sex difference
and with linguistic variation in general, concerns itself not with
 The Politics of Variation 31

generic he, terms of insult and other things that are 'in the
language', but only with variations in people's linguistic
behaviour.
Sex differences exist for the linguist, then, when a linguistic
feature (for instance rising intonation or the word darling) is used
significantly more by one sex than the other: or when people's
linguistic norms and standards are different for women and for
men (for instance, it is less acceptable for women to swear).
This statistical concept of difference defines all pre-feminist
(and some feminist) research into women's talk. It might seem
obvious that in order to study the speech of any group, one would
observe and record its members and then produce a full descrip-
tion: but in the sociolinguistic paradigm the favoured method is
in fact comparative, requiring the researcher to describe most
varieties only in terms of how the differ from other varieties (in
practice, from a white, male, standard middle-class norm). Thus
from Jespersen in 1922 to Labov in 1972, women's speech was
dealt with only in so far as it diverged from men's.
This norm-and-deviation research framework, as I shall shortly
argue, is an important factor in producing both stereotyped
findings about women, and stereotypical explanations. Yet the
emergence and dominance of this framework is not surprising,
since the sociolinguists with their hidden norms are the natural
heirs of a long tradition in the study (and indeed the lay
discussion) of linguistic variation - a tradition we might call the
'anecdotal' or 'folklinguistic' in which the speech of subordinate
groups is represented first as different, and then as deviant, from
the standard. This tradition goes back to Antiquity: where sex
differences are concerned, it persists both inside and alongside
modern sociolinguistics.

THE ANECDOTAL TRADITION: WOMEN AND

FOLKLINGUISTICS

'Folklinguistics' is the name given to that collection of beliefs

about language which are accepted as common sense within a
society. These beliefs serve both to regulate linguistic behaviour,
and to explain it to the ordinary language user; some of them are
fairly accurate, some quite false. Within folklinguistics it is
32 Femz'nism and Lz'nguz'stz'c Theory

possible to isolate 'stereotypes', popular pictures of the speech of

particular groups. Once again, these representations may be
accurate in linguistic terms (for instance, the well-known
observation that Birmingham accents are 'nasal') or they may be
wrong (for instance, the idea that Cockney glottal stops are
'sloppy'). The value judgements implied in many stereotypes fit
rather well with the relative power and social prestige of the
groups concerned.
One way of muddying the waters in sex difference studies is not
to make a distinction between false stereotype sex differences and
ones for which there is good evidence. Many linguists apparently
do not realise how far 'value-free' observations are in fact
conditioned by a long history of misogynist folklinguistics. This
ignorance is all the more unfortunate because stereotypes of
women's talk are so inaccurate; and even the most ludicrous
beliefs return to haunt 'empirical' sociolinguistics.
An amusing early example of the kind of stereotype we are
discussing comes from Jonathan Swift's 1712 'Proposal for
Correcting the English Tongue'. Swift asserts that some sounds
are typical of men, others of women:

... if the Choice had been left to me, I would rather have
trusted the Refinement of our Language ... to the Judgement
of the Women, than of illiterate Court-Fops, half-witted Poets
and University Boys. For it is plain, that in their Manner of
corrupting Words, Women do naturally discard the Conson-
ants, as we do the Vowels ... more than once, where some of
both Sexes were in Company, I have persuaded two or three of
each to take a Pen, and write down a Number of Letters joined
together, just as it came into their Heads: and upon reading
this Gibberish we have found that which the Men had writ, by
the frequent encountering of rough Consonants to sound like
High Dutch, and the other by the Women, like Italian,
abounding in Vowels and Liquids. Now, although I would by
no Means give Ladies the trouble of advising us in the
Reformation of our Language, yet I cannot help thinking, that
since they have been left out of all Meetings, except Parties at
Play, or where worse Designs are carried on, our Conversation
hath very much degenerated. 2

Everyone can laugh at this, and no one now would take it

seriously. But equally silly proposals are to be found in serious
 The Politics of Varz"ation 33

works of linguistic scholarship that date from our own century.

The paradigm example here, invariably slated by feminist
commentators, is Otto jespersen's extraordinary chapter on 'The
Woman' in his 1922 book, Language: Its Nature, Development
and Ongin. 3 Although Jespersen is sometimes accused of
misogyny, his chapter reveals him as a gallant rather than a male
chauvinist; and it is remarkable not so much for its anti-feminism
as for the obvious conviction of its author that no evidence is
required to back up his assertions about the way women speak.
Readers will simply recognise, Jespersen implies, that women
really do speak more softly than men, use diminutives like teeny-
weeny, construct their sentences 'loosely' and leave them
unfinished, all the while jumping from topic to topic.
These characteristics predicated of women were evidently not
chosen at random. Jespersen is caught between his fantasies (soft-
spoken, retiring child-women) and his prejudices (loquacious but
illogical bird-brains) to produce a sexist stereotype which is still
recognisable sixty years on. Cartoon humour still revolves around
women's verbal incontinence ('When two wives get together, who
has the last word?' enquires Andy Capp) though recent findings
indicate that men actually talk a great deal more than women
when both sexes are present at the same time. Illogical women
who can't keep to the point also surface in media representations
of folk-wisdom ('You might as well try to knit fog as follow what's
in a woman's mind' says a character in Coronation Street). As for
teeny-weeny and its childish analogues, I have heard them placed
firmly in the female lexicon as recently as July 1982.4
Stereotypes, however false, tend to persist for as long as they
reinforce important social inequalities. So long as women are
subordinate to men, their language has got to be characterised as
indicating natural subservience, unintelligence and immaturity.
While men dominate women in mixed groups by limiting their
opportunity to talk, our folklinguistic beliefs must include the
untruth that women talk incessantly.
Unfortunately, not all inaccurate stereotypes have been
rejected even by feminists. The first book to alert linguists to the
political implications of sex differences, Robin Lakoffs Language
and Woman's Place, itself creates a new stereotype. s According to
Lakoff, women use more tag-questions ('approval-seeking'
constructions of the 'that'll be all right, won't it?' type) and
more rising intonation, which according to Lakoff indicates
uncertainty. They use intensifiers like really and very more than
34 Feminism and Linguistic Theory

men do, and approximators such as a bit, not exactly. The

picture that emerges from Language and Woman's Place is of a
subservient way of talking where nothing is asserted outright, and
everything is qualified.
Lakoff is a feminist, and she explains the differences she
discusses in political terms. Women use approval-seeking
constructions because they need to seek approval, and fear that
men would find a bare statement threatening or offensive.
Nevertheless, Lakoffs picture is a stereotype for which she
presents no real evidence, and other women who have tested her
hypothesis against a quantity of speech data have found it
wanting on several counts.
Dubois and Crouch, for instance, tested the hypothesis that
women use more tag-questions than men. In their sample, on the
contrary it was men who used more than women. 6 (Tag-questions
will be discussed below, for they provide a good example of what
needs to be avoided in sex difference research.)
Why do stereotypes like this persist in modern 'scientific'
linguistics? Partly it has to do with the methods used. Lakoff was
trained in the Chomskyan tradition, which, as I pointed out in
Chapter 1, rejected the idea of collecting a sample corpus,
advising instead that the analyst examine her own intuitions
about language. Obviously this is not satisfactory when your
analysis involves large-scale generalisations about the linguistic
behaviour of groups; and there is a problem about asking others
for their intuitions. It has been found that people consistently
misreport their own behaviour when they are asked to describe it.
They report usages that owe more to the stereotype than to the
truth.
Feminists too have strong folklinguistic beliefs about women's
speech. This is not itself surprising: what is striking, however, is
the extent to which feminist beliefs resemble those put forward by
linguists whose political views are the opposite of feminist. It
sometimes seems as if feminists took their stereotype straight from
the pages of Jespersen, and this raises the interesting question of
the origin and importance of folklinguistic beliefs generally.

INVESTIGATING FEMINIST FOLKLINGUISTICS

In the feminist workshops, semmars and group discussions on

 The Polz"tics of Variation 35

language that I have attended, several folklinguistic assertions

have been made over and over again. From them I have put
together a profile of women's speech with the following six
features:
1. Disfluency (because women find it hard to communicate in a
male language).
2. Unfinished sentences.
3. Speech not ordered according to the norms of logic.
4. Statements couched as questions (approval-seeking).
5. Speaking less than men in mixed groups.
6. Using co-operative strategies in conversation, whereas men
use competitive strategies.
This is how feminists believe women typically speak, and there
is a tendency to make these attributes the basis of an authentic
'women's style' which would be positively valued. As with many
other things, so with language: women are saying 'we do it
differently, and our way is just as good as yours'.
Revaluing the things that are distinctively female may well be a
significant and necessary political act. However, there is little
point in revaluing a speech style that exists only in the
folklinguistic imagination. Therefore it seems rather important to
investigate whether women do in fact leave their sentences
unfinished, couch their statements as questions, use co-operative
strategies and sO-on: and to look a little more closely at what such
characteristics mayor may not mean.
In the discussion that follows, I have used real data from
surreptitiously recorded conversation to illustrate the difficulties
of investigation and interpretation. Some of the data comes from
a discussion on sex education between two male friends;7 some of
it comes from a discussion on nuclear disarmament between two
female friends; and the remainder comes from a tape on which I
recorded the opinions of six women on male and female speech
styles.
I t is clear from the six points of the folklinguistic profile (above)
that feminists are more interested in large-scale abstractions (how
women communicate, express themselves) than in the sociolingu-
ist's orderly matrix of micro-variables (e.g. the incidence of
particular vowel qualities or grammatical constructions). They
make impressionistic comments like the following:

Women try and make it interesting, you know, make you

36 Feminism and Linguistic Theory

want to listen, men just bash on regardless, they expect you to

listen.
Men are so competitive when they talk, it's not like what I
would call a conversation, it's like - well, a competition.
There were some women there who were really well informed
politically, one woman really talked a lot - God, isn't it awful,
if it's a woman you say to your friends, didn't so-and-so talk a
lotI And men, well it happens all the time, some man just goes
on and on and you never say a thing ...
'I used to really worry about saying things in a logical order,
because the men would say I wasn't putting it in a logical order
,

But this raises two enormous problems for linguistic re-

searchers. First, what should we look for by way of evidence for
impressionistic labels like 'logical order', 'competitiveness' and
'making it interesting'? Secondly, if we find the more straight-
forwardly identifiable elements of the profile in women's speech
(unfinished sentences for instance), how are we to check the
accompanying feminist explanations? These problems can be
illustrated by taking each of the six elements in turn.
Disfluency, meaning pauses, hesitations, false starts, rep-
etitions and so on, is relatively easy to quantify in a piece of
linguistic data. However linguists do not agree on what it actually
means. Some of them see pausing as a strategy speakers use to
plan what they are going to say next, and argue that those who do
not pause often are thoughtless or unintelligent (this argument
has been used about working-class speakers by Bernstein, and
about women by Jespersen). Others believe that disfluencies are a
sort of mistake that people have special ways of repairing; while
yet others believe that discontinuity in speech is a stylistic device
that people use for effect to convey particular messages. So if
there is a sex difference in fluency, there is still a problem of
interpretation.
Unfinished sentences, similarly, present difficulties before we
even get to sex differences. For the complete sentence is not in
fact the most usual unit of conversation, and we are misled by the
conventions of writing into asserting the supremacy of something
which is not the norm. Typical utterances in a conversation use
 The Politics of Variation 37

pausing and intonation to make their meaning clear, rather than

'perfect' sentential syntax:

and also - also the only way they could probably swing an
election behind them, that would be the issue ...
well what do you think about sex education do you think that
er it er I mean there's been a great hooha about it recently
hasn't there ...

These utterances are not at all confusing, but they are also not
much like the sentences one learns to write at school. Nor is it
clear what producing a high proportion of utterances that 'tailed
off would mean in terms either of cognition or of social
confidence.
Logic in speech is a very inexact notion, and could in addition
refer either to the content of what is said, or to the sequencing of
what is said ('logical ORDER').I suspect that linguists like
Jespersen have the first possibility in mind when they accuse
women of illogic, while feminists, in rejecting male speech
structures, are thinking more of the second. But in any case, what
do we mean by logic?
Consider two extracts from the data.

(1) B: I must say I tend to be . . . I mean I ... you know I

do talk quite openly to my pupils which is a little
daring of me because the situation in Cyprus is dif-
ferent from here . . . I mean people are a bit narrow
in that respect you know they don't like people to talk
about it too openly - but I do because I think it's
important but the trouble is that - erm - that's not
really systematic in the sense that I do it but how
many people - how many other people do it you see.
A: Yes it's er it's an enormous problem actually because
as soon as you start to make a special thing about it
then immediately I think you're creating the wrong
atmosphere, especially for sex.
(2) C: I'd sell my soul to get nuclear weapons out of Britain.
D: Well I don't think you're ever going to get it except
by civil disobedience because it's part of what a
government's all about, they have to defend the
38 Feminism and Linguistic Theory

country. I don't think getting the Labour Party to

resolve for unilateral disarmament changes anything
... they just won't put it into practice they can't ...
you look at the record of governments there's no way
they can stop defending us.
C: I think that there's some hope now with the Labour
Party because they've um compromised themselves
politically so far that they wouldn't -

The second extract might be judged in many ways more overtly

logical than the first. Its structure could be shown like this:
- nuclear weapons should be got rid of (opinion);
- the only way to do it is through civil disobedience (suggestion);
- because governments have to defend their country, so party
political promises of disarmament change nothing Uustification
of suggestion);
- look at the record of governments (evidence for justification);
- nevertheless there is hope through getting the Labour Party to
commit itself (disagreement)
- because it would be politically disadvantageous to break a
promise Uustification).
The argument progresses without a break and has all the
hallmarks of logic (e.g. careful justification, the bringing in of
evidence, the overt statement of opinion and disagreement).
The first extract is less simple.
- I talk to my pupils openly about sex (statement);
- this is daring (evaluation);
- because in Cyprus they aren't keen for people to be that open
Uustifica tion).
- But I think it's important Uustification of the first point);
- the trouble is I don't do it systematically (evaluation of the first
point);
- but no one else does it at all Uustification putting evaluation in
perspective) ;
- sex education is an enormous problem (opinion);
- because making a thing of it creates the wrong atmosphere for
discussing sex Uustification of opinion).
Speaker B's logic does not proceed in an unbroken chain but
returns to deal with the same point (the first point) in a number
 The Politics of Variation 39

of different ways. Although A starts with yes, indicating that he

will take up B's points, in fact he is off on another tack altogether,
and what he says is connected to what has gone before only
loosely, by the general subject of sex education.
So although there is logic in this extract (and there's no reason
why speakers should not change their tack, since this gives
conversation its vitality), it is not so clearly logical as the other
extract.
As these conversational samples show, logic in conversation
does not have to consist in making each utterance follow rigidly
from the last. Complex interconnections, doubling back on
yourself and so on, can be easy to understand in context.
The philosopher H. P. Grice once devised a framework for
talking about logic in conversation. 8 He suggested that speakers
and hearers follow a number of conventions known collectively as
the 'co-operative principle', namely: give as much information as
is required, and no more; give true information or information
for which you have evidence; be relevant; don't be vague,
ambiguous or prolix. If a speaker breaks one of these maxims in
an obvious way, either her utterance will be examined more
closely by the hearer, or else the hearer will derive some hidden
inference. In other words, if people deviate from the maxims,
their hearers still try to make the contribution sensible and
logical. They assume the speaker had a good reason for her
deviant behaviour, and they set out to discover what it was. So
long as hearers believe speakers are being relevant, truthful,
informative and'so on, conversation will flow in a logical way. If
hearers suspect speakers are not being relevant, truthful and
informative, perplexity will ensue. Hence in this sample, the
sequence

how many other people do it you see

Yes it's an enormous problem

is perfectly logical to the participants, though out of context the

connection may not he clear. On the other hand, in the
conversation from which extract 2 is taken we have this:

C: Was she at Styal

D: She came out yesterday
C: She was at Styal
40 Feminism and Linguistic Theory

D: no she was in first Drake Hall.

C does not take 'she came out yesterday, as a logical reply to 'was
she at Styal' because it is not informative enough. She therefore
checks up by repeating her question as an assertion, which D then
denies. To the observer it might seem logical to infer from 'she
came out yesterday' that the answer Yes is pi-esupposed: but for
some reason, C does not make this 'logical' inference, and she
turns out to be right.
What needs to be pointed out here is the uselessness of applying
inflexible notions of logic to so flexible an instrument as
conversation. All conversations viewed in context (i.e. taking into
account the shared knowledge of the participants and the
situation in which they are talking) exhibit strong logical
connections, and on the other hand, all speakers and hearers
make use of essentially extra-logical devices, particularly the
affective cues carried in tone of voice.
It is hard to see that looking for systematic differences in
people's construction of spoken texts would be a feasible or useful
insight into sex difference. So what is the source of the belief, held
in common by chauvinists and feminists, that women's speech is
less logical than men's? At this point we may consider two
suggestions, to which later chapters will return. The first is that in
some modes of discourse - public speaking for example - a very
explicit logical sequencing might be more appropriate than it is
in ordinary casual conversation. These are precisely the speech
modes used less by women than by men, because of restrictions on
women's participation in formal speech events.
The second is more interesting. Since the logic of ordinary
conversation is determined by the participants and may be
implicit in the knowledge they share rather than being objectively
inspectable from the outside, groups with little shared knowledge
will find it hard to follow each other's conversation. Men and
women, with their differing roles and experiences, do not have
perfect shared knowledge, and the interactions of one sex are
mysterious to the other. As usual, however, the more powerful
group (men) have the right to make authoritative value-
judgements, and thus to label women's talk 'illogical', backed up
by the old stereotype of inherent female irrationality in every
sphere.
Cheris Kramarae discusses the possibility that women have to
 The Politics of Variation 41

become attuned to the dynamics of male conversation, whereas

men do not have the same obligation vis-a-vis women_ 9 In that
case the bafflement caused by a lack of shared knowledge would
be peculiar to the male sex.
It is interesting to note that subordinate groups are very often
accused of speaking in an 'illogical' way. Thus educational
psychologists in the 1960s contended that the speech of Black
American ghetto children was 'a basically non-logical form of
expressive behavior'. Linguistic analysis revealed that the speech
in question had different structural rules, which simply had not
been understood by the white educators. Yet because it is the
prerogative of the powerful to evaluate everyone else, the
'illogical' label could be appended and used to disparage Black
American speakers.
The search for logic in language is, in itself, rather uninter-
esting. In some ways language is very much a logical system, so
that a totally illogical language could not be an efficient
communication medium; while on the other hand it is hard to
imagine a natural language logical in the style of the
propositional calculus. But the uses to which the lay concept of
logic in language has been put are political, and for that reason
feminists must be concerned to debunk it.
Perhaps the most widely held feminist idea about women's
speech is that whereas men compete in conversation, women use
co-operative strategies. This is by no means easy to investigate.
However, certain investigators have tried to devise frameworks in
which to look at how people establish hierarchies and make
decisions through talk.
For instance, R.F. Bales designed such a framework for the
analysis of people's interaction in conferences. 1o Bales trained
observers to code everything that was said with labels such as the
following: requesting opinion, requesting information, putting
forward opinion, putting forward information, expressing
agreement, expressing solidarity, expressing disagreement,
expressing antagonism. There are at least two major difficulties
with any framework of this kind. The first is a coding difficulty.
Situations might very well arise where the observer was unsure
which of two categories to put some utterance into. Bales used a
number of analysts and (presumably) looked for agreement
amongst them, but this suggests he was investigating the
metalinguistic abilities of his observers at least as much as he was
42 Feminism and Linguistic Theory

investigating the linguistic behaviour of the conference

participants.
Secondly, there is a problem with the sensitivity of the
framework to situations. Suppose one hypothesised that co-
operative conversationalists would produce a lot of requests for
opinions (bringing others into the discussion) and express
solidarity or agreement frequently. How could the investigator
allow for the possibility that opinions may be requested
antagonistically (for instance, by picking on someone who has no
desire to speak) or that in many kinds of interaction it is normal
for agreement or solidarity to be expressed as part of a more
general strategy of criticism or contradiction?
In this particular sample, it appeared that the women speakers
were notably less co-operative (in so far as that term can be
defined; as I have just pointed out, a lot of qualifications and ad
hoc judgements have to be made when we try to apply any
general framework). While both sexes requested approximately
equal amounts of information and opinion, expressed a good deal
of agreement and no antagonism, women interrupted more,
disagreed more, and back channelled less.
The reason for this unexpected finding seems fairly obvious. It
happens that in the data specimen I chose, the men are discussing
a subject on which they agree. The women, in contrast, do not
agree with each other. A number of facts in the situation
determine whether the participants will be 'co-operative' or
'competitive', notably, if banally, whether they agree, whether
they like each other, what they are trying to accomplish in talk
and so on. Conversation is a highly contextualised phenomenon,
and to generalise about it on the basis of so gross a variable as
speaker sex is unwise.
Why, then, does the stereotype of competitive men and co-
operative women seem so ingrained in feminist folklinguistics
(and indeed in linguistics, in so far as women's conversational
strategies have been studied at all)? Are feminists deluding
themselves about women's less aggressive nature? Are they
picking up and revaluing the stereotype men have of women as
'good listeners'? Perhaps both these things play a part in the
creation of this particular stereotype. But I think there is one
other factor, and that is the extremely co-operative style which
has been institutionalised in feminist gatherings. In such
gatherings it is conveyed to participants that they should not
 The Politz'cs of Variation 43

interrupt nor raise their voices to silence others, that solidarity

should be expressed frequently, that women must give way to
each other rather than competing for the floor, and so on. Long
silences are tolerated.
This style was worked out painstakingly to avoid advantaging
relatively more 'articulate' (i.e. middle-class and educated)
women. That it is a feminist and not a female norm is suggested
by remarks my women informants have made to me about it:

'I had a lot of trouble not interrupting, I felt everyone was

thinking I couldn't keep my mouth shut.'
'It struck me the minute we started, all the silence and
letting people finish.'

Although there are obvious political justifications for the 'feminist

style' I am talking about, it is interesting that many feminists
justify its peculiarities differently. They tell you it is a style that
'suits women better' or gets away from 'male ways of speaking'. In
other words, though the history and anecdote of early second-
wave feminist suggests a difficult and painful process of working
out a suitable style, under considerable pressure from women who
had not been trained to speak in public, this process has already
been obliterated and the style has become naturalised. Once
again, the power of the folklinguistic in explaining and regulating
usage is unmistakable.
Linguists too may have helped to further the idea, if not of
women as co-operators then at least of men as competitors. It is
very noticeable that studies of all-male interaction (for instance
the classic Harlem studies of William Labov) tend to look at
large-scale ritual performances within hierarchical groups,
whereas studies of women focus on small groups in intimate
conversations. This is surely simplistic. Even if fewer women than
men participate in groups and gangs, is it really true that men do
not have intimate conversations? Or are sociolinguists
subconsciously influenced in their choice of data by their own folk
and folklinguistic beliefs about the activities and social
organisation typical of women and men?
The anecdotal tradition is not, of course, without interest. On
the contrary, it is essential that a feminist account of language
take into consideration people's beliefs about male and female
44 Feminism and Linguistic Theory

speech, for the prejudice is often more powerful than the fact.
However, as many linguists have pointed out, a study of sex
differences cannot rely on anecdotes and stereotypes. It must
substantiate claims about women's usage with empirical evidence
(not forgetting that women of different ages, regional, class and
ethnic backgrounds cannot be dealt with in one breath) and that
evidence must be based not on what men say women say, or what
women say they say, but on what they really do say .
. The branch of linguistic enquiry which claims to approach sex
difference in this empirical spirit is modern sociolinguistics. We
must now ask whether sociolinguistic studies have improved on
anecdotal accounts of women's speech.

WOMEN AND SOCIOLINGUISTICS

The major approach to social differences in speech is the large-

scale sociolinguistic survey. A representative sample of the target
population (determined by the sociological techniques of
stratified or random sampling) is interviewed, and recordings are
analysed. Linguistic features of various sorts are then correlated
with social characteristics such as age, gender, ethnicity and
class.
This kind of survey has been done in a number of urban areas,
including New York City, Norwich, Tyneside and Belfast. The
usual focus of interest is 'social stratification', the correlation of
how often certain items of grammar or pronounciation are used
with socioeconomic class. For instance, in New York there is a
stratification pattern for the alternation of that and dat. The
lower your socioeconomic class, the more likely you are to use the
non-standard pronounciation dat rather than standard that.
Stratification is also observable in situations of varying formality:
the more formal the occasion, the more the standard will be used
by members of every socioeconomic group.
It has been argued (e.g. by Pellowe et al.) that stratification
mania does nothing but confirm stereotypes. II Investigators are so
keen to draw the magic stratification graph, they fail to look
beyond the handful of features that produce it. And these
features are, of course, the obvious ones that everyone already
knows. Pellowe believes that in concentrating on the blindingly
 The Politics of Variation 45

obvious, linguists risk missing complex variation which is equally

important.
Feminist linguists have also criticised the stratification
approach:

Both the theory and the methodology are based on the implicit
assumption that the communicative experience of white
middle class males is prototypical ... the experience of women,
other ethnic groups and classes are treated as deviations. 12
Nowhere is the tendency to simplistic norm-and-deviation
models clearer than in one theory that has emerged from
sociolinguistic survey findings, social marker theory (Scherer
and Giles).13 In effect, this theory says that people's speech
contains features that mark them out as belonging to particular
social groups. Some of these features are so consciously
attended to by hearers that they figure in popular stereotypes,
while others have to be isolated by the linguist.

In principle this has a certain merit, since it acknowledges that

there is variation beyond the stereotype and that the correlations
language-users make between linguistic and social indices are
very complex. Yet in practice social marker theory is afflicted
with the same shortcomings we have seen in surveying:
researchers do concentrate on the variations that have an overt
significance for language users, and they do tend to define a
marker as a feature distinguishing its user's speech from an
implicit norm. The recognition that someone is a white middle-
class male, for instance, is assumed to depend on the absence of
markers.
There are three objections to the norm and deviation frame-
work of sociolinguistics. First, researchers notoriously find what
they set out to find, and if all they are looking for is gross and
salient differences perceived by everyone, that is all they will find.
Secondly, if there is an implicit norm, non-standard varieties will
tend to be described as collections of deviations rather than as
integrated systems. Thus the question of how far they are or are
not integrated systems in their own right cannot even be
addressed. Third, the researcher working in this framework will
feel the need to explain why women or blacks deviate from
men/whites, but not vice versa. In other words, it is thought that
non-standard varieties have special qualities requiring explan-
46 Feminism and Linguistic Theory

ation, whereas white male speech is entirely unremarkable.

It is the matter of explaining that I want to concentrate on. For
not only is there intellectual confusion in deciding just what
variations need an explanation, the explanations themselves are
frequently very objectionable. As Jenkins and Kramarae observe,
'Gender, ethnicity and class are seen as "demographic variables"
which can be controlled and accounted for, often by using ad hoc
explanations based on cultural stereotypes. '14

THE SOCIOLINGUISTIC EXPLANATION OF SEX

DIFFERENCE

Sex differences in language are particularly susceptible to being

treated as deviations requiring explanation, for a number of
reasons (all of which rest ultimately on the notion that however
large the female population may be - and in our culture it is
more than 50 per cent - it is men who are normal, not women).
Women, unlike blacks or workers, do not in most western
cultures form a separate community. This marginalises them
from the start, since although they are at least half the
population, in patriarchal society they can easily sink out of sight.
In particular, no large public community of women draws
attention to their existence as a group. Connected to this is the
fact that male and female varieties of language are not normally
analysed as products of differing histories and cultures, since
women's history and culture is thought to be subsumed in men's.
This factor distinguishes the language of women from the
language of minority ethnic groups and many English regional
varieties, which have an acknowledged tradition and often a
flourishing culture too.
But in any case, how far can we speak of 'women's languages'
or 'genderlects' in the same way we can talk of 'Y orkshire dialect'
or 'black English'? Before we turn to English, it is helpful to look
at a case where linguists thought for many years they were dealing
with separate male and female languages. It is from cases like this
that we get the concepts still used now in the explanation of sex
difference.
 The Politics of Variation 47

Carib

The community that raised this problem for linguistics is the

Carib Indians. Here women and men were reported in 1665 to
use different phonologies (sound systems) and lexicons (vocabu-
lary). The missionaries who originally observed this phenomenon
gave a historical justification. According to them, the island had
been invaded by the neighbouring Arawak tribe, who extermin-
ated the male Caribs and kept the women only for reproductive
purposes. Both sexes retained the languages they had spoken
originally, mothers passing Carib on to their daughters, while
fathers taught Arawak to their sons.
Later linguists, however, preferred an explanation based on
gender role rather than biological sex. Noting a gender-based
division of labour, analysts argued that women and men had
developed different lexicons because they did different things.
Jespersen elaborated this with a description of 'primitive society':
while men, out hunting, developed a language of sparse and
economical remarks designed to co-ordinate activity, women
indulged in idle chatter at home. Jespersen here sets a high
standard for sexist non-explanation, but his emphasis on role has
become standard in sociolinguistic explanation.
Social marker theory distinguishes between two types of sex
marker, the sex-exclusive (where certain features are used only by
one sex, as in the Carib case) and the sex-preferential (where
features are common to both sexes, but are more likely to be used
by one sex than the other). In the case of sex-preferential markers
in particular, there may be a strong correlation with occupation
rather than with being a man or a woman per se.
A simple example from our own culture is 'motherese', the
language women use to young children, which has a distinctive
lexicon and features like high pitch, loudness, slowness and
exaggerated intonational contours. It is a sex-preferential variety,
used much more by women than by men. But this is just because
women, rather than men, take care of children. In so far as men
have anything to do with childcare, they too can be observed to
use motherese.
Sex-exclusive markers may reflect either a very strict division of
labour or the operation of taboos on speech (usually the speech of
women). There are societies, for instance, where if women want
48 Feminism and Linguistic Theory

to use words that sound similar to the names of tabooed male

relatives, they must resort to euphemism and circumlocution.

The case of English

In English, we tend to find sex-preferential rather than sex-

exclusive markers. The main difference identified in socioling-
uistic surveys on both sides of the Atlantic is that women's speech
in every socioeconomic class deviates less from the prestige
standard than men's. This applies both to grammatical form and
to pronunciation.
Obviously, a difference of this type does not lend itself to
explanations from either history or occupation, and several other
reasons for it have therefore been advanced, based on supposed
characteristics of females. Four in particular are very common:
conservatism (whereas men innovate, women stick to the
traditional standard forms); social climbing (women are more
sensitive than men to the social meaning of speech, and imitate
prestige usage in order to elevate their social status); feminine
identity (it is feminine to 'talk like a lady', i.e. like a middle-class
speaker) and covert prestige (in fact, non-standard speech is
considered a sign of masculinity, and males cultivate it, so that no
real prestige attaches to the more standard speech of women).
These four explanations of why women's speech deviates less
from the prestige standard beg a number of questions and raise a
number of problems. It seems necessary to ask, for instance,
whether they are adequate accounts of the observed facts;
whether they embody covert value-judgements downgrading
women, and whether any alternative explanations are possible.
It is immediately obvious that the four proposals are not all
compatible with each other. For example, women could not be
both conservatives who did not innovate, and social climbers who
sought to elevate their status by imitating prestige usage.
Moreover, either of these possibilities - conservatism or social
climbing - would require further explanation. Why are women
more conservative than men? Or alternatively, why are they more
aware than men of social norms, and more ambitious/ conformist
in their attitudes?
Certain other puzzles arise in relation to women's conservatism.
Linguists who claim that women are conservative must believe
 The Politics of Vanation 49

this is a consequence of femininity itself, rather than

subordination: for the groups most associated with the direction
of linguistic change (i.e. the most innovative groups) are
specifically subordinate groups, like the young and the lower
middle class, who are said to suffer from linguistic 'insecurity'.
Labov, the New York City researcher, has pointed to lower
middle-class women in particular as trend-setters. Their pattern
of style-shift (changing the way they talk when they become aware
of it in more formal situations) suggests they are very sensitive to
prestige markers in speech, and Labov speculates that they are in
an especially good position to disseminate changes, since they can
pass them on to their children. This supports the 'social climbing'
theory rather than female conservatism. Moreover, cross-cultural
evidence suggests that conservatism is not a universally feminine
trait. In some patriarchal peasant societies, Madagascar for
instance, it is women whose use of language breaks traditional
rules.
The link between prestige usage and perceived femininity is
more promising. It is anecdotally attested in many communities
where one finds agreement that certain speech styles are all right
for boys, but you wouldn't want your daughter/sister/girlfriend
talking like that. 'Like that' in this context means swearing and
coarseness, but also, and especially, pronunciation. A strong
regional accent, which in England signifies either membership of
or sympathy with the working class, is more acceptable in men
than in women. This ties in with the idea of 'covert prestige'
invented by the Norwich researcher, Peter Trudgill. Trudgill
found that working-class speakers actually valued non-standard
speech as a symbol of masculinity, and he advocated that we
should reverse our class-based notion of prestige for these
speakers to take account of the truism that anything men do is
automatically more prestigious than anything women do. In
Trudgill's opinion, for most working class speakers the prestige of
standard English is overridden by the need to identify as a real
man.l"
If it is true that non-standard speech connotes masculinity,
then we would expect to find femininity being constructed in
deliberate opposition, as it is in so many other areas. This would
explain the strong pressure put on women to talk like ladies, and
the contempt reserved for those who intrude on the realm of the
masculine with their strong regional accents.
50 Feminism and Linguistic Theory

VALUE JUDGEMENTS

It is very noticeable that the most popular explanations of

women's more standard speech, conservatism and social
climbing, are also very sexist, couching the qualities predicated of
all women (itself a sexist strategy) in terms which downgrade
them relative to men.
For instance, take the idea that men are innovative and women
conservative. In our culture, innovation is valued, so this
proposition implicitly praises the behaviour of men: the same
thing could have been expressed as women's respect for tradition
and men's cavalier disregard for it. In Madagascar, where things
are exactly the other way around and it is men who are conser-
vative, this is precisely how it is expressed. Men are guardians of a
revered tradition, women the upstarts who can't be trusted to
speak well in public. 'The use of direct speech, such as that of
women, is associated with a loss of tradition. '16 Unsurprisingly,
'Men alone are considered to be able speech-makers'Y The
notion of covert prestige is also pervaded with sexism, though it
does have the virtue of recording a prejudice that really exists.
Trudgill is really saying that where women's language has to be
defined as more prestigious than men's we should change the
definition of prestige. The Madagascar example shows he is right:
societies frame their definitions of prestige in accordance with the
social and sexual hierarchy.
One possibility which is rarely addressed by linguists is that the
finding itself, that women deviate less from the prestige standard,
may be partly an artefact of the methodology used. Two aspects
of sociolinguistic methodology particularly invite a critique.
The first of these is stratification. Sociolinguistic surveys are
obviously dependent on some kind of stratifying procedure to
determine which speakers belong to which socioeconomic class,
and the criteria are normally occupational - unskilled versus
skilled manual work, white collar versus professional. But these
criteria are not usually applied to married women. Because the
family is taken as the unit of stratification, and the key to the
family'S status is the occupation of its 'Head', married women are
usually assigned to classes on the basis of their husband's
occupation.
It has been argued by the feminist sociologist Christine Delphy
 The Politics of Variation 51

that this stratification procedure is a grave defect, and persists

only because it succeeds in obscuring essential mechanisms of
patriarchy. IS It places men and women in the same class and thus
affirms their parity, yet this parity, which is only a by-product of
marriage in any case, conceals the economic dependence of wives
on their husbands.
Delphy's critique has certain implications for the sociolinguist.
Although the distortion may not be very great, it is possible that
stratification methods which assign women to classes according to
their husbands' occupation distort the picture. The differences
between men and women may partly reflect the fact that if one
used other criteria for the stratification - educational attainment
for instance - married couples might not turn out to be parallel
at all.
The other problematic area is accommodation to the inter-
viewer. It is well known that people speak differently in more and
less formal situations. In a formal situation like an interview they
will use a more standard variety than when they are not
monitoring their behaviour so carefully. The style produced in
casual situations is known as the 'vernacular', and contains fewest
prestige features. Sociolinguists are more interested in this style
than in any other, but it is extremely hard to collect samples of.
Being interviewed in a sociolinguistic survey is not a casual
situation, and for most speakers the interviewer will be a social
superior rather than a peer with whom the vernacular might be
used. Therefore researchers have difficulty in eliciting anything
less formal than an interview style.
Labov, the New York researcher, used a number of strategies
to try and elicit the vernacular. He asked questions about
traumatic experiences in the hope that informants would get
carried away and forget to monitor their speech. He recorded
incidental talk with family members who did not know the tape
was running. And when he studied the language of Harlem
gangs, he used a young, black fieldworker, since he knew the
black vernacular was unlikely to be used with a white middle-class
academic.
Since Labov, it has been considered desirable to minimise
social distance between subject and interviewer (in a number of
surveys local interviewers were used). Yet no one has extended
this precaution to gender, or investigated whether it makes any
difference. In talking to men, it could be argued, women are
52 Feminism and Linguistic Theory

always talking to their superordinates, and perhaps the

vernacular they might use with each other is just very seldom
elicited.
It should be clear by now that while sociolinguistic surveys
produce rather more reliable findings than the anecdotal
tradition, the explanations they put forward to account for those
findings have a tendency to be ad hoc and sexist. How then do
feminists reinterpret the findings of sex difference studies? Can
we be sure sex-linked variation has any political significance at
all?

DIFFERENCE AND DOMINANCE: THE IMPLICATIONS

OF LINGUISTIC VARIATION

For some linguists, the varieties spoken by men and women are
rather like regional dialects or age-linked varieties: they reflect
the differing socialisation of women, the fact that women interact
more with each other than with men, and the existence of
particular feminine identities or gender roles which generate their
own norms of speaking and behaving. This 'subcultural' view of
sex difference (espoused for instance by Kramarae,19 and Maltz
and Borker,20) mayor may not stress the position of female
subcultures in the power hierarchy. Some commentators point
out that women are not entirely free to develop their own
subcultural norms, because the femininity they are socialised into
and seek to express is to a large extent male-defined, while others
give the impression that power is merely an extraneous factor
grafted on to the subcultural ecology.
Other linguists, however, see a much more direct link between
the position of women in particular societies and the actual
differences between their speech and men's. Sex differences for
them must be given a primarily political explanation, and the
speech style forced upon women must be shown as inherently
oppressive.
Proponents of this view (e.g. Lakoff)21 describe female speech
as a function of women's linguistic insecurity, which is ultimately
a reflection of their social marginality and need to be seen as
deferring to male norms. The phonetician Caroline Henton, for
instance, sees this insecurity as the key to understanding
 The Politics of Variation 53

differences which on the face on it have no obvious relation to

oppression at all: differences in the pitch and vowel-quality of
women and men that are either too great or of the wrong kind to
be explained by any sex difference in vocal-tract anatomy.22
Henton's explanation is that women perceive a norm which
they then overshoot in their production of vowels. She sees the
fact that pitch and vowel quality are not under conscious control
as very important, since it indicates how deep the insecurity goes;
she also points out that since these differences are instrumentally
measurable, they cannot be the misreported products of stereo-
types held either by the informant or the analyst.
Although systematic non-innate differences in something so
fundamental as vowel-quality do indeed call for an explanation,
and that explanation is unlikely to have anything to do with
conscious motives and attitudes, no amount of instrumental
measuring can tell us what the explanation is. Henton's account is
more acceptable to feminists, and indeed to scientists, than the
folkphonetics of Swift, but the problem of what causes or
motivates the differences remains.
But is this the crucial problem in the politics of linguistic sex
difference? It may be interesting to discover the origins and the
meaning of differences but difference itself is not the key to
women's oppression through language. The key is value-
judgement; the way difference is perceived, the consequences it
has. In looking at these things we need make no distinction
between real differences and folklinguistic stereotypes: indeed we
can bring them together to produce an integrated account of
what sex differences mean in our society.
We may consider, for instance, the fact that women are some-
times discriminated against ostensibly because of the way they
speak. In some jobs they must be more 'well-spoken' than a man
would have to be; in others, especially broadcasting jobs, they
may be told that female voices are too 'tinny' or 'high' and that
they 'lack authority'. It has even been suggested that the
reluctance of juries to believe women in court is partly because
their range of intonational and pitch contrasts (which is greater
than men's) suggests hysteria.
Yet it would be native to suppose these complaints have any
real linguistic substance. Some of the unpleasant quality of
women's broadcast voices (supposing it exists) may be due to
filters on recording equipment which turn up the treble (as male
54 Feminism and Linguistic Theory

voices require) rather than the bass. Audio equipment was not
designed for women. But pleasantness to a large extent resides in
the ear of the listener, and predictably enough, there is
widespread prejudice against women's voices.
As for 'hysterical' and 'lacking in authority', it is inconceivable
that these judgements can ultimately have anything to do with
pitch. If men talked in higher voices than women, low voices
would be said to lack authority. The point is that in our culture,
anything that marks a speaker out as female becomes a cause for
complaint and a proof of inferiority.
Linguistic sex-differences act simply as a badge of femaleness,
and are valued negatively quite irrespective of their substance.
Explaining that women's high pitch is learned/can be authorit-
ative/is really very pleasing to the ear will have no effect on the
irrational process by which everything 'female' is pejorated,
whether it actually reflects women's behaviour or not.
A particularly inept sociolinguistic study reported in Smith
1979 will serve as an illustration. 23 An experiment was set up to
investigate whether, in a courtroom, male and female speech
styles would be found equally credible by a jury. Two versions of
the same narrative were proposed, one a 'male' version using the
pronunciation, tone of voice and grammar supposedly typical of
men, and the other a 'female' version similarly treated. The
female version was judged less credible even when delivered by a
carefully trained man, and the experimenters therefore
concluded that the lack of credibility lay in the language and not
the speaker.
What they overlooked, however, was simply that the 'female'
version did not reflect women's actual speech style, but only the
stereotyped style people ascribe to them. It may well have failed
to reflect any real speaker's style, and its lack of credibility could
have been predicted from that alone.
We have already examined the tendency of feminists to make
unwarranted statements about female speech style, and to believe
in simplistic stereotypes. It appears that in addition they tend to
accept the normal valuations of various linguistic features as signs
of powerlessness, deference and so on, concluding that men have
forced women to learn an inherently powerless speech style which
reinforces their actual powerlessness. The truth is that although
women may be forced to learn a style of speaking that different-
iates them from men and identifies them as women, the labels
 The Polz'tics of Variation 55

which condemn this style embody not obvious truths but value
judgements which would be applied to anything female. As the
saying goes, 'a woman's place is in the wrong', and this is as true
of language as it is of anything else.
Let us take the much argued case of tag-questions as an
example for discussion. Many feminists are familiar with Lakoffs
assertion that the tag-question is characteristic of female speech
because it is 'approval-seeking', half-way between a statement
and a question. It moderates the offensiveness of a woman's
telling a man something straight out.
In fact, it turned out on investigation that women's attachment
to the tag-question could not be confirmed empirically. Dubois
and Crouch found men using it rather more. 24 Dale Spender was
very concerned about this finding, and tried to explain it away:
she pointed out that since it is hard to tell tentative tags (that'll be
all right, won't it?) from forceful ones (don't do that again, will
you?) maybe women used the tentative ones and men the forceful
ones. Spender also observed that no one tried to explain the new
finding that men used tag-questions more, by saying they were
seeking approval. 25
In fact, we can surely account for the second point in terms of
what the first implies, and the account is a revealing one. Out of
context, it simply is not possible to state once and for all what a
tag-question means. It could be tentativeness, forcefulness,
checking information (the concert starts at eight, doesn't it?)
patronising someone (that can't be the case, can it?) and any
number of other things. Whatever function we ascribe to tag-
questions or any other grammatical form, can only be a rational-
isation after the event. If that is true, it is not surprising that
when men are found using tag-questions the rationalisation does
not include approval-seeking. The cultural stereotype of men tells
us they do not need to seek approval.
Lakoff was not, as Spender implies, attempting to prove that
women's speech lacked effectiveness, but that it contained
tangible evidence of their need, which she took for granted, not
to be seen as over-assertive by men. The mistake lies in equating
linguistic forms with extra-linguistic needs in a one-to-one
correspondence: tag-questions do not always express a need for
approval, and the need for approval is not always expressed by
tag-questions.
The argument here can be summarized in three points.
56 Feminism and Linguistic Theory

l. Women do not use more tag-questions than men.

2. Even if they did, it would not necessarily mean they were
seeking approval, since tag-questions have a range of uses.
3. In any case, women's use of tag-questions will always be
explained differently from men's, since it is cultural sex
stereotypes which determine the explanation of linguistic
phenomena, rather than the nature of the phenomena them-
selves.
In this chapter I have tried to show why paying attention to the
metalinguistic (what people say about talk and how they analyse
it) is a crucial part of understanding the significance of sex-
difference and the ways in which linguists have dealt with it. In
Chapter 4 I propose to look at a different aspect of metalinguistic
politics: the way in which grammarians and linguists have
projected a malelfemale dichotomy on to the languages of the
world, and their attempts to use grammar as a tactical weapon in
the battle of the sexes.
4 False Dichotomies:
Grammar and Sexual
Polarity
Theoretically, it is possible
to classify any phenomenon as
male or female.
Man'elouise JanssenJurreit
Thought has always worked by
opposition,
Speech/Writing,
High/Low ...
Does this mean something? Is
the fact that logocentrism subjects
all thought - all of the
concepts, the codes, the values -
to a two·term system, related to
'the' couple man/woman?
Helene Cixous

Yin and yang; animus and anima; as the epigraphs to this

chapter point out, a tendency to classify the universe by an
opposition of male and female principles recurs in patriarchal
thinking. But is the opposition such a true and fundamental one?
A basic insight produced by the feminist theory of this century
(we owe it to Simone de Beauvoir1 and it has been developed by
Luce Irigaray)2 is that women in patriarchy are constructed as the
Other - as whatever men are not. If man is active, woman is
passive; if he has the phallus, she simply lacks it. Femininity is
masculinity inverted.
Irigaray points out the reductive inaccuracy of this concept
when applied, say, to female sexuality. Women are different
from men, but not opposite to them, and the binary oppositions
which locate them at one end of a male/female polarity are
artificial, reflecting both the exclusion of women from the

57
58 Feminism and Linguistic Theory

making of what counts as knowledge (philosophy and science, for

instance) and the dominance of one particular sort of science. For
as Jane Gallop observes, in a scientific model 'otherness is
suppressed to preserve the theory's consistency. Theory's
authority is guaranteed by its consistency. '3
This chapter is about the projection of the male/female
opposition on to language and especially on to the analytic
construct, grammar. I want to consider whether it is a natural or
reasonable practice: whether it has ill effects for women: how
feminists have responded to it and whether their criticism is well-
founded.
The binary opposition has a special status in linguistics. The
reader will recall that Saussure placed great stress on contrast as a
principle of linguistic structure, and that he drew a number of
binary distinctions himself (langue/parole, synchronic/dia-
chronic, syntagmatic/paradigmatic). Since Saussure, many other
linguists have found the two-way contrast useful in their analyses
of linguistic systems from phonology upwards, and it has become
institutionalised. Students learn early to look for binary
oppositions rather than, say, three-term contrasts, in language,
::.tnd may even be told what some theorists maintain, that to do so
is a universal property of the human mind. After all, why should
these contrasts keep turning up if they are not somehow 'natural'?
It is useful to separate out the two questions which arise from
this account. The first is whether we should reify (treat as a thing)
the binary opposition as so many writers seem to do. Do opposi-
tions exist in the language to be discovered by linguistic science,
or are they invented as a handy way of analysing language? In
other words, is the binary opposition principle a claim about
linguistic structure or a claim about what analytic strategies most
naturally come to the linguist?
The second is a nature/nurture question. If there really is a
tendency to think in opposites, whether or not it reflects the
existence of those oppositions in the world, is this tendency innate
in the human mind, or is it inculcated by upbringing and
education? When we look at the reasons why linguists have
asserted that the binary distinction is innate, they seem far from
convincing. The first reason is that since the binary opposition is
basic to language, this must reflect a property of the mind: which
is surely circular, as it takes for granted the very point (that
languages really are organised around two-way contrasts) which is
 False Dichotomies 59

at issue. The second reason derives from information theory, a

subject centred on studying the transmission of messages. In
information theory, a maximally efficient communication system
employs two-way contrasts because they are easy to process,
involving only one decision, whether an item is 'x or not-x'. But
however useful this idea may be when applied to machine lang-
uages, it is not an adequate account of human communication.
In any case, linguistics habitually equates very different sorts of
opposition. There are choices in language that fit the 'x or not-x'
pattern, but there are others that do not. This is especially
obvious when we analyse languages at the level of semantics
(meaning).
The semantic model most obviously dependent on binary op-
positions is called componential analysis. The aim of it is to
reduce the meaning of a word to a series of supposedly primitive
features for which it has either a plus or a minus value. So the
word animal, for instance, would be given the features + animate
and - human. Some exponents of this theory believe that the pri-
mitive features are innate, and that making binary classifications
of this type is an important element in human cognition. 4
It is clear that when we assign features like animate ness and
humanness the choice is an either/or one (although it is inter-
esting to speculate on how speakers would assign intelligent
robots). But consider another pair of opposites, old and young.
Componential analysis would handle them with one feature,
± old. But although they are dictionary antonyms, the qualities
they denote are not really opposed: rather, they are on a
continuum. Just because a middle-aged person is not young
( - young) she is not necessarily old.
The linguist who equates oppositions where yes and no cover all
the possibilities with those where degree must be distinguished
has been fooled by the language. Because she knows (has learned,
in fact: most children do a lot of formal learning of antonyms) the
words are 'opposites', she does not bother to check that their
meanings are.
The male/female opposition obviously does not fall into this
category, since anatomically speaking, every animate being must
be either male or female. But it is not at all obvious that the
feature ± male captures any awareness of this on the part of a
speaker when she defines various words. In other words, do the
creatures either side of the gender division have opposite qualities
60 Feminism and Linguistic Theory

that are relevant to people's conceptualisation of them, and to

their use of words?

PLUS AND MINUS MALE: NEGATIVE SEMANTIC SPACE

In Man Made Language, Dale Spender takes issue with the

componential analysts and claims that their practices show the
relegation of women to a 'negative semantic space' in the English
language. The practice she cites is the common way of repres-
enting words with a female component in their meaning, such as
girl and wife, not as + female but as - male.
Spender rightly remarks that this is sexist (after all, it is equally
correct to represent males as - female; all that matters is that one
duality should define all the possibilities and that everything
should have either a plus or a minus value in respect of it). But
she seems to feel that the bias is not only in the analyst but
actually in the language, implying that women are classified by
speakers of English as the negative of men. 'One of our funda-
mental rules for making sense of the ... world is ... that the
male represents the positive while the female, necessarily then,
represents the negative.'5 It is possible to adduce a good deal of
evidence in support of this claim, and arguably the practices of
componential analysts constitute just such evidence. But we are
on shakier ground if we claim it is a rule of language itself that
female terms have a negative meaning. This is to reify mere
(though sexist) terminology. It is woolly thinking to imagine that
linguistics has discovered a Law instead of acknowledging that
- male is a theoretical invention reflecting the bias of its creators.
There is no claim implicit in the use of - male that male is the
norm and female the exception. It does not matter which end of a
polarity is + and which -, though of course the choice reflects
our conventional priorities (which are not, or not only, linguistic).
To say that the language embodies value judgements based on
these plus and minus categories is a theoretical vulgarity even the
most chauvinist linguist would repudiate.
The argument which is germane here is one we have already
touched on: how accurate is it, both objectively and as a
representation of lay knowledge/belief, to construct a binary
opposition between male and female built into the meaning of
 False Dichotomies 61

words? Might this not be a false dichotomy which persists, and

not only in language, because of the amount of political energy
which has been invested in it?
Yet it might be said that although the dichotomy is an over-
simplification, the structure of language forces it upon us.
Language's natural binarity predisposed us to pair words like
male and female and to treat them as opposites. The theme of
this chapter is the way in which grammar and its practitioners
have fooled us - sometimes deliberately - by convincing us that
artificial and theoretical biases are really inherent in the
language. To the extent that grammar does this, it actually alters
the way we look at the world, and feminists must expose the
ideological task it is carrying out. Opposites are a good place to
start, since it is not hard to see they are a blatant example of
'fixing'.
Readers will probably recall learning at school the concept of a
lexical opposite and a list of pairs exemplifying it. It would be
very strange if these pairs were innate in anyone's mind, since
they are extraordinarily conventional, heterogeneous and in
many cases opaque. Who as a child could have said why black
and white are opposites? Day and nz'ght, though more obvious,
are quite different, and sometimes (e.g. in the context of clothes)
evening is the opposite of day. There are polar opposites like
North and South, and pseudo-opposites like long/short. The
principle of duality, rather than being innate, has to be etched on
the conceptual apparatus of the child.
Lexical antonyms, then, cannot be generated from a single
principle, do not hold good from context to context and vary
from culture to culture. Therefore it is an exercise in futility to
use them as representations of what people know about word-
meaning, or to claim they are innate. The emphasis placed on
learnt'ng antonyms suggests that the urge to dichotomise is
secondary indoctrination rather than native habit, typical not so
much of ordinary talk but of the systems our societies have to
teach us, of logic and dialectic. And even if dichotomy were a
strong tendency in human thinking, it would hardly follow that
particular pairs were innate. It seems unlikely, also, that
dichotomised thinking is engendered by language and the binary
nature of linguistic meaning, because models like componential
analysis which are based on binary choices are very bad at
representing what people know about the differences between
62 Feminism and Linguistic Theory

related words. Simple 'x/not-x' choices are very rarely

appropriate.
So it seems pointless for feminists to pursue the sexism of
componential analysis: rather, they should dismiss it, and its
system of oppositions. as a fundamentally wrong account of
meaning. The male/female opposition itself, of course. remains
of cultural and political interest to feminists. As Helene Cixous
points out, it is felt in many societies to be the fundamental
dichotomy: but the explanation lies mostly outside linguistics.
reflecting a very general and conscious patriarchal policy of
constructing a sexual dichotomy in every area of human
experience.

THE CASE OF GENDER: GRAMMAR AND SEXIST

IDEOLOGY

Perhaps the oldest and most familiar example of a male/female

dichotomy in language is the grammatical category of gender.
Gender has a long history of provoking controversy among
grammarians, philosophers. philologists and social theorists: not
surprisingly, it is often discussed by feminists in the contemporary
debate about sexism in language.
Yet despite centuries of argument, the issues (linguistic and
political) that gender raises seem to have been confused rather
than resolved with every fresh attempt at explication. What
follows here is an attempt to clarify the notion of gender and to
deal with the historical and current controversies from a feminist
point of view. I particularly want to demonstrate that the
supposedly objective intellectual pronouncements of traditional
grammarians and linguistic scientists alike have always
participated in, and been coloured by, the general ideology of
male superiority and sexual difference.

What is gender?

Like a lot of our traditional grammatical terms, gender

originated in the linguistic scholarship of Greece. It is usually
credited to the sophist Protagoras. The word itself is derived from
 False Dichotomies 63

a word meaning 'class' or 'kind' and was used to divide Greek

nouns into three classes, traditionally labelled masculine, feminine
and neuter.
Greek is one of the many languages which possess what is
known as grammatical gender. The three-way classification
masculine/feminine/neuter does not reflect any common· sense
division of word meanings into male, female and inanimate, but
rather the fact that nouns can behave in three different ways
when it comes to the agreement of adjectives, the choice of
article, replacement by a pronoun and inflectional patterns (word
endings). Masculine, feminine and neuter are merely labels for
formal properties of words, and do not reflect word meaning at
all (thus in German, the word for a girl is neuter, while that for a
turnip is feminine).
Not all languages exhibit this kind of noun classification.
Some, like Chinese, have nothing we could label gender at all,
whereas others, like the African Bantu languages, have more
than three types. English is an example of a further type, usually
called natural gender. What this means is that only those words
which refer to something with a biological sex can be masculine
or feminine. The vast majority of words will be neither (or, if they
refer to people, will have the ability to be both, which is called
common gender: words like friend and driver have common
gender). It is easy to see why English speakers regard this
arrangement as natural: it appears to have its roots in common-
sense reality, whereas the foreign habit of assigning a gender to
anything from a turnip to a sideboard strikes us as nothing more
than a bizarre morphological whim.
Feminists, contemplating such usages as 'Man is unique among
the apes in that he grows a long beard, and it is to this that he
owes his superior intelligence'6 have pointed out that English
gender is natural only if you are a man. This observation marks a
step in the right direction: some aspects of, say, pronoun
replacement (e.g. that ships and cars are she) cry out for a
cultural rather than nature/grammar explanation. But what is
really needed is a critique of these notions subsumed under the
heading of gender. What does gender have to do with sex, and
why do linguists work hard at obfuscating that issue? How natural
is natural gender, how common is common gender and how
grammatical is grammatical gender?
64 Feminism and Linguistic Theory

Grammatical gender

Grammatical gender has inspired innumerable articles and

speculations from the ancient Greeks to the modern linguists, but
it seems not to have preoccupied English-speaking feminists
(Dale Spender, for instance, dismisses it in two sentences as a
preamble to her remarks on English gender)_ It is clear, however,
that grammatical gender is a feminist issue, if only because
remarks on its origin and nature have so frequently been part of
male arguments against women's rights. The German feminist
Marielouise Janssen-Jurreit sums up a long historical debate like
this: 'For two thousand years there has been an unresolved
question: whether word endings of nouns and pronouns, as well
as articles, are an extension of the qualities regarded as male or
female. '7 For a number of reasons, contemporary linguists reject
the idea that there is any necessary connection between
grammatical gender and sex. Lyons is typical in asserting that
while gender does reflect some natural semantic classification of
the world, by shape, colour or texture, for example, the crucial
factor is 'not necessarily sex'.8 But Janssen-Jurreit's survey of the
debate makes it quite clear that previous scholars, from the Greek
sophists to the nineteenth-century philologists, would not have
agreed with Lyons. On the contrary, it seems to have been
generally believed that grammatical gender is assigned on the
basis of natural sex difference, and clear exceptions to this rule
were thought aberrant, in need of explanation and possibly even
reform.
Thus Protagoras, the pioneer of the gender concept, made
various attempts at linguistic reform, claiming that gender should
be assigned on some logical and consistent basis. For this he was
ridiculed by Aristophanes, and the attempt was not a success.
Nevertheless, the idea behind it persisted. The Germanic
philologist Grimm, for instance, addressed the question of what
abstract criteria determine the gender of a word. Grimm's theory
was that grammatical gender is a later stage of natural gender, a
more advanced form, as it were, of the common-sense
classification which takes into account only biological sex.
According to this theory, languages develop grammatical gender
when their speakers pass from mere recognition of male and
female creatures to the postulation of abstract male and female
principles in whose terms everything and anything could be
 False Dichotomies 65

classified. Grimm spelt out exactly which qualities went with

which gender. He felt (the girl and the turnip notwithstanding)
these qualities were present in the meanings of words: 'The
masculine means the earlier, larger, firmer, more inflexible, swift,
active, mobile, productive; the feminine the later, smaller,
smoother, the more still, suffering, receptive.'9
Having used stereotypes of male and female qualities to explain
grammatical gender, writers in some cases turned to gender for
guidance on the 'natural' attributes of womankind. According to
Janssen-Jurreit, even pro-feminist Theodor Hippel believed
women were irrational on the grounds that the German word for
reason is masculine in gender. (Presumably Hippel would have
allowed that in France, where the word for reason is feminine, the
men are irrational!) That this type of argument could even be
proposed demonstrates the close correspondence which was taken
for granted between grammatical gender distinctions and sex
differences.
Another point which was taken for granted was the relative
value of the three genders (masculine, feminine, neuter) and it is
clear that some communities in their use of language exploit this
ranking to encode the inferiority of women, or of some women
relative to others. In Konkani, an Indian language, women of low
status because of youth or widowhood are referred to by neuter
pronouns, only married women getting the courtesy of the
feminine. 1o We may link this phenomenon with the frequently
attested reluctance of successful French and German women to
accept titles such as 'Stadtssekretlirin' which have been feminised,
on the grounds that these downgrade them: ' ... the successful
Madame prefers to be Ie Docteur, Ie Professeur, I'Ambassadeur
and Ie Philosophe, even with the succeeding il which rs required
in formal texts.'ll Whatever the origin of grammatical gender, it
cannot be true that it has nothing to do with sex, since a long lay
and linguistic tradition that it is a matter of sex clearly allows it to
be pressed into service as a marker of patriarchal values and
relations.

Natural gender

So far our interest has focused on the claim that grammatical

gender is actually natural, reflecting immutable realities of sex
66 Feminism and Linguistic Theory

difference. Recently, however, this claim has been turned on its

head in an exactly symmetrical fashion, as English-speaking
linguists attempt to deny that 'natural' gender has anything to do
with the realities of a sexist society.
The classic statement of this position comes from the linguistics
faculty at Harvard University. When theology students objected
to the use of male pronouns to refer to God, seventeen members
of the faculty wrote to the Harvard Crimson with the following
statement about what they wittily dubbed 'pronoun envy'.

Many of the grammatical and lexical oppositions in a language

are not between equal members of a pair, but between two
entities one of which is more marked than the other .... For
people and pronouns in English the masculine is unmarked
and hence is used as a neutral or unspecified term. . .. The
fact that the masculine is unmarked in English (or that the
feminine is unmarked in the language of the Tunica Indians) is
simply a feature of grammar. 12

It is difficult to imagine why these distinguished academics would

have thrown their considerable intellectual weight behind such a
statement unless they found the issue threatening. Even without
the ominous noises of axe-grinding, however, it would be
necessary to dismiss the learned gentlemen for the essential
incoherence of their argument.
Marking theory is the newest red herring in the gender debate.
Even feminist scholars like Dubois and Crouch fall under its
technical spell, and rather sneeringly remark of their non-linguist
sisters,

they may be forgiven for failing to understand the theory of

marking, which explains that he can be unmarked for sex in
certain contexts but marked for it in others. Even now,
statements about 'the essential absurdity of using the same
symbols for all the human race in one breath and for only half
of it in the next' betray ignorance of this principle widely
operative in languages. 13

It is therefore worth looking at marking theory and at its

application to gender. Marking - or the use of the terms marked
and unmarked - implies that linguistic elements are like the
 False Dichotomies 67

animals on Orwell's farm: some are more equal than others. Since
the rise of a linguistics centrally concerned with universals, the
marking claim usually involves asserting that some elements will
always be found earlier or more frequently than others. An
example is that the vowel lui (as in boot) turns up in the
languages of the world more often that Iyl (as in French lune).
Any language that has Iyl will also have lui, whereas the reverse
is not true.
It is hard to see how this claim could apply to gender, even if
we allow it is a real category and not a linguist's construct. For
some languages have the masculine unmarked, others like Tunica
the feminine. The genders do not form a hierarchy across
languages as lui and Iyl are said to do. So the claim about
gender-marking must be language specific. What, in that case,
does it imply?
By looking at various precedents for the use of marked and
unmarked in linguistic analysis, it can be seen that linguists
commonly make use of three criteria to justify labelling some
variant unmarked relative to other marked variants. The first of
these is precisely the ability to be used generically, in a way that
includes or subsumes marked variants. But in the case of gender
this is an entirely circular claim: he is unmarked because it is
generic, but it is generic because it is unmarked. We are no
further on with the explanation. What factors determine that
something can be used generically, and are they entirely
linguistic?
The second criterion for labelling something unmarked is
relative neutrality of meaning. But in the case of gender what can
this be but a social value judgement? To whom does he sound
more neutral than she, and why?
Thirdly, we have the criterion of frequency of occurrence. If
something is found more frequently than its alternants, linguists
are liable to say it is unmarked by comparison with them.
It is at this point that we should recall the grammar book
dictum that English gender is not grammatical but natural, that
is, dictated by semantic rather than formal characteristics of
words. If this account is correct (and there is no reason to suppose
that English has gender in the same way French and German
have it, nor that Dubois and Crouch are unaware of this) we have
no business explaining the relative frequency of masculine and
feminine pronouns in terms of universal, or even arbitrary,
68 Feminism and Linguistic Theory

grammatical rules. If gender is fixed extra-linguistically by sex-

reference, then the occurrence of more masculine than feminine
pronouns, the generic, neutral and unspecified male, requires an
extra-linguistic justification too.
And the actual explanation delivers the coup de grace to the
Harvard faculty. One of the most thoroughly documented pheno-
mena in feminist research has been the rise of sexist practices in
prescriptive grammar, and it is this which turns out to be at the
bottom of unmarked he. Since at least 1553, when one Thomas
Wilson asserted the precedence of masculine nouns and
pronouns, grammarians have been attempting to eliminate the
tendency still present in ordinary speech to use they as a singular
for generic or unspecified referents (as in 'you can't blame a
person if they get angry about sexist grammar'). John Kirkby's
Eighty Eight Grammatical Rules of 1746 stated that the
masculine is more comprehensive than the feminine, and this
view found its way into the statute books by 1850, when he was
held legally to stand for she (though this kind of formulation
always leaves open the possibility that women will be excluded
from the definition, for instance when they claim the rights of a
person or cittzen).
In all these grammatical treatises (which are given a full
treatment by Ann Bodine)14 there was an appeal not to the laws of
language but to those of nature: the generic masculine was said to
have the virtues of 'naturalness' and 'propriety'. And while
generic they has never been eradicated from speech, it has been
stigmatised as non-standard, incorrect and unacceptable in
writing.
Unmarked he is indeed a feature of grammar - of prescriptive
grammar, reinforced by male grammarians for avowedly ideo-
logical reasons. The whole affair illustrates the ill-effects of
pseudo-scientific linguistics: for if the Harvard faculty had
applied scholarship and sought out the historical and cultural
practices determining English gender rules, they would not have
made the mistake of reifying their terms.

Common gender

The phenomenon of common gender - where a noun can be

either masculine or feminine - is found in both grammatical and
 False Dt'chotomt'es 69

natural gender languages. In English, for instance, person,

consumer and teacher have common gender, while in French
enfant and personne, though grammatically masculine and
feminine respectively, require adjectives, etc. to agree on the
basis of sex-reference. But how common is common gender?
Although it is obvious that words like person are capable of either
a masculine or a feminine interpretation depending on the
contexts, there are some respects in which common gender nouns
exhibit a sexual asymmetry.
A random foray into any newspaper will eventually reveal
usages like the following: 'FOURTEEN SURVIVORS, THREE
OF THEM WOMEN ... ', or 'PEOPLE ARE MUCH MORE
LIKELY TO BE INFLUENCED BY THEIR WIVES THAN BY
OPINION POLLS.' It will never reveal, on the other hand,
'FOURTEEN SURVIVORS, THREE OF THEM MEN ... ', or
'PEOPLE ARE MUCH MORE LIKELY TO BE INFLUENCED
BY THEIR HUSBANDS THAN BY OPINION POLLS.'
Is the common gender basically masculine then? It is not, of
course, impossible to say 'my grandmother is a survivor of two
world wars' or 'Marianne and Elinor are very strange people'. But
whereas it is possible, and in some cases apparently obligatory, to
make a special case of women, this is not possible when one is
talking about men.
Maria Black and Rosalind Coward, in an article criticising
Dale Spender, argue that this phenomenon derives from a more
general 'discursive practice' within our culture, whereby men
have available to represent them a discourse of general human-
ness, as well as a discourse of masculine sexuality, while women
have available to them only one discourse in which their sexuality
is paramount. 15 Men may efface their masculinity, but femininity
can never be effaced. This produces the effect that women are an
exception to the male norm.
The problem for feminists is to make men confront and take
responsibility for their sexuality, as women must theirs. For Black
and Coward, gender asymmetries are not the product of gender
rules in language, but of historically evolved ways of representing
things, or cultural discourses as they are termed, which define the
nature and limits of femininity and masculinity. One attacks
these discourses primarily by becoming aware of them and by
developing rival discourses (ways of representing) that people will
eventually incorporate into their own method of dealing with
reality.
70 Feminism and Linguistic Theory

A rather more banal explanation of gender anomalies is simply

that many speakers using words like friend on particular
occasions do not 'have women in mind'. Since they are not
thinking of females they have no need to select a form
appropriate to talking about females. Only the pseudo-generic
nature of most words and the non-specificity of most contexts fool
us into imagining that most utterances do not explicitly exclude
women.
Whether or not one agrees with Black and Coward, gender is
evidently a candidate for serious re-examination by feminists. It is
clear that our ideas about gender and sex affect the way the
category is used in lay language, and the way it is analysed by
linguists. Some combination of the linguistic and the metalinguis-
tic produces the kind ~f regularity Black and Coward call a
'discourse': and this regularity cannot be understood in linguistic
terms alone.
All of this gives the lie to the linguist's assertion that gender and
sex have nothing to do with each other. Analyses of gender are
coloured by ideas about sex, and may be invoked in defence of
ideas about sex. The repeated assertion that these gender
phenomena are just part of the language, open only to a technical
and apolitical explanation, simply serves to obscure the
ideological and prescriptive nature of what grammarians do.
lt is in general true that linguists shy away from the ideological
implications of their analyses, since to admit these exist is to
question the objectivity of science, or else the scientific status of
linguistics itself. One of their methods for distancing themselves
from political questions is the maintenance of theoretical fictions
like langue, through which they are reassured that they are
dealing with a system and not with the willed acts of speakers who
inhabit the real world of social and power relations (thus we get
the absurdity of the Harvard faculty, Dubois and Crouch et ai.
telling off speakers for wanting to make 'unnecessary' changes to
the language, which is somehow seen as existing in its own right,
independently of those who speak it). Another distancing
strategy, as we have seen, is to ignore history.
But those of us who live in cultures with a long history of
grammatical analysis cannot escape our history, and the fact is
that our linguistic tradition is overwhelmingly a prescriptive one.
Linguistic science does not begin on a tabula rasa, and just as we
take traditional categories like noun and tense for our descriptive
 False Dichotomies 71

tools, so we have taken value judgements, prescriptions and

myths. We must be prepared to examine linguistics for evidence
of these and to work out their ideological implications,
abandoning the idea that linguistic science can be unbiased.
Grammar, and all other metalinguistic practices, cannot but be
implicated in the oppression of women through patriarchal
scholarship.
In this chapter we have been dealing with the sexism of
grammar, attempting to separate out sexist attributes of language
from the sexism projected on to language in the analytic responses
of linguists. This has not always been easy, for metalinguistic
practices sometimes turn out to have had considerable influence
on the language itself. Bearing this in mind, we must now tum to
'the language itself, and to feminist ideas about linguistic reform.
5 Making Changes: Can
We Decontaminate Sexist
Language?
At a deep level. changes in a
language are threatening because
they signal widespread changes
in social mores.
Casey Miller and Kate Swift

In the last chapter I looked at the sexism which is expressed in

linguists' analyses of language. arguing that although its effects
are not insignificant, this sexism lies essentially in the metasystem
used by linguists rather than in the language itself. This chapter,
however, concerns the sexism that feminists have found more of
an everyday problem, since it is. so to speak, in the language
itself. Most of all, this chapter concerns the theory and the
practice of feminist linguistic reform.
The sexism of languages (as usual I shall be dealing primarily
with English, but languages vary in the type and degree of sexism
they display) is a subject invented and researched by feminists.
The ideological framework they have used is simple and explicit:
briefly, they start with the hypothesis that the lexicon,
grammatical structure, etc. of a given language will contain
features that exclude, insult or trivialise women, and they set out
to identify the features in question. Some researchers posit
underlying mechanisms of language change to account for
asymmetries; others concentrate, as I did in the previous chapter,
on prescriptive processes; some are interested in language as a
sociological research tool, relating the changing definitions and
uses of words to the differing forms of women's subordination;
while others see it as their task to suggest changes that will
eliminate or modify offensive forms.

72
 Making Changes 73
Nor has this kind of 'verbal hygiene' been ignored. Many
people who would never have thought about the matter just a few
years ago believe now in the existence of 'sexist' and 'non-sexist'
language, and expressions designated 'non-sexist' are turning up
more and more in the usage of the media (an informal look at one
day's newspapers, for instance, yielded an item on whether to
replace generic he with he or she or they, a reference to angry
young men and women and the word spokeswoman in a news
report). Such awareness and willingness to change could not have
come about without pressure from the women's movement.
Of course, many institutions and individuals - perhaps most -
continue to use sexist language, and to defend its use. Their
argument in doing so, however, has had to change. Instead of
denying that a male bias exists, they pretend to object to change
on the grounds that one should not tamper with grammar, that
non-sexist forms are aesthetically inferior or even, as a last resort,
that any willed change in language automatically ushers in 1984:
The feminist attack on social crimes may be as legitimate as it
was inevitable. But the attack on words is only another social
crime - one against the means and the hope of communication. '1
Once again, we have this notion of 'the language' as a hallowed
institution whose traditions may not be queried. 'Words' may be
attacked independently of their users, and this will be disastrous
because it will render communication impossible. This picture of
language as something external, independent and disinterested
stops us asking whose language it is, whose traditions will be
under attack if the conventions are changed. In this chapter,
questions like these must be asked. It is not good enough to shrug
our shoulders and say that male bias in usage is purely
grammatical, and that therefore it does not matter.
Obviously, it does matter to feminists. Most of us are now
thoroughly aware of the ways in which English insults, excludes
and trivialises us (universal male pronouns, misogynist insult
words, patriarchal personal names, trivialising suffixes for women
in professions (authoress), girl used in contexts where boy would
be unacceptable, words like blonde standing for the wholt
woman; etc., etc. and there is no need to rehearse them all oveJ
again in this book.
What does need to be discussed, though, is precisely why all
this is so offensive. Is it just an unpleasant reminder that men see
74 Feminism and Linguistic Theory

us either as scapegoats or as non-entities? Or is it positively

harmful? Can we eliminate it through linguistic reform, and if so,
should we?
Questions like these are especially interesting because feminists
themselves do not agree on the answers. Although there may seem
to be a consensus, the united front soon turns out to be an
illusion: most feminists believe that sexist language is a bad thing,
but they believe it for very different reasons. A particularly
important difference is between those who consider language
'symptomatic' and those who consider it 'causal'.
The 'symptomatic' camp believe that sexist language is a
symptom, a piece of rudeness which may well be quite
unintentional. To the extent that it is the product of carelessness,
ignorance and laziness, it can be cured by the linguistic reformer.
The reformer works by (a) drawing the speaker's attention to the
offending form, and to the underlying prejudice of which it is a
symptom, and (b) suggesting a non-sexist alternative which the
speaker, now made aware, can substitute for it.
Casey Miller and Kate Swift, authors of a comprehensive guide
to non-sexist English, represent admirably the attitudes of the
'symptomatic' tendency. 2 For them, sexist language is an
outdated excrescence which everyone but a few reactionaries
would dearly love to be rid of; mere force of habit is the only thing
that props it up. Since Miller and Swift subtitled their earlier
book (Words and Women)3 New Language in New Times, it
seems they take the optimistic view that we are now living in a
post-feminist world, and that their job is to help language catch
up with society. Miller and Swift are strong on common sense and
the 'facts':

The public counts on those who disseminate factual in

formation ... to be as certain that what they tell us is as
accurate as research and the conscientious use of language can
make it. Only recently have we become aware that
conventional English usage ... obscures the actions, the
contributions and sometimes even the very presence of women.
Turning our backs on that insight is an option, of course, but it
is an option like teaching children the world is flat.4

In other words, sexist language is to be condemned because it

distorts the truth: once aware of this startling fact, right-thinking
 Makz"ng Changes 75

people will immediately proceed to self-criticism and reform.

Purged of its prejudices, our language can indeed in the mouth of
a 'conscientious' user, disseminate 'accurate' information.
I have no wish to belittle the important work of Miller and
Swift, especially the detail in which they have worked out non-
sexist alternative usages so that the most unimaginative writer, if
well-meaning, can eliminate gross bias without gross inelegance.
However, the stance of those who advocate non-sexist writing for
the reasons Miller and Swift do is a theoretical reformism which
leaves an enormous amount to be desired. One of the aims of this
chapter is to produce a critique of theoretical reformism and of
the assumptions about language that lie behind it.
Recently, the reformism of the 'symptomatic' camp has been
explicitly criticised by other feminists. Dale Spender, for
example, is a well-known supporter of the idea that language
causes women's oppression rather than being a symptom of it. It
is through a language that trivialises, excludes and insults us that
we come to know our subordinate place in the world.
The 'causal' tendency has also extended the boundaries of what
counts as sexism in language. Miller and Swift have a well-defined
set of targets for reform: generic masculine pronouns, sex-
differentiated job descriptions, gz"rl used of adult women and so
on. Dale Spender would insist that all words are sexist, since their
meanings are fixed by men and embody male misogyny.
The question of whether linguistic sexism is a cause or an effect
of women's oppression, and the problem of defining its boun-
daries, ultimately links up with the debate on language and
reality, who controls language and who is alienated from it. I
shall be examining that debate in detail in the next two chapters.
In this chapter, however, I shall confine myself to the sorts of
linguistic phenomena that worry both reformists and radicals:
usages that are always and obviously sexist, and which might
conceivably be the targets of organised reform campaigns. I
particularly want to look at the ways in which feminists are hitting
back at sexism informally, through private 'reclamation' and the
coining of new terms, and institutionally, through demands for
reforms in lexicographical and journalistic practice, etc.
Where sexism in language is concerned, feminists tend to
proceed in two ways. When a problem area is identified, they are
concerned both to draw out the political and historical im-
plications of linguistic facts, and to consider changes in their
76 Feminism and Linguistic Theory

linguistic practice. These changes are often very inventive: words

may be 'reclaimed' either by revaluing their connotations or
reviving obsolete definitions, or they may undergo changes in
spelling or morphology. Sometimes feminists wage war on a word,
while at other times they introduce one. This subversive feminist
metalinguistics, a product of the wish to understand and mani·
pulate language, can be illustrated with a number of specific
examples.

INSULTS: VERBAL VIOLENCE AGAINST WOMEN

Many commentators have noted that more words are available to

insult women than men; and that generally speaking, taboo
words tend to refer to women's bodies rather than men's. Thus for
example cunt is a more strongly tabooed word than prick, and
has more tabooed synonyms. Even words like bugger and arsehole
whose reference is male are insulting because they connote homo·
sexuality, which is not only taboo in itself but associated with
femininity as well.
The asymmetry continues. There are terms for women collec-
tively as sexual prey such as ass, tail, crumpet, skirt andjlash. No
such terms exist for men, nor do we have male analogues for slag,
tart, nympho or pricktease.
Presumably this has something to do with the double standard
and the heterosexual state of play in general. Women are thought
not to have sexual desires, and if they either show that they do
have such desires, or refuse to meet the needs of men, they will be
censured. On the other hand they are vilified as prostitutes. Julia
Stanley has observed that for English-speakers the prostitute is
'paradigmatic woman'. Male prostitutes have no comparable
richness of terminology associated with them, for after all they are
far from being paradigmatic men.
To say that the asymmetry of insult terms 'reflects reality',
however, would be banal. We need to consider whether general
linguistic processes bring it about, and whether its effects on
women are significant.
It has been argued by Muriel Schulz that the asymmetry we are
considering comes about because of a systematic process of
language change called 'semantic derogation'. 5 Terms like tart
 Making Changes 77
and harlot have developed from non-insulting unisex words (tart,
for instance, was once an endearment like honey or sweetie).
When they became associated with women rather than men, they
acquired negative connotations and eventually came to mean
prostitute.
This process of pejoration can be seen at work in a number of
male/female pairs. Whereas the male terms connote power,
status, freedom and independence, the female, which in many
cases used to be parallel, now cannotes triviality, dependence,
negativity and sex.
For instance, bachelor (positive, independent, sexual libertine)
is opposed to spinster (ugly, sexless and frustrated). When the
positive aspects of being single came to be associated with women,
the term spinster seemed so unsuitable that bachelor girl had to
be coined.
Other examples of semantic non-equivalence are governor
(powerful, ruler) and governess (poor woman looking after
children); master (competent or powerful man) and mistress
(sexual and economic dependent); tramp (homeless man) and
tramp (prostitute woman).
The suggestion that politically motivated processes operate
systematically in language change is an interesting one for
sociolinguists, for it confirms their belief that language change is
not a random but a socially significant occurrence which may be
discussed in a 'scientific' way. Feminists, however, are likely to
find it more interesting that we can reconstruct the history of
patriarchy, at least to a small extent, through the history of
words.
Yet the feminist analysis of insult terms would be missing
something if it stopped there. The existence of so many insulting
words for women, many of them meaning the same thing, has a
significance over and above what it tells us about cultural beliefs.
It is, in fact, itself a form of social control. We can make an
analogy here with pornography (since the word pornography
means 'pictures of prostitutes', perhaps pornoglossia would be a
good name for the language that reduces all women to men's
sexual servants). Feminists have always seen pornography as a
symbol of men's desire to objectify and humiliate women. It
depicts the woman as an object to be abused, reduces her to body
parts and dwells explicitly on rituals of punishment. But more
recent analyses of pornographic images (for instance Andrea
78 Feminism and Linguistic Theory

Dworkin's Pornography: Men Possessing Women)6 stress that

these images are violence against women, with effects similar to
those of physical violence on women's self-image and attitude. To
see degrading images and know men seek them out for pleasure,
teaches us that we are despicable, expendable objects. It teaches
us that men want to hurt us, and that we had better be afraid.
The same is true of sexual insults. They are verbal violence
against women, expressing both our essential qualities in
patriarchy (repositories of sexuality, prostitutes) and male
woman-hatred, which makes women afraid. Since cunt and slag
are bandied about even more often that the cock and the fist,
this violence is no trivial matter, but a source of male power and a
means whereby women are daily humiliated.
In several conferences and sessions on language, feminists have
discussed this problem and asked how to cope with verbal
violence. One solution which has been canvassed for some words
is 'reclaiming' - that is, reinvesting a word with a more positive
meaning. The word dyke, for instance, a disparaging term for
Lesbian women, has been rehabilitated to some extent, and there
have been suggestions that the same could be done for cunt.
Women point out its connection with words like cunm'ng which
carry the idea of power and magic, while others simply say they
ascdbe positive value to the word cunt because it denotes the most
female and potent area of their bodies.
Two problems arise with the reclamation approach. First,
there is a content problem, for although some words are suitable
for celebration, since they refer to revalued conditions of life such
as Lesbianism, female anatomy, spinsterhood and so on, many
other words are not suitable. Reducing ourselves to body parts
(e.g. by referring to women as cunts) could never be a
compliment to our feminist selves. Nor should we glorify the
sexual dependence of prostitution.
The second problem is one of intent. An important part of the
meaning of an insult is the intention behind it, or more precisely,
what the receiver takes the speaker's intention to have been. We
all recognise that what men mean by cunt and dyke is violent and
contemptuous. Just as black people may call themselves nigger in
friendship (though many, of course, would never do so) without
eliminating the racism of the word when white people use it, so
we can reclaim certain words amongst ourselves without touching
their status as insults in the mouths of men.
 Making Changes 79

LINGUISTIC HERSTORY: RECLAIMING AND

REJECTING

It is not only insults that feminists have subjected to close

examination and found wanting. Their relationship with many
words, and even spellings, demonstrates their consciousness of
their meaning and history. Many feminist writers turn frequently
to etymological dictionaries to find out when a word entered the
language, whether it was coined or borrowed from another
language, what it meant and how it has changed.
For instance, in the discussion of the word cunt which I
mentioned above one woman said that for her, the word and its
synonyms conjured up pornographic stories, and therefore she
had always used vagina. However, she had recently discovered
that vagina came from Latin, where it meant 'where you sheath
your sword'. She found this so offensive that she had abandoned
vagina as well. And apart from this, she felt she had learned from
etymology something about the history of sexism which she did
not know before.
In other cases, however, history is deliberately ignored. The
word history is a good example: feminists often respeak/write it as
herstory. This reflects the idea that history means his-story (so
that herstory becomes the female equivalent). In fact it comes
from Latin historia which has no connection with the English
word Ms. Similarly, women is often spelt by feminists wimmin, so
that the - men element does not appear in it, even though this
element is not actually pronounced.
Linguists find this kind of thing irritating (there is no doubt
that any attempt to start a 'herstory' course at Harvard would
once again cause havoc in the faculty of linguistics) because it is
inconsistent - sometimes history is counted as relevant, sometimes
not - and in any case they tend to dislike the un-Saussurean view
that linguistic history is at all salient for speakers of current
English. For feminists, however, the main consideration in using
or not using forms like herstory ought to be political. Herstory is
an excellent word in many contexts pointing out with wit and
elegance that most history is precisely the story of men's lives;
while wimmin might be universally applauded as a clever piece of
spelling reform, had it not become associated with the unpopular
'extremism' of the women's movement.
80 Feminism and Linguistic Theory

The creative use of linguistic structure and linguistic history is a

characteristic of much feminist writing. Mary Daly's classic
Gyn/Ecology is a good example of the 'reclaim and rename'
approach: as well as playing on words with great felicity, Daly
tries to reclaim obsolete meanings of familiar words like glamour,
haggard and spinster. She shows that women's power has been
erased from definitions (for instance, the kind of glamour now
associated with film stars is a far cry from the magical powers
op.ce denoted by the word glamour) and wonders why feminists
cannot, by the same token, 'wrench back some wordpower' by
conscious redefinition. 7 I shall consider this problem (for it is
obvious to me that we cannot just redefine in the same way that
those with institutional power have done) later in this chapter.
Apart from the 'content and intent' problems we have already
touched on, there is a further problem with the work of Daly -
many women find it elitist and unreadable. Constant wordplay
and extensive terminological definition are not immediately
accessible devices, and feminists need to consider very carefully to
what extent they are politically productive. (Indeed, as we shall
see in later chapters, this is a debate of some importance in
current feminism.) Certainly they are not applicable in everyday
speech, however well they work in writing.

GAPS IN TERMINOLOGY: SAYING IS BELIEVING

Feminists are hopeful that old words can be given new meanings.
But equally, they are hopeful that we can make new words up to
name the things that have so far remained unlabelled. For they
remind us that if the first principle of sexist language is that
female words must be negative, the second is that positive aspects
of femaleness should remain unnamed.
In the introduction I referred to the idea of 'naming' to which
Adrienne Rich and Dale Spender give so much emphasis. To
these writers, it sometimes seems nothing exists until a specific
label is hung upon it. I have already said that this strikes me as an
extreme claim (and I shall be arguing later that it demonstrates a
misunderstanding of the nature of language). Labels may give
some sort of social validity to experiences, but the lack of labels
does not render any experience ineffable. However, certain
 Making Changes 81

terminological lacunae are frequently discussed by feminists

because they demonstrate that what naming has been done, has
been done from a male viewpoint.
The paradigm example of this sort of discussion concerns terms
for sex and sexuality. It is striking, for instance, that what to
women is often the most satisfying part of heterosexual love-
making is called foreplay. For the namer, obviously, this activity
comes 'before' the real thing, i.e. penetration, and thus the
namer must be a man. Indeed, the word penetration betrays
male origins; women would have called it enclosure! As things
stand, most words for the sexual act (itself a revealing term: since
when was there only one sexual act?) make it into something men
do to women: fuck, poke, screw. At school I was taught that the
word lover was not appropriate for women, since it denoted
activity, and for the passionate women in Racine's plays we were
to use the term mistress.
The male slant in the lexicon of sex would be very difficult for
anyone to deny, and feminists have made it clear that here, as in
other areas, they are dissatisfied. Beyond the sociological analysis
of language lies the tantalising possibility of reform: but we must
now look at a number of reasons why thorough-going institutional
reform of the English lexicon is difficult for any progressive
political movement to bring about.

SPREADING THE WORD: THE GATEKEEPERS

A number of processes are common in the history of words.

Words are lost: others are invented, either made from old words
joined together in new ways (like palimony and denationalise) ,
borrowed from foreign languages (like creche) or created from
the coiner's imagination. The meanings of words (by which I
mean the senses listed in dictionaries) do not stay constant for
ever, but gradually change. One reason for this is that people do
not learn most words from dictionaries but infer their meanings
from hearing them used in particular contexts: we may all differ
slightly in our beliefs about what words 'really mean'. If enough
people infer from reading the word prevaricate that it means
'stall, play for time' (to take a recent example discussed in the
newspapers), that meaning will challenge the one prevaricate is
82 Feminism and Linguistic Theory

given in dictionaries, It will be no good telling speakers that

prevaricate 'really means' 'lie' rather than 'stall', because the
meaning of words is ultimately a matter of the way the
community uses them in talk, Unless they are compulsive users of
dictionaries, this will be determined by contextual inference, and
meaning will be inherently unstable,
However, it is not true that the process of semantic change
operates magically, untouched by anything but the collective
mind, Whether new meanings and new words catch on depends
to some extent on what means exist to disseminate them, The
mass media are powerful in this respect, and so is education
(which is currently fighting a rearguard action against the falling
together of pairs like infer / imply and dt'st'nterested / uninterested),
The dictionary, too, though it cannot be an exhaustive record of
what any word 'means', has a certain role to play by legitimating
some definitions over others,
The point is that conscious linguistic reform by feminists, or
even 'natural' change deriving from women's changing
experience and consciousness, is not simply left to take its chance
with other social forces affecting language is a free-market
competition for semantic supremacy, Any new terms feminists
come up with, in order to be institutionalised in the official and
public domains of language use, have to pass a number of
'gatekeepers' - the media, education, lexicography - who are
very far from being neutral.
The media have sometimes been our allies in spreading new
terms and thus new concepts - sext'sm, sexual harassment,
battered wives, male chauvt'nism, They have shown no such
enthusiasm for unwaged, double loaded, heterosexism, male
vt'olence or patn'archy, and it is significant that this second set of
terms is less widely known than the first outside the women's
movement. Moreover, the media have played a large part in
pejorating some words that were meant to improve women's
position, such as Ms, person and even femt'nist,
Education is a considerable force in retarding the growth of
non-sexist language, Prescriptive rules about generic he are
enforced as 'grammatical' and many teaching materials still
exemplify everything feminists would want to take issue with,
The dictionary is in many ways less influential, since for the
majority of English speakers it is an irrelevance, Its sexism,
however, is clear. It fails to invest feminist words and definitions
 Making Changes 83

with permanence, official sanction and authority, and it contains

many negative and offensive definitions of females. Progressives
of all stripes should be more aware than they are of the biases that
affect the compiling of dictionaries.
It is frequently claimed that dictionaries, like linguistic
enterprise in general, are descriptive rather than prescriptive.
They merely record the way people use words, without fear or
favour. This has sometimes served as an excuse for including very
offensive definitions: those who protested at the inclusion of 'Jew'
as a verb meaning 'cheat financially' on the grounds that such a
definition promoted anti-semitism got short shrift from the
disinterested lexicographers in pursuit of English usage. But how
accurate is this picture of unbiased scholarship?
The most important bias of dictionaries is toward the written
rather than the spoken word. Lexicographers do not begin the
quest for current usage in the street or on the bus, but in libraries.
Even then they begin their search with literature rather than say,
comic books, graffiti or political pamphlets. The consequence is
that the coinings of dictionaries are the coinings of those who
write literature - middle-class men. The vitality of home and
street vernacular is simply ignored. Common words like skive and
moonlight are either absent from dictionaries, or else they are
given a special marking as 'colloquial' or 'dialectal'.
The implication here is that the words of educated middle-class
speakers are somehow not dialectal (though in fact, if Yorkshire
miners speak 'a dialect' so do BBC newsreaders). What diction-
aries assume is that while the less privileged will want (need?) to
look up the words used by their betters, the reverse is unlikely to
be true. No one needs to look up a word it would be beneath her
dignity to use. Clearly, then, dictionaries reflect the prejudices of
the ruling class.
This is also shown by the definitions they include (and those
they do not). Dictionaries in any case foster the illusion that words
have a limited number of meanings which can be listed out of
context; but worse still, the ones they list tend to be ideologically
loaded though they masquerade as objective. For whom, for
example, does the word woman mean 'weak and lacking in
vigour'? Which groups in society concur with the definition of
unfeminine (as in 'unfeminine hair') as 'not characteristic of
women'? What woman considers her clitoris as 'a rudimentary
sexual organ in females, analogous to the penis'? (All these
84 Femz"nz"sm and Lz"ngwstz"c Theory

definitions came from ordinary dictionaries that are currently on

sale, and so probably in current use by schoolchildren, scrabble-
players, etc.)
Dictionaries speak only for some people, and their authority is
political, not grammatical. This does not mean they are valueless,
of course, for arguably those who aspire to educated middle-class
usage require a reference book. But they are not the objective and
exhaustive record they claim to be. Until we have more control
over metalinguistic processes and practices like education and
lexicography, we will find it hard to disseminate innovations and
changes that we think are desirable, or even see words defined in
anything like the way we use them. Feminist reformers, in other
words, are put at a disadvantage by the reactionary nature of
prescriptive institutions.

GENDER REVISITED: MAKING WOMEN VISIBLE

The best-known aspect of sexism in language is what feminist

linguists call 'he/man language': the use of male pronouns as
generic or unspecified terms, as in 'no one would do that if he
could help it', and the use of man and mankz"nd to mean the
whole human race. A great deal of effort has gone into making
institutional changes in this area, since many feminists consider
pronouns an important subliminal influence on people's per-
ception of women as secondary or marginal.
Miller and Swift sum up what is wrong with the universal male:
'What standard English usage says about males is that they are
the species. What is says about females is that they are a sub-
species. '8 Experiments in linguistics reveal that when faced with
generic man women consciously exclude themselves from the
reference. We equate man specifically with males. For many
commentators this fact implies that language is more than just
trivially offensive: it is able to persuade its speakers that women
do not exist. This misleading impression is what reformists set out
to correct by means of non-sexist language.

THE MYTH OF NON-SEXIST LANGUAGE

Non-sexist language is language which excludes neither women

 Making Changes 85

nor men. It involves recasting words and sentences so that all

terminology is neutral. For instance, in a non-sexist formulation
mankind becomes humanity, craftsman becomes artisan, space-
man becomes astronaut, forefathers become ancestors and chair-
man becomes chairperson or chair. They is employed as a
singular indefinite pronoun ('no one would do that if they could
help it') and generic pronouns are avoided where possible, either
by recasting a sentence (e.g. 'pick up baby when he cries' could
become 'always pick up a crying baby') or by pluralising ('pick up
babies when they cry').
This kind of language is less overtly offensive than the kind it
replaces. Nevertheless there are plenty of reasons to suppose that
it is ineffective in the sense that it does not really bring women
into people's mental landscape at all. The reformists feel that
words like spaceman have a special place in the lexicon of
prejudice. Because spaceman incorporates the word man, whose
meaning has narrowed (become more specialised) from meaning
'person' to meaning 'male person', it strongly suggests a male
referent. The implication of spaceman is that women cannot fly
rockets, walk on the moon, etc. But once the linguistically
marked male element man is removed, the argument runs,
people will not think male any more. The possibility will exist that
women can fulfil the new role, non-sexistly designated astronaut.
But what if the word astronaut, despite having no overt
markers of maleness, is used by most people as if it too were male-
only? There is a large amount of evidence that this is in fact what
happens with words that are not linguistically gender-marked.
Consider, for example, the following extracts from newspapers:

The lack of vitality is aggravated by the fact that there are so

few able-bodied young adults about. They have all gone off to
work or look for work, leaving behind the old, the disabled, the
women and the children.
The Sunday Times
A coloured South African who was subjected to racial abuse by
his neighbours went berserk with a machete and killed his next-
door neighbour's wife, Birmingham Crown Court heard
yesterday.
Guardian

In these examples, two phrases without overt linguistic marking

86 Feminism and Linguistic Theory

of gender are used as if they could only be applied to men: able-

bodied young adult and next-door neighbour. In the first case,
the disabled, the old and the non-adult are clearly excluded by
the meaning of the words which make up the phrase. But why are
women excluded? In the second case, since apparently the
murdered woman lived next door to her attacker, why was she not
'the next door neighbour' but just 'the next door neighbour's
wife'?
Evidently many language-users, when saying or wntmg
common terms that might in principle refer to either sex, simply
do not think of them as referring to women. The words are
neutral on the surface, but masculine underneath. Non-sexist
language guidelines are a verbal Sex Discrimination Act, in that
they legislate on the form of words without being able to alter the
meaning. They are a purely cosmetic measure which enables us to
see justice being done without really doing us justice.
From where the reformists stand, this must seem very odd. For
as we have seen, Miller and Swift et al. regard non-sexist language
as a necessary and long overdue corrective which will make our
speech and writing more accurate: The point is not that we
SHOULD recognise semantic change, but that in order to be
precise, in order to be understood, we must.'9 Why then should
anyone be perverse enough to use words which are not inherently
misleading, such as netghbour and adult, to falsify reality?
Sociologists of literature have pointed out that exactly this
question can be asked about popular fiction, for instance the
romantic novels bought by large numbers of women. 10 Although
most readers would declare these novels realistic in mode, the
heroines in fact have attributes not granted to their real-life
counterparts. The nurse or secretary in a novel can always
maintain a lifestyle superior to anything her real-life salary would
allow: she also marries a higher-class, higher-income, more
educated man at the end of the book. Is this a deliberate
distortion? Is it a sinister plot to misrepresent the world and
sustain female false-consciousness?
Feminist discussion on this point stresses the propaganda
function of romantic novels, the part they are meant to play in
sustaining sexist ideology and their real power to mislead. Hall,
however, considers this view naive, simplistic and politically
unsophisticated. Hall suggests that literature is not really about
faithfully depicting actual states of affairs. Its referent is not the
 Makz'ng Changes 87

real world so much as the belief-system prevalent in particular

parts of it. Far from misrepresenting the world, then, popular
literature mirrors the ideological universe of its readers. If this is
true of symbolic and representational media generally, it would
be entirely beside the point to criticise language for being
'misleading' as to the state of affairs which obtains in the world.
Language is not a limpid pool through which the truth may be
glimpsed, but a way of representing, a vehicle for 'discourses' and
'ideology' .
Cultures not only tolerate but in many cases seem to demand a
contradiction between what people can see for themselves and
what they believe to be true, or right. For instance, the notion
that women cannot do heavy work (which carries high rates of
pay) ought not to cut any ice with women who regularly lift heavy
children and stones of shopping, but it does. Women who clean
up after incontinent elderly relations ought not to entertain the
oft-repeated observation that some jobs are too unpleasant and
dirty for women to do, but apparently they accept it. Everyone
knows that many women nowadays either choose or are forced to
support children on their own: yet women will still say they do not
believe in equal pay, because men have families to support. It is
not convincing to claim that because women believe incorrect
and sexist propositions, they must been misled about the facts.
Miller and Swift believe that language change is threatening
because it 'signals widespread changes in social mores'. II In fact,
few people notice language change, but reform, deliberate
intervention for some stated purpose, brings out the Colonel
Blimp in many people. This is not necessarily because it 'signals
widespread changes in social mores'. Observers in developing
cultures report that language lags behind social change, often to
the extent that expert 'language planners' are employed. 12 What
institutional language reform really signals is an agreement on
the part of those who have power to recognise a new 'discourse' or
way of understanding things, which challenges the appearance of
immutable truth previously enjoyed by the old one.
The outcry which so often attends the demand for linguistic
reform comes from those who do not want to be shaken out of the
old way of looking at things. If these people are numerous and
powerful, strong conservative forces come into play and reform
does not succeed.
The cause of such conservatism is only partly anti-feminism,
88 Feminism and Linguistic Theory

fear of social and political upheaval. Resistance to language

change is also related to the way in which people conceptualise
language itself as a fixed point in the flux of experience,
providing names (their essential correctness guaranteed by
history) for phenomena that would otherwise elude our collective
grasp. Thus the anthropological linguist Sapir could speak of
'that virtual identity ... of word and thing which leads to the
magic of spells'.13 And the conservative philosopher Roger
Scruton, talking about feminist proposals to reform language,
insists, 'Each of us inherits in language the wisdom of many
generations. To mutilate this repository of human experience is to
mutilate our most fundamental perceptions. '14
In the remainder of this book I shall be addressing myself far
more to the nature of language itself, and I shall be arguing that
this particular view of it is erroneous, an illusion from which
people take comfort, but in the end untenable as a theory of
language.

POSITIVE LANGUAGE

It should be clear by now that while I feel as excluded as any

other woman by 'he/man' language, I cannot put my faith in the
non-sexist alternatives which pay only lip-service (literally) to my
presence in the world. What sort of changes would I like to make?
The strategy I believe is helpful is the one used in this book -
every indefinite or generic referent is feminine. I and several
other linguists use this practice all the time. When questioned by
people who find it odd, we reply that we are practising positive
discrimination through positive language. If it comes naturally to
men to say/write he, obviously it comes naturally to me to use
she. In a non-patriarchal world, would we not tend to visualise
someone rather like ourselves?
I have no illusions that positive language will change the world.
More women will not take up science just because scientists are
referred to as she. But what might be achieved is a raising of
people's consciousness when they are confronted with their own
and others' prejudices against saying she.
It is still true that women have difficulty in using positive
language. (One woman said she could not use it in essays because
 Making Changes 89

her subject was theology: on being asked by another woman in

the group whether she saw God as male or female, she replied,
'Neither: I see him as an absolute supreme Being!') It seems
particularly odd to refer to engineers and astronauts as she when
everyone knows that the vast majority of them are men; yet when
similar untruths are perpetrated by he (for instance, when it is
used to refer to teachers or hospital patients, most of whom are
women) we do not notice.
From time to time, the possibility is mooted of inventing new,
sex-indefinite pronouns for the English language (E, tey and per
have all been suggested). It is interesting to consider what might
happen to such a pronoun in common usage (if indeed it ever
caught on).
As we have seen, 'neutral' words tend to make people assume a
male referent. So perhaps the new pronoun would be masculin-
ised, and a feminine variant coined. More likely, however, the
new pronoun would go the way of the suffix - person, whose
short history is an object lesson to all reformers. Words like
chairperson and sportsperson were supposed to be sex-neutral
replacements for chairman and sportsman, but in fact they are
only ever used for women: Cecil Parkinson is Chairman of the
Tory Party, but Joan Ruddock is Chairperson of CND.
It seems that a peculiar odour attaches to the suffix - woman
even in the humblest of contexts, and - person is being used as a
sort of euphemism: 'Of course full justice to a steamed pudding
can only be done by a true trencherman. The term is used ad-
visedly, for I have never encountered a feminine trencherperson
whose curves could easily expand to accommodate a second
helping' (Sunday Times). In this case, speakers have refused to
accept a feminist reform, and indeed have used the letter against
the spirit. We are left with a net loss, for if men are'spokesmen
while we are spokespersons, the presence of women is not being
drawn explicitly to anyone's attention. If reclaiming language
works, and if it is thought desirable, the word woman should be
at the top of the list for reclamation.
As many writers have shown, languages and their history are
invaluable resources for feminists in their analysis of society. But
reform on a wide scale is more problematic, and it is especially
unhelpful when it proceeds from simplistic theories about the
workings of language in general. As we have seen, supporters of
'non-sexist' language believe that language exists to represent
90 Feminism and Linguistic Theory

reality as neutrally! objectively as possible. Therefore, in the

interests of accuracy we should strive to include the female half of
the human race by replacing male terms with neutral ones. But
the 'reality' to which language relates is a sexist one, and in it
there are no neutral terms. Words cannot be brought before some
linguistic United Nations for definitive judgement; this one is
sexist, that is neutral, the other is feminist. Words exist, the
theories of linguists notwithstanding, only when they are used.
Their meanings are created (within limits, certainly, but pretty
elastic limits) by a speaker and hearer in each uniquely defined
situation.
For feminists this may prove to be a bitter pill. It means that
when we proclaim certain items positive, rehabilitated and so on,
we can have no authority outside our own narrow circle, unless
the means exist for us to influence the usage of others (and even
this is only possible up to a point). In the mouths of sexists,
language can always be sexist.
6 Silence, Alienation and
Oppression: Feminist
Models of Language (I)
. . . thinking is most mysterious, and by
far the greatest light we have upon it
is thrown by the study of language.
This study shows that the form of a
person's thoughts are controlled by
inexorable laws of pattern of which
he is unconscious.
Benjamin Lee Whorf
Without language thought is a vague,
uncharted nebula.
Ferdinand de Saussure

Sexist language teaches us what those who use it and disseminate

it think women's place ought to be: second-class citizens, neither
seen nor heard, eternal sex-objects and personifications of evil.
Within the perspective I have labelled 'theoretical reformism',
our feminist response is clear. We must expose the 'falseness' of
this language, and refuse to tolerate its continued use, providing
where necessary a set of neutral, and thus inoffensive, alter-
natives.
But in Chapter 5 we began to see how this uncomplicated
viewpoint failed as a theory of language - as an account of what
language does and how it does it. Many feminists now feel a need
to go beyond theoretical reformism, developing a more sophisti-
cated analysis of the place of language in culture, and thus in the
oppression of women. For these radical theorists there is no
neutral language: the entire system, since it belongs to and is
controlled by men, is permeated by sexism through and through.
Moreover, male language is a species of Orwellian thought-
control, for these theorists believe it is through language that we

91
92 Feminism and Linguistic Theory

construct our reality. Those who define the limits oflanguage can
make us see things their way.
It is evident that this more radical view of the nature of
language has become very influential in the women's movement.
A rhetoric of silencing, of appropriation and alienation, pervades
much recent writing. Feminists are convinced that language, or
the lack of an authentic (non-male) language, profoundly affects
women's ability to understand and change their situation.

The fact is that the female saying 'I' is alien at every moment to
her own speaking and writing. She is broken by the fact that
she must enter this language in order to speak or to write. As
theTis broken, so also is the Inner Eye, the capacity for
integrity of knowing/ sensing. In this way the Inner Voice of the
Selfs integrity is silenced: the external voice babbles in alien
and alienating tongues .
. . . Overcoming the silencing of women is an extreme act. I
When we become acutely, disturbingly aware of the language
we are using and that is using us, we begin to grasp a material
resource that women have never before collectively attempted
to repossess .... Language is as real, as tangible in our lives as
streets, pipelines, telephone switchboards, microwaves, radio-
activity, cloning laboratories, nuclear power stations. We
might hypothetically possess ourselves of every resource on the
North American continent, but as long as our language is inad-
equate, our vision remains formless, our thinking and feeling
are still running in the old cycles, our process may be 'revol-
utionary', but not transformative. 2

In this chapter I want to examine the radical theories upon which

remarks like these are based. Three approaches in particular will
be dealt with: the 'dominant and muted' model of Shirley and
Edwin Ardener; the 'man made language' theory of Dale
Spender; and the psychoanalytic model developed in the wake of
Lacan (which will be discussed in a separate chapter).
While all of these approaches are different, one of the points I
shall be trying to make is that they have a certain amount in
common. They rely, in fact, on three basic principles which are
taken as axiomatic.
First, all three approaches display some degree of linguistic
 Silence, A lz'enation and Oppression 93

determinism. Language is held to be the primary means by which

we make sense of the world, placing significant constraints on our
thought and our perceptions. Secondly, it is assumed that men
control language, just as they control all other resources in a
patriarchal society. It is men who decide what words will mean
and who will have the right to use them. That is why language
enshrines a male (and misogynist) view of the world.
Thirdly, radical theorists feel that women are put at a
disadvantage as speakers and writers. This disadvantage can
manifest itself in two ways. On the one hand, women may use the
male· controlled language, whose meanings are fixed according to
men's experience: if they do, they falsify their own perceptions
and experience by putting everything into a male frame of
reference. This is alienation. On the other hand, women may try
to discuss their experience in an authentically female way. In this
case, they soon encounter the lack of a suitable language, and fall
silent.
The radical feminist view, then, is of women who live and
speak within the confines of a man-made symbolic universe. They
must cope with the disjunction between the linguistically-
validated male world view and their own experience, which
cannot be expressed in male language. Indeed, since language
determines reality, women may be alienated not only from
language but also from the female experience it fails to encode.
In Chapter 8 I intend to put forward a communicationally
oriented alternative theory which does not rely on these axioms of
determinism, control and alienation. In this chapter and
Chapter 7, however, I want to prepare the ground by explaining
radical theories and pointing out their limitations, particularly
the questions they leave substantially unanswered. We can begin
by looking critically at the background to the crucial problems of
linguistic determinism and language control.

DETERMINISM

Linguistic determinism - the idea that language determines

perception and thus reality - is an important part of current
feminist linguistic theory. However, from the linguist's point of
view this raises a number of difficult questions, and needs to be
94 Feminism and Linguistic Theory

justified rather than assumed. In the section that follows I

consider the relation between language and reality, and trace the
origins of determinism in linguistic theory.

Language and Reality

The debate on linguistic determinism takes place within the

context of a particular view of what reality is. If we believed that
reality or the world simply existed 'out there' and that we
passively registered it as a series of images with. names attached,
the question of the effects of language on our perceptions would
hardly arise.
In fact, however, modern theories take it for granted that we
ourselves play a part in creating 'reality'. Millions of stimuli
impinge upon us at every moment: if we were like blank screens,
passively receiving every stimulus, our minds would contain an
undifferentiated, meaningless chaos. In order to make sense of
the world, therefore, we must pay attention selectively, actively
choosing, classifying and interpreting incoming stimuli.
The crucial question for the linguist is what part the language a
person learns to speak plays in this interpreting/classifying
process. To illustrate that this is a real problem, we might
consider two opposing possibilities. The first is to see language as
a tool we use, a servant of thought. Suppose we use thought (for
want of a more precise term) to interpret incoming stimuli,
influenced by factors such as the environment, our personal and
cultural history, the work we do, etc. Language encodes or
expresses this perceived reality. It is simply a medium. In this case
there is no great problem about language. It reflects social
conditions, and changes in response to social changes.
The second possibility is that language acts like a straitjacket, a
ready-made classification which our experience must be forced
into, like the Ugly Sister's foot into the deceptively alluring
patriarchal glass slipper. The language user engages not with the
'real world', but with a version of it already filtered and, given
that powerful groups control it, distorted, by the language. In
this case language is a problem. It effectively creates our
perceptions of reality, and they can produce a repressive and one-
sided picture.
Feminists and many other progressives have for some time
 Silence, A lienation and Oppression 95

tended towards the second possibility rather than the first. Their
theories derive from the work of earlier writers, especially Lacan
(and thus Saussure), Whorf and Sapir. We must now turn to
those theories and ask two questions about them: first, what they
say and what arguments exist for and against them; and second,
how far the feminists who have taken them up remain faithful to
their original meaning and spirit.

The roots of determinism: Saussure and Lacan

Saussure is often claimed as the founder-member of the

determinist camp. Certainly he rejected in the Cours the
simplistic first alternative, that language simply gives form to
ideas:

Psychologically, our thought - apart from its expression in

words - is only a shapeless and indistinct mass. Philosophers
and linguists have always agreed in recognising that without
the help of signs we would not be able to make a clear-cut,
consistent distinction between two ideas. Without language
thought is a vague, uncharted nebula. There are no pre-
existing ideas, and nothing is distinct before the appearance of
language. The characteristic role of language with respect to
thought is not to create a material phonic means for expressing
ideas, but to serve as a link between thought and sound. 5

But how far Saussure can really be called determinist is open to

question. Even if he is correct in asserting that thought without
language is formless and vague (can aphasics, deaf-mutes and
animals with problem-solving abilities 'think' and is their
'thought' an uncharted nebula?) and that there are no innate or
pre-existing ideas, he seems to be saying not that language
determines thought but that the two are inseparable.
For neo-Saussureans like Lacan, the determinism of Saussure
lies in his doctrine that the sign is arbitrary, functioning only in a
system of signification. The whole continuum of experience is
segmented in an arbitrary way by the signifiers of the language;
and since for Lacan our entry into the world of experience is
effected by learning a language, that arbitrary classification is the
one that becomes our reality. As Marks and de Courtivron
96 Feminism and Linguistic Theory

observe, for the Lacanian 'meaning is located not in the thoughts

of the enunciator but in the system of signs itself. 4 The .world into
which we are socialised is not a world of the things themselves,
but a symbolic order fixed by the linguistic system.
This need not be accepted without question. Learning
language could be a matter of imposing one (linguistic) symbolic
order on a prior order with different, or multiple, determinants.
One linguist, Ian Griffiths, speaks of language exploiting a 'jigsaw
principle' (i.e. words are cut across the pattern one would
naturally use to make sense of reality, just as one tries to cut a
picture in non-obvious ways when one is making a jigsaw puzzle).
He observes,

The fundamental Saussurean misbegotten structural semantic

fallacy is that words represent an amorphous undifferentiated
reality ... the jigsaw principle on the contrary requires an
intrinsically highly structured pre-patterned reality across
which the lexical order can cut. 5

Saussure certainly did locate meaning outside anyone

individual language user. Langue for him was a collective
phenomenon, 'the social side of speech, outside the individual
who can never create nor modify it by himself. 6 However, it is
unclear whether he would have subscribed to all the implications
of this manoeuvre. We must return later to the rigid separation of
meaning and the individual from which so much Lacanian theory
springs.

The roots of determinism: Sapir and Whorf

Within linguistics, the doctrine of determinism is associated not

with Saussure but with the American anthropological linguists
Sapir and Whorf. Indeed, the idea itself is often labelled
'Whorfian' and known technically as the 'Sapir-Whorf
hypothesis' .
Sapir-Whorfian determinism reflects a rather different set of
preoccupations from those of Saussure and the semiologists. Its
starting point is not the linguistic postulate that each language is
a system of differences but the empirical observation of various
cultures. The question for Sapir and Whorf, both of whom
 Silence, A lienation and Oppression 97

studied native American culture extensively, was why the

populations they observed seemed to have such different
perceptions of 'the same' realities. Their answer was that
linguistic differences determined differences in world view.

Human beings are very much at the mercy of the particular

language which has become the medium of expression for their
society .... The fact of the matter is that the 'real world' is to a
large extent unconsciously built up on the language habits of
the group. No two languages are ever sufficiently similar to be
considered as representing the same social reality. The worlds
in which different societies live are different worlds, not the
same world with different labels attached. 7

It is significant that Sapir uses the word 'unconsciously': he

implies that to reconceptualise the world outside your own partic-
ular language is difficult because speakers have no awareness of
the kinds of distortion language causes. Whorf, too, stresses the
unfreedom of the individual: 'no individual is free to describe
nature with absolute impartiality but is constrained to certain
modes of interpretation: even when he thinks himself most free'. 8
Yet Sapir did not feel, as later feminist enthusiasts of determinism
appear to do, that what is unconscious must remain always
hidden and unchallenged. He implied that it was the mark of an
advanced culture that it could generate forms of knowledge that
transcend the tyranny of grammar: 'as our scientific experience
grows we must learn to fight the implications of language'. 9
Sapir-Whorfian determinism is rather different from the
Saussurean brand. For the Saussurean, language is an autonom-
ous system which replaces the world of things, an intermediary
between brute reality and human perception. The more anthro-
pologically inclined Whorfian, however, sees language as a mode
of action that interpenetrates with experience to the extent that
words are things.
Whorfian determinism also stresses different things from those
which interest the Saussurean. For instance, Whorf is much more
concerned with patterns of sentence structure, the subtle distinc-
tions expressed by grammatical particles, etc., than with the
meaning of individual signs or words. 1O For Whorf the
implications of grammar escape the language user much more
easily than those of lexical meaning (and to this the post-
98 Feminism and Linguistic Theory

Chomskyan linguist would add that syntactic patterns are learned

in a much more rigid and less idiosyncratic way). For Saussure,
however, they would have been a matter of parole and thus not
determining in the same way.
It might well be asked what evidence Sapir and Whorf put for-
ward for their particular deterministic views. As I have already
observed, their work was mainly on American Indian communi-
ties, and the most famous Whorfian study is Whorfs own study of
the Hopi.
Whorf focused on a number of things about the Hopi, partic-
ularly their very un-European conception of time (they did not
see that one could have a quantity of time, or cut the time
dimension into segments). Whorf explained this peculiarity with
reference to the Hopi language, which has no tense system.
The most obvious question this raises is whether Whorf did not
put his linguistic cart before his conceptual horse. In other words,
perhaps the Hopi, for historical and economic reasons, developed
a way of life in which the European concept of time was an
irrelevance; and therefore grammatical tense never entered their
language.
This point can be made even more clearly with reference to the
well-known observation that Eskimos (Inuit) have a multiplicity
of terms for snow. Although having a vast number of terms for
something may help you to perceive its gradations in an ordered
way, clearly the terms must have been invented in the first place
because they were useful and necessary in the circumstances of
Inuit life. Discussion of this point is not just nitpicking .. A
Whorfian outlook is useful as a corrective to the common-sense
view of language as a transparent medium for expressing
thoughts and perceptions; but it is important to define the limits
of determinism.
Two questions in particular seem pertinent. First, if signifiers
do cut up the conceptual universe differently in different
languages, is this completely arbitrary or does it reflect facts
about a culture which are obviously non-linguistic? (the Inuit are
a good example here: for it would surely be very remarkable if
instead of having a lot of words for snow they had a selection of
terms for horse-hide). Second, how great is our ability to see
around the categorisation scheme of our language? It is relatively
easy or relatively hard to make interpretations that are at odds
with your language?
 Silence, A lz'enation and Oppression 99

These questions are relevant to the feminist debate on whether

language reflects or causes women's oppression. If language
segments the conceptual universe in accordance with non·
linguistic cultural norms (e.g. sexual differentiation of an
extreme kind, and the devaluing of women) then it is not
inculcating a world view but obeying the dictates of one. And if it
is not difficult to reject the cultural assumptions built into your
language then even its power to reinforce those norms must be
limited, and we must look for non-linguistic reasons why they
persist.
A number of arguments have been marshalled against the
Sapir-Whorf hypothesis. One is a belief in universals of language:
if you believe that languages are at bottom all very similar,
because all speakers have the same mental apparatus, it would be
awkward to have to admit that the undeniably diverse
perceptions of different communities were determined by their
languages. Universals have been an article of linguistic faith ever
since Chomsky, and serious engagement with Whorfian ideas has
consequently ground to a halt.
The second argument, whose proponents do not necessarily
have to believe in linguistic universals, concerns translation and
language learning. Once again, no one denies that there will be a
certain amount of cultural relativity. Since speakers of different
languages do undoubtedly inhabit differing worlds to some
extent, they will certainly have diverse ways of interpreting reality
which complicate the process of translation from one language to
another. Lyons representatively observes that 'true bilingualism
implies the assimilation of two cultures'.n
If one adopted an absolute determinism, however, translation
and second-language learning would have to be made out more
difficult than is actually the case. Consider, for example, the case
of a Hopi learning English. Until she understands our conception
of time she will not be able to use the tense system correctly: but
in a totally Whorfian world she cannot understand our
conception of time without mastering the English tense systeml If
we are not to be caught in this kind of paradox we must set some
limits to our belief in determinism.
Finally, there is an argument from the creativity of language
users, which leads to language change. It is clear that as
conditions alter, speakers can and do modify both their frame of
reference and their language. Nor do they do this simply by
100 Feminism and Lingwstic Theory

adding new elements to the existing stock: often the meanings of

old elements shift. Remoulding of our linguistic classification
system is constantly being undertaken. Perhaps, then,
determinists err in believing, as George Orwell put it, 'that
language is a natural growth and not an instrument which we
shape for our own purposes' .12

CONTROL

If the first assumption of feminist theories of language is

determinism, the second is control. Radicals believe that the
deterministic powers of language may be exploited, and routinely
are exploited, by the privileged groups who control language for
political ends. This proposition of control would be more
precisely expressed as two claims. The first is that powerful
groups can appropriate language in the same way they can
appropriate, say, financial resources; while the second is that by
doing so they are able to exercise subliminal control over their
subordinates and maintain their own power. Both these claims
will be fully discussed within the critique presented in Chapter 8:
what is of interest at this point, however, is the nature and status
in our culture of the idea that language functions essentially as a
means of political control.
Within linguistics there has never been a debate on control
comparable with the debate on determinism (perhaps because
determinism has been rather unpopular, and control is seen to
depend on it to a great extent). It might then seem all the more
remarkable that feminist theories take control, or more precisely
male control, for granted. In fact, of course, the idea of control is
too entrenched in liberal humanist thought to need any
validation from linguistic science. It has passed into the realms of
accepted wisdom.
The 'control through language' argument is normally deployed
in the ongoing debate about democracy and totalitarianism.
Commentators equate totalitarianism with a language under
strict control (some even go so far as to suppose that a 'corrupt'
language gives rise to totalitarianism); conversely, the defence of
language constantly undertaken by writers, critics, philosophers
and journalists is put forward as a defence of democratic values.
For the liberal humanist, a totalitarian state is not just one where
 Silence, A lienation and Oppression 101

you may not say what you like (e.g. because you will be put in
prison and forced to recant) but more terrifyingly, one where you
literally cannot say what you like, because the state or other
repressive agency has control of the words you would say it in.
A good example of this type of argument is to be found in a
Times article by the philosopher and Tory pundit Roger Scruton,
titled (with an obligatory bow to Orwell, the major populariser of
the whole idea during this century) 'How Newspeak leaves us
naked',u The article purports to review a Soviet dictionary, the
Novosti Press Agency's Short Guide to Political Terms, and thus it
is basically an attack on state control of the language: since
Scruton is a Conservative, however, it also attributes linguistic
'imperialism' to English political groupings, especially the
women's movement.
Scruton opens by demonstrating that he believes in determin-
ism, and in conscious conspiracies to subvert language by left-
wing extremists: 'If you want to control the world, first control
language: such has been the unspoken maxim of revolutionary
politics in our century.' His major aim is to show that there has
been such a conspiracy of Soviet lexicographers to undermine the
'true' meanings of words like democracy and liberalism, so that
anything other than Party dogma becomes literally unspeakable:

Its purpose is to forestall refutation by securing an unshakeable

bridgehead in language. Communism ... has tried to ensure
that the words of politics can be used only to express dogmas of
its own- not because it believes those dogmas to be true, but
because it wants lies to take the place of truth.

Scruton typically takes it for granted that there is an unshakeable

bridgehead in language. Once you have corrupted the system,
you have taken away all your adversaries' resources: 'all rival
creeds are to be appropriated and devoured; no words will
express them, since no words will be available that have not been
enslaved by falsehood.'
Although presumably radical theorists would violently disagree
with Scruton's views on the Novosti Guide, they would be on
remarkably shaky theoretical ground in doing so. For after all,
what is the difference between the apocalyptic view expressed by
Scruton and the feminist belief that men have appropriated the
language? The progressive and the reactionary are locked into the
same view of their object.
102 Feminism and Linguistic Theory

Another preoccupation the reactionary shares with the

progressive is with the supreme value and importance of
language. Scruton observes, 'one is struck by the outrage that
communism has committed against language'. Feminists similarly
feel that language is extraordinarily powerful and important:
'how can we conceive of a revolutionary struggle that does not
involve a revolution in discourse?'14 Language is not unimportant,
but nevertheless there is something very distasteful about such
language-olatry. Outrages cannot be committed against
languages but only against their speakers. Marks and de
Courtivron, comparing the French and American traditions,
observe that few English-speaking feminists 'can believe in the
reduction of reality - oppression, suffering - to language',l5 It is
important to consider, as linguistically-based radical feminism
becomes more popular, whether this common-sense American
attitude does not have a good deal to recommend it.
Just as a feminist theory of language must engage very seriously
with the problem of linguistic determinism, confronting and
dealing with the difficult questions it raises, so it must engage
with the notion of control. As I have been trying to show, the idea
that language, or more specifically meaning, can be manipulated
to obliterate dissent is very ingrained, so much so that to imagine
an alternative model of language and its relation to politics is
almost impossible. Nevertheless, feminists must ask a number of
questions about the accepted model. How do men effect control
of languages and how does it work in practice? Are there limits to
it? How important is it relative to other forms of control? Finally,
what view of language does it presuppose to talk about male
control, and is this view tenable?
Determinism and control are the twin foundation stones on
which feminist theories of women's oppression, alienation and
silence are built. It is time now to look at some of those theories,
paying attention to the forms of determinism and control they
postulate, and not forgetting the general background problems
we have been discussing.

DOMINANT AND MUTED: WOMEN'S REALITY, MEN'S

REPRESENTATION

The 'dominant and muted' model proposed by anthropologists

 Silence, A lz'enation and Oppression 103

Shirley and Edwih Ardener has been very influential; although it

is not itself the product of radical feminism, it is cited by such
radical theorists as Dale Spender and Cheris Kramarae in their
own analyses. The basic premise of the dominant/muted model is
that while every group in a society generates its own ideas about
reality at a deep level, not all of these can find expression at a
surface level because the 'mode of specification' or communi-
cative channel is under the control of the dominant group. In the
case of men and women, women are in this relatively less
articulate position: they are, in the words of the Ardeners, a
'muted group' whose reality does not get represented.

This dominant model may impede the free expression of

alternative models of the world which subdominant groups
may possess, and perhaps may inhibit the very generation of
such models. Groups dominated in this sense find it necessary
to structure their world through the model (or models) of the
dominant group, transforming their own models as best they
can in terms of the received ones. 16

So briefly, the Ardeners believe that women do generate a reality

of their own, but have no means of encoding it linguistically.
They have to perform a sort of translation into the male mode of
specification.
The Ardeners have been at pains to point out that linguistic
silence is not the defining characteristic of a muted group: They
may speak a great deal. The important issue is whether they are
able to say all they would wish to say, where and when they wish
to say it. '17 The point is that for whatever reason, muted groups
such as women do not generate a mode of specification, and their
expression of their experience is structured by someone else's
language: 'the muted structures are "there", but cannot be
"realised" in the language of the dominant structure. '18 Three
questions seem to me to arise from this account of what Shirley
Ardener has called 'differing orders of perception'. For the
anthropologist they may be of marginal interest only, but for the
linguist, or the feminist, they are vital.

The problem of determinism

It is clear that the dominant/muted model IS not radically

104 Feminism and Linguistic Theory

deterministic, for after all it is axiomatic that 'the muted

structures are there', and this would seem to rule out any notion
that language determines thought or perception. It is interesting
to note, though, the degree of equivocality the Ardeners display
on this question, particularly in Shirley Ardener's later paper
'The nature of women in society'. Here she observes, 'Words
which continually fall on deaf ears may, of course, in the end
become unspoken, or even unthought. '19 The hint of determinism
is even stronger a page later when Ardener enquires, 'Are they
[women] able to think in ways which they would have thought
had they been responsible for generating the linguistic tools with
which to shape their thoughts?'20
There is a vagueness here, or maybe even a contradiction. For
if we are to think of language as shaping people's thoughts, given
that there is no language to encode what women think, why
should the muted structures be 'there' at all? Should not male
domination of language ensure that everyone's model be the
same? And if on the other hand language does not determine
people's thoughts (so that alternative models would be predicted
by the theory) we are left with the rather perplexing question of
why women, and other muted groups, can generate underlying
models of reality but cannot generate a mode of specification to
express them.

The problem of control

It is this problem which seems to me at once the most important

and the most intractable within the Ardener model. It simply is
not clear that a linguistic reason exists why women who are able
to generate a model of reality independent of the male model
cannot also generate independent ways of representing their
reality linguistically. Indeed, it is not even clear that women do
not in fact generate languages with which to represent both their
reality and their separate female identity. The paper in which
Edwin Ardener initiates the whole discussion actually deals with
such a language, the secret Liengu (mermaid) language of
Bakweri women. 21 Other examples exist, and the sex differences
now coming to light may also reflect women's conscious efforts to
tailor language to their own models.
The implications of remarks made by both the Ardeners is that
 Silence, A lienation and Oppression 105

women's inarticulacy results from men's control over language.

However, it is also possible to find an alternative suggestion to the
effect that men do not control meaning at all. Rather, women
elect to use modes of expression men can understand, because that
is the best way of getting men to listen .

. . . there are dominant modes of expression in any society

which have been generated by the dominant structures within
it. In any situation, only the dominant mode of the relevant
group will be 'heard' or 'listened to'. The 'muted groups' in any
context, if they wish to communicate, must express themselves
in terms of this mode, rather than in ones which they might
otherwise have generated independently.22

So in this case we are discussing a social rather than an individual

or mental suppression of female language: it is not that women
are unable to encode their experience, but that to do so is socially
unproductive and politically inexpedient.
The linguistic data on all sorts of muted groups tend to suggest
that they do generate specific modes of language use, but engage
in 'code switching' in order to function in societies where they are
subordinated. Women together can talk about their experience
and may even share (like the Bakweri mermaid women) ritual
channels of communication which men cannot understand or
may not use. Women with men, however, will indeed practise
translation, for as Shirley Ardener says, 'Unless their views are
presented in a form acceptable to men ... they will not be given
a proper hearing. '23 No one could possibly claim, of course, that
this was not itself an important disadvantage. It is obvious, for
instance (and the Ardeners point it out more than once), that
having to translate from one code into another handicaps women
in the public ___domain where linguistic exchanges are regulated by
conventions and traditions of an essentially man-made nature.
However, it is important to grasp the difference between saying
on the one hand, that women lack the means to express their
world view in language and are thus muted in society, and saying
on the other hand that women are muted because the kind of
language they use is unacceptable to men. To make the first
assertion is to claim that women have a linguistic problem; to
make the second assertion is to say that the problem is not one of
language but one of power.
106 Feminism and Linguistic Theory

The problem of power

The question of power is the one the Ardeners appear to shy away
from. They seem reluctant to consider that dominant groups are
dominant for some reason. Shirley Ardener observes, ' ... the
present way of distinguishing a dominant from a muted ...
model does not impose upon us an obligation to talk in terms of
"domination by men" or "the oppression of women" where this is
taken to be a purposeful male activity'. 24 This reluctance to face
the facts which are detailed in her own work is less trivial than it
might appear. 25 Talk of 'structures of dominance' rather than 'the
oppression of women' is consistent with the idea that muting is a
technical and linguistic phenomenon caused by structural factors
in a society, namely the ability of dominant groups to retain
control of language. If alternatively we were to regard muting as
a partial and situationally determined constraint placed by
women on their own self-expression as a survival tactic because
men reject their communicative channel, we would expect to see
some kind of coercion, some means of inculcating female
constraint and negative value judgements on women's talk,
available to dominant groups by virtue of their general position of
power. The social rules and taboos used in many cultures to
silence women (see Ardener25 and Zimmerman and West26 ) fit this
description exactly, and it is a pity that the dominant/muted
model does not go into the implications of these practices more
thoroughly.

LINGUISTIC DEVELOPMENTS: CHERIS KRAMARAE

In her book Women and Men SPeaking, the linguist Cheris

Kramarae attempts to formulate linguistic hypotheses based on
the dominant/muted model, and to examine the evidence for
each. The hypotheses she thinks follow from the Ardeners' theory
are these:
1 Women have difficulty in public speaking (because male
language is mandatory).
2 Men have more difficulty understanding what women mean
than vice versa (because members of muted groups have to be
 Silence, A lz"enation and Oppression 107
aware of dominant and muted models, whereas the reverse is
not true).
3 Women express dissatisfaction with dominant modes of
expression and search for alternatives. 27
Kramarae concludes that women are not less fluent than men in
public speaking (a rather strange conclusion, one would have
thought) but that perhaps this is because they have developed
superior skills to compensate for the need to translate from female
to male models. On the other two hypotheses she is more definite,
feeling that there is evidence to support both. She points out that
in anecdote, sociological survey and clinical practice alike, men
continually express their inability to discover what women think
or want. Women express no such misgivings about men.
But what kind of support does this give to a specifically
linguistic theory? Given that most women depend for their
material and emotional support on intimate knowledge of what
men want, it seems an unsurprising observation. Kramarae
certainly produces little evidence that it has anything to do with
the ways in which the sexes express themselves linguistically.
On the last hypothesis, that women are dissatisfied with
dominant modes of communication and are searching for
alternatives, Kramarae relies heavily on the attitudes of present-
day women's groups. As she points out, feminists have usually
done away with formal speaking arrangements and the rhetoric of
traditional male-dominated politics. Women writers, too, have
made particular genres their own (especially the novel and the
diary).
The dominant/ muted model of the Ardeners has a lot to offer
feminists, but it also contains much that is vague and misleading.
The Ardeners clearly show that man-made rules and institutions
limit the settings and situations in which women may effectively
speak, and they suggest that muted groups are obliged either to
translate into dominant modes of speech or be ignored. They
avoid the most problematic excesses of linguistic determinism and
place a realistic emphasis on the point many theorists overlook,
that women do actually have a model of reality and, far from
being passive and silent victims of muting, they may in fact
develop ritual channels for dealing with their experience.
What the model really fails to show is that muted groups lack a
language, and that dominant groups are able to appropriate all
108 Feminism and Linguistic Theory

linguistic resources. The alternative suggestion, that women

communicate adequately with each other but are institutionally
constrained/negatively judged in the public (male) arena, is
much more plausible.

MAN MADE LANGUAGE: PATRIARCHY AND THE

POWER OF DEFINITION

Dale Spender, whose book Man Made Language must be the

best-known text on women and language, is at once more radical
and more forthright than the Ardeners. Whereas the Ardeners'
notions of determinism and control are largely implicit, often
vague and unexplained, Dale Spender is not afraid to spell out
her belief in determinism and her ideas on how men control
language. The Man Made Language theory exemplifies all the
three points I mentioned earlier. According to Dale Spender, it is
through their control over meaning that men are able to impose
on everyone their own view of the world; women, without the
ability to symbolise their experience in the male language, either
internalise male reality (alienation) or find themselves unable to
speak at all (silence).

Determinism

Determinism is not so much argued as assumed, though the usual

case for it is set out at length:

Language is our means of classifying and ordering the world:

our means of manipulating reality. In its structure and in its
use we bring our world into realisation, and if it is inherently
inaccurate, then we are misled. If the rules which underlie our
language system, our symbolic order, are invalid, then we are
daily deceived.
Yet the rules for meaning, which are part of language, are
not natural; they were not present in the world and merely
awaiting discovery by human beings. On the contrary, they
had to be invented before anything could be discovered, for
without them there is no frame of reference, no order, no
possibility for systematic interpretation and understanding. 28
 Silence, A lienation and Oppression 109

But although initially the rules had to be invented, it is clear that

for Spender they are not invented afresh by each generation of
language users. They exist as a fixed system which is learnt by
children in the course of their socialisation: ' ... these rules have a
habit of becoming self-validating and self-perpetuating, regard-
less of any misapprehensions on which they may have been
based. '29
In Spender's symbolic order one does not construct meaning,
one 'enters' it: and since it embodies the male world view, having
originally been made by men. we unconsciously build up ways of
seeing and thinking which fit in with such a view:

While at one level we may support or refute the myth of male

superiority - it being a matter of political choice - at another
level we are unaware of the way in which it structures our
behaviour and forms some of the limits of our world. With the
crucial underlying rule that the world can be divided into plus
male and minus male categories we have seen the construction
of patriarchal order. It is a symbolic order into which we are
born, and as we become members of society and begin to enter
the meanings which the symbols represent, we also begin to
structure the world so that those symbols are seen to be
applicable: we enter into the meaning of patriarchal order and
we then help to give it substance, we help it to come true. 30

It is noticeable that Spender draws on a number of different

strands of determinism. The concept of a symbolic order which
we enter is Lacanian; but elsewhere Spender insists on her debt
to Sapir and Whorf.

On the one hand there is the evidence that not all human
beings are led to the same view of the world by the same
physical evidence and on the other hand is the explanation -
namely the Sapir-Whorf hypothesis - that this is because of
language. It is language which determines the limits of our
world. which constructs our reality. 31

Yet she is much more extreme than Sapir. apparently believing

that the constraints imposed by language cannot be escaped:

Human beings cannot impartially describe the universe

because in order to describe it they must first have a
110 Feminism and Linguistic Theory

classification system. But paradoxically, once they have that

classification system, once they have a language, THEY CAN SEE
ONLY CERTAIN ARBITRARY THINGS. 52

This very strong determinism is the background to all Dale

Spender's ideas about language.

Control

Spender's central claim, however, is that men control language

and in particular, meaning. Women have been systematically
excluded from the making of meanings. This is explained with
reference to a particular view of how knowledge is constructed,
put forward by Dorothy Smith:

This is how a tradition is formed. A way of thinking develops in

this discourse through the medium of the printed word as well
as in speech. It has questions, solutions, themes, styles,
standards, ways of looking at the world. These are formed as
the circle of those present builds on the work of the past. From
these circles women have been excluded ... deprived of the
means to participate in creating forms of thought relevant or
adequate to express their own experience or to define . . . their
situation and concerns. They have never controlled the
material or social means to the making of a tradition among
themselves or to acting as equals in the ongoing discourse of
intellectuals. 55

It is alleged that men build up a tradition of received wisdom -

facts, theories, ways of seeing and interpreting - by checking
their contributions with other men. No one in the charmed circle
is likely to ask himself, or anyone else, what women think and
whether pieces of knowledge concerned with women strike the
female population as fair and accurate.
As an account of female non-participation in academic/
intellectual life, this is fair enough. One glance around any
university, one look at any intellectual TV or radio broadcast,
will reveal that charmed circles contain precious few women. It is
also impossible to forget that until recently men were able to deny
us, by statute, entry to the professions, to higher education,
before that to any education, even simple literacy. It is true that
 Silence, A lienation and Oppression III
the material means (money, educational resources) and the social
means (freedom from male persecution and ridicule) never have
existed for women to participate in academic life on an equal
footing with men: indeed they still do not exist.
It is surely not true, though, that women have never had a
tradition of their own. Of the thousands burned as witches, for
instance, many were punished precisely because they possessed
traditional knowledge which was denied to men. But in any case,
what is at issue is not Smith's highly institutionalised 'knowledge',
which can indeed be regulated by legal provisions and financial
considerations, but language. It is not at all clear that Smith's
analysis of the construction of meanings within the intellectual
tradition will do as a model of how the meanings of language are
constructed in everyday interaction. Everyday language is not the
same thing as the ongoing discourse of intellectuals, and Spender
needs to explain how the same processes of exclusion and
validation can go on in both. This is a matter I shall return to in
the next chapter. Nevertheless, the man-made language theory
states that men and men alone define meanings from their
vantage point of difference and dominance. For women, these
meanings are false and alienating, as Spender tries to show by
considering the word motherhood.
According to Spender, men have built a 'positive' meaning into
the word motherhood which makes it impossible for women to
discuss, or even mentally acknowledge, any negative experiences
they might have of being mothers. The male-ordained meaning
of the word ensures that 'unhappy motherhood' is a contradiction
in terms. The clash between what a woman knows of motherhood
and what she knows the word means forces her into silence. Her
own meaning is outside the norms of language; might her
experience be abnormal also? If she speaks the truth, will she be
labelled as unnatural and deviant as the phrase 'unhappy
motherhood?'
Dale Spender believes that the entire lexicon can be analysed
like this: every single meaning is literally man-made, and,
inevitably then, words encode a male point of view which is often
at odds with female experience. This is the source of alienation
and silence; in a man·made language you either see things
through male eyes or you reject existing words, silencing yourself.
Spender believes that men encode in language not only their
world view but their conviction they are superior to women. This
112 Feminism and Linguistic Theory

is achieved by an all-embracing semantic rule that anything to do

with women becomes negatively marked and pejorated. Women
and men thus learn to perceive the world through a haze not just
of maleness but of misogyny.
The great unanswered questions in this model of meaning are:
how is meaning made, and how is meaning learned? It surely
cannot he a matter of men writing dictionaries and women
looking up words in them, since that is not how anyone learns to
speak. But if men can infer meaning directly from their own
experience, the theory denies women that possibility. If women
constructed meanings according to their experience they would
have ideas about words very different from men's, and it is crucial
to this theory that in fact they should learn the alien male
definitions, which force them to internalise a male world view.
How then do they do it?
Spender would presumably fall back here on the neo-
Saussurean notion of 'entering' meanings which are already
available to your society as a fixed symbolic order, rather as you
might learn the traditional folktales of your culture. Yet this just
pushes the problem back in time: how was the symbolic order
constructed and by whom? Ultimately Spender's answer must be
'by men' and some account of this must be given.

Power

The mystery of why women have not encoded meanings based on

their own experience deepens when we consider Spender's
apparent belief that they can and should start to do so
immediately as a political strategy. The blurb of Man Made
Language enthuses, 'once women expose the falseness of male
meanings and encode their own, language and society can assume
new forms, and women can move towards autonomy and self
determination'. In Spender's work it is always rather unclear
whether women are without linguistic resources because they are
without power (in which case they cannot change language
without changing their status first) or whether they need
linguistic resources before they can increase their power.
Although Spender gives considerable weight to the Whorfian
notion that a patriarchal reality is constructed primarily through
a patriarchal language, sometimes she hints that linguistic
systems are purely superstructural:
 Silence, A lienation and Oppression 113

'Any exposure of the false nature of male superiority, while not

a direct attack on male power, is an indirect attack which
undermines it. If and when sufficient members of society no
longer act in a manner which acquiesces in that superiority and
permits it to go unchallenged ... that power will need to be
defended or transformed. '34

In this formulation Spender reminds us that male superiority is

not male power, but the justification of it. Language is a means of
upholding not power but the notion of superiority that makes
power look natural or fair. Undermine language and you force
men to defend their power in some other way, or to change its
nature; you expose the myth of superiority for what it is, thus
inciting women to attack male power, now revealed as non-
natural and unfair.
This is a far cry from Rich or Cixous, who consider women's
'meanings' transforming in themselves. It is a shrewd account of
why we should dismantle cultural prejudices as well as sexist
institutions. So it is hard to see why Dale Spender lands herself
with the apparatus of linguistic determinism, which does indeed
confuse the power with the myth that justifies it.
These Whorfian/radical feminist versions of determinism, male
control and female alienation belong, both intellectually and
politically. to the Anglo-American tradition. In the next chapter
we must consider a rather different framework, where problems
of femininity and male power are located in relation to a radical
concept of what language is, and what part it plays in human
affairs.
7 Feminist Models of
Language (II): Semiology
and the Gendered
Subject
One is born in a language and
the language speaks us, dictates
to us its law ...
Helene Cixous
All Western discourse presents
a certain isomorphism with the
masculine sex.
Luce lrigaray

In recent years there has been much more widespread awareness

than ever before of a serious challenge to the linguistic and
anthropological tradition of Shirley Ardener, Dale Spender and
their colleagues throughout the English-speaking world. That
challenge comes from the discipline I have been calling
'semiology', and what it claims to offer through a synthesis of
Freud, Marx and Saussure, is an explanation of the human
subject him- or her-self.
Trying to explain semiology to a non-specialist audience,
especially in relation to mainstream positivist linguistics, is a
project fraught with difficulties, not just because semiology is
complex and sophisticated with its own extensive terminology,
but also because the shift of viewpoint necessitated by any such
explanation may well do the less immediately graspable paradigm
a covert injustice. As Coward and Ellis, themselves the authors of
an introductory text, observe,

The flood of translations and of introductions in layman's

terms (i.e. within another conceptual framework, according to

114
 Feminist Models of Language (II) 115

another view of the world) is witness to more than just a desire

to neutralise this work. It shows that the traditional disciplines
... are dissatisfied with their own methodology and want to
import a controlled dose of new concepts that can cope with
'problems' that have arisen. These traditional disciplines
appropriate concepts in a piecemeal fashion to paper over their
cracks, when in reality the cracks bear witness to the weakness
of their theoretical foundations .... The result is ... traditional
British eclecticism.!

It must be all or nothing, Coward and Ellis imply. The reader

must try to enter into the semiologist's universe, and cannot ex·
pect to judge what she finds there by the same standards she has
used outside. Simplification will always be over-simplification,
and explanation 'in layman's terms' risks being called reductive
by those more familiar with the framework.
Nevertheless, it is essential to give some account of semiological
theory. Its recent concern with the construction of sexual identity
in the developing subject, and its insistence on a linguistic basis
for that process of construction, makes it extraordinarily relevant,
a theory that speaks to feminist concerns. And indeed, it has been
appropriated by several feminist tendencies. No discussion of
feminism and linguistic theory would now be complete if it did
not pay close attention to these developments.
As a relatively easy way into the sophistication of theories based
on Marx and Freud as well as Saussure, we may consider some of
the specific disagreements between semiologically oriented writers
and Anglo-American sociolinguistics. These emerge rather clearly
from a review article on Dale Spender's Man Made Language that
appeared in Screen Education under the authorship of Maria
Black and Rosalind Coward. 2 In taking exception to Spender'S
account of language and women's oppression, Black and Coward
reveal preoccupations typical of the semiological approach. They
object particularly to Spender'S 'simplistic' notion of power, and
to her empirical and mechanistic interpretation of meanz·ng.

POWER

The dispute about power, is not, perhaps, strictly linguistic, but

116 Feminism and Linguistic Theory

it does emphasise the important fact that Anglo-American and

French writers typically embrace very different concepts of
feminism. Dale Spender's is what might be termed 'radical
feminism': it posits a social organisation in which all men have
power over all women, deriving from their right to exploit women
materially in marriage, and from the use, actual or threatened, of
institutionally sanctioned physical violence. 3 Black and Coward's
feminism owes more to later marxist models (and particularly to
the ideas of Louis Althusser and Michel Foucault) in which social
organisation is a complex matter structured by ideologies - sets of
representational practices which construct power relations
without needing to resort to obvious coercion. Although the
sexual power relation is constructed to men's advantage, theorists
in the Black and Coward mould would think it vulgar to
implicate all men in its construction, or even in the exercise of
power over all women. They would stress the extent to which we
are all constructed, and the many different dimensions of power
simultaneously operative in one society. It is it). this context that
we can understand Black and Coward's objection to Spender
assuming 'that one group literally has power over the other'. 4

MEANING

Both Dale Spender and the Ardeners contend that women

generate (or would generate) different meanings because their
experiences of the world are different. The problem is that men
control the processes by which meanings are encoded in
language, and therefore language represents only male
experience, excluding female 'meanings'.
For Black and Coward, it is simply incoherent to talk about
meaning in this way, as if it simply arose from experience,
innocent of language or of ideology. Meanings for them are made
possible by language, and language structures, rather than
expressing, the individual's experience:

[Dale Spender's] understanding of language is problematic,

assuming as it does that 'meanings' derive from clear-cut
groups, generated by their different social experiences ....
Meanings reflect the individual's experience of reality and are
 Feminist Models oj Language (II) 117

simply expressed in language. The communality of meanings

... is seen as an effect of groupings of individuals with
structurally similar experiences. 5

For Black and Coward on the contrary, 'experience and identity

cannot be seen as the origin of meaning, but as its outcome'.6
This is the ultimate Saussurean determinism, which sees
experience and indeed the individual herself, as a product and
function of the institutionalised system of signs. 'Language has a
material existence. It defines our possibilities and limitations, it
constitutes our subjectivities. '7 This proposal that meaning does
not arise from experience initially strikes most people as running
entirely counter to common sense. It is, however, an integral part
of semiological approaches which 'decentre the subject', that is,
deny there is any unique stable core of personality, human
nature, etc. from whose experience meanings could come.
As we have seen, subjectivity is the condition of a social being,
aware of her separate existence and the laws of her culture. In
contemporary theory of the subject, the process by which this
subjectivity develops in children is discussed in terms of
psychoanalysis and linguistics; therefore we must look more
closely at the psychoanalytic and linguistic account of subject
development to be found in the work of Lacan, and at the
development of Lacan's ideas by other writers.

PSYCHO ANAL YSIS

It might well be asked why feminist theory should draw on

psychoanalytic concepts. There is no doubt that psychoanalysis as
a practice has been very oppressive to women, requiring them to
'adjust' to the stereotype of a passive, powerless and sexually
masochistic femininity. Psychoanalysis as a theory, however, is
currently popular among feminists because it seems to offer us,
through its account of the construction of the self in family
relations and the unconscious mind, an understanding of how
subordination can be internalised deep in our personalities.
Moreover, it is centrally concerned with the forging of sexual
identity and with the extreme importance of the sexual in all
aspects of mental life. This too has important implications for a
118 Femz'nism and Linguistic Theory

politics that stresses sexuality, as feminism does.

Anti-humanism and the Freudian unconscious

The important thing about Freud's work, according to Freudian

semiologists, is that it made possible a revolutionary new concept
of what a human being is. We think of ourselves as stable and
unified entities who 'have' experiences, personalities, sexualities
and so on, but this comfortable 'humanism' is just as illusion.

For the psychoanalyst the human subject is a complex entity, of

which the conscious mind is only a small part. Once one has
accepted this, it becomes impossible to argue that even our
conscious wishes and feelings originate within a unified self,
since we can have no knowledge of the possibly unlimited
unconscious processes which shape our conscious thought.
Conscious thought, then, must be seen as the overdetermined
manifestation of a multiplicity of structures which intersect to
produce that unstable constellation the liberal humanists call
the 'self. These structures encompass not only unconscious
sexual desires and unconscious fears and phobias, but also
conflicting material, social and ideological factors of which we
are equally unaware. It is this highly complex network of
conflicting structures, the anti-humanist would argue, that
produce the subject and its experiences, rather than the other
way round. s

If the monolithic human personality is dethroned by Freud's

discovery of the unconscious, the idea of a natural heterosexuality
is similarly shattered by his findings on psychosexual develop-
ment. Rather than possessing at birth an essential male or female
nature, with concomitant attraction to the other sex, in Freud's
scheme of things the developing child forges its sexual identity,
a mere and precarious contruct, in the crisis of its relationships
within the family. Thus Freud cleared the way for a radical and
materialist theory of the human subject as something not unified
and static, but created and l'ecreated in a constant process of
conflict and contradiction.
But what has this to do with language? The connection lies in
the notion of a 'symbolic order' or set of meanings which define
culture. In contemporary theory of the subject this symbolic
 Feminist Models of Language (II) 119

order is not only to be understood in the same way as a

Saussurean langue, that is as a system of differences, it is very
frequently equated with language itself. Coward and Ellis
explain: 'Because all the practices that make up a social totality
take place in language, it becomes possible to consider language
as the place in which the social individual is constructed. '9
It turns out, then, that the term 'subject', designating the social
individual in whom semiologists are interested, is a clever pun: for
the social individual as well as being the 'subject' of her own
perceptions and of the semiologist's enquiry, is a subject in the
other sense, 'subject' to the authority and prescriptions of some-
one or something. That something, according to semiological
theory, is precisely the symbolic order, or language. We are all
subject to the laws of language, which exist before we are born,
and our task in childhood is to insert ourselves into that order so
we may secure a place to speak from. If we fail, we become
psychotic. Subject development, then, is the process by which the
child claims a place in the symbolic order. And that order is
reminiscent of Saussure's langue, 'outside the individual who can
never create or modify it by himself'. 10

Lacan

Lacan, who as we have seen (Chapter 2) brings the insights of

Saussure to bear on Freudian theory, addresses himself to the
question of how children take up their place in the symbolic
order. He asserts that child development is a several-stage process
of 'splitting'. First the child must learn that it is separate from the
mother's body; in the mirror stage it recognises its own reflection,
further internalising the categories of self and other. Finally, it
must learn to separate itself as speaker ('I') from itself as spoken to
('you') or mentioned in the speech of others ('s/he'). Obviously
this final splitting of the speaking and the mentioned/ addressed
subject cannot come about without language, which teaches a
whole set of norms and orientations in relation to which the child
must locate itself. Language learning is a socialising as well as an
individuating process.
From our point of view, Lacan's most important claim is that
male and female children enter the symbolic order differently:
their relations to language differ. The reason for this difference
120 Feminism and Linguistic Theory

lies in the crucial significance of the phallus. To understand why

Lacan sees the symbolic order as ruled by the phallus, we must go
back to Freud's account of psychosexual development and his
ideas on civilisation. Freud, it will be recalled, believed that
children were inherently bisexual; but because civilisation
requires reproductive sexuality for its continuation, there must be
some process by which the child comes to define itself as male or
female, identifying with one sex and (preferably) desiring the
other. How does this state of affairs come about?
The answer, according to Freud, lies in the child's stormy
relations with its parents. Building on clinical observation, Freud
was led to posit an 'Oedipus complex', a stage of development in
which the child entertains feelings of sexual desire for its mother
and feelings of rage and jealousy toward the father, its rival for
the mother's affections. This state of affairs is resolved by the
'castration complex', a further stage, in which the phallus plays a
crucial symbolic role. It is the castration complex that ac·
complishes the differentiation of male and female subjects. Boys
overcome their sexual desire for the mother through their fear
that the father will castrate them, and then go on to identify with
the father, possessor, like themselves, of the phallus, and bearer
of the law. Girls, on the other hand, must recognise that they are
castrated, and will later seek to replace the lost phallus by having
children. For both sexes, the crucial fact of which they must be-
come aware in order to perceive and understand sex different-
iation is whether they - and their parents - have the phallus or
not, The law of culture, the symbolic, forces every subject to take
up a position on one side or the other, having or not-having, even
if in this non-biologistic account 'Anyone can cross over and in-
scribe themselves on the opposite side from that to which they are
anatomically destined'. 11 In other words, the castration complex
need not always result in 'normal' sexual development (i.e.
identification with the subject position of one's anatomical sex)
but it must always result in the taking up of one position or the
other.
According to Lacan, the child before its castration complex is
in the order of the Imaginary. Its close relation to its mother
means that it perceives no lack, no incompleteness, absence,
difference or unsatisfied desire. Awareness of these things comes
only when a third term, the phallus, is introduced by way of the
castration complex. This precipitates the child's transition from
 Feminist Models of Language (II) 121

the Imaginary to the Symbolic: and it is that transition, together

with the meaning of the phallus within it, that we must consider.
Because of its function in the castration complex, the phallus is
given two very powerful meanings. One of these is lack, for it
symbolises the loss of the mother's body. After the prohibition of
incest and threat of castration there can never again be the
closeness of mother and child that existed before the introduction
of the third term. The other phallic 'meaning' is the Law of
patriarchy, a social order in which incest is prohibited and
castration threatened by the father. So on one hand the phallus
symbolises loss, lack and desire, while on the other it symbolises
power and the social order.
For Lacan, it is precisely awareness of lack that impels a child
toward language. The idea that words can stand for things can
only be grasped when the child has some concept of something
missing or absent. Thus there can be no language until the
mother/child dyad is broken, and language depends on the
introjection of the phallus. This is why Lacan claims that the
symbolic order is dominated by the phallus.

ANATOMY AND DESTINY

Feminists have traditionally been very critical of Freudian

phallocentrism. They have asked, for instance, why children
should conceive the fear of castration, and tried to equate penis
envy with girls' envy of the power accorded men in society. So it
does perhaps need to be pointed out that castration and penis
envy in psychoanalytic theory are meant to be u~conscious
fantasies, not rationally motivated inferences consciously made by
the child. Indeed, they are the inverse of each other: those who
have the phallus (boys) nevertheless perceive that others do not,
and so they fantasise they will lose it; whereas girls, similarly
aware of their own lack and male possession of the phallus,
entertain the opposite fantasy. The term phallus is meant to
denote an object in fantasy, as opposed to penis which denotes a
real organ, and the Lacanian insists on a careful distinction
between the two.
Nevertheless, it can hardly be claimed that the phallus is not,
in some sense, the penis. Although we can 'explain' the apparent
122 Feminism and Linguistic Theory

oddness of castration fears by claiming they are unconscious and

not motivated by any occurrence in actual social life (for instance,
a real threat of castration by one's father), there is one real-life
event that is crucial in triggering anyone's awareness of difference
and unconscious fantasies about the phallus. That event is the
perception of visible difference. To entertain fears of castration
or envy of the penis, a child must somehow know that the genitals
of the other sex do not resemble its own. It must register this fact
and attach some initial significance to it. In the final analysis
then, anatomy cannot be disregarded.
This should not in itself be problematic for feminists, for after
all no one has ever denied the anatomical differences between the
sexes, nor their obvious importance in determining what our
experience of the erotic is going to be like. What does seem
problematic is the particular significance give the female genitals
in this account. Why should the child, comparing male and
female genitals, decide that the boy 'has' something that the girl
'lacks'? Why does the child already have the 'binary opposition'
mentality that constructs the world into sets of opposites? Any
such presupposition in psychoanalytic theory seems like prima
facie evidence for the charge of phallocentrism; (and the case has
indeed been argued by Luce Irigaray, whose work we will look at
later in this chapter).
The Lacanian commentator Jacqueline Rose replies that the
arbitrary interpretation follows from the value of the phallus as a
signifier, which is fixed by the symbolic order itself: ' ...
something can only be seen to be missing according to a pre-
existing hierarchy of values.... What counts is not the
perception but its already assigned meaning ... '12 The locus of
explanation returns, therefore, to language and the symbolic,
which alone invests difference with significance.

LACANIAN LINGUISTICS: THE SYMBOLIC ORDER

So far, we have been considering the general framework of

Lacanian ideas about identity and sexuality. We have seen why
Lacan believes children come to language, and why he believes
that the order of language is phallic, so that those who do not
possess the phallus are marginal to it. We must now examine
Lacan's view of language itself.
 Feminist Models of Language (II) 123

First, it is important to note a slight confusion which can easily

arise around the notion of the symbolic order. Sometimes in
Lacanian writing (which is not the same thing as Lacan's writing)
this appears to mean the totality of social and cultural practices,
including language, while at other times it seems to reduce to
language itself, which is alleged to create other cultural practices.
The two possibilities invite rather different responses (since the
second is more extreme in its determinism than the first) and
while we can leave the matter open for the present, the ambiguity
should be borne in mind.
Essentially, Lacan's view is a Saussurean one: language is a
system of signs defined by their difference from one another. But
Lacan differentiates himself from Saussure by pointing out that
the signifier is more important than the signified. Rather than
representing a signified in a one-to-one correspondence, a
signifier refers to the whole inventory of other signifiers from
which it is distinct. Only with reference to this chain of
associations do we work out what anything means. As Coward
and Ellis explain,

Language is seen to have the dizzying effects of a dictionary:

each word, definition by definition, refers to all the others by a
series of equivalents; every synonymous substitution is
authorised. Language results in tautology, without at any
moment having been able to hook onto any signified at all. n

The obvious question here is why in that case speakers imagine

they are making meaningful remarks that others can interpret,
rather than free-associating or playing Chinese Whispers. Lacan
meets this objection by speaking of 'points de capiton,' points at
which the hypothetically open-ended chain of signifiers actually
becomes closed. These 'points de capiton' are contextual
phenomena, artefacts of the moment, the situation or the
relevant social norms. In Saussurean terminology, they are the
limitation imposed by parole on langue.
At this point we seem to have encountered a puzzle. Lacan (like
many other neo-Saussureans) denies any strong indissoluble link
between one signifier and one signified (i.e. one form and one
meaning) and therefore concludes that meaning is a function of
context, intralinguistic associations, social structures and so on.
Yet on the other hand he insists that language itself is cons#tutive
of the social order; and furthermore, that both language and the
124 Feminism and Linguistic Theory

order it constitutes pre-exist the individual, whose identity and

experience is moulded by the need to insert herself into a
preconstituted social reality.
This raises a problem about the phallocentricity of the
symbolic order. For if the authority of the phallus, or more
generally male authority, is guaranteed neither by inherent
anatomical superiority nor by the language, where does it come
from? As the third term that breaks up the mother-child dyad,
the phallus may always be the spur to language acquisition: but
why should it mean what it does mean in the symbolic order?
Apparently Lacan has no answer. He asserts what we all know,
namely that the symbolic order (taken this time to mean the
totality of signifying processes, social, cultural and linguistic) is
patriarchal. Inserting oneself in our culture is a matter of
submitting to patriarchy. But his theory of language and sexual
identity offers no explanation why the symbolic order is
patriarchal. It can deal with matters of sexual differentiation, but
it cannot deal directly with sexual power. Possibly, of course, it
was never intended to. But in that case the feminists who have
taken it up (Cixous, for example) have been grievously misled as
to what its significance is, and they have overestimated its
political usefulness.
In fact it seems to me that despite all their disclaimers, despite
their awareness of the contextuality of meaning, Lacan and the
Lacanians do indulge in a covert Saussurean determinism, which
allows meanings to be fixed (though liable to 'slippage') by the
linguistic system. For the important concept in Lacanian
accounts of language acquisition is the idea of insertz"ng oneself in
a pre-existing order: and it is difficult to see how anyone could do
this unless the order, and the meanings it made available, were
fixed and stable, produced outside the individual and enjoined
on her as the price of entry into human society.
The neo-Saussureans claim to have improved on Saussure by
rejecting the idea of a disembodied system inhering in society and
substituting at the theory's centre the actual speaking subject.
But it turns out that this subject is constituted precisely by
locating herself in relation to a disembodied system that inheres
in society. We are all engaged in reproducing the reality language
creates: as speaking subjects, we are subject to fixed symbolic laws
not of our own making. It is this subjection to symbolic law that
Black and Coward have in mind when they claim that meaning
 Feminist Models of Language (II) 125

produces the individual's experience and not the other way

around. And although this view is precisely the opposite of Dale
Spender's, it raises, interestingly, exactly the same problems of
how meanings can be fixed and whether language does determine
reality.
For a number of feminists, Lacan's theory is taken to be an
explanation of women's oppression. Women are constructed in
the ___domain of a male sign, and therefore they are, as Lacan put
it, 'excluded by the nature of things, which is the nature of
words'.14 The nature of their oppression is once again alienation
from the symbolic, being forced to take up a marginal position in
the order that circumscribes what it is to be human.
Before looking at the general problems of determinism, control
and alienation, we must look at the way in which Lacanian ideas
have been used by feminists or at least (since in France the term
feminist is often seen as a tacit acceptance of the patriarchal
male/female dichotomy) by radical women theorists such as the
philosopher and critic Julia Kristeva, and the psychoanalyst Luce
Irigaray.

JULIA KRISTEV A

Kristeva's work is a development rather than a crItique of

Lacan's. She believes that femininity is constructed by the mode
of entry into the symbolic order, and her major contribution, so
far as the subject of this book is concerned, is to spell out the
radical rejection of anatomy as destiny entailed by Lacanian
linguistics, together with the implications of that rejection for
feminists.
Kristeva discusses the pre-Oedipal Imaginary stage before
language acquisition, in an interesting way. She suggests that
before the symbolic order there is a semiotic order linked to oral
and anal drives which flow across the child. The 'pulsions' of
these drives are gathered in a chOTa (which means,
approximately, a receptacle). Later, when the child takes up a
position in the symbolic order as a result of the castration
complex, the contents of the chora will be repressed, but its
influence will nevertheless be discernible in linguistic discourse
through rhythm, intonation, gaps, meaninglessness and general
126 Feminism and Linguz'stic Theory

textual disruption. Indeed some discourses, like art, poetry and

madness, draw on the semiotic rather than the symbolic aspects
of language.
For Kristeva, there is a possibility of choice about what subject
position the child takes up (male, fully integrated into the
symbolic, or female, marginal to it). It is not a matter of biology
but of identification with one parent or the other. Thus
femininity, defined as marginality to the symbolic order, is open
to men as well as women. Anyone taking up a feminine subject
position retains strong links with the pre-Oedipal mother figure,
and their language shows the influence of the chora to a marked
degree. It is this ability of the chora to break through rational
discourse that gives, for instance, French modernist poetry its
distinctive quality:

The modernist poem, with its abrupt shifts, ellipses, breaks

and apparent lack of logical construction is a kind of writing in
which the rhythms of the body and the unconscious have
managed to break through the strict, rational defences of
conventional social meaning. Since Kristeva sees such meaning
as the structure that sustains the whole of the symbolic order -
that is, all human social and cultural institutions - the break·
down of symbolic language in modernist poetry comes to
prefigure for her a total social revolution. IS

Here we see how important is Kristeva's emphasis on femininity as

non-biological - for the poets she deals with - Mall~rme and
Lautreamont - were, of course, men.
But if it is one's position in the symbolic that makes one
masculine or feminine, the identification of these categories with
biological maleness and femaleness is exposed as a trick.
Femininity and womanhood are not the same, but patriarchy
makes them appear identical. And because women, inaccurately
but unproblematic ally identified with feminine subjects, are
marginal to the symbolic order, they come to represent the
boundary between symbolic order and imaginary chaos. As Toril
Moi observes,

It is this position which has enabled male culture sometimes to

vilify women as representing darkness and chaos - the chaos of
the chora or of the imaginary, one might add - and sometimes
 Feminist Models of Language (II) 127

to elevate them as the representatives of a higher and purer

nature than men. 16

Kristeva sees different kinds of feminism as embodying different

attitudes to the symbolic. Liberal, equal. rights feminism
demands for women an equal rather than a marginal place in the
symbolic. Radical feminism extols the distinctively feminine,
which means it rejects the value placed on the symbolic
altogether. A further stage, which Kristeva speaks of as the 'third
generation', would reject the very opposition of male and female
as metaphysical, and attack the whole notion of sexual, or even
general, identity.
We have not yet got beyond the radical stage, and it remains
necessary both to reclaim and to proclaim the feminine within
language. But unless we are aware that the masculine/feminine
dichotomy is itself to be questioned, our struggle will not get
beyond the inverted sexism which comes from keeping the same
old categories and merely swapping round the positions of those
who occupy them. Because in Kristeva's theory language indeed
'defines our possibilities and limitations ... constitutes our
subjectivities', there can be no escape from the authoritarian and
phallocentric symbolic order. The alternative to the symbolic is
psychosis. But the order can be subverted from within, and that
subversion for Kristeva is a revolutionary act.

LUCE IRIGARA Y

Luce Irigaray is a psychoanalyst trained in the Lacanian

framework. During the 1970s, however, she mounted a relentless
critique of Lacan, focusing especially on questions of femininity
and language. She was expelled from Lacan's school after the
publication in 1974 of her book Speculum de l'autre femme.
lrigaray's main objection to Lacan is that his system cannot admit
any kind of plurality, either in sexuality or in language (which for
her, as for him, is closely linked with sexuality). This leads to
women's otherness or difference being denied, and their mere
oppositeness asserted.
To Lacan's famous pronouncement, 'the unconscious is
structured like a language', Irigaray asks 'which language?' For
128 Feminz'sm and Linguistic Theory

Lacan there can only be one, and women's relation to it is

negative. For Irigaray, women have a language of their own,
related to their sexuality and imagination. This language is not
merely repressed, like the signifiers of Kristeva's chora, but
actually suppressed, denied existence at any level. lrigaray
believes that this suppression, and therefore the question of
language, is fundamental to patriarchy,

The question of language is closely allied to that of feminine

sexuality. For I do not believe that language is universal, or
neutral with regard to the difference of the sexes. In the face of
language, constructed and maintained by men only, I raise the
question of the specificity of a feminine language; of a
language which would be adequate for the body, sex and the
imagination ... of the woman. A language which presents
itself as universal, and which is in fact produced by men only, is
this not what maintains the alienation and exploitation of
women in and by society?17

What Lacan conceptualises as a lack in woman, Irigaray regards

as difference. Women's language, if it were not suppressed,
would be different from men's in two major ways. The first
(which reminds one of the feminist-folklinguistic rejection of logic
and the complete sentence) is syntactic: 'it has nothing to do with
the syntax we have used for centuries, namely ... subject,
predicate or subject, verb, object. The female sexuality is not
unifiable. '18 The second also reflects the alleged plurality of
feminine sexuality, and relates to meaning:

There will always be a plurality in feminine language. And it

will not even be the Freudian 'pun', i.e. a superimposed
hierarchy of meaning, but the fact that at each moment there
are always for women at least two meanings, without one being
able to decide which meaning prevails, which is 'on top' or
'underneath', which conscious or 'repressed'. For a feminine
language would undo the unique meaning, the proper
meaning of words, of nouns, which still regulates discourse. 19

This rejection of semantic determinacy (about which I shall have

more to say in the next chapter, for it is of paramount
importance) is a challenge not only to Lacan, but more
 Feminist Models of Language (II) 129

important, to Saussure and the whole apparatus of structuralism.

Where words have more than one meaning, and no meaning is
more basic than any other, the one-to-one correspondence
between signifier and signified which guarantees the unity of the
sign is broken down. Irigaray is obviously aware what a drastic
step she is taking here, for she remarks that Lacan never went far
enough: 'There was I think one further step to take: to question
linguistic theory itself, viz. structuralism, and more generally, any
formalism. '20
Unlike Lacan, then, Irigaray specifically rejects linguistic
determinism, and holds that there need be no fixed order of
meanings for the subject to enter. At present, however, males do
have control over the language and its meanings. This is not
necessarily anything mysterious: men simply silence women, often
(as in lrigaray's own case) by quite literally taking away their
platform. 'Women are not allowed to speak, otherwise they
challenge the monopoly of discourse and of theory exerted by
men. '21 This silence is all too obvious, lrigaray claims, in the
differing behaviour of male and female schizophrenics. Whereas
men typically produce linguistic symptoms for the analyst to work
on, women schizophrenics find it more difficult to articulate their
illness, and suffer it therefore in the form of psychosomatic pain.

PROBLEMS: LANGUAGE AND THE BODY

In her concern for a 'women's language' lrigaray is very typical of

many feminists in France and, increasingly, in Britain and the
USA. Influenced by the Lacanian dictum that 'women are
excluded by the nature of things, which is the nature of words',
these women make a strong connection between language and
sexuality, language and the body, and they have urged that
writers take up the quest for a signifying system appropriate to
the feminine body. Such a language would be able to express
women's experience: whereas all current western discourse
presents, as Irigaray phrases it, 'a certain isomorphism with the
masculine sex'. The rallying-cry is 'write your body': that is,
women are urged to find a form of writing in which the specific
rhythms of the female body and the conflicting forces within the
female unconscious can come out.
130 Femz"nz'sm and Lz"ngu£Stz"c Theory

So it is worth considering whether the idea of a relation

between language and the body is a useful one. To some feminists
it appears, on the contrary, as dangerous biologism that plays
into the hands of the patriarchs. This argument is convincingly
presented in the first number of a French theoretical journal,
Quest£ons jem£n£Stes:

To advocate a 'woman's language' ... seems to us ... illusory .

. . . It is at times said that women's language is closer to the
body, to sexual pleasure, to direct sensations and so on, which
means that the body could express itself directly without special
mediation and that, moreover this closeness to the body and to
nature would be subversive. In our opinion, there is no such
thing as a direct relation to the body. To advocate a direct
relation to the body is therefore not subversive because it is
equivalent to denying the reality and the strength of social
mediations, the very same ones that oppress us in our bodies.
At most, one would advocate a different socialisation of the
body, but without searching for a true and eternal nature, for
this search takes us away from the most effective struggle
against the socio-historical contexts in which human beings are
and will always be trapped. 22

Or as Mary Jacobus pithily puts it, 'if anatomy is not destiny, still
less can it be language'. 23
Even leaving aside the problem of language and the body,
however, the psychoanalytic framework itself appears to raise
serious problems. Some of these are the same problems we have
seen in the work of Shirley Ardener and Dale Spender, but others
are more specific to the context of psychoanalysis. It is possible, of
course, to take an entirely sceptical view of Freud and all his
works. Many systems of talking about mental life do not give the
same place as Freud does to sexuality and even to the
unconscious: and as the materialist feminist Christine Delphy
remarks, why should anyone swallow whole the 'extravagant
claim of psychoanalysis to be, not a system of interpretation of
subjectivity, but subjectivity itself. I will not accept that objecting
to the theory of psychoanalysis is synonymous with disinterest in
its object. '24 In other words, Freudian theory is not to be taken as
a given: we are entitled to ask if it is correct. Such empiricism may
be foreign to the Lacanians, but it certainly was not foreign to
 Femz"nz"st Models of Language (II) 131

Freud, whose observations stood or fell by the quality of clinical

evidence he was able to produce in support of them.
If we want to take Lacanian and Freudian ideas critically on
board, however, two questions arise immediately. The first of
these concerns the place of the social in child development. For it
seems quite obvious that the construction of what in our culture
passes for masculinity or femininity is not simply a drama that
unfolds within the child's private mental life. It is also to some
extent a matter of what feminists have long referred to as
.conditioning' . Studies have repeatedly shown that girls and boys
are treated quite differently from birth, and that appropriate
behaviours are explicitly taught. This socialisation may have
nothing to do with sexuality per se (though I shall argue in a
moment that even this would be an over-simplification) but it is
extremely relevant to the social category of gender.
This is not to suggest that a crude social conditioning model
must be superior to a psychoanalytic account of child
development. On the contrary, the importance of infantile
sexuality and of mental life in general, can hardly be denied even
by theorists implacably opposed to Freud. The point is that
neither account will do on its own. We cannot leave socialisation
out. We need to know what part it plays in child development
and also how it relates to the psycho-sexual processes described by
psychoanalysis. For instance, are the Oedipus and castration
complexes human universals, or are they specific to some
particular kind of family organisation (e.g. the nuclear family)? If
women did not bear sole responsibility for childcare, would loss of
the maternal body retain the status Lacan accords it, or will the
fact that women bear children always dictate a close
psychological identification between mother and child?
The second problem relates to language. We need to know, for
instance, how Lacanian accounts of language acquisition can be
related to the large body of work on the subject by linguists and
psychologists. As matters stand, there is little evidence that
children's acquisition of language is significantly affected by their
sex. Nor does there seem to be much evidence for the alleged
marginality of women as speakers and writers, which is something
I shall return to. Indeed, it is difficult to see how, in the Lacanian
or Kristevan framework, one would evaluate the theory that
feminine subjects are marginal to the symbolic order: if
femininity is defined as marginality, and a writer's identification
132 Feminz'sm and Linguz'stic Theory

as feminine can only be inferred from her/his writing itself, the

theory seems endlessly circular and impossible to check. (It is only
fair to point out that Lacan and Kristeva would not accept this as
an objection to their theories, for theirs is an anti-empiricist
tradition: nevertheless it strikes me and many other linguists as a
problem.)
There is one final, explicitly political problem which can
usefully be mentioned here. As we have seen, psychoanalytic and
anti-humanist approaches downgrade the notion of 'experience'
which has previously been (and in the Anglo-American tradition,
remains) central to feminist politics. But in denying that
feminism can be the politics of experience, it could be argued
that the semiologists miss, or even deliberately conceal, an
important difference between women and men. This is precisely
the exp~i:ience women have of being sexually oppressed by men.
It would be wholly unreasonable to argue, I think, that awareness
of one's subjection (in the case of women) or power (in the case of
men) could make no difference to one's sexuality.
But if we accept that one's actual status as powerful/powerless
sexually is an important formative experience, what becomes of
the dogma that men too can be feminine subjects and feminists?
It has to be discarded at once, for power in society is as a matter
of established fact assigned not on the basis of identification
choices, but simply on the basis of biological sex. And there is
thus no choice about whether you are an oppressor or one of the
oppressed.
In the final analysis, then, it seems to me that radical feminists
must be intensely suspicious of theories that invalidate experience
and make the subordinate status of women a consequence of
something other than their mere womanhood. These theories all
too often turn out to be a cover for the role of men in upholding
male power, or else an apology for the participation of men in the
feminist movement. In other words, a wholly theoretical and non-
biological concept of gender, sexuality and so on makes it too easy
not to emphasise the real power differences between the sexes.
Diana Leonard makes the following acute observation:

It is always a joy to hear men ... asserting that which side you
are on is not a question of whether you exert or suffer
oppression, but of what values you hold. And it is classic to
turn the accusation of sexism and biologism back on the
 Feminist Models of Language (II) 133

victims, as if hostility to the oppressor is the same as hostility to

the oppressed, and as if it were radical and revolutionary
feminists who had invented the crazy idea that having or not
having a penis/phallus should be the source of significant
differences in the way in which society treats 'individuals'. 25

Psychoanalytic approaches to sex difference and women's

oppression have undoubtedly introduced stimulating ideas to the
feminist debate, but in evaluating them we must bear Diana
Leonard's warning in mind.
It will be clear by now that although the Ardeners, Spender,
Black and Coward, lrigaray and Kristeva diverge in important
ways from one another, they all subscribe to some degree of
determinism, to the idea that men control language and
(especially) to the notion that women are alienated from it to a
degree that men are not. They stress the basic in authenticity of
women's language at present, the difficulties women have in
talking about their experience, or sexuality. They trace back to
this linguistic disadvantage important elements of women's
subjection, and recommend either subversion of symbolic
language or the forging of an authentically female code.
In Chapter 8 I shall take issue with this view of women,
language and oppression. I shall be asking how, and to what
extent, the resources of language are used to construct and/or
maintain the power of men over women.
8 Beyond Alienation: An
Integrational Approach
to Women and Language
I have come to believe over and over again that
what is most important to me must be spoken, made
verbal and shared, even at the risk of having it
bruised and misunderstood. That the speaking
profits me, beyond any other effect.
Audre Lorde
The central indeterminacy of all communication is
indeterminacy of what is meant.
Roy Harris

This chapter articulates a view of language rather different from

any we have looked at so far. In it I am critical of those theorists
who claim that men control language and language controls us,
whether or not that claim is dressed up in the complexities of
psychoanalysis. On the one hand, theories of alienation and
in authenticity seem to me misguided because they locate us in a
linguistic utopia which never has existed, and never will exist: on
the other, I find them politically dubious because they are remote
from the lived experience of women, and indeed reject the
validity of much of that experience.
I do not deny that books such as Dale Spender's have had a very
positive response from women who felt they recognised the
processes of muting and misunderstanding discussed in Man
Made Language. What I do deny is the idea that, in order to
explain these things, we have to resort to notions of alienation,
male control, negative semantic space and negative entry into the
symbolic, female in authenticity (a new and pernicious sort of
female inadequacy) and the creaking apparatus of linguistic
determinism. These notions mystify language and demoralise

134
 Beyond Alienation 135

women. The only feminist theory they engender IS an arid,

intellectuals-only theory of no hope.
Marks and de Courtivron, in their introduction to a collection
of French writings, admit that semiology, though stimulating,
may turn out to be unproductive politically:

'Within the structuralist vision ... there is no liberation from

oppression, but there is an immense energy released by the
attempts to analyze and demystify the structures that
determine and oppress us. 'I

Is this really enough?

Some feminists believe it is not. In 1976, at a conference on
patriarchy, a paper was read which gave an account of Lacanian
psychoanalysis and its relevance to feminism. 2 Published
alongside that paper and others from the conference was a
critique written later by the Dalston Study Group. This critique,
titled 'Was the patriarchy conference "patriarchal"?', emphasised
the hostility and dissatisfaction produced at the conference by the
psychoanalysis paper, and pointed to the source of that
dissatisfaction:

'It felt ironical, then, to arrive at a women's conference and

feel defined negatively in relation to it; to listen to papers being
read about women's silence and women having no social
language, which itself made us passive and silent. '3

The Dalston Study Group objected not only to the content of

the paper but to the language it was written in. In their opinion,
it was exactly that sort of language women found most alienating:
the authors of the psychoanalysis paper had reproduced the
process by which academics are so often able to exclude less
privileged groups. As the Dalston women put it,

The language used ... had the effect of making large numbers
of women feel inadequate, stupid or angry ... the process we
identify in education as a process of socialisation which often
makes women, blacks, working-class people, etc., unconfident
and suspicious of intellectual work, and makes them doubt the
strength and potential of their own language. It also
136 Feminism and Linguistic Theory

perpetuates the split between the undervalued day-to-day

language of such groups ... and the impoverished depersonal-
ised analytical language of intellectuals. 4

In this chapter I want to explore the possibility that the

linguistic mechanisms used to oppress women are like the ones
being alluded to here, rather than the ones discussed in Chapter
7. The first puzzle that needs to be solved before we can look at
women's day-to-day language and the strategies used to devalue it
is this: why do the pontifications of intellectuals, and particularly
those intellectuals whose business it is to talk about language,
concern themselves so little with everyday speech and
communication? It is quite obvious, for instance, that linguistics
has nothing much to say about these things (the devaluing of
ordinary language is institutionalised in the split between langue
and parole, competence and performance) and from this
shortcoming we get many of the misconceptions on which theories
of alienation are built. So as a preliminary to my analysis of
women's oppression through language, I propose to examine the
linguist's conceptions of language, meaning and communication
and to criticise them on a number of grounds.

WHA T DO WE MEAN BY 'LANGUAGE'?

It has often been pointed out that the word language has at least
two meanings. It can refer to a general human faculty like
'cognition' or 'sense perception', in which case it is a fairly general
and abstract term; alternatively it can refer to a particular entity
like 'English' or Swahili'. In a great deal of lay and theoretical talk
about language, these two senses are not very rigorously kept
apart.
In addition to this possible confusion, there are two ways in
which scholars can look at languages. On the one hand, 'English'
may be regarded as an institution, on the other, as an object.
If language is an institution, the task of the scholar must be to
codify its conventions in things like the dictionary (for
conventions of pronunciation and meaning), the grammar book,
the style sheet, the thesaurus and so on. Some languages are more
thoroughly codified than others. The ministry of Dutch culture in
Belgium, for instance, regularly issues a definitive account of
 Beyond Alienation 137

what words and constructions are to count as proper usage for the
Flemish population. Civil servants and university students are
expected to comply with ministerial directives in their use of
language. Although English is not subject to a centralised
authority of this kind, nevertheless it has many of the trappings of
an institution: dictionaries, books of standard usage, etc., whose
prestige is very high.
Those who subscribe to the institutional view of language are
quite clear that 'English' is a cultural artefact which needs to be
regulated and protected from abuse (for instance, debasement by
foreigners and the lower orders). 'Language' is just a kind of
shorthand for 'correct usage' or 'the language of the most
educated/privileged speakers'. Communication is taken for
granted: it is assumed that people know how to communicate and
only need to be taught the most elegant way of doing it.
Feminist reformists like Miller and Swift, whose targets are the
'institutional' trappings of language (the journalistic style manual
and the editor's guidelines) appeal to these same notions of
clarity, elegance and accuracy. They differ from the sort of person
who writes to the newspapers about bad grammar/sloppy speech
only in their ideas of what is correct usage. For them accuracy
means including women.
Linguistics pretends to be above this kind of judgement, which
is selective, prescriptive and therefore unscientific. For the
linguist, language is not an institution but an object, to be
abstracted for observational purposed from the circumstances in
which it is used and the people who use it. Whether it is
Saussurean sign-system, where terms enter into relations with
each other and no relations with anything outside, or the
idealised grammar rules of transformational models, the linguist's
language has no users and no uses. The institutions which grow
up around it are ignored (though it would be quite feasible to
study them without descending to prescriptivism) and the
question of communication is sidestepped. Studying language as
an object rather than an institution or a process causes linguistics
to exclude much that is not only interesting but crucial.

WHAT DO WE MEAN BY 'MEANING'?

Even if communication were on the linguistics agenda, it is

138 Femz"nz"sm and Lz"nguz"stz"c Theory

doubtful that anything illuminating could be said about it within

the orthodox framework. To ask about communication, after all,
is to ask how hearers ascribe meaning to what speakers say: and
the linguist has a very simplistic notion of meaning.
In her autobiography, Simone de Beauvoir describes a private
world of prelinguistic experience, which words, when she comes
to use them, can render only imperfectly. 'White was only rarely
totally white, and the blackness of evil was relieved by lighter
touches; I saw greys and half-tones everywhere. Only as soon as I
tried to define their muted shades, I had to use words, and found
myself in a world of bony-structured concepts. '5 Having admitted
the necessity for language, however, de Beauvoir fell into the
opposite error of assuming there was no meaning outside the rigid
definitions given her by parents and relatives.

As I had failed in my efforts to think without recourse to

language ... I assumed that this was an exact equivalent of
reality; I was encouraged in this misconception by the grown-
ups, whom I took to be the sole depositaries of absolute truth:
when they defined a thing, they expressed its substance, in the
sense in which one expresses the juice from a fruit. So that I
could conceive of no gap into which error might fall between
the word and its object; that is why I submitted myself
uncritically to the Word, without examining its meaning, even
when circumstances inclined me to doubt its truth. 6

Both these ideas are presented as the unsophisticated theories of a

child, and both make the same mistake: they assume that words
express private experience perfectly, and that there can be no
argument about what they mean. Adults, on the other hand,
rarely believe in practice in absolute truth and exhaustive
definition. They learn that words may easily lie and distort, that
they are vague, ambiguous and sometimes inadequate. When
they fail, of course, there are always other words with which we
may attempt clarification: even this will never be a complete
success. The representation by words of experience is partial in all
senses of that term. This is a truism of everyday communication
that linguistics has never come to terms with.
The model of communication which has dominated western
philosophy and linguistics is, as Roy Harris points out, essentially
telementatz"onal. Language is envisaged as a means by which a
 Beyond Alienation 139

speaker can transfer a thought from her own mind to her

hearer's.7 This is possible because the speaker and hearer share a
code, a set of correspondences between forms and meanings.
Given something she wishes to convey, a speaker is able to select
the form associated with her chosen concept; the resulting
utterance will be decoded by the hearer in the same way, by
matching it up to a concept - the same concept - in her own
mind.
But this can only work if there really is a unique one-to-one
correspondence between forms (words, say) and meanings, a cor-
respondence reproduced in the mind of everyone who speaks the
same language. If that correspondence were absent, we could
never be sure we had really understood what anyone said, since
the incoming acoustic signal would not necessarily evoke the same
concept which had spawned it in the speaker's discourse
originally. Linguistics assumes that there is such a one-to-one
correspondence, and that the unique pairings of forms and
meanings in each language are precisely what the linguist has to
work out. Hence the Saussurean sign, in which form (signifier)
and meaning (signified) are so fused together that dividing them
is impossible. Orthodox linguistic models hold

that the ideal community will have a language in which all the
basic units must be determinate and all the rules which govern
their combinations and interpretations must be determinate.
For otherwise there is no possibility of a common codebook for
the whole community ... the expressions of a language in some
sense have to be determinate in respect both of 'form' and of
'meaning'. That is to say, there must be fixed rules ... for
assigning the correct interpretation to anyexpression. s

It is this assumption that Harris labels 'the language myth'. It

leads to the uncritical submission to the word recalled by Simone
de Beauvoir: it leads to linguistic determinism and the tyranny of
the sign. Fortunately for human communication, this model of
meaning is false.
Harris suggests replacing the language myth with a 'demyth-
ologised' or 'integrational' linguistics which would embody quite a
different way of looking at language and a radically different
theory of meaning. Integrational linguistics would acknowledge
two crucial things about language. The first of these is that
140 Feminism and Linguistic Theory

language cannot be sealed off from every other form of social

behaviour, nor abstracted from the dimensions of time and space
to which all such behaviour is inevitably subject. As Harris says,
'Human beings inhabit a communicational space which is not
neatly compartmentalised into language and non-Ianguage.,g
Locating speech events in time and space (which means
abolishing the Saussurean dichotomies of synchrony / diachrony
and langue / parole) is extremely important to an integrational
analysis. By paying attention to the whole context in which
speech occurs, the analyst would be spared the necessity of
postulating invariant correspondences of form and meaning: she
could allow what we all know, that words are used and
understood differently by different speakers at different times
and in different situations, and she could refer to the specifics of
context in order to explain that variation.
The second thing is that language-using is a creative process.
There is pretty well no limit to the novel situations humans may
encounter, and therefore the communicational demands which
may be made on them are almost limitless. To meet those
demands, speakers and hearers engage in a constant renewal of
language. They employ words in a flexible, innovative and often
playful way, and are creative in their interpretations of other
speakers. In a linguistics which takes account of this, Harris
observes, 'It would no longer be necessary to reduce speaker and
hearer to mere automata, handling pre-packaged messages in
accordance with mechanical rules. '10 The orthodox linguistic view
of meaning sees language users looking up meanings in some
internalised dictionary. The integrational view, on the contrary,
sees them creating meaning in specific contexts, negotiating
where necessary in order to achieve as fully as possible their
communicational aims.
In practice, indeterminacy of meaning is acknowledged by
structuralist and post-structuralist writers in their notion of
'deconstruction', a sort of reading that challenges established
forms of criticism by denying that the 'true meaning' of any text
can ever be arrived at. Deconstruction is a process of making text
yield a multiplicity of meanings; none is counted more basic than
any other, and there is no end to the process. Such open-
ended ness is exactly what the integrationalist approach to
language proclaims. But for an integrationalist it is quite baffling
to find practitioners of deconstruction claiming that creativity is
 Beyond Alienation 141

made possible by the existence of a decontextualised and fixed set

of signs, langue. This ossified structure, with determinacy of form
and meaning at its core, is exactly what integrationalists would
want to dispense with on the grounds that it denies all creativity
and thus renders the sort of communication typical of human
activity utterly incomprehensible.
This dispute is not trivial for feminists. Rejecting the notion of
langue and the determinate sign means rejecting the feminist
positions set out in Chapter 6. There can be no linguistic
determinism, no control of meaning by men, no privilege of the
phallus as signifier and thus no alienation.

MEANING, UNDERSTANDING AND ALIENATION

We may now return to the experience described by Simone de

Beauvoir: 'I saw greys and half-tones everywhere. Only as soon as
I tried to defin.e their muted shades, I had to use words, and I
found myself in a world of bony-structured concepts. '11 This
frustration at being unable to make language express the exact
nuances of experience is something we have already encountered
in women's own testimony: 'Sometimes when I am talking to
people I really feel at a loss for words .... A vast number of the
words I use all the time to describe my experience are not really
describing it at all.'12 Audre Lorde, a poet, notes that we speak of
our experience only 'at the risk of having it bruised and
misunderstood' .IS Difficulty in finding words and difficulty in
being understood are often spoken of by women as signs of their
alienation, the proof that 'this is the oppressors language' with
meanings and limits defined by men. The solution is to create a
new language in which women can express their meanings and be
understood.
But if we accept the idea that meaning is complex, plural and
ultimately perhaps impossible to pin down, the new language
solution appears utopian. There will never be a perfect fit
between private experience and linguistic expression, and there
will never be perfect mutual understanding. Sociologists of
language have long been familiar with the idea that participants
in any kind of talk take for granted a degree of comprehension
which, when you look more closely at the interaction, cannot ever
142 Feminism and Linguistic Theory

be quite justified. Thus the problems of expressing oneself and

being understood are not exclusively women's problems. They are
built into all interaction and affect all speakers. Which is not to
say that women do not suffer to a greater degree than men: for
the causes of their particular problems, however, I think it is
necessary to consider the whole social situation of women and
men, and not just their relative positions in an abstract symbolic
order.

The notion that perfect mutual understanding - telepathy - is

not the normal or the ideal outcome of speaking, frightens and
confuses many people. It is clear that without the indeterminacy
that stops us communicating telepathically we would not be able
to adapt our language to the novel situations we need it for;
imperfect communication is the price we pay for a creative and
flexible symbolic system. But this important insight frequently
meets with a great deal of resistance. If we cannot ever really
understand each other, are we not trapped in our own private
worlds with no hope of making contact? And is this not the
ultimate nightmare of alienation?
This fear, and the comforting certainty that perfect
understanding z's possible, goes very deep. For instance, a well-
known myth of human prehistory refers to a time when humans
did understand each other perfectly, and this understanding
conferred enormous power on them. God not unnaturally saw
that power as a threat and destroyed it by destroying the unity of
human language.

And the Lord said, Behold, the people is one, and they have all
one language; and this they begin to do: and now nothing will
be restrained from them, which they have imagined to do. Go
to, let us .go down, and there confound their language, that
they may not understand each other's speech. So the Lord
scattered them abroad from thence upon the face of all the
earth. 14

The perfect understanding which supposedly results from

'speaking the same language' is seen here as an essential
prerequisite for any kind of collective action. God, in replacing
linguistic unity with linguistic diversity, undermined the power of
the builders. Feminists have their own version of the Tower of
 Beyond Alienation 143

Babel story. They feel that men have undermined women by

confounding their language, the language of their bodies, their
unconscious, their desire or their experience. In order to act
together, an authentic language of women must be forged. If
there is no common language, there can be no true collective
action.
I do not feel, however, that the view of meaning I have put
forward excludes the possibility of collective action, nor does it
negate the communication that obviously does occur between
individuals. Rather, it says that if we are ever to understand the
nature of collective action and interpersonal communication, we
must first acknowledge its inherent difficulties and limitations.
Until we do acknowledge that communication is to some extent
an everyday triumph, until we get rid of our fantasies of what it
never can be, we can hardly study it at all, but will be content
either to avoid the issue or to take it for granted.
Where does this leave the feminist theories of language and
oppression we have discussed in this book? I suggest that it leaves
us with three propositions corresponding to the three feminist
axioms of linguistic determinism, male control and female
alienation.

1. Linguistic determinism is a myth. Where there is no

determinacy, there can be no determinism. In a system where
language and linguistic acts are integrated with non-linguistic
acts and social life generally, language can be only one of the
multiple determinants of any individual's perceptions and
experience. An important determinant it may be, but it cannot
be privileged to the extent that both Saussurean and Whorfian
theories privilege it.
2. Male control over meaning is an impossibility. No group has
it in their power to fix what expressions of a language will mean,
because meanings cannot be fixed, and interpretation will be
dependent not on the authority of some vast internal dictionary,
but on the creative and ultimately idiosyncratic use of past
experience and present context.
Learning to communicate and to participate in social life is
something which both male and female children do. They do it
by actively interacting with their environment and the people in
it, and thus they construct - rather than learn - meanings that
are highly contextualised, dependent on that environment and
144 Feminism and Linguistic Theory

those people, subject (as the environment is) to variation and to

change.
It would be very surprising if this learning process did not
exhibit sex-linked differences. Girls and boys, after all, are very
specifically socialised into female and male gender roles; one
would expect them to construct meanings which were different
not only idiosyncratically, because each individual has a different
experience, but more generally, because in patriarchal societies
males and females are allow!!d a different range of experiences.
Perhaps, then, we may talk to some extent of male and female
meanings. But we cannot speak of women being socialised into
male meanings, or of both sexes being socialised into patriarchal
meanings (except in the sense that their experience is one of living
under patriarchy). Meanings have to be constructed by the
individual language user (in this way language is radically unlike,
say, folk tales or table manners) and any child who does not learn
to construct meanings out of her own interaction with the world
cannot be said to possess language at all.
3. Female alienation from language does not exist in the form
postulated by the theories we have considered (it should not be
denied either in theory or in practice that many women do feel
extremely alienated in some modes of language use). Since
language is a flexible and renewable resource, and since girls
must come to grips with it as their socialisation proceeds, there is
no reason in principle why language cannot express the
experience of women to the same extent that it expresses the
experience of men.
In saying this, however, I do not wish to deny that women have
real problems in speaking and being heard. Although I reject the
usual explanations of them, I believe that the means do exist for
men to oppress, silence and marginalise women through
language. The sources of silence and oppression are what I want
to turn to now.

LINGUISTIC OPPRESSION: WHAT IT IS THAT MEN

CONTROL

One consequence of the integrational approach to language is

that the linguist has to take seriously the fact that languages are
 Beyond Alienation 145

not used in a social and political vacuum, i.e. she must recognise
the institutional aspects of language I have already mentioned in
this chapter.
In every society, one finds laws, rituals and institutions which
regulate language (especially its more public modes) in particular
ways. As we saw in Chapter 3, it is not always easy to separate
these 'metalinguistic' or 'discursive' practices from language
proper, since there is a constant interaction between the two.
From an integrational standpoint it is not even worth trying to
exclude the metalinguistic: institutional phenomena are part of
what the linguist must be concerned with.
If we look closely at the regulatory mechanisms which grow up
around languages, it is clear that they are rather closely
connected with the power structures of their society. The
institutions that regulate language use in our own society, and
indeed those of most societies, are deliberately oppressive to
women. Men control them, not in the rather mystical sense that
they are said to control meaning, by making esoteric semantic
rules or possessing the vital signifier, but simply because it is the
prerogative of those with economic and political power to set up
and regulate important social institutions.
Language, the human faculty and communication channel,
may belong to everyone; because of the crucial part it plays in
human cognition and development, it cannot be appropriated.
But the language, the institution, the apparatus of ritual, value
judgement and so on, does not belong to everyone equally. It can
be controlled by a small elite. As Trevor Pateman remarks,

Language, though the socially produced means of thought, is

not socially controlled. Increasingly control over the develop-
ment of language and its use is held by state institutions,
including mass-media and monopolistic private enterprise, as
in journalism and advertising.... The semiologists have
sometimes failed to appreciate the possibility and existence of
class or other minority control over language. 15

If we acknowledge the importance of institutional control, the

crucial question is: how is male control over metalinguistic
processes manifested, and what effects does it have on women? I
want to consider this now, and true to my integrational aims I
shall not be limiting my remarks to linguistic phenomena alone.
146 Feminzsm and Linguistic Theory

It is impossible to understand the practices that regulate women's

relation to language except with reference to gender roles and
regulatory mechanisms is general.

'HIGH LANGUAGE' AND WOMEN'S SILENCE

Cora Kaplan, in a short but influential essay, makes a point that

has since become the received view about women's oppression by
linguistic institutions; the point is that women are denied access
to the most influential and prestigious registers of language in a
particular culture. 16 That is to say, everything defined as 'high'
language (for instance, political and literary registers, the register
of public speaking and especially ritual - religious, legal or social)
is also defined as male language. Kaplan observes,

The prejudice seems persistent and irrational unless we

acknowledge that control of high language is a crucial part of
the power of dominant groups, and understand the refusal of
access to public language is one of the major forms of the
oppression of women within a social class as well as in trans-
class situationsY

If it is in their relation to high language that women are linguis-

tically disadvantaged, it seems that we must ask three questions:
what are the registers that men control, how do they gain and
keep control of those registers, and why does male control
constitute a disadvantage for women? I propose to explore these
questions by focusing on particular areas of language use and
linguistic control. First there is the area of written language and
women's relation to it. An investigation of this shows how a denial
of language can constitute a denial of knowledge and of certain
kinds of consciousness. Then, there is the problem of bureaucra-
tic/institutionallanguage. Recent work on interethnic communi-
cation demonstrates how very tightly controlled norms are used to
define subordinate groups as inadequate communicators (and
thus to make them inadequate). Finally, we must examine the
exclusion of women from public and ritual speech, investigating
the extent to which femininity has been produced as incompat-
ible with the sphere of rhetoric.
 Beyond Alienation 147

WOMEN AND LITERACY

Literacy - the ability to read and write - may seem to us a

'natural' concomitant of all language, but in fact it is a relatively
recent technology (6000 years old, whereas human culture is at
least 30,000 years old) and like most technologies, male-
dominated throughout its history. The definitive account of
women's relation to literacy has yet to be written, but research
currently available gives feminists food for thought.
Of the approximately 800 million illiterates in the world today,
the majority are women: indeed the higher a country's over-all
rate of illiteracy, the wider the gap between women and men. IS
Literacy has been a mostly-male phenomenon in the developed
world historically even when it meant being able to write your
own mother tongue; in the centuries when it meant knowledge of
a superordinate 'learned' language such as Latin, classical Arabic
or Sanskrit, literacy was effectively a sex-exclusive marker.19
Why are women so frequently illiterate? The short answer is, of
course, that where the education required to produce literacy is
not available to everyone automatically (and indeed compul-
sorily), women do not get it. Poor countries and poor families
cannot afford to educate children whose destiny is marriage and
domestic labour rather than the bureaucratic, commercial,
scholarly and scientific work literacy facilitates. However, there
must be more to it than this, since although the same argument
applies to working-class and peasant men (whose work does not
demand literacy any more than women's does) they are more
likely than their wives and sisters to be able to read. The exclusion
of women from education and thus from literacy is not simply a
side-effect of economic circumstances, but a way of keeping the
female in her place - dependent and domesticated, both
physically and mentally. We can see this more clearly if we
examine what one expert has to say about the difference literacy
makes to a person's attitudes.

the more literate people are, the more willing they are to accept
and work for improvements in their societies. Their sense of
'personal efficacy' is increased: that is, they gain confidence
that they are able to do something about their own lives ....
They become more willing to reason for themselves, less willing
to take opinions on authority.20
148 Feminism and Linguistic Theory

It is evident that the maintenance of compliant femininity is

threatened by these qualities. Literacy has always been associated
with democracy and with modernisation in less developed
countries, and just as the ruling class fear these things if a literate
proletariat emerges, men defend their own power by denying
women this crucial democratising technology.
In western European culture, women have been literate in their
own language for many years. What they missed through being
denied access to education was literacy in Latin, the language of
learning and culture in western Europe. This effectively insulated
women from a great deal of knowledge, and possibly also from a
particular, and particularly influential, way of thinking. Ong
remarks,

Writing ... serves to separate and distance the knower and the
known and thus to establish objectivity. . .. Learned Latin
effects even greater objectivity by establishing knowledge in a
medium insulated from the emotion charged depths of one's
mother-tongue, thus reducing interference from the human
lifeworld and making possible the exquisitely abstract world of
medieval scholasticism and the new mathematical modern
science. 21

This does not mean, of course, that science and mathematics are
inherently male discourses founded upon a male language. Girls
have proved their abilities in Latin and in science. What is
important is that these discourses were h£storically masculine and
specifically denied to women. Ultimately it is that history which
determines women's current relation to science, even if rather
indirectly and obscurely. Saussure was wrong in supposing that
the speaker/writer has no awareness of history at all: most of us
are aware of linguistic tradition, custom and practice, and may
well be insecure or tentative users of discourses we (or our
ancestors) had to fight to get into. This has nothing to do with the
language itself and everything to do with the way we were defined
in relation to it.
To sum up, then, denying women access to particular registers
of language (written and learned language) by denying them the
necessary education and persuading them they do not need/merit
it affects both their state of knowledge and their state of
consciousness. Even where male-controlled registers later become
 Beyond A lz"enation 149

available to women, women may be negatively defined in relation

to them, especially if the registers in question are prestigious, and
negative attitudes may persist for a long time. (Thus no one now
thinks it remarkable that Anglophone women can read and write,
but it is still remarkable that they should write poetry or
mathematical treatises.) The power of negative attitudes is
something we must return to in our discussion of institutional and
bureaucratic language.

INSTITUTIONAL LANGUAGE: COMMUNICATING IN

URBAN SOCIETIES

If literacy is a problem for the underprivileged in less developed

countries, bureaucracy is a major linguistic headache for the
underprivileged of modern western cities. In two recent books
advocating a new approach to communication ('interactional
sociolinguistics')22 the linguist John Gumperz points out how
important 'communicative skills' have become with the growth of
state and other bureaucracies (health, education, employment
and tax services, for instance) in modern industrial society. More
and more frequently, individuals in their everyday lives are
having to negotiate linguistic interactions with these bureau-
cracies, and the result is that

The ability to manage or adapt to diverse communicative

situations has become essential and the ability to interact with
people with whom one has no personal acquaintance is crucial
to acquiring even a small measure of personal and social
control. We have to talk in order to establish our rights and
entitlements. . .. Communicational resources thus form an
integral part of an individual's symbolic and social capital. 23

The individual needs to be able to interact effectively with

institutions and their representatives. Since it is the individual,
for the most part, who wants something from the encounter,
'effectively' will mean 'in conformity with the norms of the
institution'. Those who cannot express themselves in a way the
bureaucracy finds acceptable (or minimally, comprehensible) will
be disadvantaged.
150 Femz'nism and Lz"nguz"stz"c Theory

Gumperz and his associates have produced a good deal of work

on 'crosstalk' or, in plain language, misunderstandz"ng between
individuals whose norms of interaction are different. Their work
has focused mainly on interactions between bureaucrats of
various kinds (social workers, clerks, personnel managers) and
Asian speakers of English, but the two main points that emerge
from it are equally applicable to other ethnic minorities, working-
class speakers in certain situations and, of course, women talking
to men.
The first point is simply that socially distant individuals
(especially those differentiated by ethnicity) do not share rather
subtle strategies for structuring and interpreting talk, and this
results in misunderstanding which can be frustrating for both
parties and seriously disadvantageous for the Asian trying to get a
job or a Social Security cheque. The second point, which is rather
less explicit in Gumperz's books, is that bureaucracies use their
experience of interethnic misunderstanding to generate represen-
tations of Asians as defective or inadequate communicators -
representations which derive from racist stereotypes and reinforce
racism. It is important to note that the right to represent and
stereotype is not mutual, and that the power asymmetry here has
serious consequences. Undoubtedly the Asians have their own less
than complimentary ideas about the gore (white people), but
these are the ideas of people without power. They do not serve as
a base for administrative procedures and decisions, nor do they
get expressed routinely in mass media: whereas institutional
stereotypes of Asians do inform procedures, decisions and media
representations. If Asians are defined as inadequate language
users, they become de facto inadequate (crudely, no one listens
any more to what they are actually saying, simply filtering it
through the negative stereotype) and as Gumperz points out, in a
modern industrial society this can have disastrous consequences.
It is important for feminists to ask whether women have the
same ability as men to interact with people 'with whom one has
no personal acquaintance' and 'adapt to diverse communi-
cational situations'. If not, why not? Are women routinely mis-
understood by men who have power over them, with dis-
advantageous results? Do men represent women as inadequate
communicators, thus reducing their precious 'symbolic and social
capital'? The question of misunderstanding between women and
men has been addressed by two of Gumperz's associates, Daniel
 Beyond A Henation 151

Maltz and Ruth Borker, and they believe that the subcultural dif-
ferences between Asian and white speakers are paralleled by
male/female differences. 24 Using the available literature on
children's talk and play patterns they argue that females and
males in western culture do in fact form separate subcultures,
and that this significantly affects male/female interaction.
Women and men do not attempt to do the same things in the
same way when they talk, and thus there is likely to be a rather
poor fit between what the speaker intends and what an opposite-
sex hearer picks up.
It seems likely that there is a good deal in this idea, though
Maltz and Borker do depend heavily on the competition versus
co-operation stereotype which I have already criticised (Chapter
3) and which is itself inspired by a lot of rather dubious literature.
My main reservation is that the sociolinguistic analysis of sub-
cultural differences ought to include far more discussion about
the political structure superimposed on these differences. In other
words, Maltz and Borker say very little about the power of male
definitions of female speakers, the use of such definitions to
exclude women from certain registers and devalue their contri-
bution to others. This omission is what I shall try to make good in
the remainder of the chapter.

DISCOURSES AND REGISTERS

I have already discussed the exclusion of women from written and

learned 'registers' (i.e. kinds of language appropriate in content,
style and tone to a particular ___domain of use, say 'scholarship' or
'legal documents' or 'religion') and pointed out that whereas it
makes no sense at all to speak of women not possessing 'language'
it is quite in order to say that, for historically specific reasons,
they may be forbidden to use certain registers at particular times
and in particular places. Registers of language historically
created by men very often represent women as marginal or
inferior, and may well continue to do so even after women have
begun to use them (in this book we have already looked at the
registers of news reporting and lexicography, and while these
practices were undoubtedly masculine originally, they have long
been open to women without any noticeable diminution in their
152 Feminism and Linguistz'c Theory

sexism). Partly this conservatism reflects the importance of

tradition, 'custom and practice' in institutions. The conventions
codified in style-books, rule-books, standing orders, editing and
sub-editing manuals are quite literally handed down from
generation to generation of professionals. They are part of a pro-
fessional mystique, sanctified by history and enforced very often
by the authoritarian training and advancement procedures of
hierarchical organisations (the Civil Service and political parties,
for instance). Partly, however, it reflects more general ideological
matters. This point has been made very forcefully by semiologists
like Roland Barthes (whose work on the ideological determinants
of literary style in France remains a classic demonstration),
Michel Foucault, Michel Pecheux, Colin McCabe, Maria Black
and Rosalind Coward.
Semiologists refer to what I have been calling 'registers' as
dzscourses (McCabe representatively defines a discourse as a set of
statements formulated on 'particular institutional sites of
language use').25 Each discourse needs to be understood in
relation to its own conventions (thus TV chat shows and court
cases exemplify different linguistic norms) and its functions in
society.
Maria Black and Rosalind Coward, in the article on Dale
Spender I have already discussed, explicitly say that discourse and
not language (by which they mean langue) is the proper place for
feminists to concentrate their efforts.

Linguistic systems ... serve as the basis for the production and
interpretation of related utterances - discourses - which effect
and sustain the different categorizations and positions of
women and men. It is on these discourses and not on language
in general or linguistic systems, that feminist analyses have to
focus. 26

By concentrating on discursive regularities (for instance, the use

of generic masculine pronouns or the linguistic representation of
women in the reporting of rape trials) we will discover more about
the relation between language use and patriarchal ideology to
which not only men but a great many women also subscribe.
What is stressed both in the semiologists' approach and in my
own register- based approach is the materzalz"ty of the practices in
question. Thus rather than posit, as Dale Spender does, a
 Beyond Alienation 153

historically ubiquitous and unobservable operation by which

males regulate meaning through an underlying semantic rule
pejorating words for women, linguistic materialists look for the
historical moment and circumstances in which a particular
practice arose and the specific group who initiated it or whose
authority and interests maintain it. We rarely find that a practice
is initiated/maintained by all men (an exception might be the
practice of pornoglossic intimidation discussed in Chapter 4) or
that it extends into every linguistic register. The negative relation
of women not to 'language' or 'meaning' but to various discourses
is a variable and piecemeal affair.
Nevertheless, it seems to me that ultimately this piecemeal
linguistic disadvantage must be related to the general roles and
representations of women in their various cultures. Not every area
of language use is regulated by obvious and consciously invoked
conventions, and we must now turn to the part played be folklin-
guistic value judgements and gender-role expectations in
silencing women and representing their speech as inadequate.
We must consider in particular the Dalston Study Group's
assertion that women's day-to-day language is undervalued even
by women themselves, and that the disadvantaged 'doubt the
strength and potential of their own language'. Is women's
language in fact strong and full of potential, or is it
repressed/suppressed, impoverished and inauthentic? How could
women be persuaded to suppress or undervalue their own speech?
To answer these questions we will need to look at the creation and
regulation of femininity itself.

SILENCE: A WOMAN'S GLORY?

Some very obviously male-dominated metalinguistic practices are

the customs and traditions of public speaking, which normally
require women to be silent in public gatherings and on formal
occasions. The key to this cross-culturally widespread phenom-
enon is, as Jenkins and Kramarae observe, the boundary between
the private or familial, and the public or rhetorical. 'We find that
women's sphere includes the interpersonal but seldom the
rhetorical. '27 In many societies different linguistic registers,
dialects or even languages are used to mark the private/rhetorical
154 Femz"nz"sm and Lz"nguz"stz"c Theory

boundary. And it is part of women's role generally, not just

linguistically, to symbolise the private as opposed to the public.
When this split is important in organising a society (as in most
capitalist societies) women are important in defining the
boundaries of the private.
An illustration of female marginality in the public and ritual
speech of our own culture is provided by the etiquette of the
wedding reception. Here we have a number of visible roles distri-
buted between males and females equally: bride and groom,
mothers and fathers, bridesmaids and best man, etc. Yet the
women are ritually silent. The bride's father proposes a toast to
the happy couple; the groom replies, proposing a toast to the
bridesmaids which is replied to by the best man. Men speak,
women are spoken for: here we have an epitome of women's
position as 'seen and not heard'.
Is this the same as children being 'seen and not heard?' Cora
Kaplan, in her essay 'Language and Gender', argues that it is. 28
Children are subject to restrictions on their speech in adult
company, but whereas boys are eventually admitted to public
speaking rights (Kaplan fixes this at puberty, the onset of adult-
hood and symbolically the beginning of manhood) girls are never
allowed to grow up in the same way. Their participation in
political, literary, formal, ritual and public discourse is not
tolerated in the same way that children's participation is not. This
view seems to me to be open to a number of objections, the main
one being that restrictions on women speaking are often far
stricter than those affecting children, and appear to be linked
with explicitly sexual rites of passage. I am thinking here of the
many taboos on women's speech discussed by Ardener and
Smith.29 It is not uncommon to find women being forbidden to
speak for a set period after marriage, or to find men censuring the
conduct of married women who allow their voices to be heard
outside the private house. 30 Books of etiquette and advice to
brides also warn that a married women must underline her wifely
deference with wifely silence. In other words, silence is part of
femininity, rather than being an absence of male privilege.

UNACCUSTOMED AS I AM ...

Women as public speakers suffer not only from the customs that
 Beyond Alienation 155

silence them, but also from negative value-judgements on their

ability to speak effectively at all. Whatever style a culture deems
appropriate to the public arena, women are said to be less skilled
at using; whatever style is considered natural in women is deemed
unsuitable for rhetorical use. So, for example, Jespersen thinks
indirectness typical of women's style, mentioning 'their instinctive
shrinking from coarse and gross expressions and their preference
for refined and . . . veiled and indirect expressions'. 31 This lack of
'vigour and vividness' is what makes women unfit to be great
orators. Among the Malagasy, however, things are rather
different. Here the favoured style for ritual speech or Kabary is
indirect and allusive. Women's speech is thought to be direct and
vigorous, and thus women are once again debarred from public
speaking. 32
It is necessary, as always, to treat the interaction between
actual usage and folklinguistic stereotype with the utmost care.
The excessively ladylike style 'described' by Jespersen is unlikely
ever to have been used consistently by women: it is the usual
idealisation based on the usual mixture of prejudice and wishful
thinking. But folklinguistic beliefs are never without significance,
and certainly this kind of belief, expressed in a score of passages
masquerading as description in anti-feminist tracts, etiquette
books, grammars and even feminist writings, have an effect on
how women think they speak and how they think they ought to
speak. In formal situations where speech is monitored closely,
women may indeed converge toward the norms of the mythology,
obeying the traditional feminine commandments (silence, not
interrupting, not swearing and not telling jokes).

VALUE

Folklinguistics inculcates an important set of value judgements on

the speech and writing of the two sexes. A whole vocabulary exists
denigrating the talk of women who do not conform to male ideas
of femininity: nag, bitch, strident. More terms trivialise inter-
action between women: girls' talk, gossip, chitchat, mothers'
meeting. This double standard of judgement is by no means
peculiar to linguistic matters. It follows the general rule that 'if in
anti-feminist discourse women are often inferior to men, nothing
in this same discourse is more ridiculous than a woman who
156 Feminism and Linguistic Theory

imitates a male activity and is therefore no longer a woman'. 33

This can apply not only to speaking or writing, but also to the way
a woman looks, the job she does, the way she behaves sexually,
the leisure pursuits she engages in, the intellectual activities she
prefers and so on ad infinitum. Sex differentiation must be rigidly
upheld by whatever means are available, for men can be men
only if women are unambiguously women.
This imperative leads to an attitude toward the upbringing of
women summed up in 1762 by Jean-Jacques Rousseau:

In order for [women] to have what they need ... we must give
it to them, we must want to give it to them, we must consider
them deserving of it. They are dependent on our feelings, on
the price we put on their merits, on the value we set on their
attractions and on their virtues.... Thus women's entire
education should be planned in relation to men. To please
men, to be useful to them, to win their love and respect, to
raise them as children, care for them as adults, counsel and
console them, make their lives sweet and pleasant: these are
women's duties in all ages and these are what they should be
taught from childhood on. 34

In this notorious passage Rousseau gives us an account of why this

sort of femininity must be constructed (to make men's lives 'sweet
and pleasant') , how it is constructed (by indoctrination from
childhood) and why women conform (because they are entirely
dependent on men for the things they need).
Language, like every other aspect of female behaviour, has to
be produced and regulated with this male-defined femininity in
mind. Parental strictures, classroom practices and so on are
designed to make the girl aware of her responsibility, and failure
to conform may be punished with ridicule, loss of affection,
economic and physical hardship. In short, then, we must treat
the restrictions on women's language as part of a more general
restricted feminine role. We cannot understand women's relation
to language or to any other cultural phenomenon, unless we
examine how in different societies those with power have tailored
customs and institutions so they fit Rousseau's analysis and obey
his prescription.
 Beyond Alienation 157

THE LIMITS OF CONTROL

This model of male dominance locates the linguistic mechanisms

of control both in explicit rules and well-known customs
restricting women's speech, and in the 'voluntary' constraints
women place on themselves to be feminine, mindful of the real
disadvantages attendant on failure. Since these mechanisms are
not located in immutable mental or unconscious structures,
control can only be partial, and even women's silence has its
limits.
The more radical feminist theorists have often been unduly
pessimistic because they did not acknowledge the limits of control
and silence. When Dale Spender claims,

One simple ... means of curtailing the dangerous talk of

women is to restrict their opportunities for talk .... Tradition-
ally, for women there have been no comparable locations to the
pub which can encourage woman talk; there have been no
opportunities for talk like those provided by football or the
union meeting. Because women have been without the space
and the place to talk they have been deprived of access to
discourse with each other.35

she is simply wrong. If Spender is thinking here of the 'captive

wife' alone with small children in an isolated flat, this is a
relatively recent and restricted phenomenon. Even in middle-
class Anglophone culture women's talk with each other is an
important part of social organisation: 36 in other cultures, where
segregation is often the norm both occupationally and socially,
women's lives revolve around interaction with each other.
Spender is trying to make a case for the subversive nature of
women's talk with each other, but by ignoring the age-old oral
culture of women (as she must, to argue this case convincingly)
she misleads us into accepting what is only a half-truth. Women's
talk is not subversive per se: it becomes subversive when women
begin to attach importance to it and to privilege it over their
interactions with men (as in the case of consciousness-raising).
Men trivialise the talk of women not because they are afraid of
any such talk, but in order to make women themselves downgrade
158 Feminism and Linguistic Theory

it. If women feel that all interaction with other women is a poor
substitute for mixed interaction, and trivial compared with the
profundities of men's talk, their conversations will indeed be
harmless.

WOMEN'S TALK: THE MYTH OF IMPOVERISHMENT

Recently, feminists have begun to research women's talk. The

picture that emerges from their studies is not one of silent or
inarticulate women who struggle to express their experiences and
feelings. On the contrary, it is of a rich verbal culture.'?
Moreover, that culture has a long history (if obscure: male
metalinguistic practices strike again, this time by omission). It
may be appropriate to see early women poets breaking through
silence and absence, working in a genre where they were insecure
and had no rights, but the ordinary woman speaker in her peer
group cannot be adequately treated in this way.
To sociolinguists this story will have a familiar ring to it. One of
the most celebrated achievements of sociolinguistics in the 1960s
and early 1970s was to put working-class black American speech
on the map through painstaking study of the vernacular black
speakers used amongst themselves. Before the sociolinguist Labov
and his associates undertook this research, using a methodology
specifically designed to win the informants' co-operation (see
Chapter 3), conventional wisdom among commentators on black
language was that its speakers developed silent and inarticulate
because they grew up in a linguistically deprived culture, were
seldom addressed by their parents and not encouraged to speak
themselves. The dialect they came to school speaking was labelled
'a basically non-logical form of expressive behaviour' or, in the
terms of Bernstein a restricted code. 38
Bernstein's code theory (which was developed with the English
class structure, rather than American ethnic differences, in
mind) holds that the two types of socialisation typical of the
middle and working classes respectively, give rise to differing
relations to language. The middle-class child controls both a
restricted code (roughly emotional, illogical, inexplicit and
incorrect, useful for expressing group solidarity and feeling) and
an elaborated code (which facilitates higher cognitive operations
 Beyond Alienation 159

through its logic and explicitness). The working-class child

controls only restricted code, and thus her ability to perform the
sort of intellectual tasks expected at school is limited.
The claim that black children were restricted code only
speakers, therefore, was an attempt to explain why they failed, or
underachieved relative to white children, at school. It led to a
compensatory education project in which children were taught
the appropriate elaborated code (i.e. white middle-class English).
The sociolinguist Labov showed that this project, and the
premises underlying it, were fundamentally misguided. 89 For one
thing, the linguistic features defining elaborated code turned out
to be nothing more than an amalgam of middle-class habits (like
use of the passive and the pronoun one): it was hard to argue they
had any inherent value. The features stigmatised in black English
were not failed attempts at standard English, but systematic
variations, or more accurately, parts of a related but different
dialect. In other words, it was (and is) fundamentally unclear
whether there is such a thing as a restricted code which is unable
to express complex ideas, logical relationships and so on.
In the second place, Labov demonstrated that black children
grow up in an extremely stimulating verbal culture with its own
rituals. Individuals who at school were silent or inarticulate were
very likely to metamorphose, within their peer group, into skilled
verbal performers. To unearth the rich verbal culture of black
adolescents, Labov had to go to a great deal of trouble, for they
did not willingly display it to white outsiders - which was why
successive experiments had failed to elicit anything but silence
and inarticulacy. Labov used a young black fieldworker to elicit a
wide range of data, and in analysing it he deliberately abandoned
his educated middle-class notions of correctness and formality.
Labov concluded that black children failed in school mainly
because they had no motivation to succeed. They defined them-
selves in opposition to dominant white values, and to be fully
integrated members of their peer group they had to express
disdain for formal education.
Other studies of nonstandard language users (to use the term
restricted code would be to beg the question linguistically, as I
have already pointed out) stress that there is a linguistic problem,
but not in the language itself so much as in the stigma people
attach to it. In other words, the theory of codes could be boiled
down to an essentially political truism: those who do not speak the
160 Feminism and Linguistic Theory

language of the dominant elite will find it difficult to get on.

I have dealt with Bernstein's code theory in detail because I
think there are parallels in it with the case of women, and that
feminists could learn a number of lessons from the controversy it
provoked.
Women are not in quite the same position as working-class and
black speakers. Their language is less obviously different from
men's than working-class from middle-class or black from white
varieties, and the differences are much more often below the
level of speakers' consciousness. Nor has anyone yet suggested
that women's speech variety is responsible for massive educational
under-achievement. In many respects, however, women's
language has been treated as if it were a type of restricted code.
And this evaluation has come both from feminists (who speak of
the silence and inarticulacy of women and their culture, and of
the inauthenticity with which they have been forced to express
themselves) and by old fashioned gallants and chauvinists.
Jespersen, for instance, presents the features which typify female
speech as products of an impoverished cognitive apparatus whose
shortcomings are surprisingly similar to those detailed by
Bernstein in his descriptions of restricted-code speakers. Even the
linguistic hallmarks of restricted code and women's language are
the same: a preference for conjoining over the more complex
embedding, unfinished sentences, and a heavy reliance on
intonation rather than more 'explicit' syntactic devices. 40 Writers
agree, in short, that there are various things women's language is
inadequate to express.
Except in so far as it applies to all communication, this strikes
me as a false and dangerous belief. Perhaps research now being
done on women in small groups, on female folklore and culture,
will break it down, both by showing that women have rich and
complex verbal resources, and by proving that the folklinguistic
consensus on women's speech style is inaccurate. Researchers in
this latter area should focus, as I have tried to do in this book, on
the connections and similarities between feminist and anti-
feminist folklinguistic beliefs, and on the importance of value
judgements in producing what disadvantage women do suffer as
speakers and writers.
It is also important to make explicitly the connection between
women as disadvantaged speakers and the disadvantaging of
other subordinate groups such as ethnic minorities and the
 Beyond Alienation 161

working class. Such groups could certainly learn from the WLM's
refusal to ignore questions of language and politics; on the other
hand, the WLM in making those links might be moved to revalue
certain theoretical excesses.
As the Dalston Study group observe, 'Immigrants and working-
class people too have a negative point of entry into our culture,
something no one has yet explained with reference to the
penis/ phallus. '41 If language is an important political and
personal resource, feminism cannot afford a theory that tells
women only how they are oppressed as speakers: it must convince
them also of 'the strength and potential of their own language'.
Although the nature of communication is such that men cannot
appropriate meaning nor completely control women's use of
language, they (or a subset of them) control important institu-
tions and practices. The effect of that control is to give men
certain rights over women and to hedge women around with
restrictions and myths. Its mechanisms range from explicit rules
against women speaking in public or on ritual occasions to folk-
linguistic beliefs and values denigrating women's language and
obscuring female verbal culture. These rules, prescriptions and
beliefs can be related on the one hand to femininity in general,
and on the other, to the linguistic and cultural subordination of
other oppressed groups.
Whereas the current feminist belief in determinism, male
control and female alienation offers very little prospect of
struggle and liberation, we get a more hopeful picture when we
concentrate on metalinguistic and discursive processes linked to
women's identity and role in particular societies. These processes
can be challenged much more easily and effectively than langue,
meaning, alienation and other such abstractions. Practice and
strategy are one subject to be taken up in the conclusion.
9 Conclusion: Feminism
and Linguistic Theory:
Problems and Practices
If one continues to believe in the
project of human speech, one must
move beyond a view of language as
simply or inexorably 'power over',
discourse as domination ... and toward
speech as part of an emancipatory
effort, a movement toward social
clarity and self-comprehension.
Jean Bethke ELshtain

At the beginning of this book I set out to assess the state of the art
with regard to language in feminist theory. In this conclusion we
must draw all the threads together, pinpoint the outstanding
problems and - most important - examine the implications for a
radical linguistic practice.
In the course of my discussion I have dealt with a diverse group
of theorists addressing themselves to a wide range of questions
from a large number of political, linguistic and philosophical
perspectives. So in drawing the threads together the first problem
must be this: can we assemble all these viewpoints into one
coherent theoretical framework, and is it desirable to do so?
Undoubtedly, there are many who find the diversity of
approaches in this field problematic. Faced with writers who
experiment with frameworks from all over the social sciences
(Kramarae1 ) or cobble together an eclectic approach (Spender2),
reviewers of the literature have occasionally bemoaned the lack of
any theoretical pigeonhole to put questions of language and sex
in. One recent reviewer, by insisting that 'We need a coherent
theory within which work can be done on the subject of the inter-
action between sexual gender ... and language, 3 implicitly
acknowledges that we haven't currently got one, and that this is a
bad thing. Feminists may choose to regard this as a non-problem,

162
 Conclusion: Feminism and Linguistic Theory 163

either because they scoff at all theory, or, more likely, because
they reject the implied criticism of borrowers and cobblers as
pointless purism. Many may feel, in addition, that pluralism is
healthy and stimulating, whereas orthodoxy would be depressing.
Nevertheless, it could be argued that the present state of affairs
is not so much pluralistic as simply confused. For even if we can
never come up with a synthesis of all our viewpoints and frame-
works (and this looks more than likely, given the tensions between
reformism and radicalism, or empiricism and subject-theory) it
should be possible in a pluralist world to initiate dialogue between
theorists who disagree: and as things stand this is a far from
straightforward enterprise. The areas of disagreement are still
insufficiently clear, and theorists still either unable or reluctant to
spell out what issues are at stake.
A feminist linguistic theory, in my opinion, is a theory that
links language with sex in two ways: it spells out the connection
on the one hand between language and gender identity, and on
the other hand between language and women's oppression. If we
are to have useful dialogue, therefore, all feminist linguistic
theories must make it clear where they stand on four basic
questions. First, what are we talking about when we talk about
language? Secondly, what do we mean by women's language (or
indeed men's language): how is the link between language and
gender to be understood? Thirdly, what is the relation of
language to reality? Fourthly, what is the relation of language to
disadvantage, particularly (but not exclusively) in the case of
women?
Definite answers to these four questions would provide a basis
both for fruitful debate amongst differing theoretical tendencies,
and for the development of a feminist practice in speech and
writing. The remainder of the conclusion will therefore be
devoted to examining the questions in more detail, focusing
especially on the problems that are not satisfactorily resolved by
the theories I have surveyed so far.

WHAT ARE WE TALKING ABOUT WHEN WE TALK

ABOUT LANGUAGE?

One of the most difficult terms to grasp in feminist writing about

language is the deceptively simple word language itself, since this
164 Feminism and Linguistic Theory

word is often used with a generality bordering on the vacuous.

For instance, a feminist writes that women are oppressed by
language; but the next sentence makes it clear she is speaking of
art and mythology.4 There is no apparent distinction between
literal and metaphorical uses of the term.
Similarly, we can read in feminist writing of 'the grammar of
moral discourse' and 'the language of theory'. 5 But what are we
talking about? Does 'the language of theory,' for instance, mean
'the sort of words used in theories' or 'the type of syntax used in
theoretical writings' or just 'what is said in theories'? Are women
excluded from theoretical language because the lexical
definitions of theory writers excluded our perspective. because
the syntax customary in theory writing is alien to our way of
dealing with ideas, because we disagree with what theories assert,
or b~cause our participation in theoretical discourse would be
considered unfeminine? And are all these possibilities equally
matters of language?
Because it is usually impossible to tell, there is a tendency to
take these phrases on trust, to skip over them and accept whatever
proposition rests upon their use. Yet the implications and possible
lines of attack are very different depending on whether we are
talking about words, syntax. content or custom. In other words,
unless we formulate our critique of language with a certain
linguistic precision, our theory will not tell us how to proceed in
practice with our struggle. Instead of demystifying patriarchal
language, we have produced an analysis which, if it achieves
anything. achieves only further mystification.
This is not to say that a theory of language must exclude art, or
myth, or the metalinguistic customs that regulate language use.
What matters is simply that we should all know what we are
talking about. Theory making is the paradigm example of an
activity in which our metalinguistic skill for defining terms is
relevant, and even vital. We cannot leave the definition of
language to the vagaries of context as we would in ordinary con·
versation, because we are trying for something less hit-and-miss
than the cross-purpose talk of an ordinary interchange.
It also seems likely that problems of definition have a real
bearing on apparently more intractable problems of theory. For
instance, it might be argued that the language-and-reality
problem arises very largely out of an over-wide interpretation of
the word language. Because we use language to learn about and
 Conclusion: Feminism and Linguistic Theory 165

reflect on other social phenomena, the story goes, all those social
phenomena must in the end reduce to language. The minute we
start to consider what we mean by language this assertion looks
less secure. Examining definitions leads to a healthy questioning
of assumptions.
Another crucial distinction blurred in the term language is the
difference between speech and writing. Linguists have a dogma
on this subject to go along with their dogmas on prescriptivism
and objectivity: they say that speech is more basic than writing,
and writing is no more than a graphic representation of speech.
But the reality of linguistic practice, and indeed of people's
language-use, is entirely different.
In practice, linguists treat speech as though it were a poor
imitation of writing. You can see this in the fact that linguistics
has long taken the complete sentence as its standard unit of
analysis, even though most spontaneous speech contains rather
few complete sentences. Thus complicated 'editing rules' have to
be posited to account for the divergence of speech from the
ideal and much of what the linguist hears is defined as 'ungram-
matical'. Prosodic features like stress, rhythm and intonation are
marginalised. All this points to the existence of writing as a
hidden norm rather than a second-order representation (indeed,
the claim that writing is just transcribed speech is falsified the
minute the two are actually compared). Yet many of the claims
made for and against the language of women by semiologists and
linguists alike rest on the same prejudice that speech is 'really' just
writing in a different form. Consider Kristeva's placing of the
prosodic, that is to say, rhythm, stress and intonation, in the
semiotic order while logical syntax goes in the symbolic order. For
speech this is a wholly artificial separation. Syntactic functions are
carried out to a very great extent by prosodic devices when we
talk, and writing is arguably much impoverished by the fact that
it cannot combine the affective and the logical in the same way.
One could only ever marginalise the prosodic as something extra
rather than integral to communication if one took writing, not
speech, as the norm of language use. Once that is admitted,
claims that women use intonation rather than grammar can be
revealed for what they are: meaningless. A similar fate must also
befall Luce Irigaray's claim that feminine language has nothing
to do with the syntax of the complete sentence. Unless she makes
it clear she is talking offeminine wr#ing, the obvious retort is that
166 Feminism and Linguistic Theory

no one's language has much to do with the syntax of the complete

sentence. The complete sentence is entirely an artefact of
literacy.
When we consider the demands put forward for a feminine
'language', however, it is very soon clear that the feminists
concerned are thinking only of writing. For the kind of feminine
language they envisage is literally unspeakable, however writeable
it may be. It has puns made in the spelling, the punctuation and
through diacritic makings; the structure is convoluted, needing
considerable time to produce and to process. Snatches of uncon-
nected discourses may be juxtaposed in a way that requires
careful reading. Even the less formally difficult strategies, such as
Mary Daly's constant parenthetical interrogation of etymology
and structure in the words she uses, would be either tedious or
just impossible in talk.
Of course, textual strategies need not be rejected just because
they won't do for speech. On the contrary, the medium of the
written word should be exploited for all it's worth. But isn't it a
sort of cheating, nevertheless, to talk of language and mean only
written language? To generalise about language (as Kristeva and
Irigaray do) in terms that do not apply to the vast majority of
language users and language events? Unfortunately, this criticism
is as true of linguistics as it is of Kristevan semiology. 'Scriptism' -
a prejudice against speech as compared to writing - is ingrained
in the western tradition and constitutes one of the most serious
misunderstandings of the nature of human communication so far
embranced even by linguistic science. Any radical theory of
language must be on its guard against scriptist tendencies and the
many inaccuracies they engender.

LANGUAGE AND GENDER: MAKING THE

CONNECTIONS

Once we have got clear what phenomena are to count as

'language' in our theory, we can go on to examine what we mean
by 'women's language' and how people's use of language signals or
interacts with or partly constitutes their identity as feminine or
masculine. In fact every theorist has some notion of the frame-
work in which women's identity and position is to be related to
 Conclusion: Feminism and Linguistic Theory 167

language, whether this is made explicit or not: by deliberately

making explicit the frameworks used by the various writers we
have looked at, we can isolate three main strands.

1 Subculture and gender role

Within the first approach, the important idea is of women and

men forming separate subcultures. The differing speech of the
two sexes is thus seen as a function of their differing roles; it
derives from, and expresses, a whole complex of factors associated
with social maleness/femaleness, including particular personality
traits (e.g. in our own culture, masculine aggression and feminine
passivity) and identity markers which derive from the sexual
division of labour. Adherents of this point of view include the
Ardeners, Maltz and Borker, Smith and the later Cheris
Kramarae.

2 Dominance hierarchies

The second approach focuses on one particular aspect of the

female role: its powerlessness relative to the male role. Women's
speech therefore expresses no complex or specifically female role
identity, but simply a low position in the social hierarchy. Women
use an inherently powerless style of speech, but it is not peculiar
to their sex; it may be used by anyone in a position of relative
powerlessness. Haden-Elgin representatively asserts, There is no
such thing as a specific 'women's language' .... Rather there are
dominant and subordinate language modes, and women, being a
subordinate group, are frequently found using these subordinate
modes.'6 Robin Lakoff also belongs to the 'dominance hierarchy'
camp, holding that women's language has nothing to do with
femininity per se, and everything to do with women's subordinate
status.

3 Sexuality and the body

Finally, we have seen that there are many theorists who believe
that language is related to gender identity through its intimate
168 Feminism and Linguistic Theory

connection with the body and sexual desire. Adherents of this

view (most strongly endorsed by Helene Cixous and Luce
Irigaray) usually argue that the feminine language of which they
speak does not yet exist: it is a utopian goal rather than a
describable reality.
These three contrasting explanations of the link between
language and gender imply very different political practices. If
the link is gender role (a complex of personality and occupational
traits, socially conditioned values and beliefs regarding sexual dif-
ferentiation) then it would be feasible for us to revalue feminine
styles of speech. We could make our language into a badge of our
identity as women who are proud to be women. If on the other
hand the link is nothing more than relative power, it is incumbent
upon us to jettison our powerless way of talking as something that
reinforces our powerless position in a vicious circle. We should all
take assertiveness training and start talking like men. Or if the
link really is sexuality and the rhythms of the female body, we
must reject all current models of speech and strive towards a
discourse that doesn't yet exist. In many ways, these three alter-
natives correspond to Kristeva's three feminisms: Lakoff
represents liberal feminism (make women and men equal),
Kramarae represents radical feminism (revalue the distinctively
feminine) and Cixous is the revolutionary (refuse existing classi-
fications).
For most feminists, the notion of a language close to the body
and sexual desire is still problematic, either an impossibility or
something to be found only in the 'heightened' language of
poetry. The question of power versus role is more acute, however,
For while it clearly will not do to say that women are nothing but
passive victims, that their cultural products express nothing more
complicated than sheer powerlessness, or that their language is
just like the language of all other oppressed groups, on the other
hand it will not do to formulate any kind of female identity that
takes no account of the reality of powerlessness. Powerlessness is
an important factor in women's lives and in the identities they
forge. Theoretically, then, the choice between 'dominance' and
'subculture' theories is a false choice, and similarly the practical
choice between rejection and revaluation is not clear-cut. In
conclusion, then, we need a more complex way of linking ways of
speaking and writing to our multifaceted identities as women.
 Conclusion: Feminism and Linguistic Theory 169

LANGUAGE AND REALITY

We have already seen that the issue of language and reality - of

linguistic determinism - is central to the dispute between
theoretical reformists and radicals. However, the issue itself
remains unresolved, and this requires discussion. Most theories
take up an absolute position on the question of language and
reality. Either they see language quite unproblematic ally as a
reflection of something else - usually an ill-defined 'social organ-
isation' or 'thought' - or else they see it, equally unprob-
lematically, as the source of that mental or social reality.
Moreover, the terms of the debate are set in such a way that there
cannot be any answer in between the two extremes. Opponents of
determinism appeal to notions of common sense and to free-will,
which they accuse the radicals of denying: determinists, on the
other hand, accuse non-believers of naviety and of clinging to an
essentialist conception of human nature and experience which is
inaccurate, outmoded and irrevocably bourgeois. But is there
really nothing in between? Can there be a theory that is neither
essentialist nor determinist in the radical sense?
I do not believe that language is the first cause, and I see
nothing wrong with asserting that meaning derives from
something we might call experience, as well as from immediate
context. However, that does not mean I believe in some central
core of human nature (though I might posit a core of human
cognitive abilities/rationality, given by the nature of our bodies
and brains) nor in the existence of an inner mental world
untouched by social forces. There is no pre-existent and innocent
subjectivity on which a layer of ideology is somehow to be super-
imposed. On the contrary I agree with the semiologists and with
all anti-humanists that our 'personalities', our desires, our needs,
our ways of behaving, are constructed in our interactions with the
world. These constructed elements are our real selves, and not
just some kind of false consciousness that can simply be stripped
away. I accept, also, that we are complex creatures who cannot
fully understand our own actions, influenced as they are by
factors of which we remain wholly unconscious.
What I cannot accept, however, is the privileged status
accorded language in this process of construction. Of course it
170 Feminism and Linguistic Theory

plays some part - it cannot be just a reflection of other things, for

it is basic to much of our experience - but other things are
important too: perhaps even more important, for they happen
earlier in our lives and are less able than language to become
objects of reflection and interpretation in their own right. I am
thinking of socio-familial relations; of the division of labour and
economic organisation that regulates societies; of the physical
environment; of individual genetic make-up. All these things
interact with each other and with the situations we happen to be
in to produce what I would call experience, a body of relation-
ships and interactions with the world that inform behaviour in
complex ways. The acquisition and interpretation of this ex-
perience is a never-ending process.
Those who seek to understand the workings of human societies
must resist the temptation to reduce them to a single generating
principle. Relations of production, relations with parents and
siblings, language - it is always simplistic to set up one variable as
basic. The quest for a first cause is like asking which came first,
the chicken or the egg. Luckily, it is still possible to ask how
chickens produce eggs, and how eggs produce chickens, even if
the question of which came first remains for ever unanswered.
The either/or debate in feminist linguistic theory is equally fruit-
less; and what is really needed is a radical questioning on both
sides of their taken-for-granted notions about the production of
meaning.

LANGUAGE AND DISADV ANT AGE

The premise of any feminist linguistic theory must be that

language and oppression, language and disadvantage, are
somehow connected. That language is a resource of the powerful
(or at least, that it can be used thus) and a potential instrument of
oppression is not in doubt.
But one tendency it seems to me we must counter as much as
possible is the extraordinary tendency, illustrated throughout this
book, to blame what happens to people on the sort of language
they use. This has been a popular explanation for some time:
working-class and black children under-achieve at school because
their language is inadequate, communism triumphs because the
 Conclusion: Feminism and Linguistic Theory 171

Russian language has been corrupted, women are marginal and

negative social beings because their language is ineffectual!
inauthentic. Linguistic reform is the key to all that is good, there-
fore, from equality of opportunity to democracy and Truth.
But how could language per se be responsible for disadvant-
age? Certainly it would be easy and convenient to believe it was:
under-achieving children could be given 'compensatory'
education, women could do assertiveness training, Asians could
be taught code-switching and the Russian dictionary could be
written by the United Nations. No one in a privileged position
need change, or give up, anything. We need not ask questions
about our standards of adequate and inadequate language use,
nor consider where they came from in the first place. And above
all, we need not admit that the institutions of society dis-
advantage the poor, the black and the female just because they
are poor, or black, or female. We can go on pretending that
language change (or education, or equal opportunity laws) is
enough. It seems to me that this is a particular pernicious lie
which linguists have a responsibility to expose. The use of
linguistic and metalinguistic resources to oppress others should
not be ignored; but we must acknowledge the limitation of
theories of oppression that do not go beyond the linguistic.

CHANGING OUR PRACTICE: TOWARD A RADICAL

DISCOURSE

In this book I have reached the conclusion that no perfect

feminist speech or writing style exists, nor could one ever exist.
Moreover, the changes normally demanded by feminists have
distinct disadvantages: the writing style sometimes held to be
close to feminist ideals is also particularly difficult, so that it is
open to charges of elitism, while the non-sexist language
advocated by the reformist tendency has at its core a politically
dangerous illusion, the illusion of sexual neutrality. Yet even if we
believe that non-sexist language is little better than ordinary
sexist language, the demand that people change their usage can
be liberating. For one thing, it cannot fail to draw attention to
the present masculist bias of conventional usage; but more
important, it calls into question the common-sense transparency
172 Feminism and Linguistic Theory

and fixedness of meaning. The political importance of this has

been ably discussed by Trevor Pateman. Pateman believes that
people in our society are encouraged to engage in what he calls
'idle discourse'. Idle discourse dodges meaning, and, rather like
the young Simone de Beauvoir, treats definitions as closed, not
possible subjects for rational dispute. In other words, it is a meta-
linguistically impoverished language in which meanings are static
and taken for granted.
The existence of idle discourse is not a reflection of the nature
of language. Discourse does not have to be idle, and its idleness is
only a product of people's repressive socialisation and hopeless
circumstances.

'Idle discourse is the language of the powerless who accept their

pos~tion. To the degree that the pursuit of security dominates
everyday thought and language use, I think this is because
people have decided that other satisfactions are not
obtainable. '7

Pateman appears to believe that what people do in their dealings

with words they will do in other areas of their experience. The
unthinking acceptance of other people's definitions will go hand·
in-hand with a general reluctance to question the way the world
operates. Anything that encourages people to reflect on
language, and particularly on the provisional status of meaning
and definition, is thus politically progressive. Pateman points out
that outer changes, such as the use of non· sexist language, can
affect inner attitudes ultimately, because they do as a matter of
fact change the political status quo: 'the change in outward
practice constitutes a restructuring of at least one aspect of one
social relationship ... every act reproduces or subverts a social
institution.'8 This is an important point to make when some
theorists are enquiring cynically what the difference is between
macho men and those who have learned to say 'she'. The
importance of linguistic change, and indeed of many other sorts
of reform, is precisely that 'every act reproduces or subverts a
social institution'. We always have a choice.
Pateman also points out that to make a change in the way you
speak or write is to assume a certain responsibility for your
relationship to the world and for your behaviour, in its wayan act
of the greatest political importance. 'For in my act I have asserted
 Conclusion: Feminism and Linguistic Theory 173

that I can control language; I have stopped acting as if language

necessarily controls me. '9 This is surely what Jean Bethke Elshtain
means when she refers to speech as 'part of an emancipatory
effort, a movement toward social clarity and self comprehension'.
She goes on to add, 'the project of rational speech, an eyes-open,
truth-telling passion against "the powers that be" and "the
censors within" can be one emancipatory window into the
future'. JO Radical discourse, then, is the very opposite of idle
discourse. It constantly questions the metalinguistic practices by
which idle discourse is created and encouraged and thus by which
power relationships are reproduced. It questions, also, the
stability of meaning, and asserts very forcefully that we can
change our usage by a conscious act of will.
Perhaps the most positive effect of changing our linguistic
practice will be to destroy the pernicious belief that we have to be
controlled and oppressed by our language. Once over that
hurdle, we can start learning to speak out with confidence and to
use the resources of language and metalanguage, so often denied
us or used against us, in the continuing struggle against
patriarchy.
NOTES
CHAPTER 1: INTRODUCTION

1. Philip M. Smith, 'Sex Markers in Speech', in K. Scherer and H. Giles

(eds), Social Markers in Speech (CUP, 1979).
2. Casey Miller and Kate Swift, The Handbook of Non-Sexist Writing
(Women's Press, 1980).
3. Betty Friedan. The Femz'nine Mystique (Gollancz, 1963).
4. Camilla Gugenheim, 'Man Made Language', Amazon, 4, 1981.
5. Mary Daly, GynlEcology: the Metaethz'cs of Radical Feminism
(Women's Press, 1978).

CHAPTER 2: LINGUISTIC THEORY

l. E. Marks and I. de Courtivron (eds), New French Feminz'sms

(Harvester Press, 1981).
2. F. de Saussure, Course in General Linguz'stics, trans. W. Baskin
(Fontana, 1974) p. 16.
3. Ibid.
4. Roland Barthes, Mythologies (du Seuil, 1957).
5. Noam Chomsky, review of Verbal Behavior, Language, 35, 1959.
6. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980) and Cheris Kramarae, Women and Men Speaking (Newbury
House, 1981).

CHAPTER 3: THE POLITICS OF VARIATION

1. Mercilee Jenkins and Cheris Kramarae, 'A Thief in the House: the
Case of Women and Language', Men's Studies Modified, ed. Dale
Spender (Pergamon, 1981).
2. Jonathan Swift, 'A Proposal for Correcting the English Tongue',

174
 Notes 175

Prose Works of jonathan Swift, vol. IV, ed. H. Davis (Blackwell,

1957).
3. O. Jespersen, Language: Its Nature, Development and Origin
(Allen & Unwin, 1922).
4. Lecture on 'Varieties of English' given in Oxford by the Warden of
Keble, 1982.
5. Robin Lakoff, Language and Women's Place (Harper & Row,
1975).
6. B. L. Dubois and I. Crouch, 'The Question of Tag-Questions in
Women's Speech: They Don't Really Use More of Them, Do
They?', Language in Society, 4, 1976.
7. D. Crystal and D. Davy, Advanced Conversational English
(Longman, 1975).
8. H. P. Grice, 'Logic in Conversation', Syntax and Semantics, III,
ed. P. Cole and J. L. Morgan (Academic Press, 1975).
9. Cheris Kramarae, Women and Men Speaking (Newbury House,
1981) p. 9.
10. R. F. Bales, 'How People Interact in Conferences', Communication
in Face to Face Interaction, ed. J. Laver and S. Hutcheson
(Penguin, 1972).
11. J. Pellowe, G. Nixon, B. Strang and V. McNeany, 'A Dynamic
Modelling of Linguistic Variation: the Urban (Tyneside) Linguistic
Survey', Lingua, 30, 1972.
12. Jenkins and Kramarae, op. cit., p. 16.
13. K. Scherer and H. Giles (eds) , Social Markers in Speech (CUP,
1979).
14. Jenkins and Kramarae, op. cit., p. 16.
15. P. Trudgill, 'Sex, Covert Prestige and Linguistic Change in the
Urban British English of Norwich', Language in Society, 1, 1972.
16. Elinor O. Keenan, 'Norm Makers, Norm Breakers; Uses of Speech
by Women in a Malagasy Community', Explorations in the
Ethnography of SPeaking, ed. R. Bauman and J. Sherzer (CUP,
1974) p. 142.
17. Ibid., p. 141.
18. Christine Delphy, 'Women in Stratification Studies', Doing
Feminist Research, ed. Helen Roberts, (Routledge & Kegan Paul,
1981).
19. Kramarae, op. cit.
20. D. Maltz and R. Borker, 'A Cultural Approach to Male/Female
Miscommunication', Language and Social Identity, ed. J. Gumperz
(CUP, 1982).
21. Lakoff, op. cit.
22. Caroline Henton, 'Sex Specific Phonetics and Social Realisation',
unpublished 1983.
176 Notes

23. P. M. Smith, 'Sex Markers in Speech' Social Markers in Speech, ed.

K. Scherer & H. Giles (CUP, 1979).
24. Dubois and Crouch, 0p. cit.
25. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980) p. 8.

CHAPTER 4: FALSE DICHOTOMIES

1. Simone de Beauvoir, The Second Sex, trans. Parshley (Vintage,

1974).
2. Luce Irigararay, Ce Sexe quz" n'en est pas un (Minuit, 1977).
3. Jane Gallop, 'Psychoanalysis in France', Women and Lz"terature,
vol. 7, no 1, p. 61.
4. M. Bierwisch, 'Semantics', New Horz"zons in Linguistics, ed. J.
Lyons (Penguin, 1970).
5. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980) p. 2.
6. Women'sJoumal.
7. Marielouise Janssen-Jurreit, Sexism (Pluto Press, 1982) p. 280.
8. John Lyons, Introduction to Theoretical Unguistz·cs (CUP, 1968)
p.284.
9. Jakob Grimm, Deutsche Grammatik, quoted in Janssen-Jurreit,
Sexism: the Male Monopoly of History and Thought (Pluto Press,
1982) p. 292.
10. Ibid., p. 297.
11. Anne Corbett, 'Cherchez la metaphor', Guardian, 18 Feb. 1983.
12. Harvard Crz"mson, quoted in C. Miller and K. Swift, Words and
Women (Penguin, 1976) p. 92.
13. B. Dubois and I. Crouch, 'American Minority Women in
Sociolinguistic Perspective' IjSL, 1978, p. 9.
14. Ann Bodine, 'Androcentrism in Prescriptive Grammar', Language
in Society, 4, ·1975.
15. Maria Black and Rosalind Coward, 'Linguistic, Social and Sexual
Relations', Screen Educatz"on, 39, 1981.

CHAPTER 5: MAKING CHANGES

1. Stephen Kanfer, 'Sispeak', Tz"me, 23 Oct. 1972.

2. C. Miller and K. Swift, The Handbook of Non-Sexist Writing
(Women's Press, 1980).
3. C. Miller and K. Swift, Words and Women: New Language in New
Tz"mes (Penguin, 1976).
 Notes 177
4. Ibid., p. 8.
5. Muriel Schulz, 'The Semantic Derogation of Women', Language
and Sex: Difference and Dominance, ed. B. Thorne and N. Henley
(Newbury Hall, 1975).
6. Andrea Dworkin, Pornography: Men Possessing Women (Women's
Press, 1981).
7. Mary Daly, GynlEcology: the Metaethics of Radical Feminism
(Women's Press, 1978).
8. Miller and Swift Handbook, p. 4.
9. Ibid., p. 8.
10. Cf. John Hall, The Sociology of Literature (Longman, 1979).
11. Miller and Swift, Handbook, p. 4.
12. I am indebted to Peter Miihlhausler for pointing this out to me.
13. D. Mandelbaum (ed.), Selected Writings of Edward Sapir
(University of California Press, 1949).
14. Roger Scruton, 'How Newspeak Leaves Us Naked', The Times,
2 Feb. 1983.

CHAPTER 6: SILENCE, ALIENATION AND OPPRESSION

1. Mary Daly, GynlEcology: the Metaethics of Radical Feminism

(Women's Press, 1978) p. 1.
2. Adrienne Rich, 'Power and Danger', in On Lies, Secrets and
Silence (Virago, 1980) pp. 247-8.
3. F. de Saussure, Course in General Linguistics, trans. Baskin
(Fontana, 1974) p. 112.
4. E. Marks and I. de Courtivron (eds), New French Feminisms
(Harvester Press, 1981) p. xiii.
5. Ian Griffiths, 'Speech, Writing and Rewriting' (unpublished paper)
p.45.
6. Saussure, op. cit., p. 14.
7. D. Mandelbaum, Selected Writings of Edward Sapir (University of
California Press, 1949) p. 162.
8. J. B. Carroll (ed.), Language, Thought and ReaHty: Selected
Writings of Be~jamin Lee Whoif (MIT Press 1976).
9. Mandelbaum, op. cit., p. 10.
10. Carroll op. cit., p. 253.
11. J. Lyons, Introduction to Theoretical Linguzstics (CUP 1968)
p.434.
12. G. Orwell, 'Politics and the English Language', in Selected Essays
(Secker & Warburg, 1961) p. 353.
13. Roger Scruton, 'How Newspeak Leaves Us Naked', The Times,
1 Feb. 1983.
178 Notes

14. Julia Kristeva, 'Woman Can Never Be Defined', interview with

psych et po, in Marks and de Courtivron, op. cit., p. 140.
15. Marks and de Courtivron, op. cit., p. 3.
16. Shirley Ardener (ed.), Perceiving Women (Dent, 1975) p. xii.
17. Shirley Ardener, Defining Females Oohn Wiley, 1978) p. 21.
18. Edwin Ardener, 'Belief and the Problem of Women', in Perceiving
Women, ed. S. Ardener, p. 22.
19. Ardener, Defining Females, p. 20.
20. Ibid., p. 21.
21. E. Ardener, op. cit.
22. Ardener, Defining Females, p. 20.
23. Ardener, Perceiving Women, p. ix.
23. Ibid., p. xxi.
24. Ardener, Defining Females, pp. 22-3.
25. Ibid.
26. D. Zimmerman and C. West, 'Sex Roles, Interruptions and Silences
in Conversation', Language and Sex: Difference and Dominance,
ed. B. Thorne and N. Henley (Newbury House, 1975).
27. Cheris Kramarae, Women and Men Speaking (Newbury House,
1981) p. 3.
28. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980) pp. 2-3.
29. Ibid., p. 3.
30. Ibid., p. 4.
31. Ibid., p. 139.
32. Ibid., p. 139.
33. D. Smith, 'A peculiar eclipsing: women's exclusion from men's
culture', WSIQ I, 1978, pp. 281-2.
34. Spender, op. cit., p. 1.

CHAPTER 7: SEMIOLOGY AND THE GENDERED

SUBJECT

l. R. Coward and J. Ellis, Language and Materialism (Routledge &

Kegan Paul, 1977), p. 153.
2. M. Black and R. Coward, 'Linguistic, Social and Sexual Relations',
Screen Education, 39.
3. Cf. Christine Delphy, 'A Materialist Feminism is Possible', Feminist
Review, 4.
4. Black and Coward, 0p. cit., p. 70.
5. Ibid., p. 72.
6. Ibid., p. 72.
 Notes 179

7. Ibid., p. 81.
8. Toril Moi, 'Who's Afraid of Virginia Woolf: Feminist Readings of
Woolf', 1982, p. 1.
9. Coward and Ellis, op. cit., p. 1.
10. F. de Saussure, Course in General Linguistics, trans. Baskin
(Fontana, 1974) p. 14.
11. J. Mitchell andJ. Rose (eds), Feminine Sexuality (Macmillan, 1982)
p.49.
12. Ibid., p. 42.
13. Coward and Ellis, op. cit., p. 97.
14. J. Lacan, Le Seminaire XX: Encore (Editions du Seuil, 1975) p. 68.
15. Toril Moi, op. cit., p. 14.
16. Tori! Moi, 'Language, Femininity, Revolution: Julia Kristeva and
Anglo·American Feminist Linguistics', unpublished lecture, 1983,
p.23.
17. L. Irigaray, 'Women's exile,' Ideology and Consciousness, 1,. p. 62.
18. Ibid., p. 64.
19. Ibid., p. 65.
20. Ibid., p. 69.
21. Ibid., p. 71.
22. 'Variations on Common Themes', New French Feminisms, ed.
E. Marks and I. de Courtivron (Harvester Press, 1977) p. 219.
23. Jacobus, 'The Question of Language: Men of Maxims and the Mill
on the Floss', Critical Inquiry, vol. 8, no. 2, 'Writing and
Difference', 1981, p. 207.
24. Delphy, op. cit.
25. Diana Leonard, 'Male Feminists and Divided Women', in On the
Problem of Men, ed. S. Friedman and E. Sarah (Women's Press,
1982) p. 161.

CHAPTER 8: BEYOND ALIENATION

1. E. Marks and de I. Courtivron, New French Feminisms (Harvester

Press, 1981) p. 30.
2. R. Coward, S. Lipshitz and E. Cowie, 'Psychoanalysis and
patriarchal structures', in Papers on Patriarchy, Women's
Publishing Collective (PDC, 1978).
3. Dalston Study Group, 'Was the patriarchy conference
"patriarchal"?', in Papers on Patriarchy, p. 76.
4. Ibid., p. 77.
5. Simone de Beauvoir, Memoirs of a Dutiful Daughter, trans. Kirkup
(Penguin 1963) p. 17.
180 Notes

6. Ibid.
7. Roy Harris, The Language Myth, (Duckworth, 1981) pp. 87 -8.
8. Ibid. pp. 88-9.
9. Ibid., p. 165.
10. Ibid.
11. Beauvoir, op. cit., p. 17.
12. Camilla Gugenheim, 'Man Made Language?' Amazon, no. 4, 1981.
13. Audre Lorde, The Cancer Journals, (Spinsters Ink, 1980) p. 19.
14. Genesis, 11: 6-9.
15. Trever Pateman, Language, Truth and Politics, 2nd edn (Jean
Stroud, 1980) p. 129.
16. Cora Kaplan, 'Language and Gender', Papers on Patriarchy
(WPC/PDC, 1976).
17. Ibid., p. 21.
18. John Oxenham, Literacy: Writing, Reading and Social
Organisation (Routledge & Kegan Paul, 1980) p. 3.
19. Walter J. Ong, Orality and Literacy (Methuen, 1982) p. 113.
20. Oxenham, op. cit., p. 51.
21. Ong, op. cit., p. 3.
22. J. Gumperz, Discourse Strategies (CUP, 1982) and J. Gumperz
(ed.), Language and Socz"al Identity (CUP, 1982).
23. Gumperz, op. cit., pp. 4-5.
24. D. Maltz and R. Borker, 'A Cultural Perspective on Male/Female
Miscommunication', in Gumperz, op. cit.
25. Colin McCabe, 'The Discursive and the Ideological in Film',
Screen, 19/4.
26. M. Black and R. Coward, 'Linguistic, Social and Sexual Relations',
Screen Education, 39, p. 78.
27. M. Jenkins and C. Kramarae 'A Thief in the House', in Men's
Studies Modified, ed. Spender (Pergamon, 1981).
28. Kaplan, op. cit., p. 21.
29. Shirley Ardener, Perceiving Women (John Wiley, 1975); Philip
Smith, 'Sex Markers in Speech', in Social Markers in Speech, ed.
K. Scherer and H. Giles (CUP, 1979).
30. Marielouise Janssen-Jurreit, Sexism (Pluto Press, 1982) p. 284.
31. O. Jespersen, Language: Its Nature, Development and Origin
(Allen & Unwin, 1922), p. 246.
32. Elinot Keenan, 'Norm Markers, Norm Breakers' in Explorations in
the Ethnography of SPeaking, ed. R. Bauman and J. Sherzer,
(CUP, 1974).
33. E. Marks and I. de Courtivron (eds), New French Feminisms, p. 5.
34. J.-J. Rousseau, Emile, quoted inJ. O'Faolain and L. Martines, Not
in God's Image, (Fontana, 1974) p. 259.
35. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980) p. 107.
36. Cf. Jenkins and Kramarae, op. cit., pp. 16-17.
 Notes 181

37. See M. Jenkins and C. Kramer, 'Small group process: learning from
women', WSIQ, 3, 1980', D. Jones, 'Gossip: notes on women's oral
culture', WSIQ, 3, 1980.
38. Basil Bernstein, Class, Codes and Control, vol. 1, (Routledge &
Kegan Paul, 1970).
39. W. Labov, The Logic of Non-Standard English, repr. in Giglioli,
P. P. (ed.), Language and Social Context (Penguin, 1972).
40. Jespersen, Language, ch. 24.
41. Dalston Study Group, 'Was the patriarchy conference
"patriarchal"?', Papers on Patriarchy, p. 77.

CHAPTER 9: CONCLUSION

1. Cheris Kramarae, Women and Men Speahng (Newbury House,

1981 ).
2. Dale Spender, Man Made Language (Routledge & Kegan Paul,
1980).
3. Suzette Haden Elgin, Review of Women and Men Speaking,
Language 58, 1982.
4. Mary Daly, GynlEcology (Women's Press, 1978) p. 3.
5. Jean Bethke Elshtain, 'Feminist Discourse and its Discontents',
Femt'nist Theory, ed. N. O. Keohane, M. Z. Rosaldo and B. C.
Gelpi (Harvester Press, 1982) p. 128.
6. Haden Elgin, op. cit.
7. Trevor Pateman, Language Truth and Politics Oean Stroud, 1980)
p.77.
8. Ibid., pp. 15-16.
9. Ibid., p. 16.
10. Bethke Elshtain, op. cit., p. 129.
Bibliography
Ardener, Edwin, 'Belief and the Problem of Women,' in Perceiving
Women, ed. S. Ardener (Dent, 1975).
Ardener, Shirley (ed.), Perceiving Women (Dent, 1975).
- - , Defining Females (John Wiley, 1978).
Bales, R. F., 'How People Interact at Conferences', Communication in
Face to Face Interaction, ed. Laver and Hutcheson (Penguin, 1972).
Barthes, Roland, Mythologies (Editions du Seuil, 1957).
Beauvoir, Simone de, Memoirs of a Dutiful Daughter, trans. Kirkup
(Penguin, 1963).
- - , The Second Sex, trans. Parshley (Vintage, 1974).
Bernstein, Basil, Class, Codes and Control, vol. I: Theoretical Studies
Towards a Sociology of Language (Routledge & Kegan Paul, 1970).
Black, Maria and Rosalind Coward, 'Linguistic, Social and Sexual
Relations', Screen Education, 39, 1981.
Bodine, Ann, 'Androcentrism in Prescriptive Grammar', Language in
SOciety, 4, 1975.
Carroll, J. B., Language, Thought and Reality: Selected Writings of
Benjamin Lee Whoif (MIT Press, 1976).
Chomsky, Noam, review of Skinner's Verbal Behavior, Language, 35,
1959.
Cixous, Helene, 'Sorties', trans. Liddle, New French Feminisms, ed.
E. Marks and L. de Courtivron (Harvester Press, 1981).
Corbett, Anne, 'Cherchez la metaphor', Guardian, 18 Feb. 1983.
Coward, Rosalind and John Ellis, Language and Materialism
(Routledge and Kegan Paul, 1977).
Coward, Rosalind, Sue Lipshitz and Elizabeth Cowie, 'Psychoanalysis
and Patriarchal Structures', in Papers on Patriarchy (Women's
Publishing Collective/PDC, 1978).
Crystal, David and Derek Davy, Advanced Conversational English
(Longman, 1975).
Dalston Study Group, 'Was the Patriarchy Conference "Patriarchal"?',
Papers on Patriarchy (Women's Publishing Collective/PDC, 1978).
Daly, Mary, GynlEcology: the Metaethics of Radical Feminism
(Women's Press, 1978).
Delphy, Christine, 'A Materialist Feminism is Possible', Femiriist
Review, 4, 1980.
182
 Bibliography 183

'Women in Stratification Studies', in Doing Feminist Research,

ed. Helen Roberts (Routledge & Kegan Paul, 1981).
Dubois, Betty Lou and Isobel Crouch, 'The Question of Tag-Questions
in Women's Speech; They Don't Really Use More of Them, Do They?'
Language in Society, 4, 1976.
- - , 'American minority women in sociolinguistic perspective', I]SL
1978.
Dworkin, Andrea, Pornography: Men Possessing Women (Women's
Press, 1981).
Elgin, Suzette Haden, review of Kramarae, Women and Men Speaking,
Language, 58, 1982.
Eishtain, Jean Bethke, 'Feminist Discourse and Its Discontents',
Feminist Theory: A Cn'tique of Ideology, ed. Keohane, Rosaldo and
Gelpi (Harvester Press, 1982).
Friedan, Betty, The Feminine Mystique (Gollancz, 1963).
Gallop, Jane, 'Psychoanalysis in France', Women and Literature, vol. 7,
no. 1, 1979.
Grice, H. P., 'Logic in Conversation', Syntax and Semantics, vol. III:
Speech Acts, ed. Cole and Morgan (Academic Press, 1975).
Griffiths, Ian, 'Speech, Writing and Rewriting', unpublished paper,
1982.
Gugenheim, Camilla, 'Man Made Language?' Amazon, 4, 1981.
Gumperz, John]. (ed.), Discourse Strategies (CUP, 1982).
- - , Language and Social Identity (CUP, 1982).
Hall, John, The Sociology of Literature (Longman, 1979).
Harris, Roy, The Language Myth (Duckworth, 1981).
Henton, Caroline, 'Sex Specific Phonetics', unpublished paper, 1983.
Irigaray, Luce, 'Women's Exile' (interview with Couze Venn), Ideology
and Consciousness, I, 1977.
- - , Ce Sexe qui n'en est pan un (Editions Minuit, 1977).
Jacobus, Mary, 'The Question of Language: Men of Maxims and the
Mill on the Floss', Cn"tz"cal Inquiry, vol. 8, no. 2, 'Writing and
difference', 1981.
Janssen-Jurreit, Marielouise, Sexism: the Male Monopoly of History and
Thought (Pluto Press, 1982).
Jenkins, Mercilee and Cheris Kramarae, 'A Thief in the House: the Case
of Women and Language', Men's Studies Modified, ed. Spender
(Pergamon Press, 1981).
Jespersen, Otto, Language: Its Nature, Development and On'gin (Allen
& Unwin, 1922).
Jones, Deborah, 'Gossip: Notes on Women's Oral Culture', WSIQ, 3,
1980.
Kanfer, Stephen, 'Sispeak', Time, 23 Oct. 1972.
Kaplan, Cora, 'Language and Gender', Papers on Patn'archy (Women's
Publishing Collective/PDC, 1978).
184 Bibliography

Keenan, Elinor 0., 'Norm-Makers, Norm-Breakers: Uses of Speech by

Women in a Malagasy Community', in Explorations in the
Ethnography of Speaking, ed. Bauman and Sherzer (CUP, 1974).
Kramarae, Cheris, Women and Men Speaking (Newbury House, 1981).
Kristeva, Julia, 'Woman Can Never Be Defined' (interview with psych
et po), trans. August, New French Feminisms, ed. E. Marks and I. de
Courtivron (Harvester Press, 1981).
- - , 'Women's time', trans. Jardine and Blake, Feminist Theory: a
Critique of Ideology ed. Keohane, Rosaldo and Gelpi (Harvester
Press, 1982).
Labov, William, Sociolinguistic Patterns (University of Pennsylvania
Press, 1972).
Lacan, Jacques, Le Seminaire XX: Encore (Editions du Seuil, 1975).
Lakoff, Robin, Language and Woman's Place (Harper & Row, 1975).
Leonard, Diana, 'Male Feminists and Divided Women', On the
problem of Men, ed. Friedman and Sarah (Women's Press, 1982).
Lorde, Audre, The Cancer Journals (Spinsters Ink, 1980).
Lyons, John, Introduction to Theoretical Linguistics (CUP, 1968).
- - , (ed.), New Horizons in Linguistics (Penguin, 1970).
MacCabe, Colin, 'The Discursive and the Ideological in Film', Screen,
19 Apr. 1978.
Maltz, Daniel and Ruth Borker, 'A Cultural Approach to Male/Female
Miscommunication', in Language and Social Identity, ed. 1-
Gumperz (CUP, 1982).
Mandelbaum, D., Selected Writings of Edward Sapir (University of
California Press, 1949).
Marks, Elaine and Isabelle de Courtivron (eds), New French Feminisms
(Harvester Press, 1981).
Miller, Casey and Kate Swift, Words and Women: New Language in
New Times (Penguin, 1976).
- - , The Handbook of Non-Sexist Writing (Women's Press, 1980).
Mitchell, Juliet and Jacqueline Rose (eds). Feminine Sexuality: Lacan
and the Ecole Freudienne (Macmillan Press, 1982).
Moi, Toril, 'Who's Afraid of Virginia Woolf?, Feminist readings of
Woolf' (unpubl. 1982).
- - , 'Femininity, Language, Revolution: Julia Kristeva and Anglo-
American Feminist Linguistics' (unpublished lecture, 1983). (This
paper will form part of Toril Moi's forthcoming book in the New
Accents series published by Methuen on Feminist Criticism.)
O'Faolain, Julia and Lauro Martines, Not in God's Image (Fontana,
1974).
Ong, Walter J, Orality and Literacy: The Technologizing of the Word
(Methuen, 1982).
Orwell, George, 'Politics and the English Language,' Collected Essays
(Seeker & Warburg, 1961).
 Bibliography 185

Oxenham, John, Literacy: Writing, Reading & Social Organisation

(Routledge & Kegan Paul, 1980).
Pateman, Trevor, Language, Truth and Politics: Towards a Radical
Theory for Communication, 2nd edn Oean Stroud, 1980).
Pellowe, John, G. Nixon, B. Strang and V. McNeany, 'A Dynamic
Modelling of Lingusitic Variation: the Urban (Tyneside) Linguistic
Survey', Lingua, 30, 1972.
Rich, Adrienne, On Lies, Secrets and Silence: Selected Prose 1966-78
(Virago, 1980).
Saussure, Ferdinand de, Course in General Linguistics, trans. Baskin
(Fontana, 1974).
Scherer, Klaus and Howard Giles (eds), Social Markers in Speech (CUP,
1979).
Schulz, Muriel, 'The Semantic Derogation of Women', Language and
Sex: Difference and Dominance, ed. Thorne and Henley (Newbury
House, 1975).
Scruton, Roger, 'How Newspeak Leaves Us Naked', The Times, 1 Feb
1983.
Smith, Dorothy, 'A Peculiar Eclipsing! Women's Exclusion from Men's
Culture', WSIQ 1, 1978.
Smith, Philip M., 'Sex Markers in Speech', Sodal Markers in Speech,
ed. Scherer and Giles (CUP, 1979) ch. 6.
Spender, Dale, Man Made Language (Routledge & Kegan Paul, 1980).
Swift, Jonathan, 'A Proposal for Correcting the English Tongue', Prose
Works ofJonathan Swift, Vol. IV, ed. H. Davis (Blackwell, 1957).
Trudgill, Peter, 'Sex, Covert Prestige and Linguistic Change in the
Urban British English of Norwich', Language in Sodety, 1, 1972.
Zimmerman,. Don and Candace West, 'Sex Roles, Interruptions and
Silences in Conversation', in Language and Sex: Difference and
Dominance', ed. Thorne and Henley (Newbury House, 1975).
Glossary
ACCENT pronunCIatIon.
ACCOMMODATION unconscious process by which speaker
modifies her speech to be more like the person she is talking to.
ALIENATION in feminist linguistic theories, a term used for the
inability to express your own experience because no suitable
framework/vocabulary exists. Also a technical term in marxist
writing.
APHASIA speech impairment or loss caused by damage to the left
side of the brain.
BACK CHANNELING giving a speaker cues such as 'yes', 'mhm',
etc.
BEHAVIOURISM an approach to psychology that studies only
observable behaviour and not consciousness or mental states.
Associated with 'stimulus response' explanations of activity (i.e.
particular actions are responses to specific stimuli) and condi-
tioning techniques (i.e. training animals to perform tasks by
rewarding and punishing them). Behaviorism was popular with
later American structuralist linguists, and the leading figure In
opposing it was the linguist Chomsky.
BIOLOGISM the tendency to explain things by reducing them to
biological causes - e.g. 'anatomy is destiny'.
CODE SWITCHING changing from one language or variety to
another according to the subject or the situation of talk.
COMPETENCE Chomsky'S term for the knowledge of grammatical
rules the native speaker has internalised: opposed to PER-
FORMANCE, the use she makes of the rules.
COMPONENTIAL ANALYSIS approach to breaking down word-
meanings. Words are given plus or minus values on a number of
'primitive' semantic features such as 'animate', 'human', etc.
CONJOINING linguist's term for making complex sentences using
conjunctions like AND and BUT.
CORPUS the finite speech sample a linguist works on. General
linguists gave up corpus analysis after the 'Chomsky revolution' but
sociolinguists, discourse analysts, phoneticians and psycholinguists
still make use of corpora either 'natural' (surreptitiously recorded)
or (more often) deliberately elicited under laboratory or other
formal conditions.

186
 Glossary 187

COVERT PRESTIGE the value certain groups attach to non-

standard dialect features. Proposed by Trudgill as an explanation
of why men deviate more than women from standard norms.
DECONSTRUCTION a reading process associated with post-
structuralism in which the aim is to expose contradictions and
lacunae in a text.
DEMOGRAPHIC VARIABLE something that varies in the popu-
lation (i.e. from speaker to speaker according to social categories
like age, class, race and sex) and not according to the type and
formality of the situation.
DETERMINISM in linguistics, the belief that linguistic structures
are ultimately responsible for their users' conceptualisation of the
physical and intellectual universe. Associated with Sapir and
Whorf, and also latterly with Saussure.
DIACHRONIC historical.
DIALECT cluster of lexical, grammatical and phonological features
associated with a particular regional or social group.
DIALECTOLOGY literally, the study of dialects; more usually, the
historical analysis of rural language varieties, opposed to 'urban
dialectology' which is more usually called sociolinguistics.
DISCOURSE in linguistics, language 'above the level of the
sentence', i.e. exchanges in talk; in semiology, a set of related
utterances. (Linguists define discourse syntagmatically, whereas
for semiologists discourses are paradigmatic phenomena.)
EMBEDDING making complex sentences by using what are tra-
ditionally called subordinate clauses (e.g. that fish I bought
yesterday was off) where the italicised section i~ embedded.
EMPIRICISM is science, a methodology which tests hypotheses by
experiment or observation: characterised by the use of evidence to
support theory. In philosophy, the belief that there are no innate
ideas, and all knowledge derives from experience.
ESSENTIALISM the belief that your object of study has certain
basic qualities which do not need to be explained; they themselves
will do as an explanation of other qualities. An example is 'human
nature' as an explanation of human behaviour. One important
characteristic of most semiology is its implacable opposition to all
essentialist statements.
ETYMOLOGY the study of the history and derivation of words.
FERAL wild. Used of children who grow up outside human society.
FOLKLINGUISTICS linguist's term for the beliefs of non-linguists
about language.
FORMALISM a preoccupation with the way statements are ex-
pressed; often prescribes a standard notation or terminology.
GENDER in grammar, a property of nouns with requires adjectives
and pronouns to agree with it. In feminist theory generally, socially
188 Glossary

constructed male/femaleness as opposed to biological male/

femaleness which is called SEX.
GRAMMAR the practice of linguists; the model made by linguists;
the level of language between sound and meaning, i.e. form-
morphology and syntax; rules for correct usage (prescriptive
grammar).
GUTTURAL made at the back of the vocal tract. Linguists no
longer use this term, which has become a folklinguistic insult for
Germans, Geordies, etc. lacking its original technical meaning.
IDEALISATION modification of raw data (usually by ignoring some
of it and treating the rest selectively) to make it more manageable
or regular. In linguistics the point of idealisation is normally the
exclusion of extra-linguistic context and 'insignificant' variation.
INDETERMINACY inability to be stated definitively.
INFORMANT linguist's term for the person who produces her
corpus or takes part in her experiment.
INFORMATION THEORY study of message-transmission, ef-
ficiency and processing.
INTONATION use of pitch contrasts in speech: gives grammatical
and attitudinal information.
LANG UE Saussure's term for the linguistic system which makes pos-
sible individual speech acts (PAROLE).
LEXICOGRAPHY dictionary-making.
LOGOCENTRISM primacy of the (written) word; the term is
popular with post-structuralist followers of Jacques Derrida.
MARKING THEORY theory that linguistic elements form pairs or
sets that are hierarchically ordered, one being more neutral (and
thus frequent) than the other(s).
MATERIALISM the philosophy associated with marxism. Rejects
explanations based on ideas or supernatural concepts.
METALINGUISTIC about or beyond language. Used in this book
with broader reference than in most linguistics texts, which confine
the term to a particular set of words (i.e. those used by expert and
lay speakers to talk about language, like noun or meaning). I use it
also to describe practices and institutions that are parasitic on
language, such as grammar, dictionaries, etc.
METHODOLOGY the principles and techniques of scientific
research. Increasingly synonymous with METHOD but used in
scientific or pseudo-scientific registers.
MORPHOLOGY internal structure of words.
NON-STANDARD variety of language which lacks prestige or is
not officially sanctioned. Typically a non-standard has no writing
system, is not widely diffused and is spoken by the least educated
and poorest members of a society.
 Glossary 189

PEJORATION word used for the process by which something (e.g. a

word) comes to have negative connotations.
PERFORMANCE language use: see COMPETENCE.
PHILOLOGY the historical study of linguistic texts.
PHONETICS study of speech sounds.
PHONOLOGY study of sound systems, i.e. how speech sounds are
deployed meaningfully in languages.
POSITIVISM philosophy of science dominant in 19th and 20th
centuries.
POST-STRUCTURALISM development in philosophy and criti-
cism associated with Derrida, so-called because it came after the
structuralism of theorists like Levi-Strauss and Barthes. The
'technique' of post-structuralism is deconstruction.
PRESCRIPTIVISM the use of value-judgements and notions of
correctness in grammar.
PROSODY in speech, intonation, stress and rhythm; in verse, metre.
REFERENT what in the real world a word refers to.
REGISTER variety of language appropriate in content, medium
(speech/writing/whistling/drumming/signing), style and tone to
its particular ___domain of use.
REIFICATION treating a concept as a thing.
RELATIVITY the notion that different cultures classify reality dif-
ferently (therefore that it is effectively a different reality depending
on who's interpreting it). Linguistic relativity is the theory associ-
ated with Sapir and (especially) Whorf (see DETERMINISM).
SEMANTICS study of meaning.
SEMIOLOGY science of signs based on Saussurean principles. Used
particularly for 'theory of the subject', i.e. Lacanian influenced
theory, but also for structuralist and post-structuralist film and
literary criticism, etc.
SIGN the unit of language according to Saussure, consisting of a
signifier (form) and a signified (concept) indissolubly joined
together, and deriving its significance from a contrast with other
such signs rather from its own substance, which is arbitrary.
SOCIAL MARKER linguistic feature associated with a particular
demographic or situational variable.
SOCIOBIOLOGY also called 'selfish gene theory': a biologistic
explanation of human behaviour and human social structures.
SPEECH COMMUNITY an idealisation; the term refers to any
group of people sharing the same linguistic norms, but isolating
such a group in practice is quite impossible.
STANDARD the officially sanctioned or most prestigious language
variety within a state.
STEREOTYPE in linguistics, a folklinguistic characterisation of
190 Glossary

some group's speech.

STRATI FICATION particular statistical pattern; in sociolinguistics
it is produced by dividing a sample according to socioeconomic
class and quantifying each group's use of certain stereotypical
linguistic features.
STRUCTURALISM methodology based on Saussurean principle
that systems are sets of differences. In Linguistics, refers to the
American movement of Bloomfield, Hockett, et ai.; in general,
refers to the French theorists Levi-Strauss, Barthes, etc.
STYLE variety of language associated with situational features,
particularly formality as perceived by the participants.
STYLE SHIFT moving from one style to another as you perceive
increased/ decreased formality.
STYLISTICS the study of style and register, and especially of literary
uses of language.
SYNCHRONIC not historical.
SYNT AGMATIC relationship of elements that can be combined.
Opposed to PARADIGMATIC or ASSOCIATIVE relationships,
in which the related elements can substitute for one another in the
same context.
SYNT AX the study of sentence structure.
TRANSFORMATIONAL GRAMMAR Chomsky's theory of
language. The name derives from a new sort of rule (a
'transformation') which was central to the model. TG IS
distinguished by its formalism, mentalism and commitment to
universals.
UNIVERSAL in linguistics, a feature or tendency common to all
languages.
VARIATION the heterogeneity typical in linguistic data.
VARIETY neutral term for any kind of language (i.e. dialect, style,
register, etc.).
VERNACULAR historically, a mother-tongue (as opposed to
learned languages such as Latin, which were the vehicle of culture
and scholarship but had to be learned at school, after first-
language acquisition). In sociolinguistics, the style furthest from
prestige norms, used in casual situations when monitoring is
minimal and therefore of great theoretical interest to the linguist.
Index
alienation, 6, 93, 133, 141, 144, totalitarianism and, 100, 101
186(glossary) Courtivron, Isabelle de, 9, 102, 135
Althusser, Louis, 116 covert presige, 48-9, 187(glossary)
anthropology Coward, Rosalind, 9, 69, 114,
feminist theory and, 26 115-17,119,123, 133, 152
linguistics and, 22-3 .crosstalk, 150
see also dominant/muted theory between women and men, 151
Ardener, Edwin, 92, 103, 104, 133 Crouch, Isobel, 34, 55, 66, 67
Ardener, Shirley, 92, 103, 105, 106,
114, 133, 157, 167 Daly, Mary, I, 6, 80, 92, 166
Dalston Study Group, 135, 153, 161
Bales, R.F., 41 deconstruction, 140, 187(glossary)
Barthes, Roland, 18, 152 Delphy, Christine
Beauvoir, Simone de, 57, 138, 139, on psychoanalysis, 130
141, 172 on radical feminism, 116, 178n3
Bernstein, Basil, 36, 158, 160 on stratification studies, 50-1
binary oppositions, 57, 58-62, 122 dictionaries, 75, 81, 83-4
Black English, 45, 46, 51, 78, 135, discourse, 69-70, 87, 151-3,
158-9. 161. 170 187(glossary)
Black, Maria, 61, 115-17, 133, 152 dominant/muted theory, 102-8
Bloomfield, Leonard, 190 determinism in, 103-4
Bodine, Ann, 68 developments of, 106-8
Borker, Ruth, 151, 167 Dubois, Betty Lou, 34, 55, 66, 67
Dworkin, Andrea, 77-8
Carib,47
castration complex, 120, 121, 131 Elgin, Suzette Haden, 162, 167
Chomsky, Noam, 20-2, 99 Ellis, John, 7, 114-15, 119, 123
chora, 125, 126, 128 Elshtain, Jean Bethke, 162, 173
Cixous, Helene, 57, 62, 113, 114,
124, 168 femininity, 153, 156
componential analysis, 59-62, conservatism and, 49
186(glossary) discourse and, 69
consciousness· raising, 5 prestige pronunciation and, 49
control, 93, 100-2 psychoanalytic theory of, 117, 125
limits of, 143, 157 feminism
men as agents of, 104, 110-12, America and Britain, 16
133, 143 France. 16
metalinguistic practices and, psychoanalysis and, 16, 117-18,
144-56 121, 125

191
192 Index

radical tendency in, 116 jacobus, Mary, 130

folklinguistics, 31, 187(glossary) Janssen·Jurreit, Marielouise, 57, 64,
beliefs about sex-differences, 65
31-44, 155 jenkins, Mercilee, 29, 46, 153
empirical observation of, 34-44 Jespersen, Otto, 29, 31, 33, 36, 37,
feminism and, 33, 43 47, 155, 160
Foucault, Michel, 116, 152
Freud, Sigmund, 16, 114-15, Kaplan, Cora, 4, 146, 154
117-18, 120, 130, 151 Kramarae, Cheris, 29, 40-1, 46, 52,
see also psychoanalysis 103, 106-7, 153, 162, 167,
Friedan, Betty, 5 168
Kristeva, Julia, 1, 18, 125-7, 132,
Gallop, jane, 58 133, 165, 166, 168
gender (as linguistic category),
62-70, 84, 187-8(glossary) Labov, William, 31, 43, 49, 51, 159
in Bantu languages, 63 Lacan, 17, 18, 19-20,92, 117,
in English, 63, 66-8: English 119-21, 122-5, 132
common gender, 69-70 Irigaray's critique, 127-9
in French, 65, 67, 69: French see also psychoanalysis
common gender, 69 Lakoff, Robin, 33-4, 52, 55, 168
in German, 63, 65 language
in Konkani, 64 and the body, 129-30, 1,43
gender role, 29, 40, 52, 53, 167 definitions of, 136-7, 165-6
gendered subject, 25, 114-33 passim and disadvantage, 170
generic pronouns, 66-8, 82, 84-5 as feminist issue, 2-3
Giles, Howard, 45 and oppression of women, 91, 167,
grammar, 44, 56, 58, 66, 72, 170-1
188(glossary) and reality, 7, 91-113, 164-5,
before linguistic science, 10 169-70
ideological function of, 61, 70-1 and revolution, 125-9
in linguistics, 11, 15 as sociological resource, 72-90
prescriptive, 11-12, 68, 70-1, 82 passim
transformational, 23, 190(glossary) as a weapon, 1, 100-1
Grice, H.P., 39 see also non-sexist language,
Griffiths, Ian, 96 sexist language, women's
Grimm, Jakob, 64-5 language
Gugenheim, Camilla, 5-6, 141 league/parole, 13, 14, 23, 98,
Gumperz, John, 149 189(glossary)
Leonard, Diana, 132
Hall, John, 86-7 linguistic determinism, 93-5,
Harris, Roy, 134, 138-40 187(glossary)
Henton, Caroline, 52-3 arguments against, 96, 98-100,
Hockett, Charles, 190 143, 169-70
as axiom in feminist theory, 93,
insults, 73, 76-8, 153 133
integrationallinguistics, 139-40 in dominant/muted theory,
Irigaray, Luce, 4, 57, 114, 122, 103-4, 107
125, 127-9, 133, 165, 166, in Lacan, 95-6
168 in Sapir-Whorf hypothesis, 96-8:
 Index 193

differences between Whorl 87, 137

and Saussure, 97-8 Moi, Toril, 118, 126
in Saussure, 95
in Spender, 108-10, 112 naming, 5, 80-1
see also Sapir- Whorf hypothesis non-sexist language, 4, 73, 84-8,
linguistic reform 89-90, 171
attitudes of linguists, 66, 79 alternatives to, 88-9
attitudes of speakers, 87-8 myth of, 85-6
desirability of, 74-5, 87
feminist strategies for, 76-81: new Ong, Walter J., 148
coinings, 75; 'reclaiming', 75, Orwell, George, I, 73, 91, 107
78-80, 89 Oxenham, John, 147
obstacles to, 78, 82-4
linguistic relativity, 22, 189(glossary) Pateman, Trevor, 145, 172
see also Sapir-Whorl hypothesis patriarchy, 3, 78, 108, 173
linguistics conference on, 135
anthropology and, 22-3 P~cheux, Michel, 152
psychoanalysis and, 19-20, Pellowe, John, 44
117-21, 122-9 'person', 82, 89
psychology and, 21 pornography, 77
scientific status of, 10, 16 prescriptivism, 11, 64-8, 70-1, 72,
sexism of, 2, 28-71 passim 82, 137, 189(glossary)
see also sociolinguistics prestige, 48, 50
linguistic universals, 21-2, 24, 67, covert, 48, 49
99, 190(glossary) femininity and, 49
literacy, 147 psychoanalysis, 114-33 passim
see also writing feminism and, 16, 117-18, 121,
Lorde, Audre, 134, 141 125
Lyons, John, 64, 99 language and, 19-20, 122-5,
131-2
McCabe, Colin, 152 socialisation and, 131
Malagasy, 49, 50, 155 see also Freud, lrigaray, Kristeva,
Maltz, Daniel, 151, 167 Lacan
marked/unmarked (marking theory)
66-8, 188(glossary) Questions jeministes, 130
Marks, Elqine, 9, 102, 135
Marx, Karl, 114-15 register, 146, 148, 149, 151, 152,
masculinity 153, 189(glossary)
discourse and, 69 see also discourse
non-standard pronunciation and, Rich, Adrienne, 80, 92, 113
48-9 Rose, Jacqueline, 122
media Rousseau,1-J, 156
non-sexist language used in, 73, 82
pejoration of feminist terms by, 82 Sapir, Edward, 22, 29, 97
representation of women in, 4 see also Sapir-Whorl hypothesis
metalinguistic practices, 144-56, Sapir-Whorl hypothesis, 22, 96-100,
164, 173, 188(glossary) 109, 112-13
see also, folklinguistics, discourse see also linguistic determinism
Miller, Casey, 2, 72, 74, 75, 84, 86, Saussure, Ferdinand de
194 Index

critique of, 128-9, 131-41, 148 stereotype, 44, 189(glossary)

determinism of, 91, 95, 97 see also folklinguistics, women's
Lacan and, 20, 95-6, 123-4 language
linguistics of, 10, 13-15, 58 stratification, 44, 50-1
semiology of, 14, 16, 114 structuralism, 190(glossary)
Scherer, Klaus, 45 American, 15
Schulz, Muriel, 76 as method, 18
Scruton, Roger, 88, 101, 102 in relation to semiology, 17-18
semantic change, 82-6 see also semiology
derogation of women, 76-7: subject
semantic non-equivalence, 77 decentring of, 11 7, 118
semiology, 4, 9, 16-20, 114-33, development of, 119
189(glossary) Swift, Jonathan, 32, 53
contribution to feminist language Swift, Kate, 2, 72, 74, 75, 84, 86,
study, 25 87, 137
sex-differences, 6, 28-56 symbolic order, 118, 119, 122-5
definition of, 30-1 differential entry of female and
explanation of, 46-52 male subjects, 119-20, 126
folklinguistic representation of, marginality of women in, 122,
31-44 126, 134, 142
general study of, 28 phallocentricity of, 120-1, 122,
sociolinguistic findings on, 48 124
value-judgements on, 50, 53-6
sexist language, 6, 72-90, 91 telementation, 138-9
as cause of oppression, 75 Trudgill, Peter, 49, 50
as symptom of oppression, 74
see also linguistic reform, non- vernacular, 51, 190(glossary)
sexist language violence against women, 3, 77-8,
sexuality, 118 116
and language, 127, 129-30, 167-8 language as, 77, 78, 183
silence, 5, 93
Smith, Dorothy, 110 WAVAW (Women Against Violence
Smith, Philip, 2, 54, 154, 167 Against Women) 1981
social markers, 45, 189(glossary) conference, 3
in relation to sex-differences, 47 West, Candace, 106
sociolinguistics, 6, 9, 23-4 Whorf, Benjamin Lee, 22, 24, 91,
feminist applications of, 26 97-8
feminist critique of, 44-52: see also Sapir-Whorf hypothesis
explanation, 46-50; women's language
methodology, 50-2; in authenticity of, 133, 134
theoretical basis, 44-6 indicating linguistic insecurity, 49
see also linguistics influence of on linguistic change,
Spender, Dale, 103, 133, 152, 157, 49
162 as restricted code, 160
on componential analysis, 60 restrictions on: in public, 105,
on Lakoff hypothesis, 55 106-7, 146, 154-5; in
Man Made Language theory, 3, 5, feminist groups, 107; in
60, 92, 108-13, 134, 157 patriarchy, 128-9
Whorfian influence on, 23, 109 stereotypes of, 32-3
 Index 195

value-judgements on, 50, 53-6 see also literacy

writing
and the body, 168 Zimmerman, Don, 106
and speech, 165-6