141

Does ‘Women’s Language’ Really Exist?:

A Critical Assessment of Sex Difference Research
in Sociolinguistics

Yuki Mizokami

1. Introduction
It is often claimed that in many societies there are negative
stereotypes relating to women and language. Here are two examples of
proverbs which are evidence of this.
A woman’s tongue wags like a lamb’s tail. (English proverb)
Onna sannin yoreba kashimashii. (Japanese proverb)
(= If three women gather together, it becomes clamorous.)

These proverbs imply that many people stereotypically think women are
more talkative than men, or moreover that women talk too much. Yet
there has not been one study which supports this stereotype: Studies in
laboratories, classrooms, meetings and television chat shows indicate that
women in fact talk less than men do in mixed-sex conversations.1 This
suggests that the false stereotype of talkative women is not a reflection of
reality but just a reflection of the social expectation--‘women should be
silent’--because silence has usually been considered as synonymous with
obedience in many societies.
There are other popular folklinguistic assertions of ‘women’s
language’. For example, stereotypes such as ‘women talk more politely’
and ‘women use more questions’ have been firmly believed by the public.
We cannot override these stereotypes, since unlike the former case, these
stereotypes have actually been supported through extensive linguistic
research.2
According to such research, ‘women’s language’ has the following
142 Yuki Mizokami

features: First, female speakers interrupt less than males in mixed-sex

conversations.3 Second, female speakers use more indirect speech than
males.4 Third, female speakers use more conversational support such as
‘minimal responses’ than males.5 Fourth, female speakers use features
which indicate tentativeness, such as ‘tag questions’, ‘hedges’ and other
expressions that make them sound hesitant or uncertain.6 These findings
have given rise to the idea that women speak in a cooperative way or a
polite way, whereas men speak in a competitive way.7 Yet do all women
stereotypically speak this way? Does distinct ‘women’s language’ really
exist? It seems odd that few researchers have doubted the existence of
‘women’s language’. As we shall reveal later, one of the biggest
problems in language research of sex differences is the confusion between
how women actually speak and how people think women should speak.
Since the 1970’s, as feminism developed, linguists have conducted
thorough research on differences between the sexes and they have tried to
determine the features that ‘women’s language’ contains. In
sociolinguistics, a speaker’s sex has been treated as an independent
variable. Moreover, linguistic research into sex differences, conducted
under the name of ‘science’, has regarded men as the norm and women as
a deviation in the natural premise. In other words, most of the linguistic
research into sex differences has emphasised, empirically or not, how
different women’s language use is from men’s language which is
considered to be the norm. It will be presumed that this female/male
division in research has resulted in reinforcing the female/male dichotomy
in society. In this paper, we would like to prove that alleged women’s
language is not a product of proper research but is a product of the
‘androcentric’ ideology which works to suppress women. Our aim is not
to state that women speak exactly the same way as men but it is to show
how stereotypes of women’s speech in sociolinguistics are more unfairly
exaggerated in reality.
It is important to think about why researchers have chosen to study
sex differences rather than similarities, and why they have interpreted their
 Women’s Language 143

research findings in sexist ways. We need research on research. We

shall discuss how studies on sex differences in language use are affected
by the androcentric ideology and how ‘women’s language’ is orchestrated
or even fabricated through such research. The danger of this
androcentric ideology is that it has affected even feminist researchers. In
fact, some of what feminist linguists have done can be seen as just
reinterpreting the stereotypes. Such descriptions of women’s language
articulated by linguists, who are ‘scientific’ authorities supposedly, have a
risk of becoming prescriptions of female speech. These prescriptions, or
static stereotypes of women’s language, appear to be a structure which
enhances the polarisation between the sexes. It is therefore plausible that
in future language and sex difference studies the central concern should be
to go beyond the simple female/male dichotomy.

2. Can pure ‘objectivity’ be possible?

Linguistics is usually defined as the ‘scientific’ study of language.
The procedures of sociolinguistic research as a science generally have
these four phases. First is the formulation of a hypothesis. Second is
data sampling. Third is the interpretation of that data, and the last phase
is an explanation of those results. If research is proven scientifically,
then it must be ‘objective’ in all of these phases.
At this time, it is necessary to point out that there is, “a danger of
seeing what you want to see (Swann, 1992, P.198)” in any research. It is
reported that even people who see the same thing from the same vantage
point interpret what they see differently (Chalmers, 1978). This implies
that the interpretation depends on the observer’s situation and expectations,
therefore it can be said that the result of research is virtually determined
by the way hypotheses are formed. If this is the case, then it is accurate
to speculate that if researchers wish to verify stereotypes in women’s
speech, they could interpret empirical data in whatever way meets their
expectations. Any research findings would unconsciously reflect the
researcher’s own ideologies. In this respect, ‘objectivity’ in linguistic
144 Yuki Mizokami

research may no longer make sense. We must take into account that any
‘truth’ constructed through research is a social product that cannot be
divorced from the biases a researcher starts with. The stereotypes of
‘women’s language’ may seem truthless once we subvert the
taken-for-granted perspective.

3. Early works on women’s language

Jespersen’s work, Language: Its Nature, Development and Origin
(1922) has attracted our attention as one of the rare works which looked at
women’s language in earlier times. In the chapter entitled ‘The Woman’,
Jespersen confidently asserts that, for example: Women speak more
politely than men, and have smaller and less varied vocabularies.
Women differ from men in their use of certain adjectives such as ‘pretty’
and ‘nice’, and adverbs such as ‘vastly’ and ‘so’. His claim could be
interpreted as meaning that women use a different language than men.
Note that his work is a typical example which treats one group (men) as
the norm and the other group (women) as deviant.
Feminists have criticised his work as stereotypical and sexist.
While dogmatizing on women’s characteristics of speech, Jespersen did
not collect his own data from actual conversations of women: he instead
refers to women’s dialogues in novels written by men. Feminists have
argued that because of Jespersen’s overt prejudice against women and the
insufficiency of his method, alleged women’s language in his work should
be considered as a product of the androcentric ideology. However, the
problem is not only that Jespersen’s work is stereotypical, but that these
stereotypes have been swallowed, and perhaps are still believed by the
public. Generally the problem of stereotyping is that stereotypes, which
may or may not be close to reality, tend to be justified by the fact that
many people believe them, and consequently the social prejudices hidden
behind them also tend to be justified.
Ironically, even feminist linguists have not always probed into
folklinguistic stereotypes, whilst criticising Jespersen’s work. It is
 Women’s Language 145

generally said that Lakoff’s work, Language and Women’s Place (1975)
marked the beginning of the twentieth-century linguistic interest in sex
differences. Her claim about female speech is known as the ‘Lakoff
hypothesis’: For example, Lakoff asserts that women are more likely to
use empty adjectives such as ‘divine’, ‘charming’ and ‘lovely’.
Intensifiers such as ‘so’, ‘really’ and ‘very’, and qualifiers such as ‘not
exactly’ and ‘a bit’, are more frequently spoken by women than men.
Women use more tag questions, more hedges, more rising intonations and
more polite forms than men use. Lakoff explains that these
characteristics of ‘women’s language’ are a result of linguistic
subordination: A woman must learn to speak ‘women’s language’ to avoid
being criticised as unfeminine by society. As a result, women appear to
lack authority, seriousness, conviction and confidence in their
conversation.
It is important to point out that there are some parallels between
Lakoff’s work as a feminist and Jespersen’s work as a traditional linguist.
Both of them develop exactly the same argument, for instance, on the use
of certain adverbs and adjectives in women’s speech. Lakoff’s work has
also been attacked by feminists as stereotypical and androcentric in two
points: First, she tends to use the ‘men as the norm and women as a
deviation’ framework as if it were taken for granted. She seems to
assume the existence of women’s language which is inferior to and
different from men’s language. In short, she is biased against women’s
language from the start. Second, there is a problem with her method.
She examines her own intuitions rather than a collective corpus of data.
Using this method may direct the researcher to merely describe her own
biases. Thus Lakoff’s work should also be regarded as a product of the
androcentric ideology.
Both Jespersen and Lakoff claim to describe how women converse.
However what they actually do is merely itemise folklinguistic stereotypes
which unconsciously reflect the general public’s idea about how women
should speak. Lakoff herself, unknowingly conveys to us the realisation
146 Yuki Mizokami

that sexist stereotypes are deeply-rooted even in feminists’ minds.

Nevertheless, we should not abandon Lakoff’s work as just
stereotypical. While considering her work as a product of the
androcentric ideology, Lakoff’s study is significant in two ways: (1) She
is the very first feminist researcher to deal with women’s language; (2)
Later studies on sex differences in English language have developed
through arguments based on Lakoff’s work. In other words, later works
have attempted to verify or falsify the ‘Lakoff hypothesis’.

4. Works after Lakoff

4.1. The dominance approach and the difference approach
After Lakoff, many researchers have attempted to verify stereotypes
of women’s language. Their explanations of linguistic sex differences
can be divided into two conflicting currents; “the dominance approach”
and “the difference approach”. The dominance approach sees women
as an oppressed and marginalised group, and interprets linguistic
differences in women’s and men’s speech as a reflection of men’s
dominance and women’s subordination. Lakoff’s study, which asserts
that women’s language is trivial, is categorised as a radical version of this
approach. Other typical examples of this approach are the study by
Fishman (1980) which supports the ‘Lakoff hypothesis’, and that by
Zimmerman & West (1975). After transcribing three married couples’
taped conversations, Fishman (1980) reports that the women ask questions
of any kind two and a half times more than the men. She also finds that
the women in these conversations use hedgings five times more frequently
than the men. She explains that women use questions and hedges to
facilitate conversations, because after all, men have the upper hand in any
conversation. The conversational strategies by these women are seen as
a reflection of their inferior social position. Zimmerman & West (1975)
empirically prove that women use less interruptions than men in
mixed-sex conversations. They argue that women are more likely to
allow interruptions indicating a submission to men. Researchers of this
 Women’s Language 147

approach explain their findings in terms of sexual inequalities and a

woman’s inferiority.
On the other hand, the difference approach, emerging later than the
dominance approach, emphasises the idea that women and men belong to
separate subcultures to which no pejorative value should be attached.
The sex differences in language use are interpreted as a reflection of these
distinct subcultures. Research by Maltz & Borker (1982), Jones (1980)
and Coates (1989) is categorised in this approach. Malts & Borker
(1982) argue that the features in women’s speech reflect not so much the
point of a power imbalance between women and men but that the sexes
have different norms for conversational interaction. Boys are likely to
play in large hierarchical groups in which they learn a competitive style of
speech, whereas girls tend to play in small ‘best friends’ groups in which
they learn a supportive speech style. Likewise Jones (1980) claims, in
her observation of women’s gossip, that cooperative speech strategies
arise from valuable female subcultures. Coates’ (1989) study on
women’s gossip also highlights the strengths of cooperative conversation.
Researchers of this approach explain their findings in terms of subcultural
differences between the sexes by objectively putting the same value on
either sex.

4.2. Some problems in both approaches

While we could admit value of both approaches, we could also see
androcentric traps for feminists in both options. If one takes the
‘dominance’ approach, there is a risk of degrading women as passive
victims whose cultural forms have no positive attributes. This approach
suggests that if a woman wants to gain social power, she should speak
more like a man. This implies ‘biologism’ - women are inferior because
of their sex. The problem with this approach is that it seems to accept
women’s subordinate position.
By contrast, if one opts for the ‘difference’ approach, she seems to
ignore the political questions of actual male dominance in societies whilst
148 Yuki Mizokami

posturing objectivity. This approach basically asserts that women and

men are JUST different and that is all; This could after all result in the
justification of existing sexual inequalities. This approach is more likely
to be trapped in ‘essentialism’. This option may direct researchers to set
up oppressive norms of ‘authentic’ femininity. In the end, feminists are
in a dilemma that neither approach is politically advantageous for women.
Yet the important fact we must recognise is that women and men do
not speak separate languages. Most of the works following Lakoff are
empirical studies, which generally implies that the findings are ‘scientific’.
However the problem in both approaches, first of all, lies in their
assumption: these works have been based on the ‘Lakoff hypothesis’
which is itself unjust against women. That is to say, the majority of
researchers of both the dominance and difference approaches
automatically hypothesise the same thing -- ‘there is ‘women’s language’
which is different from men’s language, and all women talk in the same
way’.
Moreover, another problem is the lack of a clear theoretical and
practical framework. Empirical works conducted so far among various
researchers are somewhat confused in terms of data amounts, research
methodology and interpretation of data. As a result, they explain their
findings of sex differences in language, however small or vague they are,
as either a reflection of male dominance and female subordination, or as
sub-cultural differences between the sexes. It is no wonder that these
researchers have verified the existence of distinct women’s language
rather than the nonexistence of it, because their findings are almost fixed
by the way they formulate their hypothesis based on Lakoff’s. It seems
that the elaboration of a more refined framework is a matter of urgency.
Should we accept without reserve the existence of alleged ‘women’s
language’ which these researchers have found through their stereotypical
hypothesis and inconsistent methods? Or, are there other ways of
interpreting data rather than correlating them with sex differences? As
we will discuss, potential sociolinguistic markers of sex might be markers
 Women’s Language 149

of other social and interactional variables at the same time, and may be
more directly related to the latter.
Sex-typed linguistic features in English have been said to be of
relative nature -- there are no sex-exclusive differences.8 Therefore
sociolinguists have aimed to quantify differences in usage of certain
linguistic features between the sexes. In so doing, researchers have
simply counted the frequency of, for example, tag questions and
interruptions in their empirical research without taking into account the
speakers’ social status, the age of the participants, or the context of the
conversation. By ‘context’ here, we mean not only the time and place
but the structure and function of a communicative event and the
relationships between its members. Note that the method of
sociolinguistics is to compare one ‘norm’ category with another category.
The ‘norm’ category has been that of white middle-class males. So the
frequency or non-frequency of a certain linguistic behaviour in female
speech has been taken to signify negativeness.
We would like to raise two questions here. One is whether the sex
of the speaker should be categorical or not. A speaker’s sex has been
perceived as an important social category that affects how we speak,
thereby researchers have tried to demonstrate clear-cut sex differences in
language use. However, the sex of the speaker is only one of a number
of social groupings supposedly related to language use. The way
someone speaks on any one occasion in fact depends on several factors
other than sex. A woman may speak differently when talking to her
mother about clothes than when discussing a problem with classmates in a
university seminar. Moreover, women and men are not homogenised
groups as there are variations within the sexes such as the middle class,
working class, aged, young, white, coloured etc. As Poynton (1985)
argues, the statistical techniques generally employed in linguistic research
so far have focused on the data of groups as wholes and allowed
researchers to ignore the range of variation within groups, which
commonly informs us that females and males are more alike than they are
150 Yuki Mizokami

different.
The other question is whether there is a static one-to-one
relationship between certain linguistic strategies (e.g. tag questions) and
specific communicative functions (e.g. tentativeness). Many researchers
have used mechanical definitions to identify particular linguistic marks.
However, specific linguistic strategies actually have extremely diverse
meanings. We must realise that defining some linguistic features is itself
an interpretative act for researchers and that the interpretation might be
different depending on the researchers’ judgements. Language and sex
difference research should not be approached as an automatic search for
specific linguistic characters. Taking these aspects into consideration, we
will start examining the validity of some popular research findings in
women’s language.

5. A critical assessment of some sex difference research

5.1. Interruption
Women are believed to interrupt less than men do, because men are
thought to dominate the conversation whereas women are considered as
cooperative speakers. This stereotype has been supported by such
researchers as Zimmerman & West (1975). They investigate casual
mixed-sex conversations on a campus and report that 96 percent of the
interruptions are produced by men. They regard interruption as a device
for exercising power and control in conversation, and a violation of a
speaker’s turn to talk. If a second speaker starts speaking at what is
considered a transition-relevance place, they define it as an ‘overlap’
which is neutral. If a second speaker starts speaking at a point which is
not considered a transition-relevance place, they define it as an
‘interruption’ which is negative.
However, there is an ambiguity in the definition of an interruption.
We could criticise that researchers who report ‘men interrupt more than
women’ use mechanical definitions to identify interruptions. There is in
fact no absolute framework for recognising an occurrence of interruption,
 Women’s Language 151

since a speaker’s ‘completion rights’ depend on several factors such as

length and frequency of speech, which are always negotiated among the
interactants in each context. In short, to claim that one has ‘observed’ an
interruption is making a subjective judgement. The following example
which West & Zimmerman (1983) interpret as an interruption by the male
speaker may be in fact interpreted as an overlap:
Female: So uh you really can’t bitch when you’ve got all those
on the same day (4.2) but I uh asked my physics professor
if I couldn’t chan[ge that]
Male: [Don’t ] touch that
(1.2)
Female: What?
(#)
Male: I’ve got everything jus’how I want in that notebook ( # )
You’ll screw it up leafin’ through it like that.
(West & Zimmerman, 1983, p.105)9

In this case, we could argue that if the male feels the female’s handling of
his notebook is destroying his organisation of it, he surely has a right to
ask her to stop immediately, without waiting for a transition-relevance
place (Tannen, 1994). Furthermore, whether an overlapping in
conversation is ‘overlap’ or ‘interruption’ cannot be determined merely by
a researchers’ judgement. Murray (1985) shows that the following
example is judged as an interruption by half of the women he investigated,
but not by the other half:
H: I think [that
W: [Do you want some more salad?

What we should recognise here is by using a ‘simple counting’

method of linguistic features, it could direct researchers to a one-sided
interpretation. Instead researchers must take account of the context of
the conversation which even includes the participants’ intention.10 The
interpretation should depend on the situation, setting and the relationship
between the speaker and listener, since the same linguistic strategy
152 Yuki Mizokami

actually has a different interactional purpose and function in different

contexts.
Suppose one, as a teacher, wears a T-shirt and jeans. If she dresses
like that among students, it would signify solidarity with the students.
On the other hand, the same T-shirt and jeans worn in a formal faculty
meeting would connote disrespect for her colleagues. This example
symbolises that the same symbols -- a T-shirt and jeans -- can signal
various meanings depending on the context.
This example may be analogous to the ambiguity of linguistic
behaviour. Researchers should abandon the assumption that there is a
fixed one-to-one relationship between a particular linguistic strategy and
its function. Tannen (1994) demonstrates that some overlappings in fact
have supportive or cooperative effects rather than obstructive effects.
Coates (1989) shows that overlappings are used as signals of active
listenership in some contexts. Thus everything that occurs in a
conversation results from the interaction of all participants. The meaning
of any linguistic behaviour is negotiated and renewed in each
conversational context. We can then speculate that if Zimmerman &
West had closely looked at the context, they might not have found a
‘males interrupt females more’ paradigm. Women might not be as
cooperative or powerless as they claim they are. In fact, Beattie (1981)
reveals that female students interrupt more than male students in his study
of university tutorials by examining the conversational contexts.
Therefore so far as the stereotype ‘women use less interruptions because
of their inferior position’ is concerned, it can be challenged as above.
This stereotype should be considered not as a scientific truth but as an
androcentric product which tries to confine women to a subordinate
position in conversation.

5.2. Tag questions

From Lakoff onwards, tag questions are assumed to be an important
character of ‘women’s language’ which seems to indicate that women lack
 Women’s Language 153

assertiveness:
(a) The economic crisis in Japan is terrible.
(b) The economic crisis in Japan is terrible, isn’t it?

Siegler & Siegler (1976) asked college students to guess whether women
or men produce the above sorts of sentences. Their result supports the
stereotype: a sentence like (b) is more often attributed to women, whereas
a sentence like (a) is more often attributed to men. This research reflects
what people’s perceptions and expectations are. It does not, however,
reveal that women in reality use more tags than men.
In fact, the stereotype ‘women use more tag questions’ has been
empirically proven to be true by researchers such as Fishman (1980) and
Preisler (1986). However, the problem is that many researchers who
study the use of tag questions11 have assumed that tag questions have only
one function, namely tentativeness which manifests a women’s inferior
social position. On the contrary, as we will see, tag questions actually
represent multifunctionality and a diversity of meaning.
Holmes (1984) argues that tags can express either ‘modal’ or
‘affective’ meanings depending on the situation. According to her
classification, the role of ‘modal’ tags is to confirm information of which
a speaker is uncertain; e.g. “You were missing yesterday, weren’t you?”
‘Affective’ tags have two roles: One is a ‘softening’ tag which is used to
indicate concern for the addressee to mitigate a face-threatening act; e.g.
“Shut the window, could you?” The other is a ‘facilitative’ tag which is
used to offer the addressee a chance to go into the conversation; e.g. “This
is a nice car, isn’t it?” Note that Lakoff (1975) only refers to the
‘facilitative’ tags. After counting the distribution of tags in a corpus by
using the above classification, Holmes (1984) reports that women and men
do not notably differ in the total usage of tags, although men are more
likely to use modal tags whereas women use more facilitative tags. More
importantly, she finds that conversational facilitators are more likely to
use tag questions. By facilitator, she means those who are responsible
154 Yuki Mizokami

for creating a smooth interaction such as leaders, teachers and hosts.

Then it can be said that the use of tags does not always depend on the
speaker’s sex as has been posited, but the speaker’s role in the
conversation.
Furthermore, Cameron et al. (1989) find that powerless participants
in unequal encounters never use affective tags irrespective of their sex.
They suggest that in an interaction, tags overall are more frequently used
by powerful participants in terms of social class, age and occupation.
Their findings contradict Lakoff’s hypothesis. These studies show how
misleading the ‘women use more tag questions because of their lack of
confidence and of their weakness’ stereotype is.

6. Some remarks on Japanese sociolinguistics

Japanese sociolinguistic research into sex differences seems to have
the same problems as English sociolinguistics: Researchers have tended to
reveal cultural stereotypes by confusion between the socially expected
speech of women and the actual speech of women.
The ‘women speak more politely’ stereotype is one of the most
popular research findings as we have referred to earlier. For example,
Hori (1986) empirically proves that women use more honorifics than men.
She explains that women cannot safely use forms which lack politeness
because nobody except children are clearly ranked below the housewife in
social status. However, we could criticise this because it is an inevitable
result brought forth by her research methodology. First, she uses the
‘questionnaire’ method. She asks her informants questions such as:
“Write a variant of ‘iku’ you usually use when asking ‘When do you go?’
(itsu-iku-ka?) to the different people you meet”. The use of this method
may lead a researcher to obtain stereotypical results because informants
are likely to give socially expected answers rather than the actual ones.
A second problem is her choice of informants. They are 256 men
ranging from 42 to 70 years old, and 271 women ranging from 40 to 62.
Most of the men are white-collar workers and most of the women are
 Women’s Language 155

housewives. There is a clear disparity between the sexes in terms of their

occupations: Working women, whose social status is supposed to be more
equal to male workers, should have been included more in the survey. A
more important problem lies in the age of the informants: They are all
middle-aged or older. Recently young people are often accused of vulgar
speech and of not knowing how to use honorifics. This suggests that
younger women may be less likely to speak the polite ‘women’s language’
than older women.12 So the research findings cannot be definitely
persuasive if the researcher overlooks the young women’s use of language.
We cannot explore this any further in this paper, but we would like to
discuss the problems of Japanese sex difference research some other time.

7. Conclusion
We have brought forward some of the methodological problems in
much of the sex difference research. Most of the research has been based
on the questionable assumption that there is a simple one-to-one
relationship between particular linguistic strategies and their
communicative functions. In so doing, researchers have correlated the
frequency of these strategies by the sex of the speakers. By using ‘men
as the norm and women as a deviation’ criteria, they have explained their
results in order to prove how women’s language use is inferior to or
different from men’s: “a lot of sex difference research was done
specifically in order to provide a scientific account of an already-assumed
female inferiority . . . research results have been used to justify particular
aspects of women’s subordination . . . (Cameron, 1992, p.36).” If so, we
should become critically aware that alleged ‘women’s language’ may not
be a scientific truth. ‘Silent’, ‘non-assertive’, ‘indirect’, ‘polite’ and
‘supportive’ women in an interaction can be seen as merely an ideal
prescribed by the androcentric ideology. It is now recognised that
perceptions of female speech and prescriptions about how women ought to
speak are the head and tail of the same coin.
In their study of courtroom language, O’Barr & Atkins (1989) confirm
156 Yuki Mizokami

that ‘women’s language’ is not characteristic of all women. We must

therefore abandon the ‘either women or men’ notion about certain
linguistic features in future language and sex difference research. A
speaker’s sex should not be regarded as an unproblematic independent
variable but should be considered as only one of the several complex
variables.
As we have seen, even feminists have been affected by the
androcentric ideology, and as a result they have conspired to reinforce
stereotypes of ‘women’s language’ against their initial intention. Many
feminist linguists have conducted their research under the name of
‘objective’ science, and in the end they have continued to inferiorise
women. In this sense, ‘objectivity’ in linguistics could be seen as one
particular form of the androcentric ideology. If researchers’ standpoints
cannot possibly be objective, then there is no longer any sense in trying to
feign objectivity. Such being the case, feminist linguists would not have
to hide their aim of raising women’s social position in future research.
Feminists should not stress the significance of disparity in sex difference
research, rather they should study how women and men are similar, and
how women differ among themselves. The ideal situation for feminists is
that female/male differences become politically and socially
inconsequential, in the same way that some human variations such as
blood types are. To reach that goal, future researchers need to go beyond
the simple female/male dichotomy in linguistic research. We now must
conclude that it is impossible and senseless to try to abstract the distinctive
female/male linguistic characteristics like a binary opposition on the
grounds that; (1) each woman and man talks differently depending on the
situation; and that (2) the meaning of a specific linguistic behaviour is not
perpetually fixed but is negotiated within the participating group in each
interaction.

Notes
1 See for example, Swacker, 1975, Bernard, 1972.
 Women’s Language 157

2 As for the ‘women talk more politely’ stereotype, see for example,
Lakoff, 1975, Ogino et al., 1985. Regarding the ‘women use more
questions’ stereotype, see for example, Fishman, 1980.
3 See for example, Zimmerman & West, 1975, West & Zimmerman,
1983.
4 See for example, Lakoff, 1975, Conley, O’Barr & Lind, 1979.
5 See for example, Fishman, 1980.
6 See for example, Lakoff, 1975, Fishman, 1980, Preisler, 1986.
7 See for example, Coates, 1993.
8 See for example, Zimmerman & West, 1975, Fishman, 1980, O’Barr
& Atkins, 1980, Holmes, 1984.
9 In this quotation, [ ] indicates that the portions of utterances so
encased are simultaneous. Numbers in parentheses such as (4.2)
indicate the seconds and tenths of seconds ensuing between speaker
turns. ( # ) indicates a pause of about a second.
10 Tannen (1994) argues as a scientific method, a researcher should play
the recording back to participants in order to solicit their spontaneous
interpretations and reactions, and also should solicit their responses
to the researcher’s interpretations. Jenkins & Kramarae (1981) also
point out that without knowing the speaker’s interpretations of speech
construction, a researcher knows very little about sociological aspects
of speaking.
11 See for example, Fishman, 1980, Preisler, 1986, Dubois & Crouch,
1975, to name just three. However, while Fishman and Preisler
support the ‘Lakoff hypothesis’ based on their own data sampling,
Dubois & Crouch find that all tag questions in their transcript are
produced by men.
12 For example, Reynolds (1990) reports that the use of ‘boku’, a male
first person pronoun, by junior high school girls is quite common in
Tokyo.

References
Beattie, G. W. 1981 ‘Interruption in conversational interaction, and its
relation to the sex and status of the interactants’. Linguistics, 19, 15-35.
Bernard, J. 1972 The Sex Game. New York: Atheneum.
Cameron, D. 1992 Feminism & Linguistic Theory. 2nd ed. London:
Macmillan.
Cameron, D., F. McAlinden & K. O’Leary 1989 ‘Lakoff in context: the
social and linguistic functions of tag questions’. In J. Coates & D.
158 Yuki Mizokami

Cameron (eds.) 1989 Women in Their Speech Communities: New

Perspectives on Language and Sex. Pp. 74-93. London : Longman.
Chalmers, A. F. 1978 What Is This Thing Called Science?: An
Assessment of the Nature and Status of Science and its Methods.
Milton Keynes: Open University Press.
Coates, J. 1989 ‘Gossip revisited: language in all-female groups’. In J.
Coates & D. Cameron (eds.) 1989 Women in Their Speech
Communities: New Perspectives on Language and Sex. Pp. 94-122.
New York: Longman.
Coates, J. 1993 Women, Men and Language: A Sociolinguistic Account
of Sex Differences in Language. 2nd. ed. New York: Longman.
Conley, J. M., W. M. O’Barr & E. A. Lind 1979 ‘The power of language:
Presentational style in the courtroom’. Duke Law Journal, 1978,
1375- 99.
Dubois, B. L. & I. Crouch 1975 ‘The question of tag questions in
women’s speech: they don’t really use more of them, do they?’
Language in Society, 4, 289-94.
Fishman, P. 1980 ‘Conversational insecurity’. In D. Cameron (ed.)
1990 The Feminist Critique of Language: A Reader. Pp. 234-241.
London: Routledge.
Holmes, J. 1984 ‘Hedging your bets and sitting on the fence: some
evidence for hedges as support structures’. Te Reo, 27, 47-62.
Hori, M. 1986 ‘A Sociolinguistic Analysis of the Japanese Honorifics’.
Journal of Pragmatics, 10 (3), 373-386.
Jenkins, M. & C. Kramarae 1981 ‘A thief in the house: women and
language’. In D. Spender (ed.) Men’s Studies Modified: The Impact of
Feminism on the Academic Disciplines. Pp. 11-22. Oxford: Pergamon
Press.
Jespersen, O. 1922 Language: its nature, development and origin.
London: George Allen & Unwin.
Jones, D. 1980 ‘Gossip: notes on women’s oral culture’. In D. Cameron
(ed.) 1990 The Feminist Critique of Language: A Reader. Pp. 242-
250. London: Routledge.
Lakoff, R. 1975 Language and Woman’s Place. New York:
Harper & Row.
Maltz, D. & R. Borker 1982 ‘A cultural approach to male-female
miscommunication’. In J. J. Gumpertz (ed.) 1982 Language and Social
Identity. Pp. 195-216. Cambridge: Cambridge University Press.
Mizokami, Y. 1995 ‘‘Women’s Language’ as the Ideological Fabrication:
A critical assessment of sex difference research in English and
 Women’s Language 159

Japanese sociolinguistics’. M.A. dissertation, University of London.

Murray, S. O. 1985 ‘Toward a model of members’ methods for
recognising interruptions’. Language in Society, 13, 31-41.
Nakamura, M. 1990 ‘Yakusha Kaisetsu’. In D. Cameron 1990
Feminism to Gengo riron [Feminism & Linguistic Theory, 1st ed.].
trans. by M. Nakamura. Pp. 305-333. Tokyo: Keisou Shobou.
O’Barr, W. & B. Atkins 1980 ‘“Women’s language” or “powerless
language”?’ In S. McConnell-Ginet et al. (eds.) 1980 Women and
Language in Literature and Society. Pp. 93-110. New York:
Praeger.
Ogino, T., Y. Misono & C. Fukushima 1985 ‘Diversity of honorific
usage in Tokyo: a sociolinguistic approach based on a field survey’.
International Journal of the Sociology of Language, 55, 23-39.
Poynton, K. 1985 Language and Gender: Making the Difference.
Oxford: Oxford University Press.
Preisler, B. 1986 Linguistic Sex Roles in Conversation. Berlin:
Mouton de Gruyter.
Reynolds, A. K. 1990 ‘Female speakers of Japanese in Transition’. In
S. Ide & N. H. McGloin (eds.) 1990 Aspects of Japanese Women’s
Language. Pp. 129-146. Tokyo: Kurosio.
Siegler, D. & R. Siegler 1976 ‘Stereotypes of males’ and females’
speech’. Psychological Reports, 39, 167-70.
Swacker, M. 1975 ‘The sex of the speaker as a sociolinguistic variable’.
In B. Thorne & N. Henley (eds.) 1975 Language and Sex:
Difference and Dominance. Pp. 76-83. Massachusetts: Newbury
House.
Swann, J. 1992 Girls, Boys & Language. Oxford: Blackwell.
Tannen, D. 1994 Gender & Discourse. New York: Oxford University
Press.
West, C. & D. Zimmerman 1983 ‘Small insults: A study of interruptions
in cross-sex conversations between unacquainted persons’. In B.
Thorne, C. Kramarae & N. Henley (eds.) 1983 Language, gender and
society. Pp. 103-17. Massachusetts: Newbury House.
Zimmerman, D. & C. West 1975 ‘Sex roles, interruptions and silences
in conversation’. In B. Thorne & N. Henley (eds.) 1975 Language and
Sex: Difference and Dominance. Pp. 105-29. Massachusetts:
Newbury House.
160 Yuki Mizokami