7.2 Gender And Language
The grounds on which issues of gender and language have been written about over the last thirty years have always been diverse. Though it is common to conveniently categorise such writings as either those of difference or dominance, such a broad view masks a complex situation. If we take Lakoff’s (1975) work on ‘women’s language’ as a starting point we can see the variety of attitudes in the (at times) conflicting interpretations of it. Lakoff identified a number of features of ‘women’s language’ (as opposed to the assumed male ‘norm’) such as ‘empty’ adjectives (divine, adorable), milder expletives, greater use of tag questions and hedges (well, erm etc.), superpolite forms, hypercorrect grammar, and (interestingly for this study) an inability to tell jokes and lack of a sense of humour (pp.53-6). Lakoff’s work is most often seen as a study of difference, that is, ‘women’s language’ as opposed to ‘men’s language’. But even here a distinction is made, with Maltz and Borker (1982:199) seeing it as one of ‘psychological difference’ in contrast to their own approach of ‘cultural difference’ (more of which below). However, Henley and Kramarae (1991:20) and Cameron (1995:33) see Lakoff’s model not as one of difference but one of deficit. That is, they see ‘women’s language’ as commonly being considered as somehow lacking when compared to the (assumed male) norm.
Lakoff’s initial work was largely intuitive and was criticised for being so. The most common form of opposition came from those feminists whose work was not speculative but was based on empirical studies which showed that in much cross-sex verbal interaction men dominate women by, for example, interrupting (Zimmerman and West 1975, West and Zimmerman 1983) and that women did much more support work than men: ‘The active maintenance of a female gender requires women to be available to do what needs to be done in interaction, to do the shitwork and not complain’ (Fishman 1978:405). This dominance view was perhaps best summed up in the title of a work by Spender (1980), Man Made Language. It should be noted, however, that Johnson (1997:9) actually places Lakoff herself within the framework of the dominance model because Lakoff speaks of women’s use of language as ‘powerless’ vis-à-vis that of men’s.
More recently these divergent approaches to gender and language came to a
head with the publication of Tannen’s (1992) popularisation of gender
and language issues, You Just Don’t Understand. Men And Women In
Conversation. This was explicitly based on Maltz and Borker’s ‘two
cultures’ model. Maltz and Borker distinguish their approach from both
Lakoff and the proponents of the dominance model: ‘We place stress not
on psychological difference or power differentials…but rather on a notion
of cultural differences between men and women’ (1982:199). Their model
in turn was derived from Gumperz’s work concerning interethnic communication
(1982), where problems of interpretation and meaning derive from interlocutors
different cultural upbringing. A key element of Gumperz’s work is to
show that when such miscommunication occurs neither of the parties is ‘right’
and blame should not be apportioned.
Maltz and Borker argue that this approach ‘can be applied to cross-sex communication as well’ (1982:196). According to them, males and females are socialised into different subcultures throughout childhood. Drawing on a wide variety of studies of children’s play and interaction they say that girls and boys use language in different ways:
1. to create and maintain relationships of closeness and equality
2. to criticise others in acceptable ways
3. to interpret accurately the speech of other girls
1.to assert one’s position of dominance
2. to attract and maintain an audience
3. to assert oneself when other speakers have the floor
They claim that (American) men and women come from these different sociolinguistic subcultures and this leads to cultural miscommunication between them (p.200).
Tannen based her 1992 work on this model and says that these different styles mean that for women conversations are ‘negotiations for closeness in which people try to seek and give confirmation and support, and to reach consensus’, whereas for men ‘conversations are negotiations in which people try to achieve and maintain the upper hand if they can, and protect themselves from others’ attempts to put them down and push them around’ (1992:24-5). She calls women’s style ‘rapport talk’ and men’s ‘report talk’ (p.77). According to this model, much talk between women and men can lead to miscommunication for which nether side is to blame. This two cultures approach (which has a familiar binary ring about it), and Tannen’s widely disseminated exposition of it in particular, has many critics, but before turning to them brief mention is made of two oft-cited examples of male/female differences in language use which are directly relevant to the discussion to come in the final section. These are the nature of the floor in interactions and the concept of politeness. The former was first mentioned in 6.3.1 in the discussion of turn-taking and receives detailed treatment below in 8.3, the latter will also receive a greater focus below in 8.7.
The main criticisms of the two cultures model are that it tends to be apolitical and, being based on an interethnic model, it simplistically conflates two distinct phenomena, ethnicity and gender. On the first point Cameron states that ‘it must be acknowledged that many of the differences that exist between the sexes are a direct result of inequality between them. Researchers must take explicit account of this and reflect on the political character of sex-difference research in a society which is still profoundly unequal’ (1988:11). Tannen, however, does not position men and women unequally but symmetrically, and thus, power disappears. Coates concludes a wide overview of the literature which takes in overlaps, interruptions, use of hedges, tag questions, commands and directives, and the use of taboo language, with the comment that ‘men dominate conversation by interrupting women, controlling topics of conversation and also by being silent’ and their differences stem ‘directly from women’s and men’s membership of a patriarchal society’ (1993:139). Crawford takes issue with the too-narrow focus of the two cultures model, claiming that it overlooks ‘which women in which social groups’ so that ‘[w]hen sex is the only conceptual category, differences attributable to situations and power relationships are made invisible’ (1995:101, original emphasis). The most virulent criticism comes from Troemel-Ploetz, who says that, when reading Tannen,
one searches in vain for concepts like dominance, control, power, politics of gender, sexism, discrimination…Concepts like patriarchy or feminism never occur, being evidently far too radical for the author. Tannen is selling political naiveté, but neither is sociology quite so naïve nor linguistics quite so apolitical as Tannen would have us believe.
To press this point further, Henley and Kramarae suggest that male-female
miscommunication is not some simple cultural by-product of gender relations.
‘The construction of miscommunication between the sexes emerges as a
powerful tool, maybe even a necessity, to maintain the structure of male supremacy’
Tannen, though, is not unaware that men dominate women in conversation and also that many see the two cultures model as a form of, in her own words, ‘copping out’ (1992:209). She goes on to say that though she is sympathetic to this view, accepting the dominance model means also accepting the view that ‘high-involvement’ speakers such as blacks and Jews are domineering in cross-cultural communication (p.209).
But this would seem to conflate gender with ethnicity. Simply stated, gender is not ethnicity. On this point, Eckert (1989:253-4) notes that they are not equivalent categories. Though gender roles mean that men and women exhibit many differences, these are constructed in such a way that they are seen as a source of attraction (see the discussion of compulsory heterosexuality above). Further, while there are parallels in the power relationships between all dominant and subordinate groups, they are practised in distinct ways. Thus:
It is not a cultural norm for each working-class individual to be paired up for life with a member of the middle-class or for every black person to be so paired up with a white person. However, our traditional gender ideology dictates just this kind of relationship between men and women.
Further, Crawford notes that, ‘the ethnic examples chosen to illustrate
the two cultures model are often selective, and conveniently naïve about
social hierarchies of race, class, and color’ (1995:104). Troemel-Ploetz
would also concur on this point, saying that Tannen’s leaving power
out of the equation is like saying that Black English and Oxford English are
simply two varieties of English with the same validity, ‘it just so
happens that the speakers of one variety find themselves in high-paying positions
with a lot of prestige and power of decision-making, and the others are found
in more low-paying jobs, or on the streets and in prisons’ (1991:498).
Indeed, one element of this last point – that concerning equal validity – is worth further consideration. Cameron (1995:37), in an attempt to understand why Tannen may have used the two cultures model so sweepingly, suggests that it quite simply is founded on the basic assumption of twentieth century anthropological and linguistic theory, which sees all cultures and languages as equal. Initially this was a radical view, which took researchers away from formulations of ‘primitive’ languages spoken by ‘savages’, and it eventually became the normal assumption of all linguistic enquiry. But when differences arise from positions of inequality, says Cameron, then cultural relativism ‘is not only theoretically naïve, it is politically damaging’ (pp.41-2).
However, this is not to construct the difference and dominance approaches
as polar opposites. Coates notes that it is oversimple to see it as an either/or
situation, and that both approaches are necessary to account adequately for
women’s and men’s language use. (1988:72-3). Crawford, too, cautions
against a naïve use of the dominance approach and calls for a ‘more
textured concept of power’ which takes all the relevant complexities
of talk and gender into account (1995:130). More recent developments, though,
have seen some researchers move away from this basic divide, seeing it as
a limitation, and we will now consider some of their criticisms.
Such a development can be clearly seen, for example, in the two editions of a collection of feminist writings on language, The Feminist Critique Of Language, edited by Deborah Cameron. In the 1990 first edition, part three is entitled ‘Dominance And Difference In Women’s Linguistic Behaviour’. In the 1998 second edition, part three is entitled ‘Talking Gender: Dominance, Difference And Performance’, this new title taking into account certain postmodern developments in feminist linguistics in the 1990s, some of which will be now dealt with.
Johnson notes that both the difference and dominance approaches share two weaknesses: firstly, they problematise women, that is, see women’s use of language as somehow marked vis-à-vis the (assumed male) norm (1997:10). (Many dominance theorists would no doubt argue that their work problematises men, or, at the very least, positions men as culpable for the inequalities that are identified.) Secondly, both approaches see gender as based on a binary opposition and that ‘speech constitutes a symbolic reflection of that opposition’ (p.11). This she sees as problematic for both language and gender. Once again origins for such a line of thought are traced back to twentieth century linguistic theory, this time to the broad tradition which has its roots in structuralist approaches to language, deriving from the Saussurian paradigm of language as a series of contrasts (p.14), here the contrast being between males and females.
Cameron, too, highlights certain problems in contemporary linguistics, lamenting that feminist scholars in other disciplines have made progress but feminist linguistics has stagnated. Part of the problem for her is that language is the phenomenon to be explained and ‘gender’ is offered wholesale as the explanation; gender itself remains untheorised, ‘it is a given; the bottom line’ (1995:39). Too often this merely leads to linguistic research routinely and unthinkingly cataloguing what men do with language and what women do with language without actually considering how gender itself is constituted (echoes of Lorber’s ‘believing is seeing’ there).
Gal is also highly critical of ‘variationist sociolinguistics’ for counting the use of certain linguistic variables as used by men and women and simply reading off power relations from them (1995:170-1). In an approach sharing much with the views expressed in the above discussion of representations, she insists that categories such as women’s speech, men’s speech, and prestigious and powerful speech are not simply reflections of (already) given speakers’ identities. Our utterances are constitutive of identity. She continues:
These categories, along with broader ones such as feminine and masculine, are culturally constructed within social groups, they change through history and are systematically related to other areas of cultural discourse such as the nature of persons, of power, and of a desirable moral order.
(p.171, original emphasis)
Cameron would agree, pointing out that though there are styles which are produced as masculine and feminine, men and women do not mechanically produce these according to their sex but instead individuals choose what they need from these styles in the process of producing themselves as gendered subjects (1995:43). As we saw above, Johnson too does not see masculine and feminine as binary opposites and she furthers her argument by saying they are, rather, mutual constructs, and so, for her, masculinity is dependent on femininity for its own definition. Thus, the construction of male heterosexuality will involve the exclusion and denial of both women and homosexual men. Such a dialectical view of gender entails that gender identities can never in fact be complete (1997:22). And at this point it is worth looking at some empirical studies to flesh out these ideas.
We start with gossip. Jones (1980) sees women as a distinct speech community and claims gossip as a part women’s oral culture. She defines gossip as ‘a way of talking between women in their roles as women, intimate in style, personal and domestic in topic and setting, a female cultural event’ (p.194). Subsequent empirical studies, however, have shown that gossip is not a gender specific language behaviour. Johnson and Finlay, for example, analyse an episode of a television football discussion programme (all male) and this leads them to move away from gossip as a female speech genre and see how men also use gossip to create solidarity with one another, that is, ‘how men use very similar discursive strategies when doing “identity work”’ (1997:142).
Cameron (1997a) studied the speech of a group of five young male white American middle-class suburbanite friends, and she defines gossip as ‘discussion of several persons not present but known to the participants, with a strong focus on critically examining these individuals’ appearance, dress, social behaviour and sexual mores’ (p.51). As a starting point she refers to Tannen’s two cultures model and notes that if one reverses the gender in Tannen’s anecdotes it is still possible to furnish a script to make sense of them (once again the role of the audience and background knowledge play their part). For example, Tannen attributes men’s reluctance to ask for directions while driving to their wish to avoid the appearance of helplessness. Cameron suggests that if it were a woman who did this, another equally plausible script would be available – women don’t like to impose on others, or they are afraid to stop and talk to strangers (p.48). What this means is that a general discourse on gender difference is used to explain the linguistic behaviour, whereas she believes it could be more useful to say that this discourse constructs the differentiation, ‘makes it visible as differentiation’ (p.48, original emphasis). She proposes that conversationalists often do the same thing, they ‘construct stories about themselves and others, with a view to performing certain kinds of gender identity’ (p.48). Taking a lead from Butler’s performative view of gender identity, where gender is ‘the repeated stylisation of the body’ (Butler 1990:33), Cameron views gendered speech acts as a repeated stylisation which congeals into making us ‘proper’ men-women. However, people can and do choose from the repertoire of different linguistic styles and ‘behave in ways we would normally associate with the “other” gender’ (p.50).
In the study of the male group Cameron found, like Johnson and Finlay, that men do gossip and she also underlines the performative point that it is what this group of men do with gossip that is of interest. These men gossip about their sexual adventures with women and about other male students whose behaviour they see as gay. ‘Their conversation is animated by entirely traditional anxieties about being seen at all times as red-blooded heterosexual males: not women and not queers’ (p.62).
A study by Pujolar i Cos (1997) considers the construction of masculinities in a multilingual setting. He studies two separate groups of young (17-23) working-class people in Barcelona, the Rambleros (six women, five men) and the Trepas (six women, seven men) (p.87). He shows how the males in the different groups created and displayed different masculinities through their linguistic choices and attitudes. Thus, the Rambleros constructed a ‘simplified masculinity’ (a cult of the body, transgressive behaviour, verbal and physical aggression) which showed a resistance to Catalan (with its politically correct associations) and the cultivation of a distinct Andalusian dialect (with its simple and common peasant associations of the migrant workers from the south) (p.104). The Trepas males, a more politicised group, chose to use Catalan and spoke with an Andalusian accent only when mocking people they considered sexist or stupid (pp.97-8).
Turning now to two studies of women’s use of language, we start first in Japan. Japanese is a language that is often referred to in the literature as one that is heavily gender marked (see, for example, Shibamoto 1985), so it is useful to consider the way people actually use it. Okamoto (1995) studied the conversations of ten female Japanese students, aged 18-20, transcribing 150 consecutive sentence tokens per speaker. Her analysis focused on sentence-final forms, each of which was identified as feminine, masculine, or neutral, where feminine were those considered in the literature to be traditionally used by women, masculine those traditionally used by men, and neutral by both. The feminine and masculine forms were divided into ‘strong’ and ‘moderate’ (pp.300-1). She found that of the 1,500 tokens used by these women, 12.3% were feminine forms (and only 4.5% ‘strongly feminine’), 18.9% were masculine forms, and 68.8% neutral (p.303). This shows that speech styles cannot be simply read off from the speaker’s gender and Okamoto stresses that other factors also play a part in speakers’ choices. She underlines that her subjects were all middle-class and so not representative of working-class speech. Age, too, is significant and she draws on a study by Okamoto and Sato (1992) to show, for example, that women in the age group 45-57 use far more feminine forms than 18-23 year olds (p.306). Further, when she showed two young women a letter from a newspaper which strongly condemned the modern trend of women using ‘men’s language’, they replied that they were not using ‘men’s language’ but the ‘language of young people’ (p.313). Other studies of women’s language use show that occupation also influences women’s speech, with homemakers using more feminine forms than students, and female office workers using more than professional or self-employed women (p.307). Okamoto underlines the fact that speakers make strategic choices to communicate certain pragmatic meanings in particular social contexts. ‘These choices…require the context-specific consideration of multiple social attributes associated with the speaker’s identity and interpersonal relationships (such as gender, age, occupation, intimacy) as well as the speaker’s knowledge and evaluation of the relevant linguistic norms’ (p.312).
Hall (1997) discusses how women telephone sex workers in San Francisco use aspects of Lakoff’s ‘powerless’ women’s language (see above) in their work identities and through this achieve economic empowerment. Hall comments,
In their interactional histories (e.g. at school, in the family), the female fantasy-line operators have received positive reinforcement for this style of discourse, and now, through additional reinforcement in the workplace, they are selling it back to the culture at large for a high price.
Some of the (stereotyped) identities they are expected to create in this linguistic performance can be seen in a list given in a training manual for operators which was reprinted in Harper’s Magazine: bimbo, nymphomaniac, mistress, slave, transvestite, lesbian, foreigner, virgin (Hall 1997:190-1).
However, as has been repeatedly made clear, sex and gender practices do not take place in isolation from others social considerations and a note of caution needs to be sounded when considering the positioning of speaking subjects. While the choices made by the women sex workers may be economically empowering and this in turn gives them a greater measure of social autonomy, Cameron (1997b) points out that in such a market women can be the sellers but they are always the goods: ‘Whatever advantage individual women may derive from developing a particular kind of language in telephone sex work, the system of meanings on which the marketability of that language depends does not advantage women collectively’ (p.31).
7.2 Gender and Language