A major topic in sociolinguistics is the connection, if any, between the structures, vocabularies, and ways of using particular languages and the social roles of the men and women who speak these languages.

  • Do the men and women who speak a particular language use it in different ways?

  • If they do, do these differences arise from the structure of that language, which would therefore be one kind of confirmation of the Whorfian hypothesis, or, alternatively, do any differences that exist simply reflect the ways in which the sexes relate to each other in that society, whatever the reason?

  • May it be possible to describe a particular language as ‘sexist,’ or should we reserve such a description for those who use that language?

  • If the answer to either question is affirmative, what could and should be done?
  • These issues generated a considerable amount of thought and discussion in the last decades of the twentieth century and many are still unresolved. They are also very emotional issues for many who have chosen either to write on them or to discuss them, and that they should be so is quite understandable. The literature on these issues is now vast; it has been one of the biggest ‘growth’ areas within sociolinguistics in recent years.

    In this lecture, I will attempt to see what some of the underlying facts are and to avoid the kinds of rhetoric and dialectic that characterize much of the discussion of ‘sexism in language,’ a topic which often seems to invite ‘large’ arguments based on ‘small’ data.

    Sex is to a very large extent biologically determined whereas gender is a social construct (but still one heavily grounded in sex, as we can see in recent publications that use the term ‘sexuality,’ e.g., Kulick, 2003, and Cameron and Kulick, 2003) involving the whole gamut of genetic, psycho- logical, social, and cultural differences between males and females.

    Wodak (1997b, p. 13) says that gender is ‘not . . . a pool of attributes “possessed” by a person, but . . . something a person “does.”’ Elsewhere (1997a, p. 4) she adds that ‘what it means to be a woman or to be a man [also] changes from one generation to he next and . . . varies between different racialized, ethnic, and religious groups, as well as for members of different social classes.’ In such a view, gender must be learned anew in each generation.

    Cameron (1998b, pp. 280–1) states that view in a slightly different way:

    “Men and women . . . are members of cultures in which a large amount of discourse about gender is constantly circulating. They do not only learn, and then mechanically reproduce, ways of speaking ‘appropriate’ to their own sex; they learn a much broader set of gendered meanings that attach in rather complex ways to different ways of speaking, and they produce their own behavior in the light of these meanings. . . .”

    “Performing masculinity or femininity ‘appropriately’ cannot mean giving exactly the same performance regardless of the circumstances. It may involve different strategies in mixed and single-sexed company, in private and public settings, in the various social positions (parent, lover, professional, friend) that someone might regularly occupy in the course of everyday life.”

    Gender is also something we cannot avoid; it is part of the way in which societies are ordered around us, with each society doing that ordering differently. As Eckert and McConnell-Ginet (2003, p. 50) say: ‘The force of gender categories in society makes it impossible for us to move through our lives in a nongendered way and impossible not to behave in a way that brings out gendered behavior in others.’ Gender is a key component of identity.

    We will look at some of the evidence that there are gender differences in language use. One purpose will be to evaluate that evidence: just how good is it? However, the main purpose is to try to discover, when indeed there is good evidence, what it is good evidence of. That languages can be sexist? That those who use languages may be sexist? That language-learning is almost inevitably tied to gender-learning? That such learning is almost always skewed in such a way as to favor one gender over the other? That change is not only desirable but possible? It is issues such as these that will be our concern.


    That there are differences between men and women is hardly a matter of dispute. Females have two X chromosomes whereas males have an X and a Y; this is a key genetic difference and no geneticist regards that difference as unimportant. On average, females have more fat and less muscle than males, are not as strong, and weigh less. They also mature more rapidly and live longer. The female voice usually has different characteristics from the male voice, and often females and males exhibit different ranges of verbal skills.

    However, we also know that many of the differences may result from different socialization practices (see Philips et al., 1987). For example, women may live longer than men because of the different roles they play in society and the different jobs they tend to fill. Differences in voice quality may be accentuated by beliefs about what men and women should sound like when they talk, and any differences in verbal skills may be explained in great part through differences in upbringing. (It has often been noted that there is far more reading failure in schools among boys than girls, but it does not follow from this fact that boys are inherently less well equipped to learn to read, for their poor performance in comparison to girls may be sociocultural in origin rather than genetic.) There is also an important caveat concerning all such studies showing differences between groups, and the two genders are just groups like any other; it is one I made earlier (p. 158) and will repeat here. For many in the two groups under comparison there will be no difference at all: the next person you meet on the street may be male or female, tall or short, long-lived or short-lived, high-voiced or low-voiced, and so on, with not one of these characteristics being predictable from any other. (Given a thousand or more such encounters some tendencies may emerge, but even knowing what these are would not help you with the very next person you meet.)

    Numerous observers have described women’s speech as being different from that of men (see Baron, 1986, Arliss, 1991, pp. 44–112, and pp. 162–207 of this book). I should also observe that there is a bias here: men’s speech usually provides the norm against which women’s speech is judged. We could just as well ask how men’s speech differs from that of women, but investigators have not usually gone about the task of looking at differences in that way. For example, in discussing language change in Philadelphia, Labov (2001, pp. 281–2) deliberately recasts his statement that ‘Women conform more closely than men to sociolinguistic norms that are overtly prescribed, but conform less than men when they are not’ to read that men ‘are less conforming than women with stable linguistic variables, and more conforming when change is in progress within a linguistic system.’ He does this so as to avoid appearing to bias his findings.

    Any view too that women’s speech is trivial (see the denial in Kipers, 1987), gossip-laden, corrupt, illogical, idle, euphemistic, or deficient is highly suspect; nor is it necessarily more precise, cultivated, or stylish – or even less profane (see De Klerk, 1992, and Hughes, 1992) – than men’s speech. Such judgments lack solid evidentiary support. For example, apparently men ‘gossip’ just as much as women do (see Pilkington, 1998); men’s gossip is just different. Men indulge in a kind of phatic small talk that involves insults, challenges, and various kinds of negative behavior to do exactly what women do by their use of nurturing, polite, feedback-laden, cooperative talk. In doing this, they achieve the kind of solidarity they prize. It is the norms of behavior that are different.

    In the linguistic literature perhaps the most famous example of gender differentiation is found in the Lesser Antilles of the West Indies among the Carib Indians. Male and female Caribs have been reported to speak different languages, the result of a long-ago conquest in which a group of invading Carib- speaking men killed the local Arawak-speaking men and mated with the Arawak women. The descendants of these Carib-speaking men and Arawak-speaking women have sometimes been described as having different languages for men and women because boys learn Carib from their fathers and girls learn Arawak from their mothers. This claim of two separate languages is now discounted. What differences there are actually do not result in two separate or different languages, but rather one language with noticeable gender-based characteristics (Baron, 1986, pp. 59–63, and Taylor, 1951b).

    Phonological differences between the speech of men and women have been noted in a variety of languages. In Gros Ventre, an Amerindian language of the northeast United States, women have palatalized velar stops where men have palatalized dental stops, e.g., female kjatsa ‘bread’ and male djatsa. When a female speaker of Gros Ventre quotes a male, she attributes female pronunciations to him, and when a male quotes a female, he attributes male pronunciations to her. Moreover, any use of female pronunciations by males is likely to be regarded as a sign of effeminacy. In a northeast Asian language, Yukaghir, both women and children have /ts/ and /dz / where men have /tj/ and /dj/. Old people of both genders have a corresponding /7j/ and /jj/. Therefore, the difference is not only gender- related, but also age-graded. Consequently, in his lifetime a male goes through the progression of /ts/, /tj/, and /7j/, and /dz/, /dj/, and /jj/, and a female has a corresponding /ts/ and /7j/, and /dz/ and /jj/. In Bengali men often substitute /l/ for initial /n/; women, children, and the uneducated do not do this. Likewise, in a Siberian language, Chukchi, men, but not women, often drop /n/ and /t/ when they occur between vowels, e.g., female nitvaqenat and male nitvaqaat.

    In Montreal many more men than women do not pronounce the l in the pronouns il and elle. Schoolgirls in Scotland apparently pronounce the t in words like water and got more often than schoolboys, who prefer to substitute a glottal stop. Haas (1944) observed that in Koasati, an Amerindian language spoken in southwestern Louisiana, among other gender-linked differences, men often pronounced an s at the end of verbs but women did not, e.g., male lakáws ‘he is lifting it’ and female lakáw. What was interesting was that this kind of pronunciation appeared to be dying out, because younger women and girls do not use these forms. That older speakers recognized the distinction as gender-based is apparent from the fact that women teach their sons to use the male forms and men narrating stories in which women speak employ female forms in reporting their words. This practice is in direct contrast to the aforementioned situation in Gros Ventre, where there is no such changeover in reporting or quoting.

    There is also a very interesting example from English of a woman being advised to speak more like a man in order to fill a position previously filled only by men. Margaret Thatcher was told that her voice did not match her position as British Prime Minister: she sounded too ‘shrill.’ She was advised to lower the pitch of her voice, diminish its range, and speak more slowly, and thereby adopt an authoritative, almost monotonous delivery to make herself heard. She was successful to the extent that her new speaking style became a kind of trademark, one either well-liked by her admirers or detested by her opponents.

    In the area of morphology and vocabulary, many of the studies have focused on English. In a paper which, although it is largely intuitive, anecdotal, and personal in nature, is nevertheless challenging and interesting, Lakoff (1973), claims that women use color words like mauve, beige, aquamarine, lavender, and magenta but most men do not. She also maintains that adjectives such as adorable, charming, divine, lovely, and sweet are also commonly used by women but only very rarely by men. Women are also said to have their own vocabulary for emphasizing certain effects on them, words and expressions such as so good, such fun, exquisite, lovely, divine, precious, adorable, darling, and fantastic. Furthermore, the English language makes certain distinctions of a gender-based kind, e.g., actor–actress, waiter–waitress, and master–mistress. Some of these distinctions are reinforced by entrenched patterns of usage and semantic development.

    For example, master and mistress have developed quite different ranges of use and meaning, so that whereas Joan can be described as Fred’s mistress, Fred cannot be described as Joan’s master. Other pairs of words which reflect similar differentiation are boy–girl, man–woman, gentleman–lady, bachelor–spinster, and even widower–widow. In the last case, whereas you can say ‘She’s Fred’s widow,’ you cannot say ‘He’s Sally’s widower.’ Lakoff cites numerous examples and clearly establishes her point that ‘equivalent’ words referring to men and women do have quite different associations in English. A particularly telling example is the difference between ‘He’s a professional’ and ‘She’s a professional.’ Other investigators have documented the same phenomenon in other languages, for example in French uses of garçon and fille.

    One of the consequences of such work is that there is now a greater aware- ness in some parts of the community that subtle, and sometimes not so subtle, distinctions are made in the vocabulary choice used to describe men and women. Consequently, we can understand why there is a frequent insistence that neutral words be used as much as possible, as in describing occupations e.g., chair- person, letter carrier, salesclerk, and actor (as in ‘She’s an actor’). If language tends to reflect social structure and social structure is changing, so that judgeships, surgical appointments, nursing positions, and primary school teaching assignments are just as likely to be held by women as men (or by men as women), such changes might be expected to follow inevitably. This kind of work does two things: it draws our attention to existing inequities, and it encourages us to make the necessary changes by establishing new categorizations (e.g., Ms), and suggesting modifications for old terms (e.g., changing policeman to police officer and chairman to chairperson). However, there is still considerable doubt that changing waitress to either waiter or waitperson or describing Nicole Kidman as an actor rather than as an actress indicates a real shift in sexist attitudes. Reviewing the evidence, Romaine (1999, pp. 312–13) concludes that ‘attitudes toward gender equality did not match language usage. Those who had adopted more gender-inclusive language did not necessarily have a more liberal view of gender inequities in language.’

    One particular bit of sexism in languages that has aroused much comment is the gender systems that so many of them have, the he–she–it ‘natural’ gender system of English or the le–lait or der–die–das ‘grammatical’ gender systems of French and German. The possible connections between gender systems (masculine, feminine, neuter) and gender differences (male, female, neither) are various. See Romaine (1999) for some observations and claims concerning these connections, e.g., her claim (p. 66) that ‘ideological factors in the form of cultural beliefs about women . . . enter into gender assignment in [grammatical] systems that are supposedly purely formal and arbitrary.’ In English such connections sometimes create problems for us in finding the right pronoun: compare the natural ‘Everybody should hand in their papers in five minutes’ to the apparently biased ‘No person in his right mind would do that.’ Again, he–she distinctions can often be avoided – sometimes clumsily, to be sure – so it probably does not follow that languages with gender distinctions must be sexist, which would also be a clear argument in support of the Whorfian hypothesis (see pp. 221–8). It is the people who use languages who are or who are not sexist; Chinese, Japanese, Persian, and Turkish do not make the kinds of gender distinctions English makes through its system of pronouns, but it would be difficult to maintain that males who speak these languages are less sexist than males who speak English!

    There certainly are gender differences in word choice in various languages. Japanese women show they are women when they speak, for example, by the use of a sentence-final particle ne or another particle wa. In Japanese, too, a male speaker refers to himself as boku or ore whereas a female uses watasi or atasi. Whereas a man says boku kaeru ‘I will go back’ in plain or informal speech, a woman says watasi kaeru wa (Takahara, 1991). Children learn to make these distinctions very early in life. However, Reynolds (1998, p. 306) points out that ‘the use of boku . . . by junior high school girls has recently become quite common in Tokyo. Girls who were interviewed in a TV program explain that they cannot compete with boys in classes, in games or in fights with watasi. . . . The use of boku and other expressions in the male speech domain by young female speakers has escalated to a larger area and to older groups of speakers.’ In polite conversation a female speaker of Thai refers to herself as dìchAn whereas a man uses phoˇm. In Thai, too, women emphasize a repeated action through reduplication, i.e., by repeating the verb, whereas men place a descriptive verb, mak, after the verb instead.

    Different languages do seem to prescribe different forms for use by men and women. To cite another example, according to Sapir (1929a), the Yana language of California contains special forms for use in speech either by or to women. However, very few are like the language of the Dyirbal people of North Queensland, Australia, who have a special language which is gender-differentiated in a rather novel way (Dixon, 1971). The normal everyday language, Guwal, is used by both genders; but, if you are a man and your mother-in-law is present, or if you are a woman and your father-in-law is present, you use Dyalouy, a ‘mother-in-law’ variety. This variety has the same phonology and almost the same grammar as Guwal but its vocabulary is entirely different. However, both genders have access to both varieties.

    Another Australian aboriginal language, Yanyuwa, spoken by approximately 90 to 150 people, has gender-differentiated dialects. The dialects use the same word stems but there are different class-marking prefixes on nouns, verbs, and pronouns. According to Bradley (1998), men use one dialect among themselves and women use the other. Men also use men’s dialect to speak to women and women use women’s dialect to speak to men. Children are brought up in women’s dialect with boys required to shift – not always done easily – to men’s dialect as they are initiated into manhood. Bradley adds (p. 16) that: ‘If individuals wish to speak Yanyuwa then they are expected to speak the dialect which is associated with their sex – there is no other alternative.’ A person can use the other sex’s dialect only in very well-defined circumstances such as story-telling, joking, and certain singing rituals. The Yanyuwa find all of this perfectly normal and natural.

    In the Dyirbal example cited above we may find an important clue as to why there are sometimes different varieties for men and women. One variety may be forbidden to one gender, i.e., be taboo, but that gender is apparently nearly always the female gender. (See pp. 238–40.) This phenomenon has been noted among the Trobriand Islanders, various aboriginal peoples of Australia, Mayans, Zulus, and Mongols, to cite but a few examples. The taboos often have to do with certain kinship relationships or with hunting or with some religious practice and result in the avoidance of certain words or even sounds in words. They derive from the social organization of the particular group involved and reflect basic concerns of the group. Such concerns quite often lead to women being treated in ways that appear inimical to egalitarian-oriented outsiders.

    When we turn to certain grammatical matters in English, we find that Brend (1975) claims that the intonation patterns of men and women vary somewhat, women using certain patterns associated with surprise and politeness more often than men. In the same vein, Lakoff says that women may answer a question with a statement that employs the rising intonation pattern usually associated with a question rather than the falling intonation pattern associated with making a firm statement. According to Lakoff, women do this because they are less sure about themselves and their opinions than are men. For the same reason, she says that women often add tag questions to statements, e.g., ‘They caught the robber last week, didn’t they?’ These claims about tag questions and insecurity have been tested by others (Dubois and Crouch, 1975, Cameron et al., 1989, and Brower et al., 1979) and found wanting: experimental data do not necessarily confirm intuitive judgments. The latter investigators did find, however, that the gender of the addressee was an important variable in determining how a speaker phrased a particular question.

    We have already seen at other places in this book instances of language behavior varying according to gender. Many of these are quantitative studies in which sex is used as one of the variables that are taken into account. As Milroy and Gordon (2003, p. 100) say, ‘Strictly speaking . . . it makes sense . . . to talk of sampling speakers according to sex, but to think of gender as the relevant social category when interpreting the social meaning of sex-related variation.’ I will remind you of a few of these studies. Fischer’s work (see pp. 162–3) showed how very young boys and girls differ in certain choices they make, as did Cheshire’s work in Reading (pp. 170–2) in an older group. Labov’s studies in New York (pp. 164–8) and Philadelphia (pp. 209–11) also revealed noticeable gender differences in adult speech. These led him to make some interesting claims about what such differences indicated, e.g., about women’s role in language change. The Milroys’ study exploring network relationships (pp. 181–3) showed certain characteristics of men’s and women’s speech: how they were alike in some ways but different in others. Jahangiri’s study in Teheran (pp. 179–80) is of interest because of the very clear differences he reported between the speech of males and females. Finally, Gal’s study in the Oberwart of Austria (pp. 205–6) showed how it is not only what women say but who they are willing to say it to that is important. We have also noted that there are often different politeness requirements made of men and women.

    Still other gender-linked differences are said to exist. Women and men may have different paralinguistic systems and move and gesture differently. The suggestion has been made that these often require women to appear to be submissive to men. Women are also often named, titled, and addressed differently from men. Women are more likely than men to be addressed by their first names when everything else is equal, or, if not by first names, by such terms as lady, miss, or dear, and even baby or babe. Women are said to be subject to a wider range of address terms than men, and men are more familiar with them than with other men. Women are also said not to employ the profanities and obscenities men use, or, if they do, use them in different circumstances or are judged differently for using them. (However, the successful American television series ‘Sex and the City’ might seriously challenge that idea!) Women are also some- times required to be silent in situations in which men may speak. Among the Araucanian Indians of Chile, men are encouraged to talk on all occasions, but the ideal wife is silent in the presence of her husband, and at gatherings where men are present she should talk only in a whisper, if she talks at all.

    Some writers are not impressed with the kinds of findings reported in the preceding paragraphs. For example, Cameron (1998a, pp. 945–6) says that these findings ‘belong to the tradition of empirical sex difference studies that do no more than set out to find statistically significant differences between women’s and men’s behavior. This research formula has proved as durable as it is dubious (not to say dull).’ She adds that this kind of work ‘deals in arcane sound changes presented through complex statistics.’ In this view, merely to observe, count, and graph linguistic phenomena is not enough. An investigator needs some kind of theory about such behavior and some ideas to test before beginning an investigation. And, possibly, as we will see, some kind of ideology to suggest the ‘right’ theory.

    In setting out a list of what she calls ‘sociolinguistic universal tendencies,’ Holmes (1998) does offer some testable claims. There are five of these:

  • Women and men develop different patterns of language use.
  • Women tend to focus on the affective functions of an interaction more often than men do.
  • Women tend to use linguistic devices that stress solidarity more often than men do.
  • Women tend to interact in ways which will maintain and increase solidarity, while (especially in formal contexts) men tend to interact in ways which will maintain and increase their power and status.
  • Women are stylistically more flexible than men.
  • It is through testing claims such as these that we are likely to refine our understanding of those matters that interest us.

  • There are differences in gendered speech, some undoubtedly real but others almost certainly imaginary. Any differences that do exist surely also must interact with other factors, e.g., social class, race, culture, discourse type, group membership, etc. In the next section we will look more closely at some possible explanations for them. We will also try to avoid examining women’s speech in relation to men’s speech as though the latter provides the norm.

    Possible Explanations

    When we turn to matters having to do with how men and women use language in a wider sense, that is, in social interaction and to achieve certain ends, we find clues to possible explanations for the differences we encounter. One analysis of how women are presented in a set of cartoons produced some interesting findings (Kramer, 1974). The cartoons were taken from thirteen issues of The New Yorker magazine published between February 17 and May 12, 1973. The analysis showed that, when both genders were represented in the cartoon, men spoke twice as much as women. In the cartoons men and women also spoke on different topics, with men holding forth on such topics as business, politics, legal matters, taxes, and sports, and women on social life, books, food and drink, life’s troubles, and lifestyle. Women spoke less forcefully than men, and men swore much more than women. Men were also more blunt and to the point in their speaking. There was also some evidence that the use of words like nice and pretty was gender-linked. Although such cartoons are not actual records of what happens in speech, they must be based on what people think happens if they are to be effective. They make use of the stereotypes we have about the speech of men and women. Let us look at some better evidence freed from such stereotyping.

    In conversations involving both men and women many researchers agree that men speak more than women do. One also found that when men talked to men, the content categories of such talk focused on competition and teasing, sports, aggression, and doing things. On the other hand, when women talked to women, the equivalent categories were the self, feelings, affiliation with others, home, and family. Women are also reported to use more polite forms and more compliments than men. In doing so, they are said to be seeking to develop solidarity with others in order to maintain social relationships. On the other hand, men are likely to use talk to get things done. However, these are tendencies only; men also try to bond and women also try to move others to action.

    Mills (2003) contests the view that women are more polite than men. She says that ‘politeness’ is not a property of utterances; it is rather ‘a set of practices or strategies which communities of practice develop, affirm, and contest’ (p. 9). Politeness requirements vary by situation and there is no overall imperative to be polite to others; we can be impolite too and other views of politeness are incorrect (see p. 276). While there may be a stereotypical, white, middle-class (and largely female) idea of what politeness is, it is not widely shared (although it is extremely influential in the literature on politeness). ‘For some women, this stereotype may be important, but for others it may be something which they actively resist and reject’ (p. 214). Politeness ‘is clearly a resource which interactants use to structure their relations with others, and they are able to be self-reflexive about their own and others’ use of politeness and impoliteness’ (pp. 245–6).

    When the two genders interacted, men tended to take the initiative in conversation, but there seemed to be a desire to achieve some kind of accommodation so far as topics were concerned: the men spoke less aggressively and competitively and the women reduced their amount of talk about home and family. A thorough review of the literature by James and Drakich (1993) showed inconsistency in the findings when fifty-six studies of talk either within or between genders were examined. What was important in determining who talked was ‘the context and the structure of the social interaction within which gender differences are observed’ (p. 281). James and Drakich add (pp. 302–3):

    “...women are expected to use and do use talk to a greater extent than men to serve the function of establishing and maintaining personal relationships (this is not surprising, as the responsibility for interpersonal relationships primarily rests with women); for example, as we have observed, women, to a greater extent than men, are expected to talk, and do talk, simply in order to keep the interaction flowing smoothly and to show goodwill toward others, and they are expected to talk, and do talk, about personal feelings and other socioemotional matters relevant to interpersonal relationships to a greater extent than men . . . what is particularly important in female friendships is the sharing of intimate feelings and confidences through talk, whereas in male friendships the sharing of activities is more important.”

    Another interesting claim is that in cross-gender conversations men frequently interrupt women but women much less frequently interrupt men (Zimmerman and West, 1975). James and Clarke (1993) looked at fifty-four studies that addressed the claim that men are much more likely than women ‘to use interruption as a means of dominating and controlling interactions’ (p. 268). They report that the majority of studies have found no significant differences between genders in this respect and both men and women interrupt other men and women. However, according to James and Clarke, ‘A small amount of evidence exists that females may use interruptions of the cooperative and rapport-building type to a greater extent than do males, at least in some circumstances’ (p. 268).

    Still another claim is that there is evidence that in cross-gender conversation women ask more questions than men, encourage others to speak, use more back-channeling signals like mhmm to encourage others to continue speaking, use more instances of you and we, and do not protest as much as men when they are interrupted. On the other hand, men interrupt more, challenge, dispute, and ignore more, try to control what topics are discussed, and are inclined to make categorical statements. Such behaviors are not characteristic of women in conversations that involve both men and women. In other words, in their interactional patterns in conversation, men and women seem often to exhibit the power relationship that exists in society, with men dominant and women subservient.

    If different behaviors are sometimes found in cross-gender communication, what do we find within same-gender groups? Coates (1996) discusses conversation among women friends. She analyzed over nineteen hours of recorded conversation among women interacting socially in small groups. Coates admits that she is no longer a ‘dispassionate investigator’ of language. She is a middle-class woman and feminist, and an ethnographer who puts women at the center of her work. She says (p. 39) that her work shows that among the groups she looked at ‘friendships with women are a constant in women’s lives.’ In such conversations women tell and exchange stories, constantly hedge what they say, use questions to invite others to talk, i.e., for conversational maintenance, and often repeat what others say. Such talk is collaborative and establishes a feeling of solidarity among those who use it.

    In still another study, this time one that used an experimental setting, Freed and Greenwood (1996) recorded and analyzed the casual conversations of approximately thirty-five minutes each of eight same-sex pairs of friends, four male and four female. They focused particularly on the use of you know and questions. The setting of each of the 35-minute conversations was manipulated so that each conversation provided a period of ‘spontaneous’ talk, one of ‘considered’ talk, and finally one of ‘collaborative’ talk. Freed and Greenwood found no differences in the use of you know and questions: ‘Women and men of the same speech community, speaking in same-sex pairs in the same conversational context, with equal access to the conversational floor, do not differ either in the frequency of the use of you know or in the number of questions uttered’ (p. 3). Women and men also use you know and questions for the same purposes. It is the linguistic task or the speaking situation that determines the style of speaking not the gender of the speaker. They add (p. 22) that ‘just as the communicative style of women has been overly stereotyped as cooperative, so too the verbal style of men has been overgeneralized as competitive and lacking in cooperativeness.’

    When we do observe gender differences in language behavior we are confronted with the task of trying to explain them. One explanation is that languages can be sexist. I will have a little more to say about this idea later. For now, three other claims are of interest. The first claim is that men and women are biologically different and that this difference has serious consequences for gender. Women are somehow predisposed psychologically to be involved with one another and to be mutually supportive and non-competitive. On the other hand, men are innately predisposed to independence and to vertical rather than horizontal relationships. There appears to be little or no evidence for this claim; it seems rather to be a clear case of stereotyping, which offers no more than a facile solution to a difficult problem.

    The second claim is that social organization is best perceived as some kind of hierarchical set of power relationships. Moreover, such organization by power may appear to be entirely normal, justified both genetically and evolutionarily, and therefore natural and possibly even preordained. Language behavior reflects male dominance. Men use what power they have to dominate each other and, of course, women, and, if women are to succeed in such a system, they must learn to dominate others too, women included. Men constantly try to take control, to specify topics, to interrupt, and so on. They do it with each other and they do it with women, who, feeling powerless, let them get away with it, preferring instead to seek support from other women. Consequently, since women are relatively powerless they opt for more prestigious language forms to protect themselves in dealing with the more powerful. At the same time the use of such forms serves to mark them off from equally powerless males of the same social class. Women may also have weaker social networks than men but they show a greater sensitivity to language forms, especially standard ones.

    Lakoff (1975) adopts the position that men are dominant and women lack power. Women may have to behave more like men if this unequal relationship is to be changed. Others share Lakoff’s view. For example, DeFrancisco (1997, p. 39) proposes that ‘power be placed at the centre of [feminist] analysis and that gender, race, ethnicity, social class, age, sexual orientation, and other social categories be examined as political tools of oppression.’ Crawford (1995) is another who declares that power relations best explain what happens when men and women interact linguistically. Her explicit goal is ‘to create a feminist social science for all women’ (p. 8). Talbot (1998, pp. 133–4) sounds a cautionary note: ‘A major determinant [of the dominance framework] is that male dominance is often treated as though it is pan-contextual. But . . . all men are not in a position to dominate all women.’ Furthermore, anthropologists have pointed out that women are never without power and effectively control some societies. Dominance clearly fails as a universal explanation of gendered language differences.

    The third claim, which does not actually deny the second claim, is that men and women are social beings who have learned to act in certain ways. Language behavior is largely learned behavior. Men learn to be men and women learn to be women, linguistically speaking. Society subjects them to different life experiences. This is often referred to as the difference (sometimes also i>deficit) view as opposed to the dominance view just mentioned.

    Maltz and Borker (1982) propose that, in North America at least, men and women come from different sociolinguistic sub-cultures. They have learned to do different things with language, particularly in conversation, and when the two genders try to communicate with each other, the result may be miscommunication. The mhmm a woman uses quite frequently means only ‘I’m listening,’ whereas the mhmm a man uses, but much less frequently, tends to mean ‘I’m agreeing.’ Consequently, men often believe that ‘women are always agreeing with them and then conclude that it’s impossible to tell what a woman really thinks,’ whereas ‘women . . . get upset with men who never seem to be listening’ (p. 202). They conclude that women and men observe different rules in conversing and that in cross-gender talk the rules often conflict. The genders have different views of what questioning is all about, women viewing questions as part of conversational maintenance and men primarily as requests for information; different conventions for linking; different views of what is or is not ‘aggressive’ linguistic behavior, with women regarding any sign of aggression as personally directed, negative, and disruptive, and men as just one way of organizing a conversation; different views of topic flow and topic shift; and different attitudes toward problem-sharing and advice-giving, with women tending to discuss, share, and seek reassurance, and men tending to look for solutions, give advice, and even lecture to their audiences. (See also Preisler, 1986.)

    Tannen (1990, 1993, 1994, 1998) is undoubtedly the best-known proponent of the claim that women and men have been raised to live in different sub- cultures. Consequently, ‘cross-cultural communication,’ Tannen’s words, can be difficult. In various interesting and entertaining accounts, Tannen has tried to show how girls and boys are brought up differently. Part of the socialization process is learning not only gender-related activities and attitudes but gender- related language behavior. We saw earlier in Fischer’s study (pp. 162–3) how very young children show that they have learned to act ‘like boys and girls.’ Gender differences in language become established early and are then used to support the kinds of social behavior males and females exhibit. It is mainly when males and females interact that the behavior each uses separately becomes noticeable. As Holmes (1992, p. 330) says,

    “The differences between women and men in ways of interacting may be the result of different socialisation and acculturation patterns. If we learn the ways of talking mainly in single sex peer groups, then the patterns we learn are likely to be sex- specific. And the kind of miscommunication which undoubtedly occurs between women and men will be attributable to the different expectations each sex has of the function of the interaction, and the ways it is appropriately conducted.”

    One consequence of such differences is that women’s speech has often been devalued by men, for, as Tannen rightly observes, her difference approach in no way denies the existence of male dominance (1993, p. 9). Tannen’s solution is an interesting one, although one not without its critics. She believes that men and women should try to understand why they speak as they do and try to adapt to each other’s styles. However, the self-help nature of her 1990 book You Just Don’t Understand might seem to thrust much of such work onto the shoulders (or tongues?) of women rather than men. Although by no means as big a best-seller as John Gray’s Men are from Mars, Women are from Venus (1992), Tannen’s book was widely acclaimed, so its message obviously spoke to many people, women in particular. As Talbot (1998) observes of the book, with its appearance of objectivity and neutrality and its stress on differences and equality, Tannen’s approach provides a ‘comfortable explanation’ (p. 139) for some troublesome issues.

    A variation of the third claim is found in the concept of ‘community of practice.’ According to Eckert and McConnell-Ginet (1998), gender issues are essentially complex and not easy to separate from other issues. They deplore (p. 485) the fact that too often,

    “Gender is abstracted whole from other aspects of social identity, the linguistic system is abstracted from linguistic practice, language is abstracted from social action, interactions and events are abstracted from community and personal his- tory, difference and dominance are each abstracted from wider social practice, and both linguistic and social behavior are abstracted from the communities in which they occur.”

    In order to understand what is happening when people acquire and use language, we must try to understand the various communities of practice in which people function. They explain this concept as follows (p. 490): “A community of practice is an aggregate of people who come together around mutual engagement in some common endeavor. Ways of doing things, ways of talking, beliefs, values, power relations – in short, practices – emerge in the course of their joint activity around that endeavor. A community of practice is different as a social construct from the traditional notion of community, primarily because it is defined simultaneously by its membership and by the practice in which that membership engages. Indeed, it is the practices of the community and members’ differentiated participation in them that structures the community socially”

    They add that various kinds of differences arise in such circumstances, including gender differences: ‘gender is . . . produced and reproduced in differential forms of participation in particular communities of practice. . . . The relations among communities of practice when they come together in overarching communities of practice also produce gender arrangements’ (p. 491). Individuals participate in various communities of practice and these communities interact in various ways with other communities. Since these processes of participation and interaction are constantly changing, there is also constant reshaping of both individual identity and any kind of group identity, including gender identity. You must learn to be a jock or a burnout, a particular kind of man or a particular kind of women, and, as we will see shortly (p. 332), any other kind of socially categorized or gendered person. Individual identity is created through interaction with others and, as interactants change, so may identity.

    If either of the last two claims is correct, we must be prepared to acknowledge the limits of proposals that seek to eliminate ‘sexist’ language without first changing the underlying relationship between men and women. Many of the suggestions for avoiding sexist language are admirable, but some, as Lakoff points out with regard to changing history to herstory, are absurd. Many changes can be made quite easily: early humanss (from early man); salesperson (from salesman); ordinary people (from the common man); and women (from the fair sex). But other aspects of language may be more resistant to change, e.g., the he–she distinction. Languages themselves may not be sexist. Men and women use language to achieve certain purposes, and so long as differences in gender are equated with differences in access to power and influence in society, we may expect linguistic differences too. For both men and women, power and influence are also associated with education, social class, regional origin, and so on, and there is no question in these cases that there are related linguistic differences. Gender is still another factor that relates to the variation that is apparently inherent in language. While we may deplore that this is so, variation itself may be inevitable. Moreover, we may not be able to pick and choose which aspects of variation we can eliminate and which we can encourage, much as we might like to do so.

    As still another example that gender differences in language may be social in origin rather than linguistic we can look at a study of norms and norm-breaking in Malagasy (Keenan, 1974). Among the Malagasy, men do not put others into situations in which they may lose face. They use language subtly, try to maintain good communication in their relationships, and avoid confrontations. They are discreet, they prefer indirectness as an expression of respect, and they are considered to be able speechmakers: men’s ‘requests are typically delayed and inexplicit, accusations imprecise, and criticisms subtle’ (p. 141). We should note that many of these characteristics of men’s speech might be associated with women’s speech in another society. Therefore, how do women speak in Malagasy?

    Women do not operate with the same set of rules. They openly and directly express anger toward others. They also criticize and confront, and men use them to do this. They can be direct and straightforward, and because they can be so, they perform tasks, such as interacting with strangers, buying and selling when these require negotiating a price, and reprimanding children, which men prefer not to perform. In this society, then, it is the men who are indirect and the women (and children) who are direct. But the most interesting fact is that it is indirectness of speech which is prized in Malagasy society and regarded as ‘traditional’ and it is the men who employ it. On the other hand, ‘direct speech . . . is associated with a loss of tradition, with contemporary mores’ and it is found among women and children (p. 142). Women are definitely inferior to men in this society too, for ‘where subtlety and delicacy [which are prized characteristics] are required in social situations, men are recruited,’ but ‘where directness and explicitness [necessary at times but not prized characteristics] are desired in social situations, women are recruited’ (p. 143). Consequently, once more we can see how the speech of the genders reflects their relationship within the total society.

    The kinds of evidence we have looked at strongly suggests that men and women differ in the kinds of language they use because men and women often fill distinctly different roles in society. We may expect that the more distinct the roles, the greater the differences, and there seems to be some evidence to support such a claim, for the greatest differences appear to exist in societies in which the roles of men and women are most clearly differentiated. Since boys are brought up to behave like men in those societies and girls to behave like women, the differences are also perpetuated.

    In societies that are less rigidly stratified and in which men’s and women’s roles are less clearly differentiated, we may expect to find a reflection of this situation in the language that is used and also, if change in society is occurring, change in the language too. This is, indeed, what we do find: men and women, and even boys and girls, exhibit certain differences in language use in such cities as New York, Norwich, Reading, and Belfast. Most of those differences can be explained by the different positions men and women fill in society. Men have more power and may be more assertive; women tend to be kept ‘in their place’ but aspire quite often to a different and ‘better’ place. Women therefore appear to be more conscious of uses of language which they associate with their ‘betters’ in society, that is, those they regard as being socially superior. They therefore direct their speech toward the models these provide, even to the extent in some cases of hypercorrection, as in the example from New York City (p. 167). Women, therefore, tend to be in the vanguard of change toward the norms of the upper classes, and lower middle-class women are at the very front.

    One consequence is that sometimes we view the speech of certain women as being hypercorrect. That too is a normative-laden concept. It assumes a correct male norm and characterizes the female norm as deviant. Once again difference rather than deviance might be a better characterization, with the difference arising from the different experiences that females and males have of the world.

    Men have power, even lower-class men. They are less influenced linguistically by others and, in the case of the lower working class, may seek solidarity through the ‘toughness’ that nonstandard varieties of the language seem to indicate. If they lead in any kind of change, such change may well be away from the norm (p. 202). Again, as I indicated earlier (pp. 205–6), women may not find appropriate the kinds of solidarity that men seek through the use of a particular language or certain kinds of language. The peasant women of Oberwart in Austria seek not Hungarian-speaking peasant husbands, but German-speaking worker husbands and, in doing so, lead the traditionally bilingual peasant population away from Hungarian–German bilingualism toward German monolingualism. Women are not without solidarity; it is just a different kind of solidarity from that of men and just as normal.

    All deliberate attempts to change or modify languages to free them of perceived sexism or make them gender-neutral are a form of language planning. Some- times the goal appears to be to force language to catch up to social change, and at other times it seems designed to bring about social change through mandating language change. Whatever it is, it requires us to accept a very Whorfian view of the interrelationship of language and culture and is subject to all the difficulties of interpretation and implementation. Here is Pauwels’ (1998, p. 228) statement of a similar position:

    “The aims of many feminist LP [language planning] efforts are to expose the inequalities in the linguistic portrayal of the sexes which reflect and contribute to the unequal positions of women and men in society and to take action to rectify this linguistic imbalance. Language action . . . is social action, and to bring about linguistic change is to effect social change.”

    Some feminists want to go further than ‘cleaning up’ the language and even deny any possibility of ‘neutrality.’ Their expressed mission is to ‘reclaim’ language for themselves (see especially Lakoff, 1990, Penelope, 1990, Sellers, 1991, and Spender, 1985). Spender adopts a Whorfian view of language (see pp. 221– 8), declaring (p. 3) that: ‘Language helps form the limits of our reality. It is our means of ordering and manipulating the world. It is through language that we become members of a human community, that the world becomes comprehensible and meaningful, that we bring into existence the world in which we live.’ However, she goes much further than Whorf, asserting (p. 12) that ‘the English language has been literally man-made and . . . is still primarily under male control’ and that males, as the dominant group, have produced language, thought, and reality. Penelope argues that women should be aware of ‘the lies of the fathers’ tongues’ and of the ‘Patriarchal Universe of Discourse.’ Her view is that women should in a sense reinvent language for their own purposes, and many feminists have indeed tried to develop their own linguistic conventions, e.g., non-competitive, non-interruptive speech, in order to ‘liberate’ women. How- ever, other feminists such as Cameron (1992) do not hold such strong views. They would require intervention into language use on a grand scale. Any such intervention would have to be based not on any rational view of language behavior but entirely on ideology.

    Language and gender studies have seen an interesting development in recent years, known by such terms as queer linguistics and lavender linguistics. These studies deal with the language of non-mainstream groups such as gays, lesbians, bisexuals, the transgendered, etc., and focus on ‘sexuality’ rather than sex or gender. In fact, a major claim is that the focus on sex or gender may have been misdirected. In their book-length treatment of sexuality, Cameron and Kulick (2003) adopt a postmodern approach heavily dependent on the ideas of Derrida, Foucault, and Lacan, and argue that a concept they call ‘desire’ should play a central role in trying to understand human behavior since ‘ “desire” encompasses more than just the preference for partners of the same or the other sex: it also deals with the non-intentional, non-conscious and non-rational dimensions of human sexual life. The unconscious and irrational aspects of sexuality may not be manifested on the surface of people’s behaviour in the same way that their behaviour displays the sexual identities they have consciously chosen (“gay,” “lesbian,” “straight,” etc.)’ (p. 140). They argue that the issues of identity and power are less important, an argument that Bucholtz and Hall (2004) reject, claiming that ‘desire’ is much too vague a concept to be useful and that issues of identity and power are not only relevant but essential in any research on such language varieties. Just what the ultimate significance to the subject matter of this chapter this concern for ‘marginalized’ groups will have is difficult to pre- dict. The research has produced some findings of interest to us, e.g., Barrett’s study discussed on p. 117, and to ignore such findings would be to fall into the trap of appearing to use ‘power’ oppressively. However, only time will tell if this will ultimately prove to be a significant development.

    It is also apparent, as Freed has indicated (2003, p. 706), that ‘despite the enormity of our research results, the public representation of the way women and men speak is almost identical to the characterization provided thirty years ago.’ Too often researchers talk only to each other, research results are either ignored or misrepresented, and stereotyped views continue to influence how people think and behave.

    My own view is that men’s and women’s speech differ because boys and girls are brought up differently and men and women often fill different roles in society. Moreover, most men and women know this and behave accordingly. If such is the case, we might expect changes that make a language less sexist to result from child-rearing practices and role differentiations which are less sexist. Men and women alike would benefit from the greater freedom of choice that would result. However, it may be utopian to believe that language use will ever become ‘neutral.’ Humans use everything around them – and language is just a thing in that sense – to create differences among themselves. Speech may well be gendered but there actually may be no easy solution to that problem.

    Further Reading

    Gender is one of the great ‘growth’ areas in sociolinguistics and there is an abundant literature. Holmes and Meyerhoff (2003) provide very useful comprehensive coverage, and other useful books are Baron (1986), Coates (1993, 1996), Eckert and McConnell-Ginet (2003), Graddol and Swann (1989), Holmes (1995), Key (1996), Lakoff (1975), Pauwels (1998), Romaine (1999), Smith (1985), and Talbot (1998). Some recent collections of papers are Bergvall et al. (1996), Cameron (1998c), Coates (1998), Hall and Bucholtz (1995), Kotthoff and Wodak (1997), and Cheshire and Trudgill (1998). See Swann (1993) for the language of boys and girls, Johnson and Meinhof (1997) for ‘language and masculinity,’ Coates (2002) for the language of men, and Cameron and Kulick (2003), Livia and Hall (1997), and Leap (1996) for still other ‘gendered’ varieties of language.