It is often argued by both academic and popular writers alike that males and females have different ways of talking and evidence certainly points to this. What is more contentious however are the reasons given for this. I intend to give an overview of the differences between male and female speech highlighting some of the studies into this area. Having described the differences I will move on to the reasons suggested for these and look at the three main arguments suggested; difference theory, deficiency theory and dominance theory before arguing that these theories tend to be too constricted in their scope and that gender is simply part of a wider theory of language differences, that of accommodation. Different use of language for reasons of sex and gender are often placed into two categories, sex-exclusive differences and sex-preferential differences. Differences which are used exclusively by one sex or the other nearly all of the time are sex-exclusive differences whereas sex-preferential differences are those which are slightly less concrete in their usage and are more a question of frequency of usage. Some of the most dramatic examples of sex exclusive language within a community can be found in American Indian languages. The classic example is Yana where an extra suffix is added to the male form of words to make them longer, for example the female for "deer" is "ba" whereas the male form is "ba-na" (Holmes' 01:151). This only really takes into consideration the speaker's sex however. As Bodine (article in Thorne + Henley'75:142) has noted: "a language in which the sex of the speaker and the sex of the person spoken to are both involved in the determination of what form will be employed could logically have four variations," namely male speaking to male (MM), male to female (MF), female to female (FF) and female to male (FM). In practise no language differentiates between all four of these possible forms (article in Thorne and Henley'75:142). Biloxi however, part of the Siouan family of languages and the language of the American Indian Biloxi tribe (web ref 1), is one language that does differentiate between three of the possible four forms. Even in Biloxi however these three forms of differentiation are only in the imperative form of verbs, in direct address, where FF and MF speech use one form and FM and MM both use other different forms (Dorsey and Swanton, 1912 cited by Bodine in Thorne+Henley'75:143). Even more dramatic differences than these however occur in Amazon Indian languages where men and women actually speak different languages rather than just using varying forms of one language. This is because in this type of community men must marry outside of their own tribe to avoid problems with incest for example and as each tribe has a different language the woman will speak a different language to the man. There have been many different investigations into sex-preferential language. One area that has been studied many times is that of male and female use of the vernacular or the standard form of a language. Trudgill's Norwich study (article by Trudgill in Thorne + Henley'75:89) covered 20 variables including the "-ng" variable (whether the "-ng" at the end of words such as "talking" was pronounced in the standard English form [?] or in the vernacular [n].) He recorded an equal number of males and females of a number of different classes speaking in 4 different contexts; reading from a word list, reading a passage, formal speech and casual speech. For the "-ng" variable and in fact "the vast majority of the other nineteen variables,"(Trudgill'75:91 in Thorne+Henley) it was observed that women throughout all the different classes and speech styles, used the standard English forms much more than men who by contrast used more vernacular forms. Tag questions are another variable that has been looked at lots by linguists. It was believed by linguists such as Lakoff that tags showed women's hesitancy and lack of assertiveness in language. By using tag questions such as "They played badly today, didn't they? " women were seen as failing to assert themselves. Hedges such as "you know" and fillers such as "um's and er's" were also identified as characteristics of female speech and female tentativeness. Whilst the (relatively little) follow up data collected does seem to show that women use more tags and hedges than men, the interpretation of this can be challenged. Whilst tags do sometimes show hesitancy, with the right prosodic markers they can just as easily be used for the opposite purpose; "You won't do that again, will you? " or "You're a nasty piece of work, aren't you? " (Goddard 2000:97). Janet Holmes has supported Lakoff's assertions that women used more tag questions than men but also showed that only 35% of tags used by women could be associated with uncertainty compared with 61% of mens' tags (Holmes' 01:290). Instead women used tags as what Holmes termed "positive politeness devices" (Holmes 01:290) such as when a teacher prompts a child's answer in class; "It's 36, isn't it? " This shows not only how tags can have different purposes but perhaps that women's language aims to be more polite than men's. Interruptions are another area that has been investigated, most famously by Zimmerman and West whose research showed that men interrupted women more than women interrupted men. They argued that men denied women the rights "to the full utilization of their turns and support for the development of topics" (article in Thorne and Henley1975:125) by using a mixture of interruptions to break off female speech and minimal responses to discourage women from carrying on with a topic they've initiated. Their experiment showed that in mixed-sex conversations men interrupted overwhelmingly more, being responsible for 100% of the overlaps and 96% of the interruptions (article in Thorne and Henley1975:116). As with tag questions, not all incidents of simultaneous speech are detrimental to the original speaker and can instead be used to facilitate conversation, for instance a person saying "uhuh" or similar to indicate agreement and encourage the speaker to continue. However Zimmerman and West went on to show that male's minimal responses had a mainly negative effect on conversation. Based on the assumption that there was nothing inherent in conversation that makes it more likely that one speaker would fall silent before replying to an utterance more than the other they found that 62% of the females' aggregate silence followed three types of events in the preceding turns: "(1) a delayed 'minimal-response' by the male; (2) an overlap by the male; (3) an interruption by the male" (article in Thorne and Henley1975:125). Long silences in conversation are not normal or indeed desirable and therefore we can say that the male simultaneous speech is having a detrimental effect on the speech of females. The reasons suggested for these sex-preferential differences are often fitted into three broad categories, deficiency theory, difference theory and dominance theory. Deficiency theory argues that females speak a different version of the language spoken by men and crucially, that this separate version is inferior to the male version. Women are seen to have an inferior status in society and their language reflects this as the styles they used led to them "colluding in their own subordination by the way they spoke" (Lakoff quoted in Holmes'01:284). However it is slightly misleading to categorise the features Lakoff identifies as showing subordination. As I showed above, forms such as tag questions can be used in many different manners and can even assert a point like the "You won't do that again, will you? " example. Other examples show tags and hedges to be more polite and facilitative to conversation than negative and uncertain. Clearly there is more at work here than has been allowed for. O'Barr and Atkins' 1980 research into courtroom language concluded that to label tags and other such variables as 'women's language' is inaccurate and that instead they were features of powerless language which can apply to both sexes. (Coates' 04:109.) Dale Spender outlined her dominance theory in her 1980 work "Man Made Language." She argued inequalities between the sexes in society were both reflected and constructed by inequalities in language. This is seen firstly by the fact that the language system was constructed by man and is biased towards males and secondly that males have a dominant speech style compared to female's more supportive and submissive style. This first idea, whilst certainly interesting is not relevant here, I am more concerned with the different way men and women use the same linguistic variables rather than how this language was constructed in the first place. The second part of the argument is essentially a repeat of the ideas of deficiency theory. Difference theory challenges the idea that women's speech is unequal to that of males. It claims that males and females have different speech goals and hence different speech styles, but that these are equal. It characterises male language as being competitive, assertive and direct whereas female language tends to be more co-operative, affective and supportive but crucially - not submissive. Despite often being seen as one of the main proponents of difference theory, Jennifer Coates' "Women, Men and Language" actually seems in places to accept that women are dominated in their interaction with men. In summing up her chapter on mixed-sex conversation she says "The research referred to in this chapter presents a consistent picture of male-female relations, with men dominating talk in a range of environments" and then goes on to comment "by and large, sociolinguistic research into mixed (sex) talk exposes the fact that men and women do not have equal rights to the conversational floor" (Coates2004:124). I find it hard to see that if you accept that men dominate the conversational floor how you can then argue that men's and women's conversational styles are different but equal. Whilst interesting in their own rights, I believe all three theories are flawed in that they don't satisfactorily realise the idea that other factors could be involved in these differences. Language is not a constant, men and women do not talk to each other the same way all of the time and therefore it is wrong to simply argue that men talk one way and women talk another. I am likely to talk to a female friend and a female employer in a different style and surely therefore other factors aside from gender must affect speech such as power as O'barr and Atkins realised. We need to look at the bigger picture of why people as a whole talk differently. Certainly some of the time, differences between male and female speech will be most influenced by the sex of the participants but we must accept that there are other factors which also influence differences. These factors are part of the interlinked concepts of power, context and accommodation. In all speech, whether consciously or not, we tailor our speech to fit the occasion. We can see an example of this by the fact that we notice when people don't do this and speak inappropriately for the situation. Sometimes it is gender stereotypes that we accommodate to most and other times it is other things. Before I go any further into this, a look at the idea of gender stereotypes and constructs is necessary. Gender stereotypes pervade people's consciousness from a young age, informing the way we act as well as speak. Whilst maybe not as blatantly obvious as times in the past it would be wrong to think these no longer exist, for evidence we need only to look at the reaction of people when these stereotypes are subverted; we think it odd and as such prove their existence. As Coates comments; "in their talk, men and women can be seen to align themselves with dominant norms of masculinity and femininity" and adds "speakers are not free agents" and that "it is more difficult to do masculinity or femininity in ways which challenge or subvert the dominant gender norms" (Coates2004:143). Trudgill's Norwich study revealed men and women trying to conform to their gender norms. When asked to say how many standard forms they thought they had used in their speech, males under-reported the number, (they said they had used less than they had) and women over-reported the number (they said they had used more than they actually had). This behaviour can be explained by the ideas of gender stereotypes. Men it would seem, prefer to be seen using the more rough and ready non-standard forms which they associate with a macho image and masculinity, whereas women wish to be seen to be using the standard forms which correspond with the stereotype of femininity. Edelsky's research (cited in Coates2004:159) into child language shows how from a relatively young age children can associate certain types of language with the different genders further proving the existence of gender stereotypes. As children grew older they got better at identifying examples of speech traditionally associated with either males or females. The fact that they couldn't identify all of these at earlier ages shows how children "learn" the appropriate gender norms as they grow up rather than being born with them ingrained, that is to say that rather than gender differences simply being part of human nature and something we are born with an innate sense of, they are actually something we are socialised into as get older. I claimed earlier that language differences are caused by people accommodating to a number of different norms. We can see gender as part of a bigger picture as in certain contexts role reversal occurs, and men will find themselves using forms associated with female talk and vice versa indicating that there is something bigger than gender causing language differences. This idea of accommodation manifests itself in many ways. If we look at the data from Trudgill's Norwich study we can see that for every class and gender, the standard form is used more for formal speech than in casual speech (Trudgill in Thorne+Henley'75:91). The only thing to change between the two situations is the formality of the setting and so we can surmise that it is the context (i.e. the formality) of the situation that made people change their language. The context of the situation is made up of many things. Precisely because it is made up of so many strands is why it is misleading to simply look at gender as being the sole reason why men and women speak differently. This interpretation of accommodation is a broadening out of the theory laid out by Howard Giles and his colleagues. Whereas Giles has often looked at specific situations where accommodation occurs I believe that it applies much more generally and that people are accommodating all of the time, often completely subconsciously. I believe Giles' basic principle "an individual can induce another to evaluate him more favourably by reducing dissimilarities between them" (Giles and Powesland1975:157) is applied more broadly across conversation as a whole. Whilst illustrating the process of accommodation well, this phrase sounds a little manipulative for my liking, I believe the process of accommodation to be much more subconscious and wide-ranging. We seem to be socialised into knowing the norms and expectations of certain situations and tailor our language accordingly to. This is why we feel awkward when people break these 'rules' or when we are presented with an entirely new situation which we don't know the 'rules' of. In conclusion, men and women often talk differently. The reason for this is large and complicated but I believe it is based on accommodation. People are constantly accommodating and tailoring their language to meet the context of the situation. There are many other factors apart from gender that affect language and that we accommodate to. Even when gender is the main factor affecting our speech it is still to do with accommodation - we are talking differently not due to differences of sex but due to differences of gender; we are accommodating to gender stereotypes. The three major theories ignore the idea that other factors could be involved. When they talk about gender stereotypes they ignore the root of the matter; yes, gender stereotypes are important but our conforming to them is part of the process of accommodation. The theorists don't acknowledge this and as such are guilty of observing an effect without truly acknowledging what is causing it. Deficiency theory for instance argues that women's language reflects their subordinate situation in society. This may be the case but this is a theory based more on the outcome of men's and women's language rather than the causes of it. In this case the norm to which people are seen to be conforming to is that of old gender stereotypes. Arguments over the positive or negative outcomes of this accommodation process are fascinating but the root of the process is ignored as are other factors which can turn gender stereotypes on their head. As such, the theories often don't recognise the fact that both sexes talk differently in different contexts and rather seem to treat language as a constant where men and women talk in the same way at all times. What is exciting about the idea of accommodation is that the things we accommodate to are constantly changing and evolving. Our idea of gender stereotypes for instance are constantly slowly changing even over relatively short time periods. Therefore we can see language differences between men's and women's speech as a product of accommodation and tailoring of language to different norms. These different norms encompass many different areas and build up the context the conversation is being held in. Sometimes this tailoring of language is mainly affected by gender and sometimes other factors affect it more. It is probable that some norms are in fact more normal than others, and affect speech more often but it is enough to simply say: Men's and women's speech is affected by a number of different norms of which gender is an important, but not absolute, part. 