5gendpolite4.doc; Oct 2002 word count: 13,276

Chapter 5: Gender and Politeness

Introduction

Given the model of gender described in the last chapter where gender is dispersed into practices and contexts/communities of practice, and is oriented to by individuals in relation to their perceptions of race and class stereotypes, rather than being located at the level of the individual, and also given the model of linguistic politeness as described in Chapters 2 and 3, where politeness/impoliteness is considered to be a matter of assessment within a context and the result of complex negotiations with perceptions of appropriacy norms and stereotypes, it is difficult, if not impossible, simply to approach the relation between gender and politeness as a question of an investigation of the production, by individual men or women of a number of linguistic features which are assumed to be unequivocally polite or impolite. What I would like to do instead is to consider the complexity of the relationship between gender and politeness, so that the common-sense nature of gender and politeness and their relation to each other is troubled. Here, I aim to analyse the way that certain practices which are considered to be polite or impolite are, within particular communities of practice, stereotypically gendered. As I discussed in Chapter 4, these stereotypes do not actually exist as such, but are hypothesised by particular speakers and hearers within communities of practices, on the basis of their representation by others, and are then negotiated with. It is this connection between gendering of practices and assessments of politeness and impoliteness which is of interest. These stereotypes of behaviour which are considered to be appropriate within particular contexts feed back into individual participants' assessments of what is appropriate in terms of their own behaviour.

Firstly, in this chapter, I analyse stereotypes of gender and politeness, and then move on to a discussion of the theoretical work on gender and politeness which I argue seems to replicate stereotypical views of women's politeness, rather than describing women's or men's actual linguistic performance or interpretative frameworks. I then consider two aspects of linguistic behaviour which have often been stereotypically associated with women: compliments and apologies, and I analyse two extracts from conversational data in order to challenge any simple view that women are necessarily always more polite than men.

Stereotypes of gender and politeness

As I have argued throughout this book, at a stereotypical level, politeness is often considered to be a woman's concern, in the sense that stereotypes of how women in general should behave are in fact rather a prototypical description of white middle class women's behaviour in relation to politeness. The teaching and enforcement of `manners' is often considered to be the preserve of women. Femininity, that set of varied and changing characteristics which have been rather arbitrarily associated with women in general, and which no woman could unequivocally adopt, has an association with politeness, self-effacement, weakness, vulnerability and friendliness, as I showed in the last chapter. This manifests itself in the type of language practices which Lakoff described as `talking like a lady'. (Lakoff, 1975:10) Rather than considering that this is an accurate representation of `women's language', this chapter will investigate other ways of thinking about the relationships between gender and politeness. What I would like to contest is the reifying of this view of the stereotypical behaviour of a group of women, and the extension of such a stereotype to all women. Women's linguistic behaviour is often characterised as being concerned with co-operation (more positively polite than men) and avoidance of conflict (more negatively polite than men). This characterisation is based on the assumption that women are powerless and display their powerlessness in language; these forms of politeness are markers of their subordination. However, as I mentioned in the last chapter, stereotypes of gender have been contested for many years by feminists and have themselves been changed because of the changes in women's participation in the public sphere. We can therefore no longer assume that everyone has the same `take' on a stereotype, or that they share assumptions with others about what a particular stereotype consists of, or even that they accept stereotypes at face value rather than for example ridiculing them. Neither can we assume that certain forms of politeness are unequivocally powerful or powerless.

Particularly in relation to gender stereotypes, politeness and impoliteness operate in different ways, so that impoliteness functions and signifies differently for certain groups of women. Behaviour such as swearing and directness which might be considered impolite in certain linguistic communities, may often be excused or condoned for certain groups of men, particularly in the working classes, whereas for middle class, particularly middle aged white women, it is may be judged to be aberrant. Indeed, the epitome of stereotypical language behaviour for males and females seems to be white working class men (direct, assertive, impolite) and white middle class women (polite, deferent, `nice' to others). At a stereotypical level politeness is largely associated within Western countries with middle class women's behaviour. This may well be because politeness itself is generally considered a civilising force which mitigates the aggression of strangers and familiars towards one another and ideologically this civilising move is often associated with femininity. [i] Masculinity, on the other hand, is stereotypically associated with directness and aggression. From the questionnaires and interviews which I carried out, middle class and working class white women themselves, particularly those who did not have paid employment outside the home, tended to find impoliteness of greater import than other groups, and slights and perceived lack of the appropriate level of politeness were often perceived by these women as a greater problem than for other groups. [ii] Although, that is not to say that other groups of women or men do not find lack of politeness important, (certain of the middle class white men I interviewed also stressed that they also valued politeness and found impoliteness difficult to deal with); however, white middle class women more than others focus on it as crucial in terms of assessing their everyday relations with others. This association between gender and class imperatives on politeness results in concerns about class separateness and gender coalescing. `Correct' forms of behaviour for women, within assessments of politeness, particularly the more formalised views on politeness which appear in books on courtesy and etiquette, are also `correct' forms of middle class behaviour. Those elements which are judged to be not polite behaviour, that is examples of behaviour which are not acceptable in `polite (i.e. middle class) circles', are those which are considered `common', `vulgar' or which are not `decent'. These terms have a particular resonance for middle class women and those working class women aspiring to middle class values. Thus, if we analyse some of the advice given in etiquette books, politeness is seen to be avoidance of behaviour which is associated with the working classes: `never, never scrape the plate, unless you want to give the impression that you are normally ill-fed as well as ill-bred ' (Bolton and Bolton, n.d. 22). [iii] Certain foods and drinks are associated with the working classes and thus, `fizzy soft drinks should not be offered as pre-dinner drinks' and `it should not be necessary to say that bottled sauces should not appear on the table' (Bolton and Bolton, n.d.94). This concern with avoiding stereotypically working class behaviour is seen as avoiding behaviour which is `bad form' and, as the reference to `ill-bred' above clearly shows, polite behaviour is that which is normally associated with middle class behaviour, which is itself characterised as `correct' and `good form'. (Bolton and Bolton, n.d. 73) .[iv]

Thus, politeness or concern for others is stereotypically associated with middle class behaviour. Furthermore, politeness is often associated within English speaking communities with being deferent which Brown and Levinson have classified as negative politeness, often associated with powerlessness, and care for others, which is associated with stereotypes of femininity. Women's linguistic behaviour, in many accounts, because it is seen as displaying powerless is characterised as hesitant and unassertive and showing negative politeness for others through what is seen to be excessive use of respect and deference. These characteristics associated with deference and positions of unequal power become associated with the `natural' behaviour of women and other subordinate groups, as Hochschild argues: `The deferential behaviour of servants and women - the encouraging smiles, the attentive listening, the appreciative laughter, the comments of affirmation, admiration or concern - comes to seem normal, even built into personality rather than inherent in the kinds of exchange that low-status people commonly enter into' (Hochschild, cited in Scott, 1990:28). Positive politeness is associated with being friendly and nice towards others, and because there is an association between white middle class women's stereotypical behaviour and niceness, this form of politeness is then associated with women' linguistic politeness in general. [v]

Courtesy is a form of behaviour which is conventionally extended from white middle class men to white middle class women and consists of behaviour such as opening doors for women, helping them on with their coats, walking on the outside on the pavement, giving up a place on the bus, and so on. [vi] This form of behaviour is anachronistic for many people in Britain, and the display of care for others within the public sphere is something which the Campaign for Courtesy, a pressure group which tries to encourage others to be more courteous, feels has deteriorated over the last twenty years (Gregory, 2000a).[vii] Even when courtesy was more accepted as a general standard of behaviour during the 1950s -1970s, it was characterised as a set of values which was contested or under threat because of the behaviour of certain groups or classes. Courtesy is generally subsumed in most people's accounts under the general heading of politeness, especially for many older people; one book on etiquette from the 1970s states: `Politeness between the sexes is particularly important. Both men and women err greatly in their behaviour towards one another. The attitude of many men is "they wanted equality, and they've got it", and this is their excuse for letting a door slam in a woman's face or failing to walk on the outside of the pavement when they are together, or paying any of the usual forms of courtesy. Women on the other hand, often adopt a rude and militant attitude towards men, and their behaviour when courtesy is shown them is sufficient to discourage even the most courteous male. Again, so many women neglect to return thanks for politeness, and behave so impertinently that any man might be forgiven for not bothering in the future' (Bolton and Bolton: n.d. 34). Courtesy is certainly something about which there is a considerable conflict between professional middle class women and professional men. At present it is very difficult to negotiate having courtesy extended to oneself as a woman without offending by refusing. For those men who open doors for women, courtesy may be intended as a display of care and respect, whereas for many feminists, it seems demeaning because it seems to suggest that the woman is incapable of opening doors, putting on her own jacket, and carrying her own bags, and it also seems to be non-reciprocal. [viii] After all, it seems to suggest that the women involved could not perform such actions as opening doors, carrying suitcases, etc, for themselves, even though that might not be the intention. For example, in a book on etiquette from the 1970s, the following sentiment is indicative of a wider set of advice to men involving displaying care for women `Going up or down stairs, the rule is for the man to keep to a lower level so that he may offer assistance if required; i.e. he follows the woman up a stairway, but precedes her downstairs'.(Bolton and Bolton, n.d. 34) This characterisation of courtesy assumes that women need protecting by men. Thus, courtesy and etiquette, like other forms of politeness, can be seen to have particularly strong class and gender associations at a stereotypical level.

Theoretical Work on Gender and Politeness

The major theoretical work on politeness and gender has been undertaken by Penelope Brown and Janet Homes and I would like now to discuss their work in order to show how each of them draws on these stereotypes of women's and men's behaviour in relation to politeness.

Penelope Brown in her work on the analysis of politeness amongst a Mayan community, argues that women in general are more polite than men (Brown, 1980; 1993). She states that `in most cultures women among women may have a tendency to use more elaborated positive politeness strategies than men do among men’ (Brown, 1980: 251). Her general model of politeness is one associated with care for others: `what politeness essentially consists in is a special way of treating people, saying and doing things in such a way as to take into account the other person’s feelings. On the whole that means that what one says politely will be less straightforward or more complicated than what one would say if one wasn’t taking the other person’s feelings into account’ (Brown, 1980: 114). This statement sees politeness as largely a matter of concern for others. She asserts that this greater use of positive politeness by women is due to power differences within this particular Mayan community, but that power differences can be seen to produce similar behaviour in other cultures: `men may assimilate more upper-class dignity and competition for power, while women, excluded from this arena, maintain solidary ties with one another’ (Brown, ibid). However, as I shall show, her results are significantly more complex than this and seem to show just how difficult it is to prove that men as a group or women as a group use politeness in similar ways.

She discusses the way that many linguists have concluded that women’s language tends to be more hypercorrect than men’s and hence more formal (Brown, 1980). As I mentioned in Chapter 4, linguists such as Trudgill (1972) claim that this is because women tend to gain prestige through appearance and linguistic behaviour, since they cannot gain status through their job or income (Trudgill, 1972). [ix] The assumption that is made is that hypercorrectness and use of the prestige variety in English can be assumed to be markers of polite linguistic behaviour, and that this type of behaviour marks an unstable or insecure social position; thus Brown argues: `it seems reasonable to predict that women in general will speak more formally and more politely, since women are culturally relegated to a secondary status relative to men and since a higher level of politeness is expected from inferiors to superiors’ (Brown, 1980:112). Here, Brown seems to be conflating politeness and negative politeness or deference, and she also seems to be assuming that an inferior social position will necessarily determine the type of language which is produced. However, she goes on to give the example of the Malagasy village studied by Elinor Keenan where women’s speech is judged to be less polite than men’s, but in this instance, this type of speech is stigmatised by the society as a whole. Therefore, here stereotypes of women's speech are assessed rather than women's actual speech. Brown's work focuses on speaker intentions and she does not concern herself overly with hearer interpretation and judgement, which is clearly crucial in concerns about status in this community.

Brown sees politeness as being concerned with questions of social standing and this she sees as being of great importance for women. For her, since relationships in general are fairly stable, politeness levels are also fairly predictable. If there is a shift in the level or type of politeness used then we are to assume that there has been a change in respect, an increase in social distance or a change of a face threatening nature. She argues that therefore most fluctuations in politeness levels are due to the mitigation of a Face Threatening Act. `Given then a range of politeness levels over a wide range of kinds of acts, we can infer degrees of social closeness and degrees of relative power in relationships. Thus, politeness strategies are a complicated and highly sensitive index in speech of kinds of social relationship’ (Brown, 1980:117). She goes on to ask why and how women are more polite than men and she suggests that `women are either 1) generally speaking to superiors, 2) generally speaking to socially distant persons, or 3) involved in more face-threatening acts, or have a higher assessment than men have of what counts as imposition’ (Brown, 1980:117). Whilst this may be correct on a stereotypical level, I would argue that in fact in relation to women's linguistic behaviour as a whole, these assertions do not necessarily hold.

In Brown's analysis of strengthening and weakening particles in Tenajapa, she asserts that the particles which weaken an utterance, hedging on its epistemic value, can be seen to play a role in negative politeness, and those particles which strengthen an utterance can be seen to play a role in positive politeness. As I have shown in earlier Chapters of this book, making this assumption is fraught with difficulties, since the interpretation of the function of particular language items can only be understood within a particular context, judged within the framework of the hypothesised norms of a particular community of practice. However, Brown asserts that in this Mayan community which is very clearly sex-differentiated, women use more strengthening particles when speaking to women (and to men) and they also use more weakening particles when speaking to men. Women speaking to women use more particles in general than men use to men. This is an important finding since Brown shows that she is aware that women do not have a simple general style which all of them use in all circumstances; rather, their choice in terms of the use of these particles depends on the assessment of context and audience. Furthermore, when making this generalisation, Brown is aware of the importance of topic and relative knowledge in relation to the number of these particles which are included within speech, and the fact that when value judgements are given it is expected that these particles occur more than when evaluations are not being made.

However, despite these provisos, she still asserts that in general women's language use is more polite than men's. She gives, as a case study, the use of the particle `ala’ (a little) which she asserts is a diminutive which seems to her to be functioning both as a negative politeness element when used by males in her examples, but also as a positive politeness element when used by females in her examples. Despite this awareness of the fact that `ala' functions differently for different groups, she asserts that the use of `ala’ is seen to be characteristic of women’s speech. Holmes, whose work I discuss later in this chapter, comments on Brown's analysis of these particles : `the particles she examined tend to occur most frequently in speech expressing feelings and attitudes, and … in her data women spent more time talking about feelings and attitudes towards events than men. It seems possible that the association of particular linguistic devices with women’s speech may reflect the fact that they occur more often in discourse types favoured by women’ (Holmes, 1995: 110). Thus, Holmes seems to be viewing women's language at a stereotypical level, in that she assumes that women more often talk about emotions than men. As I discussed in Chapter 3, the assumption that women's speech is necessarily different to that of men often leads us to draw on stereotypes of feminine behaviour rather than on women's actual linguistic behaviour.

In general, even though there are several cases where Brown's hypothesis is not proven when tested against her data, she still asserts that women and men’s speech differs significantly in relation to politeness use. Her results `contradict our initial impressionistic hypothesis, that women are positively polite to women, and negatively polite to men. Rather the data suggest that women are overall more sensitive to possibly face-threatening material in their speech, and hence use negative politeness to women as well as men, and are more sensitive to positive face wants and hence use positive politeness to men as well as women’ (Brown, 1980:129). She suggests that this negative politeness use between women is due to the fact that `there is not a dichotomising of the social world into men vs. women ... but that overall women are paying more attention to face redress than men are’ (Brown,1980:131). Thus in her conclusions, Brown is aware of the complex interaction between social status and gender and is also aware that politeness markers (if they can indeed be classified as such) have a range of different functions for those of different status and gender. However, because of stereotypical assumptions about the homogeneity of these speech communities she is unable to challenge the notion that women's speech will necessarily be speech that displays deference.

Because of her concerns to present these Mayan women as homogeneous in terms of their speech, she notes that despite the clear differences that she has described in sex-roles (men beat women, but women do not beat men; men use direct orders in public but women do not) in fact, this community is fairly egalitarian and women are not totally powerless, since they play a major role in the community both economically and socially. She herself notes that the women’s speech community in Tenejapa is not uniformly powerless, for example, that women over childbearing age tend to speak more in line with the norms for men and some of them manage to achieve some powerful positions within the community. She argues that in communities where women have no social esteem, for example, low caste women in India, their linguistic behaviour in relation to negative politeness will be different, ranging from highly honorific language to bald on record, since Brown argues that these people are treated as if they have no face. However, if politeness use is so closely linked to questions of status, those women who have some status in the community should display this in their language use, and there should be more a sense of diversity within women's language use within Brown's account.

Brown refines her position even more in her 1993 article on `Gender politeness and confrontation on Tenejapa', where she argues that within particular contexts, in this case the courtroom setting, women may use the speech forms most associated with feminine politeness, sarcastically in order to perform Face Threatening Acts. Whilst still affirming that women are more self-deprecating than men within the public sphere generally, and more positively polite than men, she analyses the way that women in a courtroom confrontation draw on these resources of feminine behaviour in order to behave aggressively to other women, ` despite the strong constraints against public displays of anger, there is an institutionalised context and mode for confrontation: a dramatised outrage played against the backdrop of appropriate norms for female behaviour' (Brown, 1993: 137). Thus even when women are not being polite, since they are being aggressive, they do this by drawing on indirectness and politeness ironically.

Her conclusions from this work are that deference prevails when people are vulnerable within a society; thus, women in such a position will use more negative politeness. Positive politeness prevails if and when networks involve multiplex relations, where relations are multistranded. Cameron comments: `Brown's argument, however, is not that politeness works differently for men and women. It is that while both sexes must make the same calculations about the same variables (e.g. social distance, relative status, degree of face-threatening inherent in a communicative act), the different social positioning of men and women make them assign different values to those variables, and therefore behave differently. If Brown had explained the women's `more polite' behaviour as a simple consequence of either their feminine gender or their powerlessness she would not have been able to explain the fact that they are differently polite to male and female interlocutors (if it were only femininity, why should there be any difference ? If it were only powerlessness, why be polite to your equals - other women - at all ?) ' (Cameron, 1998a:444) Whilst I accept that in this way Brown's work integrates a certain element of heterogeneity within her notion of women's speech, she still characterises women are largely powerless, whilst showing at the same time that within this community, there are women who exercise greater interactional power, within particular contexts.

A similar finding can be seen in Smith-Hefner's (1988) analysis of the use of polite forms in Java, where she notes that different cultures have different definitions of what counts as polite; she argues that in Java polite forms are associated with high status and with linguistic control and skill. `Where the register is linguistically complex and not everyone is able to use it effectively, mastery of the register may thus identity the speaker as distinguished or socially exemplary... Under such circumstances, we should expect that control of and access to these polite codes would show a pattern of differential distribution among speech community members often relation to variation by class and gender' (Smith-Hefner, 1988: 537). Because of this association of politeness with high status and verbal skill, although Javanese women are expected to be more polite than men within the family and this use signals their subservience (an assertion which seems to support Brown's assertion about Tenejapan women), outside the family within the public domain politeness is associated with males.

Brown and Levinson devote a section of their revised 1987 introduction to the question of gender and here Brown's argument is extended still further so that gender becomes an even more complex variable. They argue: `empirical tests of Lakoff's specific claims (that women are more polite than men) have by and large failed to substantiate them in detail … but the argument that women have a distinctive `style', due to their distinctive position in society, is still being actively pursued, despite the persistence of negative evidence (no clear sex differences found) in much of the research' (Brown and Levinson, 1987:30). They assert that rather than simply analysing data for sex differences: `in trying to understand the often very elusive and subtle differences between the language use of men and women we need to be crystal clear about exactly where and how the differences are supposed to manifest themselves' (Brown and Levinson, 1987:30). For example, we need to be clear about whether we are examining differences due to the gender of the speaker or the hearer or both. Of great importance here is the assertion that simply analysing data for gender difference is not adequate, since `we need constantly to remember the obvious but always pertinent fact that gender is just one of the relevant parameters in any situation, and is indeed potentially irrelevant in a particular situation' (Brown and Levinson, 1987:30). They are aware of the difficult relationship between gender and the other social variables which they examine in their work. For them, it is difficult to assess whether gender is at work; if we assume that gender and power (P) are the same, since all women are powerless, then we will also have to take into account social distance (D) in relation to gender:. `unicausal explanations in terms of P (that women are universally subordinate to men and therefore more polite) will not do justice to the complexities' (Brown and Levinson, 1987: 30). Thus, although Brown and Brown and Levinson try to question the assumption that women are necessarily more polite than men, their data in general seems to prove that in most circumstances this is indeed the case. This is partly because their work focuses on speaker intentions and therefore can only deal with intentions to be polite rather than stereotypical assessments of politeness by others, which may be at odds with those intentions.

Drawing on Brown and Levinson's work, Janet Holmes argues that in general women are more polite than men. Her empirical studies belong to the `difference' model of women's language within feminist linguistics, influenced by Coates (1996) and Tannen’s (1991) work on co-operative and competitive strategies. Thus, Holmes asserts that women are more polite than men, as they are more concerned with the affective rather than the referential aspect of utterances since `politeness is an expression of concern for the feelings of others’ (Holmes, 1995:4). Holmes states that she uses a broad definition of politeness, following Brown and Levinson, so that politeness refers to `behaviour which actively expresses positive concern for others, as well as non-imposing distancing behaviour’ (Holmes, 1995:5). She suggests that women are more likely to use positive politeness than men; thus for her : `women’s utterances show evidence of concern for the feelings of the people they are talking to more often and more explicitly than men’s do’ (Holmes, 1995:6). I aim to contest Holmes’ notion that women are globally more polite than men, arguing that this is in fact based on a stereotypical view of women's language. For some women, this stereotype may be important, but for others it may be something which they actively resist and reject. What is important here is the sense of the variety in the hypothesisation of the stereotype and variety in the response to that stereotype in terms of what behaviour is then considered to be appropriate.

Holmes attempts to tackle the question of whether women and men are polite in different ways. When she poses the question, `are women more polite then men?’ she answers: `it depends what you mean by politeness, and it depends which men and which women you are comparing, and it also depends on the context in which they are talking’. (Holmes, 1995:1) However, this focus on the context-specific is frequently dispensed with in her work, in order to make larger generalisations. Holmes tries to suggest that there are global similarities amongst women; thus, she asserts that women generally are more likely to be verbally fluent earlier, they are less likely to suffer from reading disabilities and aphasia, but, perhaps more importantly, she asserts that women have a different attitude to language use to men: `Most women enjoy talk and regard talking as an important means of keeping in touch, especially with friends and intimates . They use language to establish, nurture and develop personal relationships. Men tend to see language more as a tool for obtaining and conveying information. They see talk as a means to an end’ (Holmes, 1995:2). This is very similar to the position advocated by other `difference’ feminist linguists, who claim that women and men are brought up in different gendered sub-cultures and thus use language in fundamentally different ways to achieve different ends (Coates, 1996; Tannen, 1991).

Holmes bases her claims on extensive quantitative data, unlike many researchers in this field. However, even though she analyses a great deal of data, it is largely from white middle class New Zealand women and men. [x] Holmes is aware of the difficulties entailed in making generalisations about women's speech from this data: `many of the statements about the way women and men use language will inevitably appear as gross generalisations from specific studies of particular men and women in particular situations who belong to specific cultures, social classes, age groups, occupational groups, and so on. It would be possible to qualify every statement with this kind of detailed information, but it would also be tedious in the extreme. In discussing the research evidence every reference to women and men should be appropriately qualified, but for obvious reasons I have tended to refer to women and men. This is not intended to mislead or over-generalise’ (Holmes, 1995:29). Throughout her work, there is a problematic tension between the specificity of her interpretation of her data and her wish to make generalisations from that data. The problem with generalisation is that it often reflects stereotypical views or our notion of the average and does not reflect the behaviour of any real speaker.

There is also a tension between the wish to make generalisations about the language behaviour of all men and all women, and recognising that there are differences in power amongst women: `in communities where women are powerless members of subordinate groups, they are likely to be more linguistically polite than the men who are in control’ (Holmes, 1995:8). One can presume that in communities where women are not powerless, Holmes might argue that they therefore would not be as linguistically polite as powerless women. After this acknowledgement of the importance of context, she goes on to state, however, that: `when all the necessary reservations and qualifications have been taken into account, I think the answer is “yes, women are more polite than men” ’(ibid.).

Holmes associates positive politeness with women's speech. In discussing positive politeness, she states, `women are much more likely than men to express positive politeness or friendliness in the way they use language’ (Holmes, 1995:6). Positive politeness is here seen to be synonymous with friendliness, and seems part of a general stereotype about the way that women should behave. However, even she recognises that distinguishing between positive and negative politeness is difficult: `in fact, there are few speech acts which are intrinsically negative politeness speech acts. Linguistically expressed negative politeness generally takes the form of expressions or strategies which reduce the effect of face threatening speech acts’ (Holmes, 1995:154). In discussing the positive politeness strategies which New Zealand women use as reported by Pilkington, Holmes states: `for the women, being negatively polite involves avoiding disagreement. Being positively polite is being friendly, and this involves confirming, agreeing and encouraging the contribution of others. But these politeness strategies are not typical of the interchanges described ... between males. These young New Zealand men ... are quite prepared to disagree baldly and to challenge the statements of others overtly. Indeed, for this group, insults and abuse appear to be strategies for expressing solidarity and mateship, or ways of maintaining and reinforcing social relationships’ (Holmes, 1995:66). However, here there seems to be little difference in effect in what each of the groups are doing: the New Zealand women are expressing solidarity in much the same way as men but they are using different strategies. As Bergvall, Bing and Freed have argued, once a researcher decides to analyse sex difference in language, they are forced to concentrate on difference alone and this evidence is used to argue that male or female behaviour are fundamentally different (Bergvall, et.al, 1996).

This conflict between the wish to generalise and the awareness of the specificity of the material arises throughout her analysis. For example, there is a conflict between allowing for the multivalence of linguistic elements and the need to do quantitative analysis, as can be seen in the discussion of tag questions, where she notes, in a rather tautological fashion, that `men generally use canonical tag questions more often than woman do to express uncertainty and ask for confirmation, while women use tag questions more often than men in their facilitative positive politeness function’ (Holmes, 1995:85). These categories are in themselves judgmental rather than analytical. This problem also arises when Holmes then goes on to examine the use of facilitative tags by those in positions of power, since she argues that `powerful participants used more of the facilitative tags’ (Holmes, 1995:85) If she is arguing that both the powerful and women use facilitative tags more, then we should be led to assume that there is a correlation between these two groups. The problem seems to have arisen because of the use of this judgmental categorisation of tags, since `facilitative' is already an evaluation of the speech being analysed. This problem occurs throughout Holmes' work in that evaluative categories are used, and the evaluations of the participants are not considered.

Holmes attempts to move analysis to an examination of functions rather than forms; however, when she finds that men and women use a form to the same extent, she interprets their use differently, according to its function: in the use of `I think', for example, she found no difference between its use by men and women in her samples, `but there was a contrast in the predominant functions for which it was used by women and men. Women used I think as a booster more frequently than as a hedge ... whilst the reverse was true for men ... and women used I think more frequently than men as a politeness device, especially as a positive politeness device, boosting an utterance expressing agreement with the addressee’ (Holmes, 1995:94) This interpretation of the data attributes an unproblematic function to the use of `I think' which may not be shared by the participants. Again, with the analysis of `sort of’ she states: `although there was little or no difference in the data analysed in the total number of instances of these forms used by New Zealand women and men, in general women used them as politeness devices more often than men, and men used them more often than women as epistemic devices in their referential function. This pattern of female concern with affective meaning and male with referential meaning ... illustrates once again the different orientations of women and men in interaction' (Holmes, 1995:96). However, this seems to be an evaluation of women's speech, and these stereotypical views of women's speech as more polite seem to be influencing the interpretation of the data.

Furthermore, if data does not prove Holmes' initial hypothesis, then she interprets the data differently in order to make it fit; for example, in an experiment she found that boys and girls use interruption in supportive ways equally. She states that `it is interesting to speculate on why the usual pattern of disruptive interruptions by males was not evident in this data’ (Holmes, 1995:54). She then indeed speculates on why boys and girls behave in similar ways in this context, for example, that they were working in pairs and hence co-operative behaviour was demanded by the task and context, or that the males had not as yet learnt male dominance patterns. In another instance, in examining comments made after presentations by females and males, she divides the comments into supportive, critical and antagonistic. Males and females were equally supportive, males were more antagonistic than women, (which she discusses in greater detail than the other results), but the higher level of women being critical of the presentations she dismisses as determined by the context, `where it was clear that criticism had been invited’ (Holmes, 1995:47). She then goes on to make a generalisation from this that: `overall then this detailed analysis ... provides further support for the suggestion that New Zealand women tend to be more sensitive to the positive face needs of their addressees than New Zealand men. (Holmes, 1995:47) In both of these cases, when data does not fit in with her overall hypothesis of difference, it can be dismissed because of context-dependent imperatives.

Holmes interprets utterances as polite largely on a grammatical basis, even though she recognises that each lexical item has multiple possible interpretations and that there is no clear unproblematic link between form and function. Holmes focuses on grammatical features such as questions as she feels that these indicate politeness; she suggests that who asks questions most is important in gender terms, and she goes on to state that questions are one way of handing the floor over to another speaker. (Holmes, 1995:31) She is aware that questions and interruptions may be differently judged by hearers: `what is perceived as rude, disruptive and impolite by women, may be acceptable and normal in male interaction. And when women politely (according to their norms) avoid interrupting others, they may be interpreted by males as being reluctant to get involved, or as having nothing to say’ (Holmes, 1995:53). Thus, what her data is clearly showing is that politeness is an evaluation of behaviour rather than an intrinsic grammatical quality, but her model of analysis, based as it is on Brown and Levinson's, forces her into a grammatical and speech act -based analysis.

Holmes interprets her data according to stereotypes of female and male. Despite the fact that her book was published relatively recently, Holmes makes certain rather outdated assumptions about women and the private sphere, for example: `the amount of talk contributed by women and men differs in public, formal contexts, compared to private, intimate contexts ... men tend to value public, referentially orientated talk, whilst women value and enjoy intimate, affectively orientated talk. Each gender may be contributing more in the situation in which they feel most comfortable. Correspondingly, women may experience formal public contexts as more face threatening than men do, while men perhaps find private and intimate contexts less comfortable. Each gender contributes least in the situation they find less comfortable’ (Holmes, 1995:37). This may in fact be an accurate assessment of some women's and men's responses to stereotypes of what is appropriate for each sex, but it is important to see this as a stereotype rather than to assume that it is a reflection of actual behaviour. [xi]

In order to examine these problems in more concrete terms, I now analyse two of Holmes’ concerns in her book, particularly as these are ones which are most associated with the notion of women’s speech, and which seem to be most stereotypically associated with positive and negative politeness: compliments and apologies.

Compliments

Holmes argues that `a compliment is a speech act which explicitly or implicitly attributes credit to someone other than the speaker, usually the person addressed, for some “good” (possession, characteristic, skill, etc.) which is positively valued by the speaker and the hearer,’ and she also asserts that `compliments are generally regarded as the paradigm of a positive politeness strategy’ (Holmes, 1995:144; Holmes, 1986, cited in Holmes, 1995:117). Therefore, a compliment for Holmes functions as an unequivocal marker of positive politeness. But compliments can also function in very different ways depending on the context: compliments can set up a debt, or they may be intended to be interpreted negatively. Holmes is unable to see compliments as anything other than positive, but she is willing to admit that `very clearly, the relationship between the complimenter and recipient is crucial in accurately interpreting the potential functions of a compliment’ (Holmes, 1995:118). She does not analyse these relationships, however. She is aware that context plays a role in whether a compliment is interpreted as an aspect of positive politeness or not, but she is forced to assume that compliments can be interpreted globally, because of her concern with generalisation and, influenced by Brown and Levinson, universal statements. Thus, whilst she admits that: `in some relationships compliments will be unwelcome because they are experienced as ways in which the speaker is asserting superiority’, her general positive model of compliments holds (Holmes, 1995:119). She lists the functions of compliments: ironic, sarcastic, flattery, patronising, expressive of solidarity, praise, envy or admiration, but despite this very mixed list, in her data she only analyses those compliments which function to express admiration. She asserts that `provided it is not sarcastic, a compliment on someone’s appearance such as "you’re looking wonderful" is difficult to interpret as anything other than a positively polite utterance’ (Holmes, 1995:131). However, there may be range of different motivations and interpretations for such a remark: for example, the hearer might consider that the speaker is being insincere and is only complimenting because he/she wants something - i.e. that it is serving some longer term goal, or it might be interpreted as suggesting that the person does not look good at all, but that the speaker is being kind; or it might be interpreted as insinuating that the person looks better than they did before, because they looked dreadful before. In her data she does not include street remarks, where there is a gendered divide between intention and interpretation, and there is also a difference in the hypothesised intention - some women see them as compliments, some as harassment; some men assume that their remarks are intended to be positive, others may see them as a way of expressing hostility towards women without using more openly aggressive language, or as a way of showing their solidarity with other men in their community of practice. Despite this diversity of function of compliments, and clear disagreements about the interpretation of compliments, Holmes produces a total number of the compliments in her data, sorted by gender, and asserts that `complimenting appears to be a speech event occurring much more frequently in interactions involving women than men’ (Holmes, 1995: 122). On some occasions, she seems to recognise that compliments have different functions and interpretations, as ` women may regard compliments as primarily positively affective speech acts, for instance, expressing solidarity and positive politeness, while men may give greater weight to their referential meaning, as evaluative judgements, or to ... potentially negative face-threatening features’ (Holmes, 1995: 123). This is a key interpretation since she has shown how many various interpretations there are, even though she has globally assigned certain types of compliment to each sex. Although her model of politeness is focused on speaker intention, hearer interpretation necessarily asserts itself, for example, she states, `it is possible that men more than women more readily perceive compliments as face-threatening acts’ (Holmes, 1995: 124).

Holmes then tries to rationalise her assertion that globally women use compliments more than men, for, she argues that `compliments between women are most frequent ... but men compliment women more often than they compliment other men’ (Holmes, 1995:125). In this way, she, like Brown (1980; 1993), is able to see that the two sexes do not always operate in globally different ways but that there may be patterns to the way that the sexes behave according to context and the sex of the interactants. Rather than simply assuming that this difference is part of women's nature, she asserts that perhaps women are complimented more than men because ` they know women value them’(Holmes, 1995:125). She also suggests that women are complimented more, because, like children, they are subordinate.

For Holmes, men and women compliment in very different ways globally; women tend to compliment on appearance and men focus on possessions. [xii] She draws on stereotypes of males being more able to risk the interpretation of their speech as face threatening when she asserts that `compliments on possessions ...are much more vulnerable to interpretations as face-threatening acts ... since there is the possibility that the complimenter will be heard as expressing desire for or envy of the object referred to’ (Holmes, 1995:131). However, there is a similar risk of face threat when in the supposedly feminine strategy of complimenting on appearance, where a compliment such as `you’re looking very smart today’ might be taken as implying that you do not normally look smart. Compliments are extremely risky for both males and females as it is very easy for them to be misinterpreted; whilst they may be intended as positive politeness, they may be interpreted as face-threatening, for example, as overly familiar, intrusive and impolite. Because Holmes does not consider the responses to compliments, whether they are accepted, deflected or challenged, in any detail, she is unable to claim to be able to know that the linguistic behaviour she categorises as compliments function as such for the hearers.

Apologies

At the level of stereotype, apologies are often associated with feminine behaviour, with excessive apologising, particularly the use of `sorry' and self-deprecation being assumed to be characteristic of women's behaviour. However, apologies cannot be considered to be a formal linguistic entity as I noted in Chapter 2, since they can be made using a wide range of different linguistic strategies. Because of this diversity, it is possible to misinterpret apologies and assume that someone is apologising when they are not, to assume that someone apologising is in fact insincere, or to overlook an apology which is phrased in an indirect way. Thus, quantitative analysis will only analyse those elements which the analyst can recognise as clearly constituting an apology, such as those containing the words `sorry' or `I apologise'.

Holmes claims that women give and are given apologies in different ways to men. In her comparison of the numbers of apologies given and received in her data, she claims that women gave 75% of the apologies and received 73%. However, with this global picture of women as apologising more and receiving more apologies, we have to exercise some caution If we assume that those in positions of subordination apologise more, and that women are in such a position in relation to men, then we would expect more apologies from women to men because of these power differentials. She asserts: `we are likely to apologise to those who are more powerful. And we are likely to apologise more profusely and extensively if the offence is serious' (Holmes, 1995:174). However, in fact her data suggest otherwise: `men apologise twice as often to women as they did to men, regardless of the women’s position in relation to the apologiser. Interpretation must be speculative, but for men perhaps it is easier to apologise even to a woman boss or social superior than to a man’ (Holmes, 1995:175). If this is the case, that men apologise more to women than to men, then we have to revise the hypothesis that subordinates always apologise more to their superiors, since the model of power which Holmes employs cannot deal with the position of the female boss, who, strictly speaking within Holmes' terms, would have to be seen by the male employer as a subordinate because of her gender.

The statistics given by Holmes do not back up her original hypothesis that women apologise more than men, and the raw data suggest there is no real difference between men and women except that which you would expect through randomness: she notes: `the resolution of this puzzle will involve exploring other social features of apology ... part of the answer may lie in differential perceptions by women and men of verbal politeness devices ... women may regard explicit apologies for offences as more important in maintaining relationships ... than men do’ (Holmes, 1995:159). Thus Holmes, in order to try to make her data `fit' her hypothesis argues that women will tend to use clear apologies whilst men do not. She argues that men tend not to use more formal linguistic strategies `since formal linguistic strategies are generally not considered appropriate between people who are close friends’ (Holmes, 1995:162).[xiii]

Again, as with her analysis of compliments, whilst employing a model of politeness which focuses on the speaker's intentions, she is forced to try to interpret her data by calling on notions of the experience or evaluation of hearers and speakers: she argues, `it seems likely that while apologies may be experienced as admissions of inadequacy by men, that is they emphasise power differences, they are regarded by women primarily as ways of restoring social harmony and expressing concern for the other person. Additionally, it may be the case that the society as whole, both women and men, recognise the high priority that women place on politeness strategies as interactive tokens' (Holmes, 1995:176). Furthermore, she asserts that `failing to apologise to a woman is likely to cause greater offence than overlooking the need to apologise to a man’ (Holmes, 1995: 208). Thus, Holmes' analysis, whilst based on the analysis of speaker intentions finds that analysis of the differential judgement of the impact apologies on hearers is necessary. However, she does not have the means to call on the judgements of the interactants apart from her own intuitions and stereotypes which she assumes the hearers are drawing on.

Holmes is aware that stereotypes of gender-appropriate behaviour inform the use of apologies, as she states: `what society calls polite linguistic behaviour is largely based on women’s norms on interaction .... consideration for the feelings of others has been the hallmark of every aspect of women’s verbal behaviour which has been examined’ (Holmes, 1995:194). However, rather than recognising that this is in fact a stereotype, she assumes that `women's norms' exist in some tangible form. She seems to be trying to reclaim these `women's norms' from the stigma and ridicule as `redundant verbal frippery' which often surrounds them, arguing in fact that women's speech is not powerless but surprisingly influential: `despite their lack of social power, women have considerable social influence: their linguistic behaviour determines the overt and publicly recognised norms of polite verbal interactions in the community’(Holmes, 1995: 194). Evaluation of `men's' linguistic behaviour begins to enter into her account towards the end of the book, so that `polite verbal interaction is based on women’s talk, and it has a very limited place in the male public sphere. It is not that men cannot do it. Rather it seems that most of the time they choose not to (Holmes, 1995:195). She suggests that men denigrate polite behaviour since `polite behaviour is acceptable in contexts where nothing important is happening (in men’s perceptions). It has no place when important decisions are made. (Holmes, 1995:196) This suggests that those places where `nothing important is happening' are associated with women.

Because of her alignment with `difference' models of women's language, Holmes' main aim is to change a negative evaluation - women’s language is weak - to a more positive evaluation, and to show that the same language items used by men and women can be judged differently: `one (female) person’s hedge may be another (male) person’s perspicacious qualification ... the association of linguistic markers of tentativeness and a high incidence of epistemic modal devices with insecurity, lack of confidence, powerlessness, and subordinate status, is to a large extent restricted to studies of women’s language. Epistemic devices are not interpreted this way when used in scientific discourse ( which is dominated by men). There they are regarded as evidence of judicious restraint and meticulous accuracy’(Holmes, 1995: 111). Thus, when she asserts that women's language is more polite, she is also implicitly and sometimes explicitly stating, women's language is better than men's language Holmes seems to argue that `women's norms', that is, polite behaviour, lead to better interaction: `linguistic behaviour which follows women’s norms can result in better working relationships, better understandings of complex issues and better decision-making’ (Holmes, 1995: 198). She suggests that `high quality exploratory interaction is essentially collaborative,’ rather than the conflict model which currently holds. (Holmes, 1995:212) She also asserts that `the quality of a discussion is likely to improve when women get a more equal share of the talking time’ (Holmes, 1995: 212) and `on average females are much better than males at providing a favourable context for the kind of talk which is likely to lead to better understanding and cognitive progress' (Holmes, 1995: 212; 217). She suggests that if women's norms of politeness were followed in discussions then there would be an improvement in interpersonal relations, understanding and performance in general. However, she also argues that women's self-effacing behaviour, that is, over-politeness, can often lead to some groups of women being disadvantaged, (which paradoxically seems to accept the assumptions of the `male' characterisation of politeness as being where `nothing important is happening'). She suggests that: `female students are generally not getting their fair share of the talking time. They are too polite’ (Holmes, 1995:199). She also suggests, rather paradoxically, that : `language learners need to be informed about gender-appropriate ways of using pragmatic particles, and the ways women and men use speech acts such as compliments and apologies' (Holmes, 1995:208). However, this assertion needs to be questioned, since, if we are to assume that what she is arguing is that men can apologise just as much as women and in fact should be encouraged to do so, then language learners should be taught women's norms rather than men's.

Because of this positive view of politeness and women's speech in general, Holmes argues that men need to change: `many males need explicit practice in enhancing their conversational competence’ since ` what has been called `polite’ language … has also proved to be cognitively beneficial language. Linguistic politeness contributes to better understanding, and may assist people to reach better decisions. Finally interactions which involve the use of positive politeness strategies are generally pleasant and enjoyable experiences. Being polite makes others feel good. There are many sound reasons, then for recognising the value of polite speech’ (Holmes, 1995: 229). However, as I showed in Chapter 2 politeness is not always interpreted positively and may be viewed as manipulative or excessively deferent and hence weak.

Thus, whilst showing clearly that politeness is associated with women at a stereotypical level, I would argue that Holmes' analysis does not show that women in general are more formally polite than men as she asserts, but merely illustrates the difficulties of a methodology which focuses on the intentions of speakers and assumes that politeness can be recognised objectively by the analysis of formal features. Her analysis also demonstrates the difficulties of a model of gender which assumes that men and women are necessarily different and that they conform in their linguistic behaviour to gender stereotypes .

Analysis of Gender and Politeness

In analysing politeness in relation to gender, I would argue that we need to consider how to analyse the way that hypothesised stereotypes of gender are drawn upon by speakers and hearers in order for them to try and work out what are appropriate forms of behaviour. I would argue that we as analysts also need to be aware that only the judgements of the speakers and hearers about what constitutes polite or impolite behaviour can lead us to the description of polite and impolite utterances. We need therefore to be aware that utterances may have a range of different interpretations. Kharraki's (2001) analysis of Moroccan women and men's bargaining in Arabic dialect illustrates that it is possible to analyse gender difference without assuming that women are more polite than men. He shows that women in Eastern Morocco, when bargaining with shop-keepers drive just as hard a bargain as men, but occasionally they draw on different strategies. Thus, they may be judged by the shop-keeper to be just as polite or impolite as men, but perhaps restricting themselves to forms of behaviour which are stereotypically classified as more feminine. For example, in one extract from his data, a woman says to a man who is selling onions:

W: How much are these onions Hassan?
G: 60 doro
W: Oh ! Don't send your customers away ! Reduce the price a little ! On Monday they only cost 40 doro
G: There is not much profit in it hajja.[respectful term for older Moroccan female]
(Kharraki, 2001: 623)

In another example, he shows that women use very many strategies when bargaining, calling on the shopkeeper's pity:

W: Those potatoes look good, my fine fellow. How much do you want for them first?
G: 80 doro. They're very good quality.
W: We're not criticising the quality. Just see how you treat us. By God, if there were no potatoes, we and our children would die. We are very poor.
G: God help you. Just choose some potatoes. Take it easy.
(Kharraki, 2001: 625)

Because there are certain parameters within which bargaining is successfully executed, these seemingly quite forceful tactics on the part of women, insulting the shopkeeper and suggesting that he is trying to starve the women and her children, are not considered impolite by the shopkeeper. The interactional power of these women contrasts quite markedly with the stereotypes which many Westerners have of passive, deferent women within the Arab world.

To illustrate the sort of analysis which focuses on gender and politeness without assuming that everyone behaves according to stereotypes of feminine behaviour, we might consider the following extract. Here, a New Zealand white middle class middle aged woman [K] who is staying on holiday in the home of a white middle class younger British woman, [D] at the same time as D's mother, [M] a white working class middle aged woman, is visiting. K has been travelling around the world with her husband and is staying with D and her husband [F] because of a mutual acquaintance. D's husband, a white middle class young man, [F] is preparing lunch. The extract takes place in the dining room with all participants standing near the door. In this extract which takes place just before the serving of lunch, the day before K is about to leave, we can see that the context of the exchange, rather than the force of feminine stereotyping, determines to a great extent the degree to which the females here feel that they need to continue to thank K for a gift which she has given. We can also note that stereotypes may in fact inform some women's habitual form of self-presentation, whilst not constituting the whole of that person's identity:
---------------------------------------------------------------------------------------------
1K: D (.) here’s a little colourful Maori shell oh and I'm sorry we're down 2K: to the ones that haven't got nice bright colours in them(.) when you 3K: come to New Zealand you can come and pick your own off the rocks [ 4D: (laughs) just 5D: look at that= 6M: =beautiful= 7D: =isn't that gorgeous [ 8M: that's a real shell= 9K: =have you seen these ? that's for you to take home if you've got room 10K: in your bag (gives shell to M) [ 11M: oh that's lovely [ 12K: if not leave it behind 13D: how nice of you [
( 14M: oh that's lovely= 15D: =oh thanks that's really lovely 16K: have you seen the pauwa shell before ?= 17D: =no 18K: do you know them ? 19M: lovely isn't it ? 20D: never 21M: what(.)untreated are they K? 22K: no (.) well these have been polished (.) normally they've got a roughish 23K back which looks like that there= 24D: =marvellous=
( 25M: =oh that's lovely thank you very much I love the colours= 26D: =aren't they beautiful? [
( 27K: yes well this is I'm sorry but I'm sort of after a couple of years our bundle is well picked over and we're down to the(.) that's 28D: [(laughs)] 29K: got a nice inside though(.)the thing that makes them different the [ ] [ ] 30D: yes 31M: yes 32K: American abalone the American abalone has the opaly colours [ ] 33D: yeh 34K: in fact the one your mother has got is bluer than yours= 35D: =mmm I'll just go and show F 36F: oh that's beautiful (goes back to kitchen) 37M: and then the colours are in them and then they [ 38K: yes are then then if they'd been 39K: washed backwards and forwards in the tide they're just like that [ ] 40M: yes yes 41K: but they're usually broken= 42M: =oh I see= 43K: so what you've got to do is to get the fresh really fresh shells and then 44K: grind the crusts off them it's only a calcium(.) yes and they just lay on [ ] 45M: that's really lovely 46K: our rocks= 47M: =oh well= 48K: =we just use them for ashtrays and things(.)but they're unique [ ] 49M: mmm 50K: to New Zealand [ 51M: oh they are?
( 52D: no I've never seen anything= 53M: =I've never seen anything like that (.) really lovely thank you very 54M: much
( 55D: RIGHT well, I think the easiest thing to do is. if I start serving 56D: everyone.

(Data: 56.5) EXAMPLE 11
--------------------------------------------------------------------------------------------------
If we analyse this exchange simply on the level of the number of politeness markers which are exhibited here, as Holmes and Brown have done, we might argue that this exchange between a group of women is characterised by a high number of positive politeness strategies on the part of D and M, mainly in the form of repeated thanks and by a fairly high number of negative politeness strategies, in the form of apologies, on the part of K. For example:

( 25M: =oh that's lovely thank you very much I love the colours= 26D: =aren't they beautiful? [
( 27K: yes well this is I'm sorry but I'm sort of after a couple of years our bundle is well picked over and we're down to the(.) that's 28D: [(laughs)] 29K: got a nice inside though(.)the thing that makes them different the

In lines 25 and 26 M and D praise the shells in much the same way as they do later on in the interaction, for example:

( 52D: no I've never seen anything= 53M: =I've never seen anything like that (.) really lovely thank you very 54M: much

In this later example they use almost the same language as well as expressing the same sentiments. In the first example, K apologises for the poor quality of the shells.
Since social distance is fairly low and the power relations are fairly evenly balanced, they might seem to be displaying a very feminine form of speech behaviour in thanking and apologising excessively. We might assume that the gift of two shells for putting someone up might merit perhaps one or two `thank you's', but here, D and M consistently and repeatedly praise the gifts and thank K ( see Lines, 5, 6, 7, 8, 11, 13, 14, 14, 19, 24, 25, 26, 45, 52, 53). Since there seems to be a disparity between the gift and the degree to which D and M are thanking K, we can only assume, according to Holmes' model that these women are behaving in a feminine over-polite way or that they are signalling a subordinate relationship to K. However, when I discussed this interaction with D and M, they stated that what they were trying to do was to get K to sit down so that they could start lunch. They felt that that by saying `thank you' and praising the shell, they were closing down the conversation, allowing the topic to be changed to the subject of lunch, which, in line 63, D finally does. Instead of accepting the thanks and praise, K interprets their moves as a sign that they are interested in hearing more about the Pauwa shell in general and she then continues discussing the shells. D and M also stated that they found that K always talked a great deal and during her stay they had tried to develop strategies to cut short her talk. Thus, although here the strategies adopted might seem to be stereotypically feminine, because they seem to be over-polite, in fact the aim, (although perhaps relatively unsuccessful since K continues to talk), was to terminate the discussion rather than to give an excessive amount of thanks.

Furthermore, K might be seen, according to Holmes' model, to be self-effacing, using negative politeness to minimise the value of the gift that she is giving to D and M; for example, K states in lines 1-3,

1K: D (.) here’s a little colourful Maori shell oh and I'm sorry we're down 2K: to the ones that haven't got nice bright colours in them(.) when you 3K: come to New Zealand you can come and pick your own off the rocks

Brown suggests that the Mayan women in her data use diminutives more than men as a way of devaluing or limiting the claims that they make; here `little' seems to devalue the shell, and she explicitly apologises for the lack of bright colours in the shells. Later in the exchange, as noted above, she again apologies when D praises the shell, saying `Yes, well this is I'm sorry but I'm sort of after a couple of years our bundle is well picked over', referring to the fact that she has been travelling for several years around the world and the best shells that she brought with her as gifts have been given to others. She concedes that the one she has given to D `has a nice inside though', again drawing attention to the fact that it is not a good shell in comparison to others. We might assume that apologising for gifts, when power relations are fairly equal, is an indication of the low self-esteem or excessive femininity of K. However, when compared to the assertiveness, as judged by D and M, of the rest of her speech, these apologies cannot be read as simply indicating that K's speech is self-deprecating, particularly if we consider the difficulty remarked upon by D and M in getting her to stop speaking. Her speech is regarded by D and M in terms of her verbal habits, of speaking too much, rather than in terms of apologising too much, here. In fact, here we may be forced to analyse this instance of her speech style simply as an example of her trying to apologise for the fact that her small gift does not equal the hospitality that she has received from D, F and M ( a point which both D and M themselves brought up when I interviewed them); and her apologies may be a response to what she sees as excessive thanking by M and D.

Thus, rather than simply analysing data in terms of the way that males and females act differently in relation to their use of politeness, it is important to analyse what politeness, or what seems to be functioning as politeness, is being used to achieve within a particular community of practice. Here rather than seeing politeness as the sole production of one speaker, we need to see that the type of politeness used and the functions to which it is put are constructed and negotiated with in a concerted effort by all of the participants.

In the next extract, if we were to use Holmes' and Brown's model of gender and politeness, we might again assume that the sex of the participants was necessarily significant in the interaction. In a misunderstanding which is then resolved, a group of young white working class friends, a female employed as a bus driver [C], a male, her partner [T], who is unemployed and another female [A] who is employed in another town and visiting her parents, discuss their plans for an evening out. The apology in the extract is something which is worked at jointly, even though the fault seems to lie largely with only one of the participants [A]. Thus, quantitative analysis of the apologies which only analyse the production of explicit apologies by one speaker need to be questioned. Here the interactants are discussing meeting up with a group of friends at a pub:

1T: they might go for a curry after(.)that's my bet anyway [ ] [ 2A: yeh 3C: but I haven't found out yet 4C: what time I'm working tomorrow(.) so today I finished by about half 5C: past one quarter to two so the day sheet hadn't been filled out(.) [ ] 6A: yeh
( 7C: I'm going to phone them up in a minute or when we go out we can 8C: drive down and I'll jump out [ 9A: yeh is your car working?= 10C: =it's it's away at the moment = 11A =um= 12T: =at the hospital [ 13A: so how can we drive round there?= 14C: =I thought you said you had your car 15A: I haven't got a car here [ 16C: what did you say 17C: when she came in ? 18T: I said I presume you came by car 19A: [(laughs)] 20C: I thought you 21C: must have had your father's car= 22A: =you you must be JOKING 23C: no when T said I presume you've got your 24C: car and you went yeh [ 25A: well why why was I saying the other day that I 26A: was going to hire a bicycle 27C: I would just have thought you had you 28C: might have had your father's er [ 29A: no I er rather thought I might have my er 30A: father's car but er no(.) it hasn't even been suggested= 31C: =oh= 32A: =and I can't ask=
( 33C: =oh well this calls for a bit of phoning because it's gonna have to be a 34C: bus up there cause it takes about half an hour to walk up there [ ] [ 35T: quick the time table 36A: well can't we 37A: get that bus that goes from the Strand?= 38C: =yeh that's what I'm saying but they only run every half an hour and I [ ] 39A: oh I see 40C: shall have to look up the times=
( 41A: =right then sorry I've um [ 42C: oh that's alright (laughs) [ 43A: messed up all the plans (.)no(.) I 44A: thought it was a joke when you said have you got the car [ 45T: ME? JOKE? 46C: yes I know(.) I knew you didn't have YOUR car here(.) I thought you 47C: had your father's here= 48A: =no this is er [ 49C: I presume it is your father's car and not your mother's [ 50A: no it's my father's car 51A: but my mother did bring me here because my father didn't think he 52A: could stand going to see my gran (.) no if we go round there before 53A: they're going on holiday she tends to get a bit er over emotional like

(data: 89.4) EXAMPLE 12
--------------------------------------------------------------------------
In other interactions between these three friends, it is clear that there is a great deal of give and take in the interactions: T, the male, is far quieter and intervenes less frequently than the women, partly because the women were at school together and they are more clearly friends with each other (although all of them have known each other since they were at school: they were in their thirties at the time of the recording). The misunderstanding about whether A has a car with her is something which is circled around in this conversation, in lines 8-14, where C first refers to jumping (out of a car) to find out from the bus depot when she has to work, and A in line 9 tries to check that this is C's car which is being referred to:

7C: I'm going to phone them up in a minute or when we go out we can
( 8C: drive down and I'll jump out [
( 9A: yeh is your car working?= 10C: =it's it's away at the moment = 11A =um= 12T: =at the hospital [ 13A: so how can we drive round there?= 14C: =I thought you said you had your car

As in many other misunderstandings in their conversations, conflicts are resolved through the use of humour and irony, with each of the participants trying to lessen the impact of the conflict, by using direct and forceful language, which might in Brown and Levinson's model be categorised as FTAs, as well as more conventional apologies and negative politeness. Thus, when A apologises for not having a car with her and thus for having `messed up all the plans', for going out, as they now have to take the bus, C states that she in fact knew that A didn't have her car, contradicting her earlier statements that she had presumed she did have a car with her, a conventional negative politeness strategy, whereas in 45, T repeats A's words in 44 ironically, presumably to lessen the tension of the situation.

41A: =right then sorry I've um [ 42C: oh that's alright (laughs) [ 43A: messed up all the plans (.)no(.) I 44A: thought it was a joke when you said have you got the car [
( 45T: ME? JOKE?
( 46C: yes I know(.) I knew you didn't have YOUR car here(.) I thought you 47C: had your father's here=

Focusing only on the way that A explicitly apologises, which conventional politeness theories such as Holmes do, would not allow us to focus on the way that in this interaction questions of the sex of the participants is not particularly salient. Analysing only the individual's speech in relation to this apology also would not take into account the way that positive politeness or facework is being mutually accomplished, with each of the interactants contributing to the resolution of the misunderstanding with none of their faces being particularly threatened, and the functioning of their particular community of practice being contributed to partly through the conflicts which go on in their group, that is, they can present themselves as a group of friends who get on well together because they can resolve conflicts jointly, not allowing difficulties and misunderstandings to threaten anyone's face. For example, when A states in line 44, ` I thought it was a joke when you said have you got the car' T responds with irony: `ME (.)JOKE?' drawing attention to the fact that, in fact, the group functions largely though the resolving of conflict through irony, banter, mockery and joking. Even A's apology in line 41, which seems like a straightforward apology, has the ring of irony about it, especially given the use of hesitation in the middle of the utterance. Thus, within this group it is difficult to make assumptions about the linguistic performance of the individuals on the basis of sex alone, since their production of their identities and roles within this particular community of practice is not determined solely by the gendering of their individual selves. This type of analysis is therefore trying to tease out where gender is a salient feature to the way that individuals present themselves within a group, and the way that they may draw on resources of polite behaviour and also impolite behaviour, ironically, in order to resolve conflicts.

Conclusions

In this chapter, I have tried to question the theoretical and analytical work which has been undertaken on the subject of women's and men's use of politeness. Contrary to Holmes' and Brown's work which asserts a global difference between men's and women's use of politeness, however mitigated some of their generalisations are, I would like to assert that gender should not be seen as a factor which determines the production or interpretation of speech in any simple way. That is not to say that gender is not important as hypothesised stereotypes of feminine and masculine behaviour obviously play a role in the production of what participants see as appropriate or inappropriate speech. However, decisions about what is appropriate or not are decided upon strategically within the parameters of the community of practice and within the course of the interaction rather than being decided upon by each individual once and for all. The extracts I have analysed in this chapter show the difficulty of asserting that women and men speak in different ways, or of asserting that women behave in powerless ways, because it is clear that some participants use seemingly feminine strategies strategically to achieve their ends rather than to assert their subordination. They also collaborate with others to achieve certain ends within the group and thus an analysis of their individual utterances alone would not enable the way that the group functions and the interaction takes shape. Because gender and other factors impact upon the context and because gender is indeed something which participants perform and interpret in the context of hypothesised gendered stereotypes within a community of practice, it is essential to analyse gender at both the local and the structural level especially in its relation to the production and interpretation of politeness.

-----------------------
[i] Women are often viewed in certain contexts as a civilising force; for example in the white colonisation of Australia, women were regarded as either whores or as `God's police', that is, as the guardians and enforcers of civilised norms (Robinson, 1988).
This policing role is contrasted to the stereotypical Australian masculine role who is associated with the outback and implicitly freedom from constraints (Schaffer, 1989).

[ii] These stereotypes of feminine behaviour, whilst being generated from middle class women's behaviour are often extended to working class women's behaviour as I mentioned in Chapter 4 to distinguish between `respectable' and `rough' working class women, both by others and the women themselves.

[iii] It might be argued that etiquette books are generally directed to members of the working classes who aspire to be considered as middle class, rather than at the middle classes themselves, who might be assumed to know the `correct' behaviour as part of their `cultural capital' (Bourdieu, 1991).

[iv] Distinctions between U and non-U behaviour (for example, whether to use the terms: tea, dinner or supper; sitting room or lounge; napkin or serviette, and so on) which were popular in Britain from the 1940s onwards still seem to have some resonance, although for many these distinctions seem less clear than they were.

[v] Feminist linguists and activists have often criticised this stereotypical association of women with politeness and have urged women to speak more assertively. However, strong women speakers, as I argued in Chapter 4, may often be considered to be acting outside their allotted role, particularly by those who have a strong sense of `women's place' or the importance and value of femininity. Assertiveness may be judged to be aggression which is evaluated negatively for women. Anne Robinson, a British TV presenter who hosts the game show `The Weakest Link' has been judged to have a particularly `cruel and po-faced' persona (Braid, 2001). She was rated the rudest woman on TV, and journalists have termed her a `bitch' and the `Miss Whiplash of prime-time TV' (Braid, 2000).

[vi] This type of behaviour is also extended from middle class men to other men and from middle class women to other women, but at a stereotypical level, for historical reasons, because of its origins in chivalry, it is often associated with the behaviour of men to women. It certainly causes most conflict between men and feminist women; (others might argue that it is most appreciated between older groups of men and women).

[vii] There is certainly an assumption in much media reporting that it is working class men who are most prone to incidents of `road-rage' and `air-rage'.

[viii] Reindl Scheuering defines courtesy in the following way: courtesy is `the name given to practices such as a man helping a woman into her coat, leading her through crowded places or into a restaurant, opening her doors - the same as mothers do for children. See it this way and you know how men look on women' (Reindl-Scheuering, cited in Kramarae and Treichler, eds. 1985).

[ix] For a more complex view of women's use of prestige forms, see James 1996 `Women, men and prestige forms: a critical review'.

[x] This in itself is not problematic but the generalising about women and men as a whole from this data is.

[xi] Baxter, in a conference paper at the International Gender and Language Association conference at Lancaster, 2002, stated that although girls often find it more difficult to speak in public than boys initially, this is because of their perception of stereotypes about appropriate behaviour, rather than any intrinsic quality.

[xii] However, it should be noted that compliments from men to women in a work environment may be interpreted as drawing attention to gender difference when it is not, or should not be, salient.

[xiii] She also suggests that men and women tend to apologise for different infringements, suggesting that men apologise for time offences more than women: `keeping someone waiting is impolite behaviour and women tend to avoid being impolite more than men do’ (Holmes, 1995:168). She even goes on to suggest that men are more often late than women. She gives other reasons for apologies occurring more : `it is perhaps not surprising to find a predominance of apologies for accidental body contact in a group who are the main victims of sexual harassment' (Holmes, 1995:169). However, we might argue that in this respect it should be men who apologise more because they are the ones who are most likely to be accused of sexual harassment.