Abstract

Impoliteness is highlighted as a neglected area of politeness research, a consequence of the dominance of Brown and Levinson’s (1987[1978]) theory of politeness universals. Previous studies that have investigated language, gender and politeness are critiqued for an over-reliance on Brown and Levinson’s model, and for over-generalising the complex relationship between language and gender by simply cataloguing differences in male and female speech patterns. The community of practice (CofP) framework (Eckert and McConnell-Ginet 1992a, b) is presented as an alternative approach to language and gender studies. Following Mills, (forthcoming) the CofP approach is also adopted to bring a new dimension to the manner in which notions of linguistic politeness and impoliteness are conceived. In this paper, the communities of practice approach is modified in order to include the construct of media genres, and impoliteness in political broadcast interview data is analysed. The need for language and gender studies to acknowledge that gender may not be the only salient variable affecting linguistic behaviour is argued, and the CofP framework is suggested as a way forward to approach the complex relationship between language, gender and politeness.

Key Words: community of practice, politeness, impoliteness, gender identity, institutional discourse, political interviews

The topic of politeness has proved to be a popular line of enquiry for language and gender researchers in recent years [1], with interest originating with Lakoff’s (1975) anecdotal assertions that women are more linguistically polite than their male counterparts. Similar conclusions have been drawn by Brown (1980) and Holmes (1995) who offer empirical evidence to justify their arguments. At its time of publication, Holmes’ (1995) work offered a detailed analysis of linguistic politeness and gender, drawing on her own and others’ research in a variety of contexts. As Crawford (1997: 428) argues, Holmes manages to incorporate a large amount of material ‘under her politeness umbrella’, including critical reviews of influential language and gender studies conducted by Zimmerman and West (1975), Fishman (1978, 1980) and Tannen (1984, 1990). Holmes concludes that the multitude of evidence she has collected over a number of years clearly demonstrates that women are more linguistically polite than their male counterparts.

However, researchers including Cameron (1995, 1996, 1997) and Bergvall, Bing and Freed (1996) have questioned assertions such as those made by Holmes and others, where men and women are seen as having distinctive speech styles. They argue that viewing men and women in a dichotomised way results in a gross oversimplification of the complexity of language and gender. It not only ignores the diversity of speech within groups of women and groups of men, it also ignores cultural differences, and differences that may result from other social variables such as class, age and ethnicity. As Freed (1996: 55) points out, whilst ‘people generally persist in believing that...women are more polite than men’, research which continues to address such questions is both ‘misguided and naive’. She argues that researchers need to abandon frequently asked questions such as ‘what differences exist between men’s and women’s speech?’ (1996: 55), as this serves only to perpetuate stereotypes about male and female discourse.

Cameron (1995) argues that the problem lies in the persistence of both the power/dominance and culture/difference approaches to language and gender. Both approaches assume that there is a pre-existing difference between male and female speech patterns. Power/dominance theorists believe that this stems from the considerable amount of economic, social and political power men have over women in society, whilst the culture/difference researchers believe that speech differences are implemented during the socialisation process. Cameron (1995: 39) accuses the dominance approach of becoming ‘obsolete’, and the difference approach of being ‘reactionary’. Both frameworks result in the formulation of inaccurate over-generalisations such as women are more linguistically polite than their male counterparts.

Theorists, including Freed (1996), Bergvall (1996) and Cameron (1997), have turned to the notion of gender as a performative social construct, following Butler’s (1990) work, as a way of avoiding gender polarisation in language and gender studies. Butler believes that masculinity and femininity are not traits that we inherently possess, rather they are effects that are produced by the activities we engage in. She argues that ‘gender is always a doing’, and as there is ‘no gender identity behind the expressions of gender...identity is performatively constituted by the very "expressions" that are said to be its results’ (1990: 25). Freed (1996), Bergvall (1996, 1999) and Mills (forthcoming) argue that as well as adapting the notion of gender as a performative social construct, language and gender research also needs to adopt Eckert and McConnell-Ginet’s (1992 a, b) Community of Practice (CofP) approach. Bergvall (1999: 282) argues that the CofP framework enables the ‘performative construction and achievement of gendered identity’ to be examined, thus focusing ‘much needed attention on the social construction of gender’ (1999: 273).

In order to explore the ways in which the CofP approach can be of benefit to an investigation of impoliteness, it is necessary to define exactly what is meant by linguistic politeness and impoliteness by assessing current theoretical perspectives. The CofP framework will then be defined and applied to the institutional context of political interviews, with the aim of demonstrating how viewing them as a CofP can bring insights into how notions of gender and impoliteness can be conceived.

More than 20 years after it was initially published, Brown and Levinson’s ([1978] 1987) theory of politeness universals is still highly influential. Although it has been criticised on many counts, researchers continue to adopt Brown and Levinson’s definitions as the basis for their studies (see Johnstone, Ferrara and Bean 1994, Holmes 1995, Cheng 2001, Perez de Ayala 2001).

Brown and Levinson (1987: 62) base their theory on the concept of face, following Goffman (1967). They define face as the public self image that all rational adult members of society possess. Face consists of negative face, the desire to be unimpeded, and positive face, the desire to be liked/admired. All participants in spoken interaction emotionally invest in face, and it must be constantly considered. Brown and Levinson argue that, in general, it is in the mutual interest of interactants to maintain each other’s face.

If a demand or an intrusion needs to be made on another person’s autonomy, then Brown and Levinson distinguish this as a potential face-threatening act (FTA). When speakers are posed with the problem of performing an FTA, they have to decide whether the FTA should be performed on record or off record. The off record strategy enables a speaker to avoid responsibility for performing an FTA by either inviting conversational implicatures or by being deliberately vague or ambiguous. Alternatively, if the on record strategy is chosen, then a speaker can either perform the FTA without redressive action, known as going baldly on record, or they can perform the FTA with redressive action, thus paying attention to the addressee’s face needs.

Brown and Levinson expand their definition of redressive action by outlining a number of politeness strategies that speakers can use. The strategies that they set out are designed to avoid conversational breakdown which they believe will occur if speakers neglect their addressees’ face needs. They define positive politeness strategies, where the speaker pays attention to the hearer’s positive face needs by demonstrating that the hearer’s wants or needs are thought of as desirable. Conversely they define negative politeness strategies, where the speaker desires not to impose on the hearer by restricting the hearer’s actions, thus paying attention to the hearer’s negative face needs.

Brown and Levinson’s theory focuses on interaction in informal contexts, and due to the predominance of their approach, politeness in institutional contexts has been neglected. Following Lakoff (1989), Harris (2001: 453) argues that researchers should examine politeness in institutional settings, as this ‘forces us to see politeness from a different perspective’ and works to ‘foreground different dimensions’ of politeness. Furthermore, Harris (2001: 452) points out that Brown and Levinson, and research which follows their principles, concentrates only on short stretches of talk. This criticism is also noted by Mills (forthcoming), who argues that politeness needs to be viewed as ‘something that emerges at discourse level over stretches of talk’, instead of something that is merely ‘grafted on to individual speech acts’.

I agree with both of these arguments, and a reason to focus this paper on the institutional context of political interviews is motivated by the lack of attention that has been paid by politeness researchers to this setting in the past. One main purpose of the data analysis in section 6 is to demonstrate that impoliteness emerges at a discourse level, rather than in the production of a single utterance.

Brown and Levinson’s concentration on strategies which avoid the performance of FTAs leads to the area of linguistic impoliteness being overlooked. As Eelen (2001: 90) argues, their positive and negative politeness strategies ‘stipulate how to be polite rather than impolite’. Although Eelen (2001: 92) acknowledges that, as Brown and Levinson do not claim to look at impoliteness then they cannot be criticised for something they did not claim to do in the first place, he does argue that politeness and impoliteness are ‘two sides of a coin’, and therefore ‘any theory that pretends to say something valuable about one side, automatically needs to deal with the other side as well’. This argument is also made by Culpeper (1996: 350) who believes that, in order for a theory of politeness to be comprehensive, it is integral that the topic of linguistic impoliteness is addressed.

Eelen (2001) points out that by conceptualising face redress and face threatening acts, Brown and Levinson’s theory seems to be capable of accounting for linguistic impoliteness. Their theory implies that if speakers aim for conflict, then they should avoid using redressive strategies and instead use bald on-record strategies.

However, Eelen proceeds to argue that a major problem with this idea is the notion of face and face wants that is central to Brown and Levinson’s model. If speakers do not satisfy the face wants of others, then they are being ‘self-destructive’, displaying ‘almost unnatural behaviour’ (2001:101), as hearers will then not satisfy the face wants of their speakers. Therefore, despite Eelen’s initial expectations that Brown and Levinson’s theory could conceive notions of linguistic impoliteness, for it to do so would be entirely contradictory:

If face-wants were to account for impoliteness in the same way as politeness, they would need to include the want not to satisfy one’s own face-wants, which is a contradiction in terms (2001: 102).

In their section on bald on record utterances, Brown and Levinson do state that, on rare occasions, participants may perform an utterance ‘baldly, without redress’ if ‘the speaker does not fear retribution from the addressee’ (1987: 69). This may be due to the urgency or efficiency required by a situation, or in situations where the danger to the hearer’s face is very small, as in offers or requests such as ‘come in’ or ‘sit down’. Despite these acknowledgements that FTAs can be performed without redress on some occasions, Brown and Levinson’s model still does not account for the fact that their concept of face wants ‘can never explain their own non-fulfilment’ (Eelen 2001: 102).

Furthermore, Thomas (1995) points out that Brown and Levinson’s theory appears to predict that their bald-on record strategy is only employed in situations where face threat is small, due to their premise that it is generally in the mutual interests of participants to maintain each other’s face. Thomas (1995: 171) argues that Brown and Levinson have neglected the fact that there are occasions where speakers do perform utterances that are designed ‘deliberately to be maximally offensive’.

In light of this evidence, a theoretical approach to politeness that incorporates both politeness and impoliteness appears to be required in order to provide a more comprehensive explanation of linguistic politeness as a whole. Culpeper (1996) has attempted to define a theory of impoliteness that works as a parallel framework to Brown and Levinson’s model of politeness. However, the criticisms that have been highlighted above can also be applied to Culpeper’s work, as he follows Brown and Levinson’s basic principles of face, face-threatening acts, and positive and negative politeness. Further criticisms can also be applied to the manner in which Culpeper defines impoliteness in his data (see section 4).

Harris (2001) makes the point that a vast amount of criticism applied to Brown and Levinson’s work is a consequence of the fact that they are seeking to define a set of politeness universals which are applicable to numerous cultures [2]. Quoting Kasper (1990), Harris argues that the point he makes eleven years ago is probably still relevant today: researchers should not be aiming to generalise politeness principles. Instead, politeness research needs to confine itself to certain, specific contexts which have some ‘well-defined commonalities’ (Kasper 1990: 213, quoted in Harris 2001: 470). Harris accounts for this by utilising the communities of practice approach to analyse impoliteness in the discourse strategies of British politicians in the specific context of Prime Minister’s Question Time. Her work demonstrates how fruitful the CofP approach can be to analysing discourse in specific, institutional contexts.

Mills (forthcoming) suggests that following the CofP approach will radically alter the way in which we conceive notions of politeness and impoliteness. In order to demonstrate the difference a CofP framework can make to an interpretation of impoliteness, she revisits Culpeper’s (1996) data. Culpeper (1996: 359) applies his impoliteness framework to the discourse of army training officers in a documentary on female recruits. He asserts that impoliteness is a crucial part of army life; thus, a universalising theory of impoliteness needs to be developed to explain such instances. However, Mills makes the significant point that in the specific CofP of an army camp, the kind of behaviour that Culpeper has classified as impolite would not be classified as such by the female participants in his data as it is an expected norm.

Mills (forthcoming) argues that by adopting the CofP approach to studies of gender and politeness then:

Individual linguistic acts between individual speakers must be replaced by a community-based perspective on gender and politeness, which must therefore involve a sense of politeness having different functions for different people.

Mills (forthcoming) therefore argues that Holmes’ (1995) definition of ‘polite people’, i.e. the female speakers in her data, ‘does not relate those polite acts to a community which judges the acts and people as polite’. A CofP approach would avoid these kind of problems.

Eckert and McConnell-Ginet (1992a) argue that a major problem with language and gender research stems from the fact that it has abstracted both language and gender from the social practices in which they are produced, thus blurring the complex way in which they relate to one another. They assert the need for researchers to connect social and linguistic practice in order to examine how they are realised in specific communities.

Eckert and McConnell-Ginet also urge researchers to stop seeing the differences between male and female speech patterns as the major goal of research. They argue that studies which follow this pattern, such as the politeness studies of Brown (1980) and Holmes (1995), serve only to draw attention away from ‘a more serious investigation of the relations among language, gender, and other components of social identity’ (1992a: 91). They believe that the CofP approach can overcome these problems. Drawing on work by Lave and Wenger (1991), where the CofP approach is presented as a social theory of learning, Eckert and McConnell-Ginet (1992b) define a CofP as:

An aggregate of people who come together around mutual engagement in an endeavor. Ways of doing things, ways of talking, beliefs, values, power relations - in short, practices - emerge in the course of this mutual endeavor (1992b: 464).

Communities of practice can develop out of formal or informal enterprises, and members can be either core or peripheral depending on how integrated they are in a CofP. Communities of practice vary in size and quality. They can survive changes in membership, and they can come into existence and go out of existence. Eckert and McConnell-Ginet (1992a: 93) argue that individuals ‘participate in multiple communities of practice and their individual identity is based on the multiplicity of this participation’. Gender, as one part of that individual identity, is produced and reproduced in different forms of membership in communities of practice. As well as gender, an individual’s exposure or access to communities of practice is related to many other aspects of social identity including class, race, age, etc.

Eckert and McConnell-Ginet (1992a: 94) proceed to point out that gender is also ‘produced and reproduced in differential forms of participation’ in particular communities of practice and this is crucially linked to the place of such groups in wider society. They believe that the CofP approach enables researchers to focus on ‘people’s active engagement in the reproduction of or the resistance to gender arrangements in their communities’  (1992b: 466).

It is clear that to apply the CofP approach to the media-constructed setting of a political interview requires some justification. However, Ehrlich (1999) has successfully applied the CofP approach to the constructed setting of courtroom discourse in a language and gender study examining sexual harassment, and I believe that, by re-defining the concept of a CofP, then the benefits of the approach to a study of gender, impoliteness and political interviews can be emphasised.

The fact that political interviews are confrontational, competitive encounters is clearly documented by both researchers in this area (Harris 1991), and by media personalities themselves. Robin Day (1975, 1991) argues that politicians deliberately ignore the questions they are asked and insist on repeating statements irrespective of whether they bear any relation to the interviewer’s questions. In a previous study on broadcast interviews and politeness (Mullany 1999), the confrontational nature of political interviews is clearly emphasised. At first glance, the adversarial discourse patterns found in political interviews appear to be at odds with the general, harmonious principles of the CofP approach.

However, as I have argued elsewhere (Mullany 2000), Eckert and McConnell-Ginet tend to focus on the harmonious nature of the CofP approach, thus neglecting to draw attention to the fact that Lave and Wenger’s (1991) definition of a CofP can incorporate confrontational discourse as well. This point is summarised succinctly in Wenger’s (1998) later work. He outlines three defining characteristics for a CofP: mutual engagement, a joint enterprise and a shared repertoire (1998: 73). Wenger argues that the process of mutual engagement can be either harmonious or conflictual and part of the misinterpretation of a CofP as a co-operative, non-confrontational environment lies in the implicit assumption carried by the term ‘community’. He points out that because ‘community’ often carries with it ‘positive’ connotations, there is an expectation that relationships taking place within a CofP will be of harmonious (1998: 76). He denies that this is the case in every CofP, and this leads him to argue:

Connotations of peaceful coexistence, mutual support, or interpersonal allegiance are not assumed...peace, happiness and harmony are therefore not necessary properties of a community of practice...disagreement, challenges and competition can all be forms of participation (1998: 77).

Therefore, the confrontational nature of the political interview, where ‘disagreement, challenges and competition’ all frequently occur in interaction, are perfectly acceptable forms of linguistic behaviour in a CofP.

A further problem with utilising the concept of a CofP to investigate political interviews is the fact that, in Lave and Wenger’s (1991) and Eckert and McConnell-Ginet’s (1995) studies, the approach has been adopted to analyse naturally-occurring linguistic data. In contrast to this, political interviews are constructed encounters. They are a genre of the mass media, set up to produce discourse for an overhearing audience.

However, it has already been pointed out that Ehrlich (1999) has successfully applied the CofP framework to the constructed setting of courtroom interaction, where discourse is set up specifically for the overhearing jury to make a decision about a defendant’s innocence or guilt. In settings such as courtrooms and political interviews, I wish to argue that the CofP framework is still a useful one if it is modified to acknowledge the constructed setting of its discourse production. To accomplish this, I propose that a CofP be redefined as a synthetic community of practice (SCofP), to account for discourse that takes place in constructed contexts. In a SCofP, the roles of the interlocutors are pre-determined. Furthermore, the main purpose of a SCofP is to produce discourse for an overhearing audience. In general, each participant will seek to influence the audience to agree/empathise with their point of view.

Furthermore, as part of his theory of social learning, Wenger argues that ‘communities of practice are everywhere’ (1998: 6) and the learning experience within these communities of practice make up ‘an integral part of our daily lives’ (1998: 7). Just because individuals enter a constructed discourse setting does not mean that they will suddenly stop learning. For the interviewers and interviewees in political interviews, the learning experience continues as they perform their social and institutional identities in this SCofP.

Another problem with using the CofP framework to examine political interviews lies in the fact that both Lave and Wenger (1991) and Eckert and McConnell-Ginet (1995, 1999) use the approach to examine the speech of the same group of individuals over a period of time. Unlike the tailor’s apprentices Lave and Wenger observed, or the high school adolescents focused on by Eckert and McConnell-Ginet, the individuals who participate in broadcast interviews will differ from broadcast to broadcast. Although political interviewers regularly conduct interviews on specific programmes in the same broadcast network, the political interviewees change constantly and often appear on different programmes and networks.

However, political interviewees will participate in political interviews over a period of time during their political careers. Although they do not appear as often as the interviewers, interviewees learn the rules of engagement in political interviews through their own experience and through the training they receive from their political parties. They are thus aware of the norms of the SCofP they engage in as much as their interviewers, even though they appear less frequently.

Furthermore, although the individual participants do change, the pre-determined roles of the interviewer and interviewee remain constant. As a SCofP, the political interview has a clearly defined set of norms and expectations governed by the established membership roles of political interviewer and political interviewee(s). Greatbatch (1986) details the distinct roles that broadcast interviewers and interviewees occupy. He argues that, in news interviewing (in which all of the data he uses are political interviews), there is a ‘turn-taking system which specifies that interviewees should confine themselves to answering interviewers’ questions’ (1986: 441). Within the turn-taking system, the interviewers are responsible for topic control and the time management of the discourse. Wenger’s three defining characteristics for the existence of a CofP will now be applied in more detail to give clarity to the definition of the political interview as a SCofP.

Wenger’s first characteristic, mutual engagement, is evident in political broadcast interviews, with the interviewer and interviewee(s) mutually engaging with each other to produce discourse for an overhearing audience. As has already been emphasised, the aims of a CofP do not have to be harmonious. Therefore, although the interviewers and interviewees will be mutually engaging with one another, this does not mean that conflict and competition will not occur. On the contrary, confrontational discourse will be expected, with the interviewer attempting to hold the interviewee(s) to account on the perspectives and policies of their particular political party.

Wenger’s second characteristic, a joint negotiated enterprise, is also observable in political interviews. It is the job of the interviewer and interviewee(s) to jointly produce discourse for the overhearing audience. Through the production of question-answer adjacency pair sequences, the interviewer and interviewee(s) work jointly with one another in a two-way process: the interviewers ask questions designed to elicit information from the interviewees for the listening audience and the interviewees’ role is to answer these questions to best effect for both themselves as individuals and the political parties which they represent. Whilst their purposes within the interview are different, the interviewer and interviewee(s) work jointly in order to produce discourse for the listening audience.

The third characteristic, a shared repertoire, is also evident in political interviews. Holmes and Meyerhoff (1999: 176) point out that by a shared repertoire, Wenger is referring to the ‘joint resources for negotiating meaning’ that participants have available to them. In political interviews this can include strategies such as specific linguistic routines, including interviewers introducing interviewees to the listening audience, and rituals for terminating an encounter. In a database of 20 political broadcast interviews, a distinct pattern was observed [3]. All of the interviewers introduced their interviewees to the listening audience by referring to the title they have within their political party, along with their FNLN. The interviewees’ spatio-temporal relations are also made clear to the listening audience, as in the following examples [4]:

MER: The Agriculture Minister Jack Cunningham is with me

MER: In the radio car is The Roads Minister Baroness Hayman

Often, these details are accompanied by a ritual greeting that is reciprocated. For example:

MER: The Schools Minister Estelle Morris is in our Westminster studio good morning

FEE: Good morning

MER: The Leader of the Opposition William Hague is with us this morning good morning

MEE: Good morning

All of the interviewers in my database follow exactly the same ritualistic pattern for terminating an encounter. The interviewer states the FNLN of the interviewee, and then thanks them for partaking in the interview. For example:

FER: Estelle Morris thanks very much

MER: Jack Cunningham thanks very much indeed

As the interviewer and interviewee(s) have such clearly defined roles in political broadcast interviews, it becomes clear that as participants, they do not have equal access to the CofP, thus raising the important issue of power between participants. In order to illustrate this point, I wish to expand on the notion of core and peripheral membership (outlined in section 4), by defining two intersecting axes on which interviewers’ and interviewees’ membership of the SCofP can be defined.

On the first axis, both the interviewers and interviewees can be either core or peripheral in the conventional sense, depending on the amount of experience they have, either as broadcast journalists, or as politicians being interviewed on a political topic in the broadcast media. This axis can be referred to as core and peripheral membership at an individual level. Secondly, the definitions of core and peripheral can be given a more global distinction, with the interviewers being the core members of the SCofP, whilst the interviewees are the peripheral members. The interviewers are core members as they are responsible for controlling the discourse at a local management level. Furthermore, they are employed by the broadcast media as an institution, thus it is within their profession that the interview takes place. The political interviewees are the peripheral members of the CofP as they are controlled by the turn-taking system over which the interviewer has control. This axis can be referred to as core and peripheral membership at an institutional level. The SCofP approach to political interviews will now be combined with an examination of politeness from a CofP perspective, and the specific issue of linguistic impoliteness in political interviews will be focused on.

If Brown and Levinson’s (1987) model is applied to a study of political interview discourse, the FTAs uttered by participants without redressive action would be regarded as impolite by implication. However, in political interviews it is not in the interests of participants to pay mutual attention to each other’s face needs. The centrality of the preservation of face needs to Brown and Levinson’s theory means that it does not appear to account for confrontational discourse where not paying attention to the addressees’ face needs and attacking their position is a frequent and expected occurrence. Failure to pay attention to the face needs of fellow interlocutors does not result in conversational breakdown in political interviews, as would be predicted by Brown and Levinson’s theory.

Unlike Brown and Levinson’s approach, the CofP framework enables a more fluid and dynamic approach to be taken to a definition of politeness, which accounts for impoliteness as part of the overall concept of politeness. It draws attention away from a search for politeness universals, and leads instead to a detailed examination of what politeness means in specific contexts. Adopting the CofP approach means that it is the participants themselves who define what is polite and impolite behaviour against the norms they have for the specific communities in which their discourse practices take place. In a specific concentration on the CofP approach and impoliteness, Mills (forthcoming) argues:

Impoliteness is only classified as such by certain, usually dominant community members, and/or when it leads to a breakdown in relations.

For impoliteness to be evident, either the interviewer or the interviewee would need to highlight this themselves. However, as has been illustrated above, conversational breakdown rarely occurs in political interviews, and it is notable that only one example of impoliteness, as defined through the CofP approach, can be found in my database. Before analysing this example, and the potential implications it has for language and gender studies on politeness, I wish to focus on Mills’s definition that it is usually the ‘dominant’ member of the CofP who will classify impoliteness in discourse as this raises interesting issues of power operating at both a micro and macro level in the SCofP of political interviews.

At a local level of discourse management, it is the interviewers who appear to control the encounter, as illustrated in section 5. However, it is often the case that the interviewees will hold a much more dominant position in wider society by virtue of the fact that they are Members of Parliament. As Winter (1993) points out, this is noticeable through the non-reciprocal use of address terms, which is also observable in all the interviews in my database. The interviewees use FN to refer to their interviewers, whereas the interviewers will use T only, TLN, TFNLN or FNLN to refer to their interviewees, signalling the distinction between their positions of power in society as a whole. As a consequence of the discrepancy between power at local and global levels, the identification of the most dominant participant in the interaction can become blurred, and on certain occasions, interviewees can stipulate which topics they will talk about, either during the interview or beforehand, thus limiting the amount of local power interviewers have over the encounter [5]. It thus appears that as the identification of the most dominant participant can become indistinct, both the interviewers and the interviewees could, on occasions, perceive that impolite linguistic behaviour has occurred in the interview.

The only example of impoliteness found in my database of 20 encounters occurs in a M-F interview between John Humphrys (JH) and Hilary Armstrong (HA), broadcast on the 2/12/97. It is the female interviewee who accuses her male interviewer of impolite behaviour towards her by marking his linguistic behaviour as inappropriate using metalanguage and by posing a question to him, thus reversing the expectation of whose role it is to initiate the question-answer sequences. In the SCofP of political interviews, the interviewer and interviewee should mutually engage with one another through a process of joint negotiation in order to produce discourse for the overhearing audience. However, HA accuses JH of not allowing her to respond to his questions. The accusation by HA therefore shows that the norms of the SCofP, as far as she is concerned, are not being adhered to by the male interviewer.

I intend to demonstrate that impolite behaviour, as classified by the participant herself, is something that emerges at a discourse level, rather than at the level of a single speech act. HA does not judge JH’s behaviour as impolite due to a single speech act that he performs; rather it is his interactional behaviour over a stretch of talk during the interview that leads her to accuse him of not letting her have the conversational floor when she is entitled to it.

The topic of the interview is the Government’s announcement of a settlement which will determine the amount of money local authorities are to be given towards the cost of providing services. This issue has been highlighted on the news bulletin that took place fifteen minutes before JH and HA’s interview. Immediately preceding their encounter is a F-M interview with Anna Ford and a Liberal Democrat MP who argues that by sticking to Conservative spending plans, the Government will be unable to provide more finance as council tax bills will not be allowed to increase. The extract below shows JH questioning HA about the fairness of the settlement:

All of the questions posed by John Humphrys in this section are defined as antagonistic in nature. Following Holmes’ (1992) definition, antagonistic questions are aggressively critical assertions, which function to attack the interviewee’s position and demonstrate that it is wrong. They occur frequently in political interviews. Antagonistic questions in this extract take either the syntactic form of a question (lines 1-6, 8, 10) or appear in the form of a declarative which has the pragmatic function of a question (lines 13, 16-17, 21-22, 27-31, 33).

In lines 1-6, JH’s antagonistic question accuses the Government’s settlement of being unfair, thus challenging HA’s position and attempting to demonstrate that it is wrong. HA begins to respond to this, but JH successfully interrupts her answer with another antagonistic question ‘but it won’t be will it?’ [6] HA precedes to answer the question by repeating her initial utterance ‘the settlement’, but JH successfully interrupts again with another antagonistic question (line 10). For the third time HA precedes to answer the initial question commencing yet again with the repetition of ‘the settlement’. On this occasion she is allowed to proceed (lines 11-12), but JH successfully interrupts her again with the declarative ‘not all those other things there listed’ (line 13).

HA responds to JH’s declarative, successfully interrupting his utterance to agree with him (line 14). JH show his surprise by interrupting with ‘OH’. As HA attempts to explain why all of the things listed haven’t been taken into account commencing with ‘because’ (line 15), JH interrupts once more to declare that this is what Frank Dobson attacked the Tory government for doing, thus accusing HA of hypocrisy. She responds to this, accusing JH of not presenting a correct version of the facts (lines 18-20).

He interrupts her again in an attempt to demonstrate that her position is wrong (line 21), by quoting directly from Hansard. It is during this declarative utterance that HA accuses JH of not allowing her to answer his questions, thus accusing him of being impolite because, as far as she is concerned, he has broken the norms of the political interview. Her initial interruption attempt to do this is unsuccessful (line 23). However, she then proceeds to accuse JH of not wanting her to respond to the questions he is posing, interrupting him, though he does not concede the floor to her. In her accusation she reverses the expectation that the interviewer initiates the question-answer sequence by asking him an antagonistic question of her own: ‘I don’t think you want me to get a word in edgeways do you John? He denies that this is the case in his response which he repeats four times (line 24), but after he has completed his utterance he does allow HA to have a turn in the discourse without disruptively interrupting her (lines 34-42).

It is important to note that HA’s accusation is mitigated. She hedges the initial part of her utterance ‘I don’t think’, followed by a conventional metaphorical expression ‘get a word in edgeways’, and then laughs when she has completed her utterance (line 26). The joking tone adopted lessens the force of the antagonistic question, and ironically focuses on their respective roles of interviewer and interviewee, i.e. that the interviewee is expected to take up the majority of turn-taking time when responding to the interviewer’s questions. An unmitigated accusation could have led to conversational breakdown. This would be damaging both for HA and the Labour Party she represents if she is perceived to be unable to cope with JH’s aggressive style. HA thus avoids potential conversational breakdown by implicitly classifying her fellow interlocutor’s linguistic behaviour as being impolite, using metalanguage and reversing the role expectations of who should be the questioner in political interviews. JH is thus being impolite as he is not allowing HA to mutually engage with him in the joint enterprise of producing discourse for an overhearing audience.

In terms of gender, a potential explanation that could be suggested for JH’s behaviour by the culture difference/approach is that as a male speaker he will not pay attention to conventions of linguistic politeness and thus disruptively interrupts his female interlocutor on numerous occasions. The power/dominance approach could explain JH’s behaviour towards HA as yet another example of male domination, with the male interrupting the female on far more occasions in mixed-sex interaction. However, it is the female speaker who accuses her male interlocutor of displaying impolite behaviour towards her. Although her accusation is mitigated to avoid potential conversational breakdown, she nevertheless displays dominant behaviour by uttering an antagonistic question of her own which subverts the expected interviewer-interviewee power relationship at a local level of discourse management. Both the power/dominance and the culture/difference approaches would have difficulty in explaining this example of a female not paying close attention to the face needs of her male interlocutor, implicitly accusing him of inappropriate behaviour.

The CofP approach emphasises that gender is just one aspect of social identity that is being enacted in a CofP (see section 4). In political broadcast interviews, the interviewers and interviewees, as well as enacting their gender identity, are also enacting other parts of their social identity, such as their age, class and ethnicity. Furthermore, because the discourse takes place in an institutional context, they are also enacting their professional identities as either politicians or journalists, which is influenced by factors such as what political party they belong to or what broadcast network they appear on, what position they have within their party or within their broadcast network, and how long they have been an MP or a political interviewer. All of these aspects of identity can be expected to affect the discourse strategies male and female journalists and politicians use in political interviews.

By focusing on the institutional context of political broadcast interviews, it is hoped that the benefit of focusing studies of politeness on discourse in institutional settings has been emphasised. The neglect of impoliteness by previous linguistic politeness researchers has been overcome by using the CofP framework to theorise how notions of both politeness and impoliteness can be conceived. Instead of applying an alleged set of politeness universals to individual speech acts, the CofP approach demonstrates that politeness and impoliteness have different functions for different individuals depending on what kind of community they are interacting in.

The CofP approach also redefines the manner in which the relationship between language and gender should be conceived. Instead of treating men and women as monolithic categories, the focus on local practices enables language and gender researchers to stop searching for differences between male and female speech patterns and instead consider that gender may not be the only aspect of identity that is influencing linguistic behaviour in specific communities of practice.

However, this is not to suggest that gender as a variable is not significant. Although Freed (1996) points out that gender may not always be the most salient feature which affects speech patterns on all occasions, she argues that inspiration for language and gender studies should lie in the fact that, due to deeply embedded gender stereotypes that continue to operate in society, ‘there are a well-organized set of social expectations about who, women or men, will convey which social meanings’ (1996: 70) and, as a consequence, women and men have different access to specific social roles and activities.

Indeed, in the political arena, women are still heavily under-represented in the British House of Commons, and therefore appear far less frequently than male politicians on political broadcast interviews. Although there was an increase in female MPs in the 1997 General Election, taking the number of female MPs up from 9.2% to 18%, this decreased at the last election [7]. Furthermore, Walsh (1998) points out that when women have achieved the status of MP, they then face stereotyping within the political arena. She argues that there is ‘a gendered division of labour’ in political parties, and females are stereotypically seen as politicians who should ‘occupy a practical supportive role vis-à-vis the activities of visionary men’ (1998: 208). Therefore, due to this sex role stereotyping, women are far more likely to remain as peripheral members of political interviews as a SCofP at an individual level. Freed (1996) argues that as long as gender stereotyping exists:

Researchers will be obliged to consider gender (as a variable that interacts with other social variables) in the process of analysing the interaction of language and social life (1996: 70).

Adopting the CofP approach has highlighted that gender is not the only variable which affects linguistic behaviour. Instead gender should be seen as ‘a sex-based way of experiencing other social attributes like class, ethnicity and age’ (Eckert and McConnell-Ginet 1992a: 91), and also as a sex-based way of experiencing other aspects of identity, such as professional identity that are enacted when interaction takes place in an institutional context. Furthermore, adopting the CofP approach to theorise notions of politeness and impoliteness should avoid the decontextualised analyses that are a consequence of adopting Brown and Levinson’s politeness universals, thus bringing advantages to studies of language, gender and politeness.

1. I would like to thank Francesca Bargiela, Sandra Harris, Peter Stockwell and members of the Cross-Cultural Linguistic Politeness Group who have commented on draft versions of this paper. I am also indebted to the comments of two anonymous reviewers.

2. Brown and Levinson have been heavily criticised from a cross-cultural perspective. Researchers including Gu (1990), Spencer-Oatey (1992), Flowerdew (1999) and Bargiela et al. (this issue) have highlighted that Brown and Levinson’s notions of face and positive and negative politeness are culture-dependant. Although they are generally applicable to Western culture, other cultures do not share the same norms and conventions.

3. The database contains 20 dyadic political interviews broadcast on BBC Radio 4 from October 1997-February 1998. All of the data analysed in this paper are taken from this database. The interviews are taken from the Today programme, and were collected and transcribed according to the sex of the interviewer and the interviewee. The interviews were split into 4 separate cells to account for all possible interviewer-interviewee combinations by sex (F= female M=male): F-F, F-M, M-M, M-F. The interviewer is always referred to first in these pairings, and each cell contains 5 interviews. The encounters all have a political topic, and all of the interviewees are MPs apart from one female in the F-F category. Due to difficulties encountered in finding F-F political interviews in the designated recording period, it became necessary to include an interview with a female professor from the University of Cambridge that had a political topic (the effect of the Government’s education policy on their admission system).

4. The transcription conventions are as follows:

Bargiela, Francesca, Corinne Boz, Lily Gokzadze, Abdurrahman Hamza, Sara Mills and Nino Rukhadze (this issue [May 2002]), ‘Ethnocentrism, politeness and naming strategies’.

Bergvall, Victoria, Janet Bing and Alice Freed eds., (1996) Rethinking Language and Gender Research: Theory and Practice. New York: Longman.

Bergvall, Victoria, (1996) ‘Constructed and enacting gender through discourse: negotiating multiple roles as female engineering students’, in Victoria Bergvall, Janet Bing and Alice Freed eds., Rethinking Language and Gender Research: Theory and Practice, 173-201, New York: Longman.

Bergvall, Victoria, (1999) ‘Towards a comprehensive theory of language and gender’ Language in Society 28: 273-93.

Brown, Penelope and Stephen Levinson, (1978) ‘Universals in language usage: politeness phenomena’, in Esther Goody ed., Questions and Politeness: Strategies in Social Interaction, 56-311, Cambridge: Cambridge University Press.

Brown, Penelope, (1980) ‘How and why some women are more polite: some evidence from a Mayan community’, in: Sally McConnell-Ginet, Ruth Borker and Nellie Furman eds., Women and Language and Literature in Society, 111-136, New York: Prager.

Brown, Penelope and Steven Levinson, (1987) Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press.

Butler, Judith, (1990) Gender Trouble: Feminism and the Subversion of Gender Identity. New York: Routledge.

Cameron, Deborah, (1995) ‘Rethinking language and gender studies: some issues for the 1990s’, in Sara Mills ed., Language and Gender: Interdisciplinary Perspectives, 31-44, New York: Longman.

Cameron, Deborah, (1996) ‘The language-gender interface: challenging co-optation’, in Victoria Bergvall, Janet Bing and Alice Freed eds., Rethinking Language and Gender Research: Theory and Practice, 31-53, New York: Longman.

Cameron, Deborah, (1997) ‘Performing gender identity: young men’s talk and the construction of heterosexual masculinity’, in Sally Johnson and Ulrike Hanna Meinhof eds., Language and Masculinity, 47-64, Oxford: Blackwell.

Cheng, Rong, (2001) ‘Self-politeness: a proposal’, Journal of Pragmatics 33: 87-106.

Coates, Jennifer, (1996) Women Talk. Oxford: Blackwell.

Crawford, Mary, (1997) Review of Holmes’ ‘Women, Men and Politeness’ Language in Society 26: 426-429.

Culpeper, Jonathan, (1996) ‘Towards an anatomy of impoliteness’ Journal of Pragmatics 25: 349-67.

Day, Robin, (1975) Day by Day. London: William Kimber.

Day, Robin, (1991) Grand Inquisitor. London: Pan.

Eckert, Penelope and Sally McConnell-Ginet, (1992a) ‘Communities of Practice: where language, gender and power all live’, in Kira Hall, Mary Bucholtz and Birch Moonwomon eds., Locating Power: Proceedings of the Second Berkeley Women and Language Conference, 89-99, Berkeley CA: Berkeley Women and Language Group.

Eckert, Penelope and Sally McConnell-Ginet, (1992b) ‘Think practically and look locally: Language and gender as community-based practice’ Annual Review of Anthropology 21: 461-90.

Eckert, Penelope and Sally McConnell-Ginet, (1995) ‘Constructing meaning, constructing selves’, in Kira Hall and Mary Bucholtz eds., Gender Articulated: Language and the Socially-Constructed Self, 469-507, New York: Routledge.

Eckert, Penelope and Sally McConnell-Ginet, (1999) ‘New generalizations and explanations in language and gender research’, Language in Society 28: 185-201.

Edelsky, Carole, (1981) ‘Who’s got the floor?’ Language in Society 10: 383-421.

Eelen, Gino, (2001) A Critique of Politeness Theories. Manchester: St. Jerome’s Press.

Ehrlich, Susan, (1999) ‘Communities of practice, gender and the representation of sexual assault’ Language in Society 28: 239-56.

Fishman, Pamela, (1978) ‘Interaction: the work women do’ Social Problems 25:397-406.

Fishman, Pamela, (1980) ‘Conversational insecurity’, in Howard Giles, Peter Robinson and Philip Smith eds., Language: Social Psychological Perspectives, 127-132, Oxford: Pergamon Press.

Flowerdew, John, (1999) ‘Face in cross-cultural discourse’ Text 19: 3-23.

Freed, Alice, (1996) ‘Language and gender in an experimental setting’, in Victoria Bergvall, Janet Bing and Alice Freed eds., Rethinking Language and Gender Research: Theory and Practice, 54-76, New York: Longman.

Goffman, Erving, (1967) Interaction ritual: Essays on face-to-face behavior. New York: Anchor Books.

Greatbatch, David, (1986) ‘Aspects of topical organisation in news interviews: the use of agenda-shifting procedures by interviewees’ Media, Culture and Society 8: 441-455.

Gu, Yueguo, (1990) ‘Politeness phenomenon in modern Chinese’ Pragmatics 14: 237-257.

Harris, Sandra, (1991) ‘Evasive action: how politicians respond to questions in political interviews’, in Paddy Scannell ed., Broadcast Talk, 76-99, London: Sage.

Harris, Sandra, (2001) ‘Being politically impolite: extending politeness theory to adversarial political discourse’ Discourse and Society 12: 451-472.

Holmes, Janet, (1992) ‘Women’s talk in public contexts’ Discourse and Society 3: 131-150.

Holmes, Janet, (1995) Women, Men and Politeness. New York: Longman.

Holmes, Janet and Miriam Meyerhoff, (1999) ‘The community of practice: Theories and methodologies in language and gender research’ Language in Society 28: 173-183.

Johnstone, Barbara, Kathleen Ferrara and Judith Mattson Bean, (1994) ‘Gender, politeness and discourse management in same-sex and cross-sex opinion-poll interviews’ Journal of Pragmatics 18: 405-430.

Kasper, Gabriele, (1990) ‘Linguistic politeness: current research issues’ Journal of Pragmatics 14: 193-218.

Lakoff, Robin, (1975) Language and Woman’s Place.  New York: Harper Row.

Lakoff, Robin Tolmach, (1989) ‘The limits of politeness: therapeutic and courtroom discourse’ Multilingua 8: 101-129.

Lave, Jean and Etienne Wenger, (1991) Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press.

Mills, Sara, (forthcoming) ‘Rethinking politeness, impoliteness and gender identity’ in Evangelia Litosseliti and Jane Sutherland eds., Discourse Analysis and Gender Identity, Amsterdam: John Benjamins.

Mullany, Louise, (1999) ‘Linguistic politeness and sex differences in BBC Radio 4 broadcast interviews’ Leeds Working Papers in Linguistics and Phonetics 7:119-142.

Mullany, Louise, (2000) ‘The application of current language and gender theory to managerial meeting discourse’ Nottingham Linguistic Circular 15: 1-16.

Perez de Ayala, Soledad, (2001) ‘FTAs and Erskine May: Conflicting needs?- Politeness in Question Time’ Journal of Pragmatics 33: 143-169.

Spencer-Oatey, Helen, (1992) Cross-cultural politeness: British and Chinese conceptions of the tutor-student relationship, unpublished PhD thesis, Lancaster University.

Tannen, Deborah, (1984) Conversational Style: Analyzing Talk Among Friends. Norwood NJ: Ablex.

Tannen, Deborah, (1990) You Just Don’t Understand: Women and Men in Conversation. New York: William Morrow.

Thomas, Jennifer, (1995) Meaning in Interaction: An Introduction to Pragmatics. London: Longman.

Walsh, Clare, (1998) ‘Gender and mediatized political discourse: a case study of press coverage of Margaret Beckett’s campaign for the Labour leadership in 1994’ Language and Literature 7: 199-214.

Winter, Joanne, (1993) ‘Gender and the political interview in an Australian context’ Journal of Pragmatics 20: 117-139.

Wenger, Etienne, (1998) Communities of Practice. Cambridge: Cambridge University Press.

Zimmerman, Don and Candace West, (1975) ‘Sex roles, interruptions and silences in conversation’ in Barrie Thorne, Cheris Kramarae and Nancy Henley eds., Language, Gender and Society, 89-101, Rowley MA: Newbury House.