Chapter 5
Justifying injustice: broadcasters' accounts of inequality in radio
Rosalind Gill
What I want to do in this chapter is to use the discourse analytic approach developed by Potter and Wetherell (1987) to examine broadcasters' accounts for the lack of women disc jockeys (DJs) at the radio stations where they work. It forms part of a wider project concerned with the ideological features of DJs' on-air talk and how this is understood or 'read* by radio listeners, and the ways in which DJs construct their role and their audience.
One of the most striking features of popular radio in contemporary Britain is the lack of female DJs — at least during the day. When radio stations do employ women as presenters they tend to be on in the evenings when audiences have historically and consistendy been at their lowest. BBC Radio One is a good example of this: during weekday daytime programming, women are conspicuous by their absence1 whilst a small handful have been allocated night-time or weekend slots. Such inequalities in the number and status of women DJs have been well-documented (Baehr and Ryan, 1984; Karpf, 1980, 1987).
What has been less well researched, however, is how these inequalities are understood and made sense of by people working within radio. The aim of this chapter is to examine the accounts put forward by five male DJs and programme controllers (PCs) to explain the lack of female DJs both at their own stations and more generally. In this way we should learn something about how this inequality is perpetuated.
The approach used in this chapter is one which draws, with certain reservations (see Gill, 1991) on Potter and Wetherell's formulation of discourse analysis (Potter and Wetherell, 1987; Potter et al, 1990; Edwards and Potter, 1992). This work acknowledges theoretical debts to a variety of different approaches: linguistic philosophy and
76   The rhetorics of politics and identity
Justifying injustice   77
speech-act theory (Austin, 1962; Searle, 1969; Wittgenstein, 1953, 1980); semiology and post-structuraíism (Barthes, 1964, 1972, 1977; Derrida, 1978; de Saussure, 1974); ethnomethodology and conversation analysis (Garfinkel, 1967; Heritage, 1984; Wieder, 1974); and recent studies of rhetoric (Billig, 1987; Simons, 1989). From this disparate collection of work a coherent approach to language and discourse has been fashioned, which has been articulated primarily in relation to the sociology of scientific knowledge, and social psychology.
ANALYSING BROADCASTERS' ACCOUNTS
The analysis which follows uses a discourse analytic approach to study the accounts of five broadcasters from two independent local radio (ILR) stations for the lack of female DJs. In terms of female DJs neither station was atypical; one had no female DJs at all, whilst the other employed one female whose phone-in show was broadcast twice a week between 11.00 p.m. and 1.00 a.m. (For reasons of clarity I will treat these two stations as if they were one, which I shall call Radio Matchdale.) Two of the five broadcasters were DJs, two were programme controllers (PCs) and the fifth was both a DJ and a deputy PC. The interviews were conducted by a female interviewer and covered a range of topics including how the broadcaster saw his role and responsibilities, what he saw as the function of the station, his view of the audience, how much autonomy he felt he had, as well as questions about the lack of women DJs.2 What I am interested in are what Wetherell et al. (1987) have called the 'practical ideologies' through which gender inequalities in the employment of DJs are understood. The transcripts were analysed to find the broad types of accounts being offered by the broadcasters for the lack of women DJs. Five different types were identified, each organized around a particular claim, such as 'women don't apply* or 'the audience prefers male DJs*. What I want to stress is that these were not alternative accounts which were espoused by individual broadcasters. Rather, the DJs and PCs all drew on and combined different and contradictory accounts for the lack of women DJs. The analysis is divided into four sections, which deal with four of the five accounts, and examine some of the warrants which were put forward to support each broad account. The aim is not to provide an exhaustive analysis but simply to give a sense of how discourse analysis can be used to analyse this sort of material.
Accounting for inequality: (1) 'Women just don't apply*
The first and most prevalent type of account offered for the lack of female DJs was organized around the claim that women do not 'apply' to become radio presenters. Four out of the five broadcasters drew on this idea. ((.) indicates a pause in speech.)
Extract one (Goodman)
Int:     Why do you think there are so few female DJs? DJ:      (laughs) probably because they don't apply. It's, it's that literally is it.
Extract two (Dale)
DJ:     It's a more popular sort of occupation to men. We get a lot
of tapes from people who want to be DJs and they're all from
men.
Extract three (Chapman)
PC:    It's a question that. I get tapes from hopefuls on my desk every day of the week and none of them are ever women.
Extract four (Lightfoot)
PC:    I get all the applications to come in here (.) We get about
400 a year (.) We've had none from women in the last year.
Not one to be a presenter.
One of the most interesting features of these interviews is that for each of the broad explanations put forward to explain the lack of women DJs, such as the one above that women do not apply, the broadcasters spontaneously offered further accounts, often constructed around little narratives or stories. These can be understood as ways of warranting their explanations which make them sound more plausible. In the case of the claim that no women apply to become DJs, four different types of supporting account can be identified — that women are not interested in becoming DJs, that 'education and social process' does not prepare them for it, that they opt instead for jobs in journalism or television, and that they are put off DJing because 'it's a man's world'. Here I will examine two of these.
Accounting for women's non-application: 'There aren't many ... who are interested in doing it*
In the following extract the PC, Chapman, has just been explaining how Radio Matchdale recruits its staff.
78   The rhetorics of politics and identity
Justifying injustice   79
Extract five (Chapman)
PC: and it's where people come from (.) so in hospital radio there aren't many women DJs (.) there aren't many women DJs in pubs (.) there aren't many female DJs (.) especially teenage age which is when we're looking to bring people like (.) who are interested in doing it.
What Chapman is doing here is accounting for the lack of women DJs at Radio Matchdale by reference to the lack of women DJs in the station's traditional recruiting ground — hospital radio and pubs. But he does not stop there: he offers an explanation for this — 'there aren't many women ... who are interested in doing it'. In making this psychologistic claim, Chapman does two things. First, and most straightforwardly, he denies that there is any real or genuine motivation on women's part to becojme DJs. It would be interesting to discover just how common is this pattern of accounting. I want to suggest that the idea that oppressed groups do not 'really' want to change their position, is one frequently drawn on by members of dominant groups in order to justify their actions or inaction.
Second, this assertion serves to deflect criticism or charges of sexism from radio stations in general and from Radio Matchdale in particular. It gives the impression that radio stations would be happy to take on women as DJs but that they are faced with a wall of disinterest from women. The idea that radio stations are battling against women's lack of interest in DJing, and are even putting in extra effort to find female presenters is reinforced by Chapman's comment a few moments later: 'so we have to look hard'. That Chapman is looking hard for female DJs establishes his 'good faith*, his lack of sexism, and responsibility for the lack of female DJs is placed firmly on women's shoulders.
The idea that women are not interested in becoming DJs is also drawn on by Goodman.
Extract six (Goodman)
DJ: I'm sure there's a helluva lot of them out there that would be really er good communicators but have never even given a thought of doing it (.) Maybe they're doing a job that either pays more money or is more interesting to them.
This is an explanation which rests upon an implicit view of society as characterized by social mobility. It suggests that women could become DJs but have chosen to do other work. The salary and
satisfaction of a radio presenter is downgraded. In fact, women's non-application is made to appear eminently sensible and rational when contrasted with the likelihood that they are doing better paid or more interesting jobs. Again, the picture presented of women doing other highly paid and satisfying work serves to undermine the notion that women really wish to become radio presenters.
Accounting for women's non-application: 'It's a man's world'
A different explanation for women's claimed non-application is put forward by the PC named Lightfoot.
Extract seven (Lightfoot)
PC:     It's also very much a man's world so they're picked on if they
are here (.) you know a woman has got to assert herself pretty
definitely if she's working in radio.
It is clear that the phrase 'it's a man's world* is being used to refer to much more than the simple numerical superiority of males at the radio station, since it is used to explain the 'fact* that women are 'picked on'. What's interesting, however, is the fact that it is not formulated as sexism. To be 'picked on' is to be subjected to nasty and unjust behaviour, but it is the behaviour of individuals — something that can be highlighted by trying to imagine a formulation in which a radio station was deemed to 'pick on' women. The choice of this construction serves further to play down any notion of structural inequality or institutional practices.
It is significant that for the first time a feature of life within the radio station is introduced to account for women's non-application. But finally, the problem is not one for the men at the radio station, nor for the radio station as a whole to deal with, but rather it is up to each individual woman to 'assert herself pretty defimtely if she's working on radio*.
Accounting for inequality: (2) audience objections: 'It's a bit strange to have a woman talking to you*
A second type of explanation for the lack of female DJs focused on the audience's expected or apparently 'proven' negative reaction to female presenters.
80   The rhetorics of politics and identity
Justifying injustice   81
Extract eight (Dale)
DJ: Research has proven (.) and this is not mine but it's echoed by many surveys throughout the years (.) that people prefer to listen to a man's voice on the radio rather than a woman's voice. Women like to hear men on the radio because they're used to it (.) and it's a bit strange to have a woman talking to you. And men like hearing men on the radio (.) perhaps because they're just chauvinistic. Whatever the reasons, research has borne out this fact you know that people like to have men on the radio (.) and we just go along with the consensus of opinion. We do have women — Marie does an admirable job on the phone-in. We've got a lot of women newscasters so you know there's certainly no prejudice.
The first thing to note about this extract is that it came from Dale, (see page 77) arguing that the lack of female DJs can be explained by the fact that no women apply. Here, he constructs a different explanation for the small number of female DJs. Suddenly, the lack of female presenters looks less like the result of a lack of applications from women, and more like a deliberate policy not to employ women - because of audiences' alleged preference for men. In both formulations, it should be noted, the radio station is depicted as blameless - in the first because it is women themselves who are choosing not to apply and in the second because the radio station is merely serving its audience by giving it the presenters it wants. Several authors have pointed out that accounts which merely appear to be describing the world are more persuasive than accounts which seem to be motivated by particular interests or psychological dispositions of the speaker (Smith, 1978; Potter and Wetherell, 1988; Edwards and Potter, 1992). Thus, one of the problems for a speaker is to accomplish the 'out-thereness' (Potter and Wetherell, 1988) of their claims. One way this is achieved by Dale in this extract is through the discursive work being done by 'research* and 'surveys'.
Audience objections: research, surveys and more research
These terms give authority to Dale's claims. In the fint sentence alone Dale talks about 'research' and 'surveys' implying that these are separate rather than different words for the same thing: not only has research shown it, Dale argues, but it has also been echoed by 'many surveys*. The use of these terms and their associated
vocabularies such as 'proven' lend credence and a sense of objectivity to Dale's claims.
The terms also serve to distance Dale personally from the claim that listeners would prefer to listen to a man. It is constructed not as an aspect of his own beließ, not an opinion, but rather something 'out there' which 'research* and 'surveys* have 'proven'. Dale's own role, as someone involved in the recruitment and appointment of staff, in mediating between research findings and appointment policy, is completely glossed over in his talk. The research findings which 'prove* that listeners prefer male presenters and the lack of female DJs are presented as related together in a way which is totally independent of human action.
Audience objections: a 'new sexism*?
One of the most interesting features of this extract is the striking parallel with what has become known as ^new'racisť discourse (Barker, 1981). This type of discourse is dianctén^edt^nffié'tén^jcy to justify racist acts or legislation in non-racíal terms, often drawing on other values such as equality and fairness (Billig, Í988). ft is ajsQ marked by denials of prejudice, frequently accompanied by the claim that it is the liberal anti-racists who are the real racists (Barker, 1981; Billig, 1988). Perhaps the most straightforward type of denial takes the form of the 'disclaimer' (Hewitt and Stokes, 1975). Typically, a statement such as 'I'm not being racist' is followed by a 'but' which precedes the expression of something which could easily be heard as racist.
The widespread existence of denials of prejudice has led to some discussion of the possibility that there exists a 'cultural norm against prejudice' (Barker, 1981: Van Dijk, 1984; Reeves, 1983; Billig, 1988).
Racism is generally taken to be the prototypical example of prejudice; and indeed 'prejudice* is often used as if it were synonymous with racism. Yet, if we look back at extract eight we see that there are significant similarities with the 'new racist' discourse. The most obvious of these is the disclaimer — 'We've got a lot of women newscasters so you know (.) there's certainly no prejudice.' It does not take the classic form discussed by Hewitt and Stokes (1975) — it is retrospective rather than prospective — but the work it is doing in the extract in attempting to disclaim a prejudiced identity is the same as that identified by researchers studying racist talk. In the extract here, the disclaimer is reinforced by the contrasts which
82   The rhetorics oř politics and identity
Justifying injustice   83
are established between men who demand male presenters because they are chauvinist and the women who do so from force of habit and the radio station where 'there's certainly no prejudice'. It is worth noting Dale's use of the notion of 'chauvinism* and contrasting it with the term 'picked on' discussed in the consideration of extract seven. It is an interesting indication of the fact that broadcasters do have access to the notion of chauvinism, which, although not politicized in the same way as 'sexism', does at least have the merit of suggesting a pattern to discrimination. The broadcasters use this notion only to do particular work: not as a characterization of the radio station's behaviour, but rather an attitude with which Radio Matchdale can be contrasted favourably.
A further notable similarity with 'new racist' discourse is to be found in Dale's claim that 'we just go along with the consensus', where Dale presents himself as a mere victim oí other people's prejudice. This 'I'm not prejudiced myself but the audience wouldn't like it' type of accounting bears such a similarity to new racist talk that it suggests that the existence of a 'new sexism' might be worth investigating.
Accounting for inequality: (3) gender differences: 'Those things are not as advanced ... as nur as women are concerned as with men'
A third type of explanation for the small number of female DJs focused on women's putative lack of the qualities and skills necessary to be a DJ. The following extract from Chapman is an example oí this kind of account. We will examine it in some detail.
The interviewer's question is a response to Chapman's claim (see the first section of this chapter) that none of the tapes he receives from applicants are from women.
Extract nine (Chapman)
Inf. Do you think there are a set of reasons why women are put off from entering the DJ world?
PC: (...) Presenters have to have a number of skills. They've got to have ... they've got to be very very dextrous (.) they've got to be very familiar with technical equipment (.) they've got to have a personality they are used to expressing and they've got to have a good knowledge of music as well as having a good personality (.) and those things are not as
advanced in my view as far as women are concerned as with men. Urn (.) urn (.) I've got to be able to sit somebody in a radio studio and they've got to understand what they're doing kind ofthing as well as being a good broadcaster and women (.) in their whole background are not brought up in that kind of environment.
Two aspects of this extract are immediately striking. First that Chapman does not appear to be answering the question he was asked. Instead of explaining why he thinks women are put off from applying for DJ jobs, he appears to be providing a justification for not employing women: 'I've got to be able to sit somebody in a radio studio ....' In this respect his answer is defensive. The second is that his opening words are extremely formal. One important effect of Chapman's use of the passive form and of his use of a list construction (however stumbled over) of attributes needed for radio presentation is to give the impression that certain impersonal, objective and, crucially, non-gendered, criteria are applied to the selection and appointment of DJs. What Chapman is suggesting is simply that women fail to meet these (necessary) standards.
It is worth looking at this point at the nature oí the skills and qualities which are formulated by Chapman as necessary for DJs. What is striking about the list is both its inexplicitness and the fact that the skills mentioned do not seem to be tied to stereotypes about gender. DJs have got to be 'very very dextrous', 'very familiar with technical equipment', have *a personality they are used to expressing' and 'a good knowledge of music'. With the possible exception of'familiarity with technical equipment', none of these qualities seems to fit more readily with stereotypes of masculinity than femininity. Indeed, if anything, the qualities appear to match more closely stereotypes of women: it is women, who, according to stereotype, are dextrous and good at expressing themselves. The significance oí this can be highlighted by rereading the extract, substituting 'men' for 'women*.
The point is, then, that the force of the passage derives from the list itself rather than from the specific items which comprise it. The only arguably stereotypical item is 'familiar with technical equipment' which is interesting both for its vagueness and for the fact that it suggests that potential DJs should already be working technical equipment. Again, this supports the impression that Chapman is accounting for not employing women rather than for why women do not apply.
84   The rhetorics of politics and identity
Justifying injustice   85
As with other explanations, Chapman spontaneously offered reasons to account for why women lacked the skills and qualities                 í
necessary for a DJ. í will examine just one of these.
Explaining gender difference: 'Education and social process'
For Chapman there seems to be nothing mysterious about why women tail to live up to the selection criteria for DJs. He accounts for it with reference to 'lay sociological explanations' (Potter and Wetherell, 1988).
PC: Those things in education and social process are not as advanced in my view as far as women are concerned as with men (...) and women (.) in their whole background are not brought up in that kind of environment.
Although the language is vague, it is clear that an explanation is being constructed around notions of the contrasting socialization and education of women and men. I am not here concerned with the 'truth* or adequacy of such an explanation but rather with what its articulation achieves for Chapman.
One of the functions of the use of this lay sociological theory for Chapman is to provide a mitigation for women's failure to meet the appointment standards for DJs. In a society where at least one strong ideological current emphasizes meritocracy and individual success, failure can easily appear as blameworthy. In this extract the lay sociological theory provides a mitigation by offering reasons or causes for women's putative failure — 'education and social process' — which make it understandable and thus less potentially blameworthy. The notions of'education' and 'social process' are ideal for doing this kind of work since they are both extremely vague and suggest no particular agency on women's parts. Yet it should be remembered that women's 'failure' is as much Chapman's construction as the mitigation for this failure. If he characterizes women as lacking the skills and qualities to become DJs, why should he also provide a mitigation for them?
Potter and Wetherell (1988) discovered similar simultaneous constructions of blame and mitigations in Pakeha (White New Zealanders') discourse about people from the Pacific Islands living in New Zealand, and have suggested why this pattern should occur. They argue that one of the problems for speakers of producing negative claims about a group of people is that it can easily be
heard as prejudice, something (as discussed in the second section of this chapter) the speaker may be anxious to avoid. One of the ways in which the hearability of this can be reduced is 'to reduce the force of the blamings being made' (Potter and Wetherell, 1988: 64). And in turn one of the ways that this can be accomplished is by the use of a mitigation. In the current example. Chapman could easily be heard as an out-and-out sexist, arguing quite simply that women are not as good as men. By providing a mitigation Chapman reduces the availability of this charge.
A second related function of Chapman's use of lay sociological theory is to emphasize the 'out-thereness' of his characterization of women. That is, his spontaneous production of an account for women's 'failure' actually reinforces the idea that it is because women fail to meet the selection standards that there are so few women DJs. Just as the terms 'research' and 'surveys' give the impression of facticity to claims so the sociological notions suggest that Chapman is merely describing the world as it is. Chapman's independence from the object of discussion is reinforced by the regretful tone of his next remark:
Extract ten (Chapman)
Int:     Well I think that in the last say ten twenty years things have
changed (.) have PC:     Yes they've changed. But they haven't changed enough.
The implication is that the world is not the way he would like it to be, but that is the way it is — regardless of his motivation.
(A contrasting explanation focusing on 'natural' gender differences and aptitudes was put forward by the DJ, Goodman. I have discussed this elsewhere (Gill, 1991).)
Accounting for inequality: (4) women's voices: too *shrill% too Musky* and just plain 'wrong'
The fourth type of account put forward by the broadcasters to explain the lack of women DJs centred on women's voices. In making these claims the broadcasters placed themselves within a long tradition in British broadcasting. Women's exclusion from particular types of employment within the media on the basis that their voices are 'unsuitable' is now well-documented (Ross, 1977; Karpf, 1980; Kramarae, 1989).
86   The rhetorics of politics and identity
As recently as the 1970s similar reasons were being offered by the BBC for their refusal to employ women as newsreaders. Mileva Ross showed how the most pervasive arguments were that women's voices were 'too high' or 'lacked authority'. In the words of Jim Black, then editor of Radio Four:
If a woman could read the news as well as a man then she could do it. But a newsreader needs to have reliability, consistency and authority. A woman may have one or two of these things but not all three. If a woman were to read the news no one would take it seriously.
(quoted in Ross, 1977) As Ross wryly comments, did he expect us to fäll about laughing or just to disbelieve it? His colleague Robin Scott was of a similar opinion. He said it was 'unnatural* for women to read the news: 'There's always bad news about and it's much easier for a man to deal with that kind of material' (quoted in Ross, 1977).
The concerted efforts of the feminist campaigning group, Women In Media, led to a small handful of women being appointed as newsreaders by 1975. Jim Black spoke of'an awful lot of special training' which had 'come to fruition' leaving two female newsreaders to take their place alongside their fifteen male colleagues. Black commented: 'I think we have got the right mix now. I don't want Radio Four to sound all-female ... If you have two on it sounds a lot' (quoted in Ross, 1977).
All the DJs and PCs interviewed in this research found women's voices worthy of comment. Although one remark by Toller seems to be a positive one — he says that he does not think the Radio Two presenter Gloria Hunnifbrd has a shrill voice — the mere fact that he felt it worthy of comment is significant. There were no comparable remarks about men's voices. Next, we examine the rather more lengthy comments of Goodman when asked to elaborate upon his claim that women's voices are 'not right'. Extract eleven (Goodman)
DJ: As I said to you before (.) people are sensitive to voice (.) they pick up a lot in a voice. They can see it as exuding friendliness, sarcasm, angriness or whatever and if it happens to be(.) and if a woman's voice sounds grating or high (.) shrill, then that will switch them off. If it sounds dusky and sexy (1.0) unfortunately that switches them on (.) now Marie has
Justifying injustice   87
got a dusky, sexy, deep voice perfect for it (.) she's actually nothing like that when you meet her (.) she's a very sweet lady but she's not like that but people are conned totally by the voice.
The extract is similar to that discussed in the second section of this chapter in that Goodman is involved in justifying not employing women as DJs by reference to what listeners like or dislike. However, whereas on p. 80 the listeners* resistance to female DJs was characterized as 'chauvinist' or merely habitual, and the DJ presented himself as regretfully just 'going along with the consensus', here listeners* putative reservations about (some) female voices are characterized as perfectly reasonable. What could be heard as prejudice is recast as 'sensitivity'. Listeners' sensitivity, unlike their chauvinism, is not to be regretted. The radio station merely translates this sensitivity into appointment decisions.
One of the ways in which listeners* sensitivity to women's voices is brought off as reasonable by Goodman is through the subtle linking of notions of sensitivity to particular emotional or motivational states (angriness, friendliness, sarcasm) and sensitivity to particular vocal pitches. Goodman starts by asserting that people see voices as 'exuding friendliness, sarcasm, angriness or whatever' and goes on 'and if a woman's voice sounds grating or high (.) shrill then that will switch them off. The 'reasonableness' of this second phrase is effectively achieved by its ostensible connection to the first. For whilst sensitivity to friendliness or sarcasm seems admirable, 'sensitivity' to pitch may betoken prejudice.
It is worth briefly considering the way that pairs of words are used to characterize women's voices. The first thing to note is that the notion of what is 'shrill' or 'dusky' is not unproblematic: these are not neutral words to describe pitch — whatever a neutral word may be. Indeed, Goodman starts by characterizing some women's voices as 'high* but then substitutes a word which has far more richly negative connotations — shrill. To object to (or be 'sensitive' to) 'high' voices could be heard as blameworthy, but to object to 'shrill' voices seems perfecdy reasonable — it is a word which contains an evaluation (cf. Wowk, 1984).
Second, we should note the way the second word in each pair is used to add to and to describe the first — giving the impression that, for example, dusky is sexy. I want to argue that it is not insignificant that the two examples used seem to fit almost perfecdy
88   The rhetorics of politics and identity
Justifying injustice   89
with two commonly used stereotypes of women — the 'nag' and the 'femme fatale'. This is not to imply, however, that these stereotypes are somehow static and non-changing.
What Goodman seems to be doing is presenting a 'no-win' situation for women. If they sound 'grating and shrill' then that 'switches listeners off. This phrase has a fascinating double meaning. Goodman may mean simply that shrill or grating female voices displease people, turn them off. But his phrase also serves to remove all agency and responsibility for switching the radio offfrom listeners, and places it instead on women's voices. In this way people's sensitivity comes to seem perfectly reasonable; it is women's voices in themselves that do the switching off, and are therefore blameworthy.
If a woman sounds 'dusky and sexy* 'that switches them on'. One might imagine that this is exacdy what the radio station would want, but Goodman treats it ambivalently describing it as 'unfortunate', but also describing Marie's 'dusky sexy deep' voice as 'perfect for it*. This becomes explicable if we understand the 'it' for which Marie is apparently 'perfect' as her own show (which is broadcast between 11 p.m. and 1 a.m.) rather than more prime-time radio presentation. It also illuminates a further nuance of meaning for the word 'dusky' — suggesting appropriateness for nighttime broadcasting. More generally, it seems that Goodman's ambivalence about 'switching them on' is due to its sexual connotations. This would denote a level of sexual assertiveness deemed unacceptable in a woman.
Goodman's remarks about Marie are also interesting for three other reasons. At one level they serve simply as a reminder (in what may be for the speaker a critical interpretative context) that the radio station does haue a female presenter (albeit only one who is relegated to the wee small hours). By explicitly praising Marie's presentation style, Goodman reduces the hearability of sexism. This also accounts for his ambivalence: for he is both justifying the non-employment of women as DJs and attending to the possibility that he may be heard as sexist.
Second, the passage is interesting because it supports the idea raised earlier that 'dusky' and 'sexy' are tied to the notion of the 'femme fatale'. What Goodman seems to be saying is that she sounds dusky and sexy, but actually she is not — she is no 'femme fatale'.
Finally, the passage is significant because it reasserts the importance of voice — 'people are conned by the voice totally'. However, it does so in such a way as to undermine completely Goodman's earlier claim that people are 'sensitive to voices' and can 'pick up
a lot in a voice'. For listeners so easily 'conned' the notion of 'sensitivity' as a justification for not employing women who apply begins to look a litde thin.
It is tempting to suggest that the only way a woman can succeed is by sounding like a man. And indeed, this is what Goodman seems to have concluded.
Extract twelve (Goodman)
DJ:     They they build a mental picture so it's really your voice (.)
if your voice is right. For some women that can be hard
because their voice is naturally higher.
If we leave aside the considerable debate over the supposed differences in the pitch of male and female voices (see, for example, Spender, 1985) what is clear from this short extract is that the male voice is being used as the norm against which other voices are judged for their appropriateness, implicit in the extract is the idea that a low, male voice is somehow naturally right for DJs. This extract is a very good example of what has been called the 'male as norm' phenomenon (Spender, 1985; Griffin, 1985) and it is against the background of this norm that becoming a DJ can be judged 'hard' for 'some women'. Significantly, although the male voice is presented as the 'natural' 'right* voice for a DJ it is presented as non-gendered.
DISCUSSION
This chapter has examined some of the practical ideologies through which the lack of female DJs is explained and justified. What I have tried to show is that far from the broadcasters each espousing a particular attitude or advancing a specific explanation to account for the lack of women DJs, each had available a whole range of ways of accounting, which they drew on selectively in the interviews. Overall, I pointed to a pervasive variability in broadcasters' accounts which would be overlooked or suppressed by more traditional, social psychological approaches. The accounts constructed by broadcasters were flexible, inconsistent and sometimes contradictory. The claim by the DJs and PCs that no women apply, within moments of explanations by those same broadcasters about why those women who do apply are not suitable for DJing, is simply the most dramatic example of this, and poses severe problems for attitude theories and all other approaches which work with a realist view of language. Rather than seeing such assertions as unproblematic
90   The rhetorics of politics and identity
Justifying injustice   91
statements of fact, discourse analysis argues that they are better understood in terms of their discursive functions.
The chapter also looked in detail at the construction of particular accounts, examining how broadcasters attempted to accomplish them as factual or 'out-there' and discussing the way the accounts offered seemed to make the lack of women flow apparently self-evidently from the explanations. Specifically, all the accounts put forward by broadcasters to explain the lack of women DJs constructed the reasons as lying in women themsetues or in the wants of the audience. The role of the radio station was made invisible in these accounts, and discussions of employment practices and institutional sexism were conspicuous by their absence. In this way broadcasters were able to present themselves as non-sexist, whilst they simultaneously justified the lack of women at the radio station.
None of the DJs or PCs said at any point that they did not think that women should be employed as DJs. On the contrary, they were keen to point out their lack of sexism ('there's certainly no prejudice') and that they were 'looking hard' for female presenters. However, what they produced were accounts which justified the exclusion of women, in providing these accounts for why there are so few female DJs now, the broadcasters also provided justifications for the continued absence of women in the future. The ideological effects of these discourses is to perpetuate inequality within radio stations.
One potential disadvantage with this kind of discourse analytic approach is that it does not produce the broad empirical generalizations which are developed in more traditional sociological and social psychological work. Thus, the analysis here cannot be understood as identifying a universal process underlying gender discrimination in employment or even sexism in radio stations. What it does is examine the explanations put forward by a particular group of white, male broadcasters in a particular social and historical context, in the course of interviews with a white, middle-class and almost-certainly-feminist (from the broadcasters' perspective) woman. For discourse analysis, die failure to theorize universal processes is not a weakness but an inevitable consequence of the fact that explanations are always constructed out of particular interpretative resources and designed for specific occasions (Wetherell and Potter, 1988). Thus, we might expect to see different accounts put forward to explain the lack of women DJs if the broadcasters had been talking amongst themselves or had been interviewed by a male. This does not invalidate
the research but merely serves to emphasize discourse analysts' point about the constructive, action-orientated nature of language.
A further disadvantage for people thinking about doing discourse analytic work is the sheer effort involved. Discourse analysis is extremely labour-intensive, and the time taken up conducting interviews and transcribing them can be considerable. In order to produce this analysis, for example, I produced five transcripts which totalled 114 pages of typed A4 script. Added to this is the length of time it takes to learn and develop the skill of analysis. As Wetherell and Potter (1988) have argued, discourse analysis is a craft skill, and it is possible to work with an analytic schema for several days only to find that it cannot be validated by the available materials.
However, against this, discourse analysis has considerable value. It offers both a practical and theoretically coherent way of analysing a whole variety of talk and texts, taking them seriously in their own right (not as vehicles for some underlying psychological reality) and treating them in their full specificity.
Discourse analysis also constitutes a systematic approach to the evaluation of texts whose findings are open to evaluation. Reports of discourse analytic work include either full transcripts or samples of the analytic materials used so that readers are able to assess the success of the interpretations, and indeed, offer alternatives (Wetherell and Potter, 1988).
Finally, and most importandy, discourse analysis offers a new way of understanding ideology. It sees ideological discourse not as a fixed subset of all discourse which works in standard recurrent ways and is defined by its content or style, but rather as a way of accounting. It highlights the fact that what is ideological cannot be straightforwardly read off: propositions do not come with their ideological significance 'inscribed on their backs' and nor is the operation of ideology limited to discourse which naturalizes, reifies or legitimizes — or any of the other familiar modes (Thompson, 1988). Discourse analysis suggests that what is ideological is an analytic question. In the present analysis I hope to have shown that even as broadcasters declared their desire to see more women DJs, they produced discourse which was ideological — because the accounts they produced served to justify and perpetuate inequality within radio stations.3
92   The rhetorics of politics and identity
NOTES
1    Except in the new 'weathergjrľ (sic) role in which female news or weather readers or production assistants are used by male DJs to feed them witty one-liners and create an impression of relaxed banter.
2   1 would like to thank Sue Reilly for her help in carrying out these interviews.
3   I would like to thank Michael Billig, Jonathan Potter and Andy Pratt for their comments on an earlier draft of this chapter.
REFERENCES
Austin, J.L. (1962) How to do Things with Words, Oxford: Clarendon Press. Baehr, H. and Ryan, M. (1984) Shut Up and Usten: Women and Local Radio:
A View from the Inside, London: Comedia. Barker, M. (1981) The New Racism, London: Junction Books. Barthes, R. (1964) Elements of Semiology, New York: Hill & Wang.
------(1972) Mythologies, London: Paladin.
■------(1977) Image-Music-Text, London: Fontana.
Billig, M. (1987) Arguing and Thinking: A Rhetorical Approach to Social
Psychology, Cambridge; Cambridge University Press. ------(1988) 'The notion of "prejudice": some rhetorical and ideological
aspects', Text 8: 91-110. Coward, R. (1984) Female Desire:  Women's Sexuality Today, London:
Paladin. Derrida, J. (1978) Writing and Difference, London: Roudedge & Kegan
Paul. De Saussure, F. (1974) Courses in General Linguistics, London: Fontana. Edwards, D. and Potter, J. (1992) Discursive Psychology, London: Sage. Garfinkei, H. (1967) Studies in Ethnomethodology, Englewood Cliffs, NJ:
Prentice Hall. Gill, R. (1991) 'Ideology and popular radio: a discourse analytic examination of disc jockeys* talk', unpublished PhD thesis, Loughborough
University. Griffin, C. (1985) Typical Girls: Young Women from School to the Job Market,
London: Roudedge & Kegan Paul. Heritage, J. (1984) Garfinkei and Ethnomethodology, Cambridge: Polity Press. Hewitt, J.P. and Stokes, R. (1975) 'Disclaimers', American Sociological Review
40: 1-11. Karpf, A. (1980) 'Women and Radio', Women's Studies International Quarterly
3: 41-54. ------(1987) 'Radio Times; private women and public men', in K. Davies,
J. Dickey and T. Stratford (eds) Out of Focus: Writings on Women and
the Media, London: The Women's Press, Kramarae, C. (1989) Technology and Women's Voices, London: Sage. Potter, J. and Wetherell, M. (1987) Discourse and Social Psychology: Beyond
Attitudes and Behaviour, London: Sage.
Justifying injustice   93
------(1988) 'Accomplishing attitudes: fact and evaluation in racist discourse',
Text 8: 51-68.
Potter, J., Wetherell, M., Gill, R. and Edwards, D. (1990) 'Discourse analysis: noun, verb or social practice*, Philosophical Psychology 3: 205—17.
Reeves, W. (1983) British Ratial Discourse: A Study of British Political Discourse about Race and Race-related Matters, Cambridge: Cambridge University Press.
Ross, M. (1977) 'Radio*, in M. Stott and J. King (eds) Is This your Life? Images of Women in the Mass Media, London: Virago.
Searle, J.R. (1969) Speech Acts, Cambridge: Cambridge University Press.
Simons, H. (1989) Rhetoric in the Human Sciences, London: Sage.
Smith, D. (1978) *K is mentally ill: the anatomy of a factual account', Sociology 12: 23-53.
Spender, D. (1985) Man Made Language, London: Routledge.
Thompson, j. (1984) Studies in the Theory of Ideology, Cambridge: Polity.
Van Dijk, T. (1984) Prejudice in Discourse: An Analysis of Ethnic Prejudices in Cognition and Conversation, Amsterdam: John Benjamins.
Wetherell, M. and Potter, J. (1988) 'Discourse analysis and the identification of interpretative repertoires', in C. Antaki (ed.) Analysing Everyday Explanation: A Casebook of Methods, London: Sage.
Wetherell, M., Stiven, H. and Potter, J. (1987) 'Unequal egalitarianism: a preliminary study of discourses concerning gender and employment opportunities', British journal of Social Psychology 26: 59—71.
Wieder, L. (1974) 'Telling the code', in R. Turner (ed.) Ethnomethodology, Harmondsworth: Penguin.
Wittgenstein, L. (1953) Philosophical Investigations, Oxford: Blackwell.
------(1980) Remarks on the Philosophy of Psychology, Oxford, Blackwell.
Wowk, M. (1984) 'Blame allocation, sex and gender in a murder investigation', in Women's Studies International Forum 7: 75—82.