Frankie Boyle, why don't you just shut up?

Linguistic Research Seminar, Sheffield Hallam University “Frankie Boyle, why don't you stop talking?" - Combining sociolinguistic and corpus linguistic approaches to media language: The case of Mock the Week Robert Lawson (@dr_bob82)! School of English! Birmingham City University! Wednesday 2nd December Outline ❖ Women in comedy - status, prestige and integration.! ❖ Mock the Week and the ‘male bear-pit’! ❖ Data and methodology! ❖ Quantitative analysis - facts and figures ! ❖ Qualitative analysis - gender and the politics of representation! ❖ Conclusions and future directions The problem ❖ The position of women in comedy has traditionally been unstable and precarious.! ❖ They are often the target of sexist jokes, sexual innuendo, are viewed to be ‘less funny’ than men, and remain under-represented in many mainstream television comedy shows and live performances, particularly in the UK (Bird 2009; Bemiller and Schneider 2010; Hertz 2010; Kalviknes Bore 2010).! ❖ Even those women who do appear on mainstream comedy television programmes potentially face the challenge of getting their voices heard in an arena where it is often the loudest voice who wins. A case in point?: Mock the Week A case in point?: Mock the Week ❖ British comics, such as Jo Brand, Victoria Wood, Miranda Hart and Kate Smurthwaite, have commented on the difficulties of being a woman in the comedy circuit.! ❖ Structural inequalities have caused several prominent female comedians to refuse to participate in certain prime-time comedy shows, including Jo Brand, who quit Mock the Week in June 2009. ! ❖ Picture of the “‘all-purpose male oppressor’ who talks too much, interrupts and generally dominates conversations with women” (Johnson and Meinhof 1997: 11). Bearpit-ish in the extreme? (Brand 2009) A case in point?: Mock the Week ❖ Currently no empirical proof which either supports or refutes Brand’s claim that male panelists are more conversationally dominant than female panelists.! ❖ Drawing on data from season five of Mock the Week, this presentation critically examines Brand’s claim and presents a pilot analysis of conversational dominance within the show. ! ❖ We consider the extent to which conversational dominance is an effect of speaker gender or whether it is an effect of the relative degree of familiarity between speakers (Anderson and Leaper 1998: 230). Mock the Week ❖ Mock the Week is a weekly 30-minute (edited) comedy panel show which has been running since 2005 and is currently in its fourteenth series. ! ❖ Involves two teams of three panelists, with each team containing two permanent panelists (PP) and one guest panelist (GP). ! ❖ In series five (the series discussed here), the four permanent panelists were Frankie Boyle, Hugh Dennis, Russell Howard and Andy Parsons, all well-known British male comedians.! ❖ Data from ten episodes were transcribed by four transcribers (RL, UL, EF, and AT), cross-checked for accuracy. Tagging and data prep ❖ Transcribed both dialogic and monologic data! ❖ Marked up interruptions and overlaps, using…! ❖ …. XML - Extensible markup language. ! ❖ XML transcripts are easily extractable and searchable for - ! ❖ ❖ ❖ ❖ Total number of words;! Length of overlaps;! Who interrupts whom and how often;! Can be expanded to include other features. How? ❖ In QuickXML developed by Matt Gee (RDUES unit at BCU). ! ❖ Available for download @ http://rdues.bcu.ac.uk/tools.! ❖ Mac and PC compatible. ! ❖ “QuickXML allows you to edit an XML document in the same way as any other text editor, but also allows you to quickly add XML tags and attributes from an automatically generated list.” XML tagging ❖ A tag set (<markup>) which is easily expandable and can be a reusable resource by the wider academic community (when we release it…). ! ❖ Done on a turn-by-turn level.! ❖ Tags for <overlaps> and <interruptions>.! ❖ Each interruption was introduced by an opening tag <interruption by="dara" with ="hugh">.! ❖ and closed by a closing tag </interruption>.! ❖ Overlaps were indicated by an opening tag <overlap by="jo" with ="frankie">. ! ❖ and a closing tag </overlap> Example 1 Actual clip Example 2 Actual clip So what did we look at…? ❖ Quantitative stuff:! ❖ ❖ ❖ ❖ Number of turns! Number of words.! Interruptions! Qualitative stuff! ❖ ❖ Gender and stylistic repertoire ! (Politics of) representation Breakdown of season 5 Figure 1. Mock the Week, series five participants and appearances (episodes 1-10) Words, words, words 60000 n = 56,353 45000 Female Male 30000 Speaker sex Word Counts % FEMALE 2,295 (4%) MALE 54,058 (96%) TOTAL 56,353 Table 1. Word count totals (with Dara) Speaker sex Word Counts 15000 0 Word count Word count 2 Figure 2. Word count totals % FEMALE 2,295 (5.8%) MALE 36,933 (94.2%) TOTAL 39,228 Table 2. Word count totals (w/o Dara) Turns, turns, turns 2500 n = 2,324 1875 Female Male 1250 625 0 Turns 1 Figure 3. Turn count totals Turns 2 Speaker sex Turns 1 % FEMALE 125 (5%) MALE 2,199 (95%) TOTAL 2,324 Table 3. Turn count totals (with Dara) Speaker sex Turns 2 % FEMALE 125 (7.5%) MALE 1530 (92.5%) TOTAL 1,655 Table 4. Turn count totals (w/o Dara) Breakdown by speakers Male participants Turns Words Female participants Turns Words Dara O’Briain (10) Frankie Boyle (10) Hugh Dennis (10) Russell Howard (10) Andy Parsons (10) Ed Byrne (3) MarkWatson (2) David Mitchell (2) Michael McIntyre (1) Adam Hills (1) Ben Norris (1) Jimmy Tingle (1) Rhod Gilbert (1) Alun Cochrane (1) Average (m) Total 669 441 299 292 191 71 56 51 30 26 21 18 18 16 35 2,199 17,215 10,757 6,110 6,581 5,557 1,919 1,059 1,175 919 896 427 521 473 449 858 54,058 Jo Caulfield (3) Lauren Laverne (1) Jan Ravens (1) Gina Yashere (1) Fiona Allen (1) 51 23 20 20 11 1,102 326 375 332 160 Average (f) Total 18 125 328 2,295 Table 5. Distribution of turns and words per participant Breakdown by speakers ❖ Dara is highest - more than twice than Hugh, Russell and Andy. ! ❖ Frankie Boyle is the only PP who comes close. ! ❖ Andy Parsons is a bit of an outlier, producing only 191 turns in total (~19 turns per episode). But he’s a regular, which is surprising…! ❖ All of the male GPs contribute fewer turns than the average, and only Michael McIntyre and Adam Hills exceed the number of words produced. ! ❖ Female GPs mostly exceed average turn/word counts. ! ❖ Looking at the results by episode, female GPs tend to have less floor time than male GPs and rank mainly at the bottom of the scale in terms of word count. Guest vs. Panelist Figure 2. Turn count per panel member by permanent and guest status Guest vs. Panelist ❖ Guest 1 - combination of the different guest panellists contributing the least to an episode (< 20 turns p.e).! ❖ Guest 2 - those guests speaking the most (> 20 turns p.e, only one female GP included).! ❖ Frankie Boyle - most turns throughout series. ! ❖ Group guest 1, number of turns ranges between 10-20.! ❖ Group guest 2, number of turns ranges between 20-30.! ❖ Generally speaking, PPs contribute a higher number of turns than GPs. ! ❖ Male GPs sometimes come close to turn count of male PPs. Interruptions ❖ Interruptions - ‘violations of the turn-taking rules of conversation. The next speaker begins to speak while the current speaker is still speaking, at a point in the current speaker’s turn which could not be defined as the last word’ (Coates 2004: 112) ! ❖ Early sociolinguistic research viewed interruption and overlap as a substantiation of speaker dominance (cf. Smith-Lovin and Brody 1989: 425).! ❖ But possible to go further than this (as I discuss later). Overall number of interruptions 90 67.5 Interrupter - Interruptee Female - Male Male - Female Male - Male Total 45 22.5 0 M/M F/M M/F Total 8 5 84 97 % (8%) (5%) (87%) (100%) Overall number of interruptions ❖ The percentages of the total number of interruptions indicate that the majority of interruptions appear in a same-sex context.! ❖ In terms of mixed-sex interruptions, women interrupt men more frequently than vice versa.! ❖ If we weight the number of mixed-sex interruptions to the number of turns produced by men….! ❖ 0.4% (8/2,199) of all male turns are interrupted by female speakers, 0.2% (5/2,199) of all female turns are interrupted by male speakers.! ❖ Tentative support that women are more likely to interrupt men than they are to be interrupted (very small number of instances though). Type of interruptions Interrupter - Interruptee Disruptive! Neutral Supportive Total Female - Male 5!(0.2%) 0 3!(0.1%) 8!(0.4%) Male - Female 3!(0.1%) 1!(0.05%) 1!(0.05%) 5!(0.2%) Male - Male 51!(2.3%) 20!(0.9%) 13!(0.6%) 84!(3.8%) Total 59 21 17 97 ❖ Women use disruptive interruptions twice as often as men in mixed-sex interactions (although again, overall frequencies are rather low).! ❖ In same-sex male contexts, disruptive interruptions dominate.! ❖ Both male and female panellists tend to interrupt their interlocutors to take over the floor rather than to support what they have said in mixedsex interactions. Gender and stylistic repertoire ❖ ‘Gender-related variations in behaviour are influenced more by situational factors than by inherent individual differences between men and women. Relevant situational factors include characteristics about the interactions such as the number of persons, their respective genders and their relationship to one another’ (Anderson and Leaper 1998: 247). ! ❖ In this vein, our study has found that the situational role of the panellists needs to be taken into account (in addition to speaker sex) as guest panellists speak less than permanent panellists. ! ❖ Moreover, our results suggest that guest panellists and more specifically female guest panellists are granted (or make use of) the least speaking time overall. Gender and stylistic repertoire ❖ Is this differentiation part of the ‘stylistic repertoire’ (cf. Kirkham 2011; Eckert 2012) of the panellists?! ❖ To what extent do the linguistic practices of the panellists part of the discursive construction of their identities as ‘permanent’ and ‘guest’?! ❖ Closer attention to the deployment of interruption in context will shed light on the interactional and interpersonal roles such features play. Politics of representation ❖ Looking at the raw data, we see that male panelists tend to speak more (both in turns and number of words). But to what extent is this a result of editing choices? Who gets edited in/edited out? Who gets represented? What guides the editing choices?! ❖ Politics of representation potentially shaped by views on the respective abilities of males/female ‘to be funny’ (cf. Hitchens 2007; Sutherland 2013: 656; Burnett 2014). If this is the case, then prevailing views that ‘men are funnier’ will likely influence how ‘the cut’ is put together. ! ❖ BBC’s commitment to including at least one female participant on panel shows is one step towards addressing historical imbalances, but without a consideration of how editing practices might constrain the representation of female comedians, the policy is unlikely to make any significant in-roads to the issues identified at the start of this talk. Future directions? ❖ Qualitative focus on instances of interruptions, looking in more detail at interruptions in context.! ❖ Expand analysis to look at other features (lexical choice, discourse markers, linguistic accommodation etc). ! ❖ Expand data beyond series 5 of the show.! ❖ Expand data to include more shows (QI, Have I Got News for You etc).! ❖ Expand data beyond the BBC (E4, ITV etc). Future directions? ❖ More concretely, we’re in the initial stages of developing partnerships with the BBC to expand this body of research. ! ❖ As a ‘proof of concept’ paper, the research challenges some of issues raised by Jo Brand and others, but simultaneously raises other issues to do with representation etc. ! ❖ Thus, potential to develop a ‘BBC corpus’ examining linguistic practices in other areas of television (news reports, radio shows, dramas etc). ! ❖ This could be used, for example, to consult on how their editing practices might be altered/improved and how issues of gender are implicated in language. Thank you! Any questions? ! [email protected] [email protected] References ❖ Brand, Jo (2009). ‘Women - beware panel shows!’. [Online]. [Accessed 30 November 2015]. Available from: http://www.theguardian.com/culture/2009/jun/10/ television-panel-shows-jo-brand. Conclusions ❖ The total number of words (4%) and turns (5%) contributed by women is rather low.! ❖ In addition to speaker sex, however, our study has found that the variable of situational role needs to be taken into account as guest panellists speak less than permanent panellists. ! ❖ Our results suggest that guest panellists and more specifically female guest panellists are granted (or make use of) the least speaking time overall. Conclusions ❖ Concerning interruptions, we found that in mixed-sex contexts the number of interruptions is higher for female-male interactions, with female panellists interrupting men more frequently than vice versa. ! ❖ As all female participants are guest panellists, this interlinks with the tendency of guests to interrupt more than permanent panellists, indicating that it is more difficult for them to start a turn than for the permanent panellists, who are familiar with the show. ! ❖ With regard to the functions of interruptions, we found that both male and female panellists use them with interruptive functions most frequently, neutral interruptions appear exclusively in male speech, and women are more prone to use supportive interruptions than men in mixed-sex contexts.!