Discourse Roles, Gender and Language Textbook ...

Gender and Education

ISSN: 0954-0253 (Print) 1360-0516 (Online) Journal homepage: http://www.tandfonline.com/loi/cgee20

Discourse Roles, Gender and Language Textbook Dialogues: Who learns what from John and Sally? Martha A. Jones , Catherine Kitetu & Jane Sunderland To cite this article: Martha A. Jones , Catherine Kitetu & Jane Sunderland (1997) Discourse Roles, Gender and Language Textbook Dialogues: Who learns what from John and Sally?, Gender and Education, 9:4, 469-490, DOI: 10.1080/09540259721204 To link to this article: http://dx.doi.org/10.1080/09540259721204

Published online: 01 Jul 2010.

Submit your article to this journal

Article views: 497

View related articles

Citing articles: 10 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=cgee20 Download by: [Lancaster University Library]

Date: 09 April 2016, At: 07:55

G ender and E duca tion, V ol. 9 , N o. 4 , p p. 4 6 9 ± 4 9 0 , 1 9 9 7

Downloaded by [Lancaster University Library] at 07:55 09 April 2016

D iscourse R oles, G ender and L anguage T extb ook D ialogues: w ho learns w hat from J ohn and S ally ?

M ART HA A. JON ES, C entre for E nglish L anguage E ducation, U niversity of N ottingham , U K CATH ERINE KITETU, D epartm ent of L anguage and L inguistics, E gerton U niversity, K eny a JANE SUND ERLA ND , D epartm ent of L inguistics and M odern E nglish L anguage, L anca ster U niversity, U K

M uch w ork has b een done on gender stereotyping in textb ooks for foreign language teach ing. H ow ever, little has focus ed on the language of dialogues, w hich are a characteristic feature of language textb ooks. I n the light of this gap, this article describ es the autho r’s ow n study analy sing discourse roles in dialogues in three recent text b ooks for the teaching of E nglish as a foreign language. T he results sho w ed an enco uraging level of gender fairness. T he authors then exp lore w hether this fairness w as ach ieved through the creation of gender b alance in social and occ up ational roles, and suggest that this w as so in the case of these three b ooks. T hey also discus s the negative im p lications of gender- im b alance d dialogues for fem ale students 1 language practice opp ortunities. T his study has im plications for language teach ers, teach er educa tors, and w riters and pub lishers of language teach ing m aterials. ABSTRACT

W hen studying French in a `Languages for A ll’ course at Lancaster University, in a dialogue Alison played a female tourist. T wo m en played a male tourist and a guide. Alison said `Q u’ est-ce-que c’ est?’ , seven times, and `Quoi par exemple?’ and Òh, allons-y, allons-y!’, once each. The guide inform ed the tourists about various points of interest in Paris, and A lison’s male com panion asked a range of inform ed questions. Alison was not impressed. [1] A large number of studies have now been don e into gender bias in language textbook s, especially that in English as a Foreign Language (E FL) textbooks (e.g. R ees-Parnell, 1976; Hartm an & Judd, 1978; Hellinger, 1980; Porecca, 1984; O bura, 1985; Talansky, 1986; Florent & W alter, 1989). The results have been depressing: gender bias is rife in terms of both relative visibility and occupational and personal stereotyping of fem ale characters. None of the above studies was carried out this decade, and it would be encouraging to think the situation had im proved. M eta-analyses of the ® ndings of studies carried out C orrespondence: M artha A . Jones, C entre for English Lan gu age Education, U niversity of N ottingham , University Park, N ottingham N G 7 2R D, U K . 0954-02 53/97/040469- 22 $7.00

Ó

1997 C arfax Publishing L td

469


470

M . A . J ones et al.

in each of the 1970s, 1980s and (hypothetically) the 1990s just m ight provide grounds for hope. But even carrying out such further investigations m ay be to m iss the point. The studies referred to above all found m any m anifestations of sexism and, correctly, condemned them . W hat they did not do (or did on ly in a very token, super® cial or unsubstantiated fashion) w as attem pt to relate this gender bias to learning , or even learning opportunities . Y et it is essential to do this. Those teachers who embrace a philosophy of vive la diffeÂ rence , or who view gender issues as trivial or misguided, or at best peripheral, will only be convinced that they must actually resp ond to gender bias if they believe it has the potential to affect their students’ learning opportunitiesÐ and, thus, their actual learning. It is of course rarely possible to dem on strate a conne ction betw een learning and a gender-biased text, on ly to suggest a hypothetical one. In an (unpublished) study of gender bias by `Wom en in EFL Materials’ (a subgroup of `Women in TEFL’ ), the example from S tream line D ep artures (1978) of the young woman in the short skirt and the m ale student who crashes his car because he is lookin g at her was the target of several accusations of sexism in term s of the lasciviousness of the illustrations and (more arguably) because the young wom an is somehow positioned as being to blam e for the crash. Yet it is hard to identify exactly how such a text with its visuals can actually hinder girls’ and women’s language learning. It is easy to say that female learners will be alienated and hence demotivated by such a portra yal of women. Some indeed might be: we have rather found that, either despite or because of its sexism, the S tream line series is usually a great success w ith learners of both sexes, many of whom see the books as having the capacity to improve their oral ¯ uency. Com plaints about m any older textbooks tended to be not so much that women were sex objects but that they were overcontented hom emakers. W riting on English as a Second Language (ESL) textbooks for adult learners in the USA, Fairlee W in® eld Carroll notes: Adult w omen attending foreign language classes are ¼ there because the language is necessary to them for career advancement, university studies or to m ake hom es and ® nd employm ent in a second language environment ¼ W hen m any single and m arried women are part of the labor force, seeking to enter it, or acquiring foreign language skills to im prove their potentials, it is unjust to portray only housewives and future housewives in textbooks. (1978, pp. 55, 59) Indeed, it isÐ but again such portrayals may adversely affect these women’s learning, or they may not. The writers of O n B alance (W om en in EFL materials, 1991), the guidelines for the representation of women and men for British publishers of EFL book s, observe: The images and language w hich are used in teaching, and the extent to which learners can identify with them, have an important effect on how well people learn. If women are under-represented in teaching m aterials, or represented in dem eaning w ays, the women who are taught with these materials m ay learn less well. Again, the key word is m ay . W in® eld Carroll and the O n B ala nc e writers can only speculate about the cognitive effects of sexist texts for two (related) reasons (there may be oth ers). Current literary theory stresses that readers will respond in m any different ways to a given text. Accord ingly, a given reader’s interpretation of a text cannot be predicted. A language learner m ay thus unconsciously accomm odate what many would con sider gender-stereo


D iscourse R oles

471

typical images, but may alternatively contest the sexism in a text and resist any positioning of herself as, say, a contented hom em aker or som eone concerned only w ith her hairstyle and boyfriends. Having done so she may take what she can from the book in terms of the learning opportunities it offers. Alternatively, what is a repressive text to one fem ale student may in fact be a progressive one to another; for example on e w ho lives in a country where the subordination of wom en is institutionalised. All this, of course, highlights the major limitation of content analysis of a language textbook (or indeed of any other text): as Sara M ills (1995; pp. 14, 15) notes, conten t analysis is essentially static and does not allow for different interpretations; it may be valuable, but needs to be done alongside a more dynamic and essentially w ider discourse analysis. A second reason why cognitive effects can on ly be suggested is that w hat happens to a text in class is unpredictable when based only on w hat it looks like on a page of an unopened book lying on the teacher’ s desk or in the school storecupboard. W hen she uses the text the teacher m ay give it a m eaning com pletely different from that intended by the writer. A n apparently sexist text can be addressed critically by the teacher and/or learners; by the same token, a progressive, egalitarian one can be interpreted by a conservative teacher in non-p rogressive w ays. For these tw o reasons alone it is thus im possible to do more than sp eculate about the effect of gender bias in a textbook on students’ language learnin g. The relationship betw een learning opportunities , or at least practice opportunities, and gender bias in the textbook m ay, however, m erit stronger claims, as we attempt to show below in relation to textbook dialogues. T extb oo k D ialogu es

One `genre’ of a language textbook for which it is possible to hypothesise actual uses in the classroom is the dialogue , i.e. a w ritten or taped conv ersation betw een two or more people. Am ong other w ays, a dialogue can be used: (a) as an òral model’ , to be demonstrated by two or more students and listened to by other class m embers; (b) as a basis for oral, pair or group work for the w hole class, to be `parroted ’ and/or adapted or extended through such activities as role play, simulation or dram atisation; (c) as a m odel in the textbook which can be extended by the students in w riting; (d) as a m odel to be read silently in the textbook, or listened to on tape; and (e) for the oral com pletion of a gapped conversation (e.g. in a language laboratory, where dialogues can be used by individual students). These ways can be used alon e or in com bination. The literature on textbook dialogues suggests they play a threefold role in the teaching/learning process. First, dialogues help in developing know ledge of the language being learnt on the level of vocabulary, structure and language use (see Hedge, 1985); D obson (1975) points out also that in dialogues, pupils become aware of aspects of pronu nciation, and particular features of spontaneous speech such as rejoinders, ® llers and interjections. Secondly, dialogues provide a social conte xt in which to practise new language. Students thus get a framework in which to practise discourse (D obson, 1975). Thirdly (and this applies to uses a, b and e), dialogues are a means of actually practising conversational speech, particularly colloq uial expressions, speed, inton ation, and use of pragm atics. By being given practice in speaking dialogues, pupils thus develop a range of conversational skills (Rivers & Temperly, 1978; R ivers, 1981; Grant, 1987).

472

M . A . J ones et al. T A BL E I. C haracters and `® rstness’ by gen der in dialogu es of F unctions of E nglish

U nit

M ale characters

Female characters

1 2

John Guy Roy Guy Guy John K en John Guy John Guy M ichael M ichael K en John Guy K en Guy K en John John K en Guy

Sally

John N /A

Sally

Guy N /A

3 4 5


W ho starts in m ixed-sex dialogues?

6 7 8 9 10 11 12 13 14 15

N /A

N /A Sally Su e Su e

M ichael M ichael K en N /A

Su e Su e

K en Guy N /A N /A

Su e

Guy

G end er Im b alance in T ex tbo ok D ia logues

It would thus appear that dialogues are of considerable potential value in providing different types of language learning opportunities. But does gender imbalance exist in textbook dialogues as it appears to do in language textbooks in general, and if so w hat form does it take? Very few investigations of gender bias in EFL materials have in fact looked at the ways in which the male and fem ale characters in their dialogues use language in dialogues differently, either qualitatively or quantitatively. An early exception is the now disbanded ÈTHEL’ (a group of fem inist EFL teachers working in Italy, and the name of their newsletter), who in 1980 analysed several units of N etw ork (1980), and noted that in addition to gender variation with age, body language, actions, jobs and clothes, men in dialogues asked about other people’s likes and habits and described their own, w hereas wom en only asked about other people’ s, and only m en gave orders (to women). More recently Sophia Poulou (1997) exam ined two different textbooks used for teaching Greek as a foreign language to adults. In both books, in mixed-sex dialogues in which both speakers w ere `non-experts’, she found a tendency for women to ask for inform ation, m en to provide it. And in on e of the books the majority of expressive language was used by the women (in the other, little was used by either). Quantitatively, ETH EL claimed that all dialogues in all units of the textbook F unctions of E nglish (1977) were initiated by a male. In verifying this, Jane Sunderland found also that seven of the 15 dialogues were betw een either two or three men, and the rem aining

D iscourse R oles

473

m en, and the remaining eight between a m an and a wom an. T here were no dialogues betw een women only (see Table I). (see also Gupta & Yin, 1990).


P eda go gica l an d C ognitive Im p lica tions o f G end er Im b alance in T extb oo k D ia logues

W hen an analysis of textbook dialogues shows that there are imbalances between the roles of females and those of m ales, the question that one then needs to ask is w hether these im balances matter. W hat pedagogic implications are there for teachers, and for female and male learners? W e have above identi® ed ® ve possible roles of textbook dialogues in language teaching and the learnin g purposes served. W e will now brie¯ y suggest how gender imbalances in dialogues may defeat these purposes and thus what in¯ uence they m ay have on classroom goals generally. To begin with, in mixed-sex classes of equal numbers of males and fem ales, when a disproportiona te number of roles and thus a disproportionate am ount of speech is allocated to the sexes so that one sex has fewer w ords than the other in dialogues, the `silenced sex’ will have poorer practice opportun ities in playing their roles as dialogue participants. Secondly, if one sex initiates con versation m ore often than the other, the ìnitiating sex’ will end up having more active practice in the skill of initiating conversation. The relationship between language output and language acquisition is unclear: M erill Swain (1985) suggests different roles for output; D ick A llwright & K athleen Bailey are less conv inced, arguing that some learners w ill learn best by listening (1991, p. 150). But while the jury is still out, it would seem safest to ensure that output op portunities are available for everyone, perhaps especially for very speci® c social/linguistic skills such as ìnitiating conversation’. Further practice opportunities can be speculated on hereÐ and could indeed be explored. For exam ple, gender im balance in a textbook dialogue may affect learner behaviour in other classroom activities like simulation and dram atisation. W hen students are asked to dram atise a text or make a simulation of it, because of the authority afforded textbooks, they are likely to im itate the original text. As Bessie D endrinos (1992, p. 26) observes: The textbook carries a unique authority which is created and m aintained through its texts ¼ these are understood as the legitimate version of a society’ s sound know ledgeÐ the knowledge that every pupil has a primary responsibility to m aster. Frequent and regular gender imbalances in dialogue roles may thus predispose students to m ake assumptions about the gendered nature of verbal behaviour of native speakers of the second or foreign language, and they m ay accordingly continue to im itate it in classÐ probably to the disadvantage of the female students. M ale and female students alike m ay be further misled to the point w here they form ulate particular, gender-im balanced assumptions about when and how much they should speak outside the classroom. A negative cognitive in¯ uence, as indicated earlier, m ay be loss of interest on the part of those who are m arginalised by the discourse. As suggested by the w riters of O n B ala nc e (1991), fem ale students m ay be consciously or unconsciously in¯ uenced to the point where they becom e demotivated as they continue to roleplay roles which are restricted linguisticallyÐ as w ell as, often, occupationally. This marginalisation m ay in turn shape

474


these fem ale students’ expectations of disempowered roles. Such possible in¯ uences deserve and require further investigation.


T he S tudy

Given the existing gender stereotyping in language textbook s, about w hich there is a large empirical body of knowledge, and the gender bias in textbook dialogues, about which there is a very small empirical body of know ledge (ETHE L, 1980; Gupta & Y in, 1990; Poulou, 1997), we had to decide whether to look at dialogues in older, more studied textbooks, and thus com plement the existing studies of stereotyping, or dialogues in m ore recent textbooks, of which there have been fewer studies as regards gender bias. W e decided on the latter simply because the more recent books are those not only in current but also in near-future use. The purpose of our study was thus to establish whether and to what extent gender bias in dialogues obtained in three recent, popular (and thus hopefully reasonably representative) English language textbooks. W e also decided to do a quantitative rather than qualitative study, seeing this as a starting point which a follow -up qualitative study could use as baseline data. However, w e recognise that qualitative and quantitative studies m ay reveal different manifestations of gender bias. W e decided to look at dialogues intended to be spoken . If `male ® rstness’ and m ale verbal dom inance are present, dialogues intended only for listening m ay inculcate these ideas, but dialogues which are to be spoken, in addition to this, are actively, p rom oting this verbal practice in female and male students throu gh the students’ own discourse practices. T extb oo ks A nalysed

The three textbooks we chose to analyse w ere H eadw ay I nterm ediate (Soars & Soars, 1987), H otline I nterm ediate (H utchinson, 1993) and L ook A head 2 (Hopkins & Potter, 1994). O ur reasons for this selection were that all were recent, all were well known, all contained dialogues, and all were for interm ediate or pre-interm ediate students who could be expected to deal with reasonably `rich’ dialogues. All three were borrow ed from the R esources Centre of Lancaster University’s Institute for English Language Education (IELE); they could thus be expected to be found in and used by students and teachers of English at similar units. H eadw ay I nterm ediate

The H eadw ay series is authored by John and Liz Soars and published by O xford University Press. It is an English revision and extension course intended for adults at the interm ediate stage. H eadw ay I nterm ediate is part one of the course. It does not have a particular storyline, rather it has 14 different teaching units, each covering all four skills of reading, writing, speaking and listening, in that order. There is at least one oral ¯ uency activity per unit. H alf of these are role plays in which the students have to think about the roles they should play and then write out their parts. The other half of the speaking activities are discussions. T he authors write: `The advantage of role plays is that in theory, even the quiet students participate, whereas a discussion can be dominated by on e or two vociferous students’ (p. vii). The dialogues themselves are self-contained and relate to the topic of the unit in w hich they occur.

D iscourse R oles

475


H otline I nterm ediate H otline , by Tom H utchinson and also published by Oxford University Press, is a series of three books from beginner to interm ediate level. H otline I nterm ediate includes dialogues to be used for both listening and speaking practice. The storyline, focusing on the activities of a group of teenage friends, is in fact carried by the dialogues. Very often a dialogue form s part of the Language W ork section in each unit, focusing on the language structures introduced. Students are encouraged to look at the picture story, work out the grammar rules by them selves, discuss the rules, listen to the dialogue, and then practise the language orally. The focus is on ùseful expressions’, including colloquial expressions such as `lo ads of ’ and `they ’ll rip y ou off ’ . The students are encouraged to think about expressions with similar meanings to these, then read out the dialogues in pairs or groups. In the teacher’ s book there is, however, no speci® c discussion of dialogues.

L ook A head 2

W ritten by A ndy H opkins & Jocelyn Potter, and described as a partnership between BBC English, the British Council, the University of Cambridge Local Examinations Syndicate and Lon gm an ELT, with the cooperation of the Council of Europe, L ook A head is a series of four books from beginner to First Certi® cate level. L ook A head 2 is for post-elementary/ pre-intermediate students, and describes itself as taking students `beyond the Council of Europe W aystage level’ . The dialogues are intended as a way to present new language in conte xt, and to provide both listening practice and speaking practice. T he storyline, about developm ents at an advertising agency, is carried via the dialogues (though not all the dialogues are concerned with this storyline). The teacher’ s book’ s suggestions for work ing with dialogues (p. 9) include: `(4) Ask students to read the dialogue in pairs and then to change roles’ , and `(5) W ith their book s closed, pairs of students act out the dialogue, using their own w ords to supplem ent the lines that they remember.’ T hough this is not made explicit, the ® rst of these permits both m embers of m ixed-sex pairs to have a turn at the part which provides the better language learnin g opportunity, if either does; the second allows for student subversion of any gender-biased roles (it also, of course, allows for a consolidation of them). In none of the three textbooks nor the accom panying teacher’s books is there any explicit mention of gender-fairness either in use of the books or in classroom interaction in general. (This theoretical gender-blindness is characteristic of foreign language learning m aterials in general; it may not, how ever, be an indicator of gender bias in the m aterials themselves.)

S election o f D ata and M etho do logy

The data for the study consisted of all the dialogues to be used for speaking practice in all three books: 10 in H eadw ay , 26 in H otline , and 15 in L ook A head . The gapped dialogues in L ook A head were excluded, as were those in the `Progress Check’ sections, since these were clearly intended for individual writing practice. Our m ethodology was quantitative: we counted the number of m ales and females who played roles in the dialogues [2], the number of tim es females and m ales


476


initiated dialogues, the number of turns taken by male and by female characters, and the number of words spoken by fem ales and by m ales [3]. In counting the females and females who played roles in the dialogues, we distinguished between (a) characters as individuals, and (b) the number of times each character appeared, using the distinction in linguistics between `type’ and `token’ . In this distinction, `type’ refers to a class of items and `token’ to a m ember of that class: `hello ’ thus being a token of the type `greeting’. In our simpli® ed version of the distinction, `type’ referred to a particular character, `token’ to an occurrence of this character. For example, `Richard’ as a `type’ counts as one (in this case, one male character), regardless of his num ber of appearances, but R ichard as a `token’ is the total num b er of appearanc es of Richard in the text. (This use of the distinction was adopted by D avid Crystal [1986] in counting the words produced by young children.) The use of this distinction m eant that it was possible to obtain m ore than one measure for each issue. The actual research questions are listed as follow s: R esea rch Q uestions

For each book: 1a. H ow m any different male and female characters are there? (`types’) 1b. H ow many appearances of the different male and female characters are there? (`tokens’) 1c. W hat is the average (mean) number of tim es female and male characters appear? (`mean type appearance’ ) 2. H ow many mixed-sex dialogues are initiated by male and how many by fem ale speakers? 3a. H ow m any m ale turns and fem ale turns are there altogether? 3b. W hat is the average (m ean) number of turns per man-`type’ and per w oman-`type’ ? 3c. W hat is the average (m ean) number of turns per m an-`token’ and per woman`token’? 4a. H ow m any dialogue words are spoken by m ale and by female characters? 4b. W hat is the average (mean) num ber of dialogue words per man-`type’ and per wom an-`type’? 4c. W hat is the average (m ean) number of words per man-`token’ and per woman`token’? Our m ethodology did not thus have a qualitative aspect. H ad tim e permitted, it would have been interesting and perhaps revealing to look at the language functions used by m ale and fem ale characters to see if differences pertained (see Poulou, 1997). Find ings

The ® ndings for the different research questions for each book are indicated in the Tables II± XI. (Findings for each unit of each book can be found in the appendix.) S um m ary of Find ings H otline

As regards visibility, there was an equal number of female and m ale characters (® ve) in the textbook as a whole (`type’), and only a slight difference betw een fem ale and m ale

D iscourse R oles

477

T A BLE II. H ow m any differen t fem ale and m ale charac ters (`types’) are there?

B ook

N o. of different fem ale characters

N o. of different male characters

5 13 11

5 9 6

H otline H eadw ay L ook A head


T A BLE III. How m any ap pearan ces (`tok ens’ ) of the differen t fem ale and m ale charac ters are there?

B ook H otline H eadw ay L ook A head

N o. of different fem ale appear ances

N o. of m ale ap pearan ces

35 13 20

39 9 10

`tokens’ (35 female, 39 male). The difference in `mean type occurrence’ for each sex was thus also small: 7 female, 7.8 m ale. However, there w ere three gender imbalances which, even though not highly m arked, are worth mentioning. (1) M ore dialogues w ere initiated by males than by fem ales (15 : 11). (2) Fem ale characters spoke slightly fewer words: 2242 vs 2481 (448.4 per woman-`type’ , 496.2 per man-`type’; 64.05 per woman-`token’ and 63.61 per man-`token’ ). (3) M ale characters had slightly more turns in dialogues (142:13 3), the average num ber of turns per man-`type’ and per woman-`type’ being 28.4:26.6. Female `tokens’, however, scored slightly better as regards their average number of turns (3.8:3.6). Overall, then, H otline presents a relatively progressive picture of female characters (in this case mainly girls), both in terms of visibility and of discourse. M ale characters did consistently better on all variables except tw o, but only very slightly. For the variables the females did better onÐ `m ean number of turns per wom an-token’, and `mean number of dialogue words per woman-token’ Ð the differences were also extremely slight.

T A B LE IV . W hat is the m ean num ber of app earances of fem ale and m ale characters (`m ean type app earan ce’ )?

B ook H otline H eadw ay L oo k A head

M ean fem ale type app earanc es

M ean m ale type ap pearan ce

7 1.00 1.82

7.8 1.00 1.67

478

M . A . J ones et al. T A BLE V. H ow man y m ixed-sex dialogue s are initiated by fem ales an d how man y by males? N o. of dialogues initiated by fem ales

Book H otline H eadw ay L ook A he ad

N o. of dialogu es initiated by m ales

11 5 6

15 4 3


T A B LE VI. H ow m any fem ale turns an d how m any m ale turns are there? B ook

N o. of fem ale turns 133 43 67

H otline H eadw ay L oo k A head

N o. of male turns 142 27 44

T A BLE V II. W hat is the mean num ber of turn s per w om an -`type’ and m an-`type’?

B ook H otline H eadw ay L o ok A head

M ean no. of turn s per w om an-`type’

M ean no. of turns per m an`type’

26.6 3.30 6.09

28.4 3.00 7.33

H eadw ay

As regards visibility, there were 13 fem ale and nine male `types’, and the same num ber of male and fem ale `tokens’. The `m ean type appearance’ for both was thus the same, i.e. 1. Unsurprisingly, considering the greater num ber of females, there were fewer m ale turns (27) as com pared to those of fem ales (43) and females did better on àverage number of turns’ for both `type’ and `token’ (for both the ratio w as 3.3:3.0). Fem ales had the greater total number of words (302), as com pared to that of the males who had 248. The average num ber of words spoken by wom an `types’ and `tokens’ was, how ever, 29.22, w hereas that for m en-`types’ and `tokens’ was 27.55. M ales initiated four dialogues, females ® ve. Our analysis of the dialogues for speaking practice as found in H eadw ay showed differences between males and fem ales in visibility, there being more females, who also initiated conversation marginally m ore often, though as individuals males were more verbose.

L ook A head L ook A head w as in som e ways com parable. There w ere 11 different female characters, and only six males. A `woman-type’ appeared on average 1.82 times; a `m an-type’ 1.67 times.

D iscourse R oles

479

T A B LE V III. W hat is the m ean number of turns per w om an-`token’ and per m an-`token’ ?

B ook H otline H eadw ay L o ok A head

M ean no. of turn s per wom an-`token’

M ean no. of turns per m an`token’

3.8 3.00 3.35

3.6 3.00 4.40


T A BL E IX. H ow m any dialogu e-word s are spoken by wom en an d how m any by men? Book H o tline H eadw ay L ook A head

Dialogu e words sp oken by wom en

Dialogu e words sp ok en by men

2242 302 632

2481 248 501

T AB LE X . W hat is the m ean num ber of dialogue w ords per wom an-`type’ an d per m an-`type’ ?

Book H otline H eadw ay L ook A head

M ean no. of dialogu e words sp ok en per w om an -`type’

M ean no. of dialogu e words spoken per m an-`type’

448.4 23.23 57.45

496.2 27.55 83.50

Perhaps not surprisingly, then, women initiated six of the nine mixed-sex dialogues, took 67 turns com pared with the m ales’ 44, and spoke 632 dialogue words com pared w ith the m en’s 501. Though the women in L ook A head are thus very visible as `types’ and in their com bined verbal production, as individuals they do not do so well linguistically. As regards the `density’ of dialogue speech of the individual males, the men are rather more verbally visible: each wom an-`type’ has on average 6.09 turns and each m an-`type’ 7.33 turns. Each woman-token has 3.35 turns on average, and each m an-`token’ 4.40. The difference in mean num ber of dialogue words is m ore m arked: the m ean number of dialogue w ords spoken by a woman-`type’ is 57.45 and by a m an-`type’ 83.50; the m ean number of dialogue words spoken by each woman-`token’ is 31.60, and by each m an-`token’ 50.10. Though there are fewer m ale characters in these dialogues, then, they are thus clearly better developed linguistically as individuals than are the females. T A BLE X I. W hat is the m ean number of dialogue w ord s per w om an-`token’ an d per m an-`token’?

Book H otline H eadw ay L ook A head

M ean no. of dialogu e words spoken per w om an-`token’

M ean no. of dialogu e words sp oken per m an -`token’

64.05 23.23 31.60

63.61 27.55 50.10

480

M . A . J ones et al. T AB LE X II. O ccupational an d social roles, by gen der, with freq uencies, of eac h of the three textbooks

B ook H otline


H eadw ay

L o ok A head

N o. of fem ale role app earances

N o. of m ale role ap pearan ces

sister 3 3 frien d 3 29 salesw om an 3 1 `rivals’ (for Vince!) 3 girlfrien d 3 2

brother 3 2 frien d 3 32 custom er 3 1 boyfriend 3 1 m anager 3 1 son 3 1 father 3 1 partner 3 2 w ork er 3 1 interviewee 3 1 of® cer 3 1 acquaintance 3 2 policem an 3 1 colleag ue 3 1

2

partner 3 2 [dom estic] work er 3 1 m other 3 1 daugh ter 3 1 interviewer 3 1 girlfrien d 3 1 co-traveller 3 2 acquaintance 3 2 citizen 3 1 colleag ue 3 1 frien d 3 7 secretary 3 1 accounts m anager 3 1 colleag ue 3 2 driver 3 2 ward en 3 2 sister 3 1 girlfrien d 3 2

frien d 3 3 colleag ue 3 2 visitor on business 3 porter 3 1 brother 3 1 boyfriend 3 2

1

A nalysis of Find ings

The gender differences found are too small either way to be signi® cant. This is encouragingÐ but it would be interesting to establish w hy the differences w ere sm all. The reason for the relative gender balance in these discourse roles may lie in the distribution of occup ational and social roles . A gender im balance in discourse could be largely because the roles allocated to male and female characters are those which im ply relative verbosity or relative silence, respectively, or, as in Poulou’ s study (1997), èxpertness’ or `non-expertness’. As W illiam O ’Barr & Bow man A tkins (1980) found, when looking at actual language use in the courtroom , power as well as gender was a factor in the am ount of speech used. In the same w ay, if all textbook fem ale characters are secretaries and all m ale characters line managers, it w ould perhaps not be surprising if the m en spoke m oreÐ but that would be because they were line m anagers, not because they w ere men. It is therefore worth looking at the occupational and social roles in the dialogues in these three books to see if gender balance has been achieved here. The actual occupational and social roles by gender for each book, w ith frequencies (`tokens’), are given in Table X II. The num ber of different occupational and social roles (`types’) in the three books are given in Table X III. W ith the possible exception of the one male manager in H otline , and the two family roles in H eadw ay , the occupational and social roles seem fairly distributed between fem ale

D iscourse R oles

481

T A BLE X III. T he num ber of different occupational an d social roles, by gen der, for each of the three textbooks

Book


H otline H eadw ay L ook A head

N o. of different fem ale roles

N o. of differen t male roles

5 10 8

7 7 6

and male charactersÐ in contra st with the ® ndings of earlier `content analysis’ studies of language textbooksÐ and neither females nor m ales seem to have a much greater range of roles. This m ay be relevant. It may be that one way to ensure gender fairness in discourse in dialogue roles is to ensure a fair gendered distribution of occup ational and social roles .

D iscussion of Find ings

Though we are aware that a qualitative analysis m ay have revealed m ore subtle and pernicious gender biases (cf. Carroll & Kowitz, 1994; Glass, forthco ming), the ® ndings of this q uantitative investigation were largely positive. W e are of course pleased about thisÐ gender imbalance and the tendency to `think male’ is clearly som ething that can be avoided, given a little thought, som e counting, an awareness of the `think male’ tendency and a shared intention of the textbook writer and the publisher not to fall into the same trap, and probably a concern for fairly distributed social and occupational roles. The results of our analysis suggest that these authors and publishers indeed gave (considerable?) consideration to social roles, language use and gender. Our ® ndings would seem to re¯ ect the progress made and raised con sciousness of publishers and writers in the world of language teaching and perhaps in society in relation to gender representation. This awareness is now actively encouraged by W omen in E FL M aterials’ document O n B ala nc e: guidelines for m aterials w riters in E F L (1991) (of which all British ELT publishers have a copy), which does in fact make reference to counting characters in dialogues (p. 3), and to both men and w om en starting dialogues (p. 4). The publishers’ objective is of course to sell books, but if counting fem ale and m ale heads will help achieve this, then sales will not be put at risk. Customers are unlikely to ob ject to the idea of gender balance, after all (though those accustomed to seeing a preponderance of males as `norm al’ may perceive such balance as a preponderance of females!). It would seem important, m ethodologically, to use different measures when quantitatively assessing the extent of gender balance and imbalance. As we have shown, visibility in one sense does not m ean visibility in all: there may be m any female characters, but who are not presented engaging in any ìn-depth discourse,’ for exam ple, because they each appear only once. W omen characters m ay exemplify large numbers of words but m ay have few turns or m ay rarely speak ® rst. The sheer num bers of different women in L ook A head and the total number of words they speak, for example, is thus not a good indicator for gendered verbal visibility of individuals . The num ber of words spoken by wom en m ay be a rather crude measure of learning opportunities if this is not paralleled by the presence of individual female characters whose `verbal depth’ equals that of their m ale counterparts. In order to achieve equal language practice opportun ities for both


482


m ale and fem ale students in dialogues, all measures need to be considered. Sexism, after all, can be very pernicious, and can pop up where it is least expected and where it is hard to identify. The issue of how to achieve gender balance in discourse roles remains a problematic one. One way m ay indeed be throu gh ensuring gender balance not only in the num ber of female and m ale characters but also in their social and occupational rolesÐ as was apparently don e in the three textbooks studied here. The question then is one of credibility w ith textbook users, which will obviously vary with culture. The theoretical alternativeÐ that of ensuring gender balance in discourse roles but not social and occupational rolesÐ is likely, however, to lack validity, since people in the more subservient occupational roles, at least, are unlikely to speak as m uch as their `superiors’ . A logical w ay out would be to include characters of the same status throu ghout the bookÐ say, a group of factory w orkers, or ® fth form studentsÐ but this w ould inevitably m ean dif® cult restrictions on any storyline. Gender balance in discourse roles, and hence in language practice opportunities, m ay thus be achievable, one way or anoth er, but at a possible cost. T he Id eal Situa tion

Up to now the desirability of balancing discourse roles equally between fem ale and m ale characters has been seen as a `given’. However, this view perhaps requires further support. Let us take a step back, and ask `naively’ , `What should the distribution of gender roles in language textbook dialogues be like?’ , and let us com pare it w ith the situation with occupational roles. The claim has been made that occupational roles in a given textbook are often worse for women than they actually are in the society in which the textbook is used or on which it is based (e.g. Ittzes, 1978). T wo possible publishers’ and authors’ professional responses to this are: (a) attempting to make textbook occupational roles for women and m en m irror those of the society; and (b) having `positive role m odels’ in textbooks, to the extent that women are employed in a wider range and at higher levels of professions than they actually are, perhaps to the point of balancing all roles so that there are, say, the same number of m ale and fem ale m anagers and the same num ber of male and female lorry drivers. There are valid argum ents in favour of (and problem s with) each, though it is not the purpose of this article to go into these (but see O’ Neill, 1994). However, are either of these approaches valid w hen it com es to discourse roles ? In a relevant article entitled `Natural con versations as a model for textbook dialogue’, Carol Myers Scotton & Janice Bernsten (1988) argue for a closer relationship between natural conversations and textbook dialogues. H owever, are although in their em pirical investigations they uncovered gender differences (some signi® cant) in different aspects of direction giving, alon g w ith other differences, they stop short of explicitly saying these should be re¯ ected in dialogues, contenting themselves rather with the bland àll profession als involved in second language instruction should pay more attention to quantitative data from natural conversations to ® nd out what they are really like’ (1988, p. 383). In an earlier article, `The problem of applying sociolinguistic ® ndings to TESOL: the case of m ale/female language’, Elliot Judd (1983), how ever, apparently accepts the

D iscourse R oles

483


`discoursal ’ equivalent of response (a) (see earlier) without question. Starting with the `given’ that Ìt is now a widely accepted principle that language materials should be based on models that represent valid linguistic data’, he then goes on to explore ways of gathering data on gender differences in language use, and of incorporating this into language teaching m aterials. H e never questions the desirability of thisÐ though he is aware of research ® ndings in the area of gender differences in language use, and of the debate over w hether gender differences are in fact re¯ ections of differences in power: [A] crucial factor w hich must be incorporated into language materials when they are designed to include male/fem ale variations is that of the status relationships betw een the participants. M any of the differences between fem ale and male language use have been attributed to an unequal power relationship betw een women and m en in our society (Spender, 1980; K ram arae, 1981; and M cCon nell-G inet, Borker and Furman, 1980). In fact, som e have argued that m ost gender differences in English occur because of status inequalities rather than sex alone. Of course, since women are generally in lower-status positions, linguistic differences caused by lower status are m ore apparent in m ost wom en’s speech than in men’s (O’Barr and Atkins, 1980). T hus this factor m ust nec essarily b e re¯ ected if valid language m odels for m en and w om en are desired . (Judd, 1983, our italics) Tw o key words here are ìf’ and `valid’. W hile it m ay or may not be desirable for a textbook to m irror occupational roles as regards gender, to mirror the quantitative aspects of discourse roles has a clear identi® able potential to disadvantage female learners in term s of unequal provision of language practice opportunities. M uch research into gender differences in language use has found that, in mixed-sex conversation in m any contexts, men talk m ore than women, who expend a great deal of energy asking questions, taking up the men’ s topics, and providing conversational support in term s of `backchannelling’: `m m ’, `hm ’ , `really ’, and so on (e.g. Fishm an, 1983). This has not only been found to be true of speakers of English as a ® rst language, but also of adult learners of English as a second language doing group work in the classroom (Gass & Varonis, 1986; Munro, 1987; H olm es, 1994). To mirror such ® ndings would not only provide female students with fewer speaking practice opportunities than males, but would also give them practice in `supportiv e’ rather than àssertive’ language use. Judd’s suggestion has other weaknesses. W ork on gender differences in language use has also uncovered q ualitative pragmatic differences in, for example, asking for help and apologising. M ost apologies seem to be from women to other women, fewest from men to men (Holm es, 1988). Should ® ndings such as these be replicated in language textbooks? And frequency is never the whole story: in the case of com pliments to women by men, there is often a ® ne line between a `sincere’ or disinterested com pliment, interested ¯ attery, and verbal sexual harassment. Recognition of this, for both fem ale and m ale students, together with recognition that what may be sincerely intended as a com plim ent m ay not be so perceived, w ould seem far m ore important than mechanically re¯ ecting questionably generalisable ® ndings of frequencies in language textbook dialogues. Rather than uncritically mirror gender differences in language use, it is surely fairer if textbook dialogues provide com parable models and thus com parable speaking opportunities, and in addition inform students of em pirically established gender differences. There are problem s with this , of course. One is the questionable generalisability of ® ndings in the area of gender differences in language use and the im portant fact that


484


these differences are not ® xed . A nother (w hich m ay in fact obtain whether gender roles in textbook dialogues are equivalent to those outside, or not) is that some non-na tive speakers of English will have sociopragmatic problem s adopting those norm s of English which are differently gendered from their ® rst language norm s (Thom as, 1983). Y et, despite these problem s, to balance discourse roles by gender (and, if necessary, to do this by balancing occupational and social roles) w ould seem the safest, fairest starting point. Textbook discourse roles m ust thus be differentiated in essence from occupational roles as regards gender. Occupational roles cannot be shown to have a direct in¯ uence on classroom proceedings. Discourse roles in dialogues are very likely to have a direct in¯ uence, in that in a m ixed-sex class w ith equal numbers of males and females, as we have suggested, male roles w ill often go to m ale students, female roles to female students. W e are not suggesting here that students should be forced to speak in class, or indeed that a speaking opportunity will necessarily result in language learning. But discoursal dis empowerm ent throu gh a gender imbalance in discourse roles in dialogues surely has no role to play in language teaching. Our conclusion echoes that reached by Geoffrey W alford in a study of illustrations in physics textbooks. W alford notes both relative invisibility of female characters and stereotypical roles for fem ale characters, com parable to that frequently found for language textbooks. He writes: It could, of course, be argued that physics textbook s are just show ing the world as it is, and that if there are fewer girls taking physics and fewer wom en in physics-related jobs, then this is exactly the way that physics textbook s should be illustrated. Such an argument would com pletely miss the purpose of this article ¼ if we seriously w ish to encourage m ore girls to enter physics w e need to change the clear m asculine im age. It is recognised that a major part of this image is re¯ ected and projected in the day-to-day activities of physics teachers in the classroom and that this is going to be dif® cult to change. But it is possible that in the area of school textbooks, steps could be taken to ensure that the image presented of physics is one which is encouraging to young people of both sexes and not just one at the expense of the other. (1980, pp. 225± 226) W alford adds that since society is moving towards greater gender equality, physics teachers should play a role in this. So, we say, should language textbook writers. C o nclusio n a nd R eco m m end a tions

D ialogues are m odels of oral language which throu gh the learning/teaching process should enable all students to take part in conversations in the foreign language they are learning. If there are imbalances in dialogue participant roles to the point that discourse roles are not equally distributed between m ales and fem ales, then all learners may not get equally good learning opportun ities and experiences. A nd remembering that output as well as input m ay be important in foreign or second language acquisition (Swain, 1985), for som e if not all learners, this is a serious pedagogical implication which should be addressed if the aims of using dialogues in textbooks for language learnin g are to be achieved. W e have shown in this article that the authors and publishers of L ook A head , H eadw ay I nterm ediate and H otline I nterm ediate have addressed the issueÐ or, at least, have avoided the trap their predecessors have fallen intoÐ and w e hope that these and other writers and publishers will continue in this progressive vein.


D iscourse R oles

485

W e hope too that other teachers and researchers in the ® eld will carry out other, similar investigations, perhaps replicating our m ethodology with their ow n textbooks. In particular, we hope that investigations w ill be carried out into qualitative aspects of textbook dialogues, especially the speech acts exempli® ed by women and men, boys and girls. W e recognise that the very notion of `speech act’ is highly problematicÐ but `functional’ labels such as `requesting’ and ìnviting’ are often used in language textbook s, and position students as users of such discourse. W ho is inviting and who accepting (or declining), for example? Interviews w ith authors and publishers about their intentions as regards textbook dialogues would be illustrativeÐ and interviews with the student users of dialogues are likely to be fruitful too. A ® nal plea is for studies of how mixed-sex dialogues in textbooks are actually used in m ixed-sex language classroom s: do teachers tend to allocate female roles to fem ale students, and m ale roles to m ale students, or are teachers prepared to vary the distribution of the part labelled `John’ and that labelled `Sally’? And, given a very gendered dialogue and an instruction to read it aloud and/or extend it in oral pair work, what do m ixed-sexÐ or, indeed, single-sexÐ pairs of students actually do w ith it? It would also be interesting to see if gender imbalances in language textbook dialogues played a role in any male dom ination of a language classroom (dom ination in the sense of the teacher paying more attention to m ale students, or the m ale students them selves speaking more than the female students). Interestingly, though such male dominance has frequently been reported (see K elly [1988] for a m eta-analysis of studies), language classroom s seem to have been under investigated in this respect. W e hope too that teacher educators in the ® eld of language education will include discussion of gender and textbook dialogues in w ork on materials selection and evaluation, and will include gender-balanced dialogue writing in sessions on materials design. Finally, we would like to express the hope that teacher-researchers looking into gender bias in language textbooks will not treat the texts and visuals as static objects in which the language and meanings are `there’ , waiting to be revealed, but rather will consider both different possible interpretations on the part of different readers, and, on a more practical level, different uses of these texts by language teachers (see Sunderland, 1997 for an exam ination of actual teacher mediation of gendered FL texts). And, for reasons of pure expediency (since not everyone is concerned about gender bias unless this is accom panied by a convincing answer to their `So what?’ ), that they will relate gender bias to possible effects on students’ language practice and learning opportunities, and thus potentially to their actual language learning .

A ckno w led gem ents

W e w ould like to express our sincere thanks to M oham mad A lavi for helping us with layout, Keith Johnson for his constru ctive com m ents on the article, and m em bers of the Language and Gender in the Classroom (LAGIC) group, Departm ent of Linguistics and M odern English Language, Lancaster University, for discussing the content of this article on several occasions.

NO TES [1] [2]

A true story (person al com munication ). If the sex of a person w as unclear from the text or from visual clues, then this person an d her/ his words

486

[3]

M . A . J ones et al. were not included in the count. (It was not assum ed, for exam ple, that an unnam ed `m echanic’ w ou ld be maleÐ unless there was a picture of a man .) Hesitation s su ch as ùm ’ and èr’ w ere included in the word counts. C ontractions were taken as single words.

TEXT S ANALYSED H O P K IN S , A. & P O T TE R , J. (1994) L oo k A head S tudents’ B o ok 2 (Harlow, British C ouncil/BB C English /Longm an ELT /U niversity of C am bridge Local Exam inations Syndicate). H U T C H INS O N , T . (1993) H otline I nterm ediate (O xford; O xford U niversity Press). S O A R S , J. & S O A R S , L . (1987) H eadw ay I nterm ediate (O xford; O xford U niversity Press).


R EFERENCES A LL W R IG H T , D . & B A ILE Y , K . (1991) F ocus on the L anguage C lassroom (C am bridge, C am brid ge U niversity Press). C A R R O LL , D . & K O W ITZ , J. (1994) U sing concord ancing techniques to study gend er stereotyping in ELT textbooks, in: J. S U N D E R LA N D (Ed.) E xploring G ender: q uestions and im plicatio ns fo r E nglish language education (H em el H em pstead, Pren tice Hall). C A R R O LL , F .W . (1978) T he lim its of m y language are the limits of m y w orld, in: R. F R E U D E NS TE IN (Ed.) T he R ole of W o m en in F oreign L anguage T ex tb o oks: a collection of essay s, C ollec tions of d’Etudes L ingu istiques, no. 24 (G hent, Federation Internationale des Professeu rs de Lan gu es Vivantes). C R Y S TA L , D. (1986) L isten to y our C hild (Lond on , Penguin). D EN D R IN O S , B . (1992) T he E F L T ex tb oo k and I deo logy (A thens, N . C . G rivas Publications). D O BS O N , I. (1975) D ialogue s: why, w hen an d how to teach them , F orum X II , pp. 55± 63. E A S TW O O D , J., K A Y E , V., M A C K IN , R. & S T R E V EN S , P. (1980) N etw ork (O xford , O xford University Press). E T H EL (1980) Ethel in gen derlan d, E T H E L , 5. F IS H M A N , P. (1983) Interaction: the work w om en do, in: B A R R IE T H O R NE , C H O V IS K R AM A R A E & N A N C Y H E N LEY (Eds.) L anguage, G ender and S ociety (Row ley, M ass., N ew bury House). F LO R E N T , J. & W A LTE R , C. (1989) A better role for w om en in TEFL , E L T J ournal, 43, pp. 180± 184. G A SS , S. & V A R O N IS , E. (1986) Sex differen ces in nonnative sp eaker-n onnative speak er interac tions, in: R. D A Y (Ed.) T alking to L earn: conversation in S econd L anguage A cq uisition (N ew Y ork, N ewb ury H ou se). G LA SS , J. (forth coming) A comparison of the represen tation of fem ales in two editions of an EF L textbook, C R I L E W orking P aper (Lan caster, Linguistics Departm ent, Lanc aster U niversity, U K ). G R A N T , N . (1987) M aking the M ost of y our T ex tb ook (Lond on , Longm an). G U PT A , A. & Y IN , A. (1990) G ender rep resen tation in English language textbook s used in Singapore prim ary classroom : an interactional account, Lanu gu age and Education 4/ 1, pp. 29± 50. H A R T LEY , B. & V INE Y , P. (1978) S tream line D epartures (O xford University Press). H A R T M A N , P. & J U D D , E. (1978) Sexism an d TESO L m aterials, T E S O L Q uarterly , 12, pp. 383± 392. H E D G E , T. (1985) U sing R eaders in L anguage T eaching (Lond on , M ac millan). H E LLING ER , M . (1980) `For m en must work and wom en m ust w eep’ : sexism in English lan guage textbooks used in G erm an schools, W o m en’s S tudies I nternational Q uarterly , 3, pp. 267± 275. H O L M ES , J. (1988) Paying com plim ents: a sex-preferential positive politeness strategy, J ournal of P ragm atics , 12, pp. 445± 465. H O L M ES , J. (1994) Im proving the lot of fem ale lan gu age learn ers, in: J. S U N D E R LA N D (Ed.) E xploring G ender: q uestions and im plicatio ns fo r E nglish language E ducation (H em el Hem sptead; Prentice Hall). H O P K IN S , A ., P O TT ER , J. & D U V IV IE R , M . (1994) L ook A he ad T eacher’s B ook 2 (H arlow, B ritish C ouncil/BB C English /Lon gm an ELT /University of C am bridge Local Exam inations Syndicate). I TT Z ES , K . (1978) H ungarian w om en in reality , in: R . F R E U D E N ST EIN (Ed.) T he R ole of W o m en in F oreign L anguage T ex tb o oks: a collection of essay s, C ollections d’Etudes Lingu istiques, no. 24. (G hent, Federation Internationale des Professeu rs de Lan gu es Vivantes). J O NE S , L. (1977) F unctions o f E nglish (C am bridge, C am bridge U niversity Press). J U D D , E. (1983) T he problem of ap plying sociolinguistic ® ndings to TESO L: the case of m ale/ fem ale lan gu age, in: N . W O LFSO N & E. JU D D (Eds) S ociolinguistics and S econd L anguage A cq uisition (R ow ley , M A N ewb ury H ou se). K E LLY , A. (1988) Gender differences in teacherÐ pupil interaction s: a m eta-an aly tic review , R esearch in E ducatio n, 39, pp. 1± 23. M ILLS , S. (1995) F em inis t S ty listics (L on don, H arvester W heatsheaf). M U N R O , F . (1987) Female and m ale participation in sm all-grou p interaction in the ESO L classroo m, unpublished term s project, G rad uate Diplom a in TESO L (Sy dney C ollege of Ad vanced Education).


D iscourse R oles

487

M Y E R S , S . & B E R N ST EN , J. (1988) N atural conversation s as a m odel for textbook dialogue, A pplied L inguistics, 9, pp. 372± 385. O ’B A R R , W . & A TK INS , B. (1980) W om en’s Langua ge or Pow erless Lan gu age?, in: S A L M C C O N N E LL- G IN E T , R U TH B O R K E R & N E LLY F U R M A N (Eds.) W om en and Language in L iterature and S ociety (N ew Y ork , Praeger). O ’N E IL L , R . (1994) `T he case of Julia Frost’ or `T he lure of the stereoty pe’, in: J. S U N D E R LA N D (Ed.) E x ploring G ender: q uestions and im plication s for E nglish language educatio n (Hem el H em pstead , Pren tice H all). O BU R A , A . (1985) C hanging Im ages: portray al of girls and w o m en in prim ary textb ooks in K enya (N airobi, A C TS Press). P O R EC C A , K . (1984) Sexism in curren t ESL textbooks, T E S O L Q uarterly , 18, pp. 705± 724. P O U LO U , S. (1997) Sexism in the discourse roles of textbook dialogues, Lan gu age Learn ing Jou rn al 15, 68± 73. R EE S- P A R N EL L , H . (1976) W om en in the world of K ernel L essons I nterm ediate, A R E L S J ournal, 2, pp. 29± 31. R IV ER S , W . (1981) T eaching F oreign L anguage S kills (C hicago, IL, University of C hicago Press). R IV ER S , W . & T E M PE R LY , M .S. (1978) A P ractica l G uide to the T eaching of E nglish as a S econd or F oreign L anguage (N ew Y ork, O xford U niversity Press). S U ND E R LA N D , J. w ith A B D U L R A H IM , F ., C O W LEY , M . & L EO N TZ A K O V , C . (1997) G ender in language textbooks: lookin g beyond textual im balance. C R ICE W orking Paper 27, Departm ent of L ingu istics an d M odern English Lan gu age, Lan caster University. S W A IN , M . (1985) C om m unicative com petence: som e roles of com preh ensible input and compreh ensible output in its dev elopm ent, in S. G A S S & C . M A D D E N (Eds) Input in S econd L anguage A cq uisition (C am bridge, M A, N ew bury H ouse). T A LA N S K Y , S. (1986) Sex role stereoty ping in TEFL teaching m aterials, P erspectives , X I, pp. 32± 41. T H O M A S , J. (1983) C ross-cu ltural pragm atic failure, A pplied L inguistics 4, pp. 91± 113. W A LFO R D , G . (1980) Sex bias in physics textbooks, S chool S cience R eview , 62, p. 219. W O M EN IN EFL M A TE R IA LS (1991) O n B alance: guidelines for the representation of w o m en and m en in E nglish language teaching m aterials reprin ted in J. S U N D ER LA ND (Ed.) (1994) E xploring G ender: q uestions and im plicatio ns for E nglish language education (Hem el H em pstead , Pren tice H all).

Acquaintance

9 tokens 7 types

M an

9 tokens 9 types

11 12 13 14 T otals

Partner C olleague

Acquaintance Partner W orker Policem an Boyfriend Interview ee

M ale roles

B oy M an

M an Peter M an Policem an Paul M r W igmore

M ales

10 (i) (ii) (iii)

1 2 3 4 5 6(i) (ii) (iii) 7 8 9

U nit

13 tokens 13 types

M other Daughter Partner C olleague C o-traveller C o-traveller Acquaintance

M other Gill Girl W om an W om an W om an W om an

13 tokens 13 types

Acquaintance Partner W orker C itizen Girlfriend Interviewer

Fem ale roles

W om an Ann W om an O ld w om an Anna M s B rown

F em ales

5F ,4M

W om an

Girl M an

W om an Peter M an O ld W om an Paul M s B rown

W ho starts in m ixedsex dialogues?

Ap pendix: D iscourse R oles in D ialogues for th e T hree B ooks T A BL E AI. H eadw ay I nterm ediate (1987)

27

2

1 2

N /A

2 3 3 5 2 7

M ale turns

43

2

2 1

4

3 2 3 16 2 8

Fem ale turns


248

28

6 16

N /A

10 32 33 62 10 51

N um ber of m ale w ords

302

12

14 6

39

20 17 15 62 40 77

N umber of fem ale w ords

488 M . A . J ones et al.

Jam es

10 tokens 6 types

T otals

Jam es B ob M arco A lan M arco

K arl Porter

Jam es M arco

M ales

14 15

9 10 11a 11b 12 13a 13b

6 7 8

5b

3a 3b 4 5a

1 2

U nit

10 tokens 6 types

C olleague

Brother Boyfriend Friend Boyfriend Friend

Visitor on business Porter

C olleague Friend

M ale roles

Rosie Penny Teresa Rosie Becky Sally Lucy Rita Julia Anna 20 tokens 11 types

Sister G irlfriend Friend G irlfriend Friend Friend Friend C olleague Friend Friend 20 tokens 8 types

Driver W arden Driver W arden

Rita W arden Rita W arden

Asst. Teresa

Secretary Accounts m anager C olleague Friend

F em ale roles

Rita Julia Julia Teresa

F em ales

6F, 3M

Ð

Rita

Ð

Rosie B ob Teresa A lan B ecky

A sst

Ð

Ð

Julia M arco

Ð

W ho starts in mixedsex dialogues?

T A BL E II. L ook A head 2 (1994)

44

1 N /A

7 4 5 7 3 N /A

N /A

67

2 4

8 3 5 6 3 5

5

N /A

3

N /A

10

5

3 4

11

F em ale turns

N /A

3 4

N /A

M ale turns


501

8 N /A

43 23 60 116 64 N /A

N /A

94

N /A

N /A

45 48

N /A

M ale w ords

632

28 112

49 31 60 39 21 47

30

N /A

43

36

52 21

63

Fem ale w ords

D iscourse R oles

489

Int. a Int. b Int. c 1a 1b 2a 2b 3a 3b 3c 4 5a 5b 5c 6a 6b 6c 6d 6e 6f 7a 7b 8 9 10a 10b 11a 11b Totals

Unit

Brother Brother Friend Friend Friends Friends Friend Friend Friend Friends

Customer Friends Friends Friends Friends Friend Boyfriend Friends Friends

Friends

Friends & manager Friends

Friend Son and father 39 tokens 7 types

M r Scott (Vince) Vince and Terry Vince and Terry Terry and Vince Vince and Terry Vince Vince Terry and Vince Terry and Vince

Vince and Dan

Vince, Terry, M an. Terry and Dan

Terry Vince and M r Sc. 39 tokens 5 types

M ale roles

Vince Vince Vince Vince Vince and Dan Vince and Terry Vince Terry Terry Dan, Terry, Vince

M ales

Vince’s girlfriend Friend Friends Friends Sister 35 tokens 5 types

Friends Friends

K im and Rosy Rosy and A ndrea

K im Rosy Sue and Andrea K im and Rosy Sue 35 tokens 5 types

Saleswoman Friend Friend Friends Friend Friends/rivals Vince’s girlfriend Friend

Sister Sister and Friend Friend Friend Friends Friend Friend Friend Friend Friends

Female roles

Saleswoman K im K im K im and Andrea K im Andrea and Kim K im Andrea

Sue Sue and Andrea Andrea Andrea Rosy, Andrea, Kim Rosy Rosy Rosy Rosy K im and Rosy

Females

Terry Sue 11F, 15M

Ð

Vince Terry

Vince

Ð

Ð

Salesw oman K im K im Terry K im Andrea Vince Terry

Vince Vince Vince Vince Rosy Vince Rosy Rosy Terry Dan

17 9 N/A 7 2 142

1 10 2 5 6 1 2 2 2 N/A 17

4 3 8 10 5 15 2 3 3 6

6 4 13 18 1 133

1 3 3 4 4 3 2 1 N/A 5 9

3 5 9 11 8 9 2 3 3 3

337 130 N /A 110 57 2481

13 206 34 76 48 32 40 34 36 N /A 279

48 17 187 162 96 316 25 49 26 122

125 78 203 378 5 2242

34 66 48 60 40 19 28 7 N/A 116 235

102 53 143 106 80 82 21 63 75 74

W ho starts in mixedsex M ale Female Number of Number of dialogues? turns turns m ale words female words

T A B L E III. H otline Interm ediate (1993)

Downloaded by [Lancaster University Library] at 07:55 09 April 2016 490 M . A . J ones et al.