Name Your Favorite Musician: Effects of Masculine Generics and of their Alternatives in German

Journal of Languageand Social Psychology

EDITORS

Howard Giles James BradacUniversity of California University of California

at Santa Barbara at Santa Barbara

BOOK REVIEW EDITOR

Jake Harwood, University of Kansas

EDITORIAL ASSISTANT

Heather Clark, University of California at Santa Barbara

EDITORIAL BOARD

Karin AronssonLinkoping UniversityJanet Beavin BavelasUniversity of VictoriaCharles BergerUniversity of California at DavisRichard Y. BourhisUniversity de Quebec a MontrealJoseph N. CappellaUniversity of PennsylvaniaAaron Castelan CargileCalifornia State University,Long BeachRichard ClémentUniversity of OttawaNik CouplandUniversity of WalesDerek EdwardsLoughborough UniversityJohn EdwardsSt. Francis Xavier UniversitySusan Ervin-TrippUniversity of California at BerkeleyStanley FeldsteinUniversity of Maryland at BaltimoreKlaus FiedlerJustus-Liebig Universität Giessen

Susan R. FussellCarnegie MellonCynthia GalloisUniversity of QueenslandRobert C. GardnerUniversity of Western OntarioBarbara GroszHarvard UniversityTom HoltgravesBall State UniversityWendy LehnertUniversity of MassachusettsSik Hung NgCity University of Hong KongW. Peter RobinsonUniversity of BristolDerek RogerUniversity of YorkEllen RyanMcMaster UniversityGün SeminFree University of AmsterdamWilliam StilesMiami UniversityMargaret WetherellOpen University

FOUNDING EDITOR

Howard Giles, University of California at Santa Barbara

For Sage Publications: Jason Ward, Jim Kelly, and Paul Doebler

Cover Art: Tracy E. Miller

Journal ofLanguageand Social

Psychology

December 2001Volume 20, Number 4

CONTENTS

Enacting Gender Identity in Written Discourse:Responding to Gender Role Bidding in Personal Ads

LAURA L. WINN and DONALD L. RUBIN 393

Similarities and Differences in Mother-Daughterand Mother-Son Conversations During

Preadolescence and AdolescenceSHERRY L. BEAUMONT, VIVIANE C. B. VASCONCELOS,

and MONICA RUGGERI 419

Construction and Action in Food Evaluation:Conversational Data

SALLY WIGGINS 445

Name Your Favorite Musician: Effects of MasculineGenerics and of Their Alternatives in German

DAGMAR STAHLBERG, SABINE SCZESNY,and FRIEDERIKE BRAUN 464

Book Reviews

Louis-Jean CalvetPour Une Écologie Des Langues Du Monde

GRANT D. MCCONNELL 470

Nancy A. Niedzielski and Dennis R. PrestonFolk Linguistics

GRAHAM MCGREGOR 480

Pierre Philippot, Robert S. Feldman,and Eric J. Coates, Eds.

The Social Context of Nonverbal BehaviorCHRIS SEGRIN 483

Lillie Chouliaraki and Norman FaircloughDiscourse in Late Modernity:

Rethinking Critical Discourse Analysis

December 2001Volume 20, Number 4

Carla Willig, Ed.Applied Discourse Analysis:

Social and Psychological InterventionsIan Parker and the Bolton Discourse Network

Critical Textwork: An Introduction toVarieties of Discourse and Analysis

DAVID WELTMAN 487

About the Authors 494

Index 495

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001Winn, Rubin / PERSONAL ADS

ENACTING GENDER IDENTITYIN WRITTEN DISCOURSE

Responding to Gender RoleBidding in Personal Ads

LAURA L. WINNWayne State University

DONALD L. RUBINUniversity of Georgia

This study investigated ways in which gender identity is enacted within written lan-guage.Participants first supplied a self-descriptive letter that might be filed with a datingservice.Next, they responded to a fabricated personal ad posted by a potential dating part-ner. Contextual factors in this study were writing task (self-description or response to apersonal ad) and the gender role “bid” (either instrumental or expressive) of the hypotheti-cal personal ad writer. Individual difference variables were biological gender and mea-sured gender role orientation. Texts were coded for frequency of seven gender-typed lan-guage features (e.g., hedges, first-person pronouns). Writers’ own gender role schemataaffected their language use, as did their biological sex. Contextual factors were morepotent than the writers’ gender in affecting these stylistic features. Overall, writers alteredtheir styles to complement (rather than converge toward) the apparent gender role orien-tations of their interlocutors.

Recent theories depict social identity as enacted or constructedwithin interactions (see Collier & Thomas, 1988; Coupland, 1996).Individuals may choose to either emphasize or de-emphasize aspects oftheir identities in response to the characteristics of a situation. Com-munication adaptation theory (CAT) (Burgoon, Stern, & Dillman,1995; Giles, Coupland, & Coupland, 1991) holds that individuals varytheir language choices within interactions, depending on their socialgoals. Thus, speakers may choose to emphasize (or de-emphasize) par-ticular aspects of their identities as a way of aligning with (or distanc-ing from) interaction partners. In this manner, situational influencesmay have a profound effect on discourse (e.g., Hogg, 1985; Hogg &Rigoli, 1996). In fact, as Meyerhoff (1998) noted, the primary value of

393

AUTHORS’ NOTE:Earlier versions of this article were presented at the Sixth InternationalConference on Language and Social Psychology (Ottawa,Canada,May 1997) and the An-nual Conference of the Organization for the Study of Communication, Language andGender (Chicago, October 1997).

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY,Vol. 20 No. 4, December 2001 393-418 2001 Sage Publications

CAT is that it is designed to address the interactive nature of identityconstruction.

Personal advertisements for dating partners offer a rich approach tounderstanding the processes by which individuals construct theiridentities within the dating context (Coupland, 1996). Here, ad writerscontrol their initial presentations of self by means of written discourse.This written venue offers a chance to at least initially bypass some ofthe early uncertainty-reduction difficulties in mate selection (Ahuvia &Adelman, 1992; Coupland, 1996). However, the requirements of thisgenre may influence self-presentation a great deal. Writers composingpersonal ads must balance the need to be brief when describing them-selves while presenting those aspects of their identities they considermost relevant to finding partners in as vivid a manner as possible(Bolig, Stein, & McKenry, 1984; Montini & Ovrebro, 1990; Parrott,Lemeiux, Travillion, & Harris, 1997). Therefore, dating ads may beseen as a type of “distilled” social identity performance, with this writ-ten discourse as a representation of what individuals believe to be (orwish to portray as) the most relevant aspects of their personalities withrespect to dating encounters.

Gender schematic information is a particularly influential aspect ofsocial identity formation (see Bem,1981,1984,1993;House,Dallinger &Kilgallen, 1998; Yaeger-Dror, 1998). Of particular interest to the cur-rent study was the way participants might use language to framegendered identities within their writing. In her review of linguisticstudies that have investigated CAT with respect to gender,Yaeger-Dror (1998) noted that message senders will often convergewith, or match, the speech patterns of their interlocutors. However,Yaeger-Dror also noted that the social goal of the message sender isequally influential. Sometimes, these social goals can lead to linguisticdivergence rather than convergence. In some cases, divergence mayemerge because the social goal is to complement rather than match thelanguage strategies of the other interactant. The act of matching thediscourse of one’s interlocutor would seem to indicate a bid for solidar-ity; however, if one wished to express dominance, then elements ofdivergence might emerge. Studies on gender-related linguistic mark-ers demonstrate that such goals (e.g., solidarity) often override the sexof the message sender as potential influences on language style (e.g.,Aries, 1996). In fact, both types of influences need to be considered(Yaeger-Dror, 1998).

For instance, Hogg (1985) suggested that at least for oral forms ofdiscourse, the use of gender-linked linguistic markers may be subjectto considerable situational influence. In fact, although gender differ-ences in speech often occur, these are just as likely to be indicators ofgender-related role behaviors as markers of biological sex.On the basisof speech accommodation theory, Hogg predicted that there would beboth biological gender–related and gender role–related differences in

394 JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001

speakers’ behavior, related to the speakers’ convergence with the gen-der of their interlocutors. Speakers were recorded interacting in both asame-sex dyad and mixed-sex larger groups. It was expected that thedyadic situation would elicit more solidarity goals for speakers andthat the group situation, being more formal, would elicit more control-oriented strategies.

Results of this study (Hogg,1985) showed that female speakers usedmore “masculine” (i.e., instrumental) terms and fewer “feminine” (i.e.,expressive) terms in heterogeneous group situations than when theyspoke in all-female dyads. In dyads, both men and women adaptedtheir speech to the more affiliative context of the situation, whereas ingroups, both men and women used more instrumental speech, indicat-ing an awareness of the different needs of this context. Overall, menexhibited less overall speech adaptation than women, despite the con-text. These results confirm Yaeger-Dror’s (1998) contention that boththe situation and the gender of speakers are salient influences ondiscourse.

The current study was conducted as a means of further investigat-ing gender-linked use of linguistic markers in managing written iden-tity presentation. We chose the heterosexual personal ad context as away of obtaining cross-sex targeted discourse. In the following, we firstdescribe a constructivist approach to social identity, followed by theimplications of such an approach for research on gender and discourse.Next, we present relevant studies pertaining to gendered identity inwritten language. Finally, we briefly review the literature on personalads, paying particular attention to those studies with implications forthe issue of constructed gender.

CONSTRUCTION OF SOCIAL IDENTITY

Conceptualizations of social identity in the context of oral as well aswritten discourse have evolved considerably in recent times. Contem-porary identity theory regards social identity as nonessentialist,socially constructed, and enacted (see, e.g., Bohan, 1993; Collier &Thomas, 1988; Sachdev & Bourhis, 1990; Waters, 1990). Asserting thatsocial identities are nonessentialist implies that social categories canencompass substantial variation, so that one’s communication traitscannot be reliably determined on the basis of membership in a particu-lar identity category. Bipolar categorization of gender itself has under-gone considerable challenge (e.g., Bem, 1974, 1993; Freimuth &Hornstein, 1982). Likewise, many scholars within language and gen-der research hold that the dichotomous distinction between male andfemale no longer seems appropriate for describing communicationbehavior (see Kramarae, 1990).

The assertion that social identities are constructed implies that themeanings of gendered identities are not a given but are contingent on

Winn, Rubin / PERSONAL ADS 395

https://www.researchgate.net/publication/246771557_Masculine_and_Feminine_Speech_in_Dyads_and_Groups_A_Study_of_Speech_Style_and_Gender_Salience?el=1_x_8&enrichId=rgreq-9622fc65-9eca-48a9-bff3-eb1a700c82a8&enrichSource=Y292ZXJQYWdlOzI0MDI4MTI5MDtBUzoyNjU5MjUyMzI0MjcwMDlAMTQ0MDQxMjkyNDQzMQ==

history and context. Kramarae (1990) offered the example of the“strong-minded women” of the nineteenth century, who would today beconstructed as “feminists.” Contextual influences may also shift themeaning of gendered identity for individuals. For example, the experi-ence of being a woman in a room full of men may make one’s genderespecially salient. Similarly, within the specific context of heterosexualmate selection, both biological and psychological gender are especiallysalient factors influencing discourse style.

Finally, to assert that social identities are enacted is to recognizethat one creates identification through verbal performance and innegotiation with one’s interactants. Thus, one may emphasize oneaspect of identity (e.g., gender) in one context and emphasize anotheraspect (e.g., ethnicity) in another context. In fact, as much of the litera-ture on language code shifting reveals, one may emphasize orde-emphasize various aspects of identity even within the course of oneconversation (e.g., Blom & Gumperz, 1972).

Written language serves as a vehicle for expressing and construct-ing many facets of social identity, no less than does speech (Rubin,1995). Although oral discourse is accompanied by a rich array of pros-ody, kinesics, and other nonverbal signals, writing is more constrainedby its medium and also less transactional than oral speech forms. Still,the distinctions between oral and written linguistic codes are more amatter of degree than kind (Tannen, 1982). For example, the use ofpunctuation in written discourse may function similarly to the use ofintonation patterns in oral communication (Chafe, 1986). By the sametoken,written discourse may contain linguistic markers by which writ-ers convey ethnic, role, and gender identity.

Many studies of variation in written language are consistent withthis contemporary notion of social identity. For instance, studies ofwriting produced by high school–aged or college-aged speakers of Afri-can American Vernacular English (AAVE) show that these writers canbe quite adept at manipulating sociolinguistic markers. When theydeem it desirable to project a mainstream identity, they typicallyreduce the incidence of nonstandardisms and converge with norms forStandard Edited English (Zeni & Thomas, 1990). On the other hand,these writers may increase the frequency of AAVE ethnolinguisticmarkers when they wish to express alienation from mainstreamnorms (Balester, 1993) or when they seek to evoke solidarity with read-ers from their own minority ethnolinguistic backgrounds (Redd, 1995).Similar patterns of planned convergence and divergence may arisewhen gender identity is negotiated within discourse.

GENDER AND LANGUAGE

Early studies of gender and language (Edelsky, 1976; Kramer,Thorne, & Henley, 1978; Lakoff, 1975) associate female speakers with


features that hedge or blunt assertions (e.g., “maybe,” “sort of,” “Iguess”) and avoid conflict with listeners by the use of politeness formu-las (e.g., “if you don’t mind” or the use of question forms rather thanbald requests). Other “markers” commonly associated with female lan-guage include the use of double-sided arguments (e.g., “It was probablyShakespeare’s sister, but then again, some people believe it wasMarlowe”), expressions of uncertainty (e.g., “I don’t know, but . . .”), andthe use of certain vocabulary likely to be judged as trivial (e.g., fine-grained color terms such as fuchsia).

Curiously, research on language and gender remains vulnerable tocriticism for failing to take into account more dynamic models of iden-tity. For instance, both Crawford (1995) and Aries (1996) specificallytargeted studies aimed at establishing catalogues of discrete languagefeatures that discriminate between male and female speakers (e.g.,Mulac, Lundell, & Bradac, 1986). They argued that across the litera-ture, men’s and women’s language is more similar than different andthat apparent differences are quite context specific. Rather than enu-merating superficial sex differences, it would be more productive forresearchers to examine features of interaction contexts that leadspeakers to enact gender roles in particular ways: what West andZimmerman (1987) called “doing gender.”

Similarly, both Freimuth and Hornstein (1982) and Kramarae(1990) challenged the tradition of dichotomizing gender differenceswithin empirical studies. They argued, like Bem (e.g., 1981, 1993), forthe expansion of gendered identity to a more continuum-based per-spective. In fact, in a recent study, Grob, Meyers, and Schuh (1997)challenged the “dual cultures” model of gender. They found that maleand female speakers were more similar than different in their use ofpreviously gender-linked linguistic markers of tag questions, hedges,and interruptions. Grob et al. discussed the implications of a gendersimilarity model for language-related study, concluding that a morecontinuous or more complex view of gender is warranted. Because lan-guage use is socioculturally conditioned, where gender differences doexist, they are more likely due to differences in gender role schematathan to biological sex.

Within oral discourse, if women are found to use more standard dia-lect features than men, the reason may be that women are more oftencompelled by the terms of their employment to function in contextsdemanding linguistic propriety, whereas men may be less subject tothose particular workplace demands on language (Nichols, 1983). Onthe other hand, if women use “you know” more often than men, the rea-son may be that they are failing to receive the back-channel cues fromtheir listeners, affirming that they still have their listeners’ attention(Holmes, 1986). This view receives support from studies on men’s lan-guage as well. Men who by happenstance find themselves occupyinginteraction roles more typically occupied by women (e.g., a powerless


witness in a trial) are likely to enact those roles using language that isstereotypically associated with women’s speech (O’Barr & Atkins,1980). Consequently, there is considerable support for a view of genderidentity as dynamically constructed through language choice ratherthan as a static label.

GENDER IDENTITY NEGOTIATION IN WRITING

Studies of gender and written language have proliferated (forreviews, see Roulis, 1995; Rubin & Greene, 1992). Gender-linked fea-tures in writing are derived from a number of sources. For example, lit-erary critical studies of female versus male authors yield variablessuch as markers of excitability (exclamation points) and nonessentialdigressions (parentheses) or lack of conjunctive coherence, all of whichhave been subsequently subject to empirical textual analysis (e.g.,Hiatt, 1977). Similarly, analyses of business writing (e.g., Sterkel,1988) have suggested that hedges and expressions of uncertainty aregender related in writing as well as in speech.

To be sure, some of these studies are guilty of a “sex difference”approach to simply cataloguing stylistic features of men’s and women’swriting. However, others do look at gendered writing as a function ofthe interaction context from which it emerges. Johnson and Roen(1992), for instance, examined politeness strategies (stereotypicallyassociated with women’s language) in comments written by membersof peer-editing groups. They found that overall, male and female writ-ers produced similar numbers of politeness strategies. However, thenature and functions of these strategies differed not only by the genderof the writer but also by the gender of the reader. Female writerstended to use politeness-framing strategies to bracket their messages,but this was done almost exclusively when writing to other women.Johnson and Roen concluded that this pattern reveals the use of polite-ness strategies to establish solidarity between writer and reader.

Rubin and Greene (1992) also adopted a constructivist approach toexamining gender enactment in writing. In their study, individualvariation within dichotomous categories of sex was acknowledged bymeasuring individual differences in instrumental and expressive gen-der role schemata (Bem, 1981). Analytic procedures evaluated theseindividual difference variables prior to evaluating the effects offemale-male group differences. The research design examined writtenlanguage produced in two contrasting contexts: (a) informal letterswritten to well-known readers and (b) revised essays addressed tointerpersonally remote readers. A review of the literature revealed aseries of written language features that have traditionally been linkedwith gender, including markers of excitability (e.g., the use of underlin-ing or exclamation points), nonessentials (e.g., dashes and parenthe-ses), connectives (a combination of illustrators, illatives, adversatives,


causals, additives, temporals, and conditionals), and hedges (e.g., acombination of intensifiers, deintensifiers, proximals, modal adjuncts,perceptual verbs and auxiliaries of possibility, audience acknowledge-ments, first-person markers, enumerations, and sentence length/verbosity).

Rubin and Greene (1992) found that overall, women in the studytended to use more markers of excitability (e.g., exclamation points)and a higher frequency of egocentric phrases (e.g., “I believe that . . .”)within their writing than men. When asked to write in an expressiverather than argumentative context, men and women did not differ intheir use of hedges. However, when writing within an argumentativecontext,men used far fewer hedges than women.With regard to genderrole orientation rather than biological gender, writers with moreinstrumental (rather than expressive) orientations were more likely touse longer sentences and to refer to themselves in the first person.

However, the overall findings of Rubin and Greene’s (1992) studyconfirmed many of the suppositions made by critics of “sex difference”research (see Aries, 1996; Crawford, 1995). Collapsed across languagecontexts, the writing of men and women exhibited far more similaritiesthan differences.Gender typicality was difficult to establish.For exam-ple, even a feature such as anticipating a reader’s reservations (e.g., “Iknow you might think that drug testing is no big deal if you have noth-ing to worry about, but . . .), which was found to occur more often inwomen’s writing than in men’s, could hardly be called “gender typical.”It appeared in the writing of fewer than half the female writers. More-over, context seemed to have the greatest effect on some aspects of gen-der performance. For instance, women and men did not differ in theiruse of egocentric sequences (e.g., “I guess,” “I think”) when writing to adistant audience and after revision, but the invitation to write in-formally to intimate readers engendered a higher incidence of suchsubjective phrasing for women. Men’s production of this particular fea-ture was generally unresponsive to the social context of the writingtask.

Rubin and Greene’s (1992) study suffers from several flaws, how-ever. From the perspective of identity theory, perhaps the most seriousflaw is that nothing in the writing task was designed to accentuate thesalience of gender identity in particular. The elicitation task may havecalled on writers to enact aspects of identity related to friendship or tosubordinate status vis-à-vis the two audiences, but there was little callfor writers to “do gender” (West & Zimmerman, 1987) in this context.Moreover, the elicitation context did not include any element ofinteractivity or negotiation with the intended readers of the writtenmessages. At best, writers were ascribing identities to audiences towhom they had to construct themselves on the basis of rather scarcedata (Rubin, 1998). The language elicitation context selected for thecurrent study was designed to address these gaps.


Personal Ads

A small body of literature on personal ads reveals a precedent forconsidering this medium as a discourse mode representative of themate selection process (e.g., Ayres, 1992; Parrot et al., 1997). Often,these studies conduct content analyses of newspaper ads to determinewhether there are gender differences in representations in self versusother attributions (e.g., Harrison & Saeed, 1977; Willis & Carlson,1993). The process of offering one’s own set of self-descriptors whilebidding for a partner who possesses another set invokes the inherent“bargaining” component that has been a theoretical cornerstone ofwork in this area (Bolig et al., 1984; Coupland, 1996; Deaux & Hanna,1984; Harrison & Saeed, 1977; Montini & Ovrebro, 1990).

Most studies of personal ads consider only the writers’ and readers’biological sex, failing to consider more theory-driven issues of genderrole identity (e.g., Sitton & Rippee, 1986). However, the findings of afew studies suggest that gender identity (as opposed to biological sexalone) can be an important factor for the process of identity bargainingwithin the context of mate selection. For instance, Willis and Carlson(1993) considered over 800 personal ads placed in newspapers. Theyasked college students to rate (a) which ads they would prefer if theywere interested in selecting dating partners and (b) which ad was mostlikely to be “successful.” Those ads preferred mainly by men (i.e., ratedlowest by women) and those preferred mainly by women (i.e., ratedlowest by men) were labeled “sexually dimorphic,” meaning that thecontent of the ads was likely to appeal to only one gender. Those thatwere preferred by both men and women were considered to be moreandrogynous by the authors. The ads that appeared to “target” opposite-gender readers (by emphasizing the writers’ own gender traits) weremore likely to be successful. Thus, it appears that within this context,individuals privilege more stereotypical gender identity bids (see alsoThiessen, Young, & Burroughs, 1993). The authors did not consider,however, the gender-linked linguistic markers used within the letters,so it is difficult to say to what degree readers were responding togendered language when they made their choices.

Koestner and Wheeler (1988) employed Bem’s (1981) gender roleschema theory directly and examined the instrumental and expressivecontent within personal ads collected from local newspapers. Instru-mental traits tend to be more “male valued” (e.g., assertive, indepen-dent, and goal oriented), whereas expressive traits tend to be more“female valued” (e.g., warm, caring, and family oriented). Ads that ref-erenced educational and professional status were coded as instrumen-tal. The authors found that both women and men were likely to writeads that offered what they apparently believed the other preferred.Women shaped their social identities in more instrumental terms,whereas men emphasized more expressive traits. At the same time,


women were more likely to describe their desired partners in expres-sive terms, whereas men listed more instrumental qualities in describ-ing their ideal mates.

Consequently, the findings of Koestner and Wheeler (1988) suggestthat within the discourse of personal ads, individuals are likely to basetheir descriptions of both themselves and their ideal partners on theirbeliefs about what their readers may prefer (see also Greenlees &McGrew, 1994). This is consistent with CAT (Giles et al., 1991), inwhich individuals are seen to adapt their language in one of two ways.Language users may adapt by converging with their interlocutors’ lan-guage styles, or they may adapt by using styles that diverge in a man-ner that complements rather than mimics their interlocutors’ styles.Koestner and Wheeler’s findings suggest that personal ad writers tendto adapt by complementing rather than converging with the role iden-tities of their intended audiences. However, because contextual com-parisons were not available in this study, that conclusion is tentative atbest.

It should be noted that most of the studies of personal ads use archi-val data (i.e., published ads found in newspapers) and thus can onlyinfer characteristics of the language producers. However, even amongthose studies in which personal ad writers were indeed contacted (e.g.,Ayres, 1992; Jason, Moritsugu, & DePalma, 1982), the gender identi-ties of the writers are described in terms limited to the biological gen-der of the ad placer. No previous studies employ experimental manipu-lations to directly observe the ways in which writers shape theirgendered identities within the personal ad context. Also, although ithas been argued that the personal ad context performs an uncertaintyreduction function for individuals (e.g., Parrot et al., 1997), little isknown about the nature of self-disclosure on the affinity-seeking pro-cess between writer and reader. In other words, when attraction is thedesired goal, as might be supposed in the personal ad context, writersare likely to seek affinity with their readers (Bell & Daly, 1984;Douglas, 1987). However, in a cross-sex attraction scenario, are writerslikely to match (converge) or complement (a form of divergence) theperceived gender role traits of the other?

The present study was designed to examine adaptation in the gen-der identity of individuals as portrayed in their written discourse. Assuggested by CAT,we predicted that characteristics of the hypotheticalaudience would influence writers’ self-presentations. Therefore, wemanipulated the gender role orientation (instrumentality orexpressivity) of the hypothetical target audience. To maximize thechance that writers would indeed attempt to attract readers, writerswere given instructions to this effect for both letter-writing tasks.Much of the research conducted within the frame of CAT indicates thatattraction, or solidarity, in interpersonal relationships is often accom-plished via converging language styles (e.g., Giles et al., 1991).


Alternatively, a complementary language style may occur when thereis reason to believe that attraction or solidarity may best be expressedby adopting a slightly divergent language style (Burgoon et al., 1995,p. 19).

We wished to consider the possible change in language marker usedue to gender and/or gender role orientation separately from thoseeffects due to the writing task alone (i.e., addressing an indeterminateversus a specified audience; see Rubin, 1998); consequently, we posedthe first two research questions below. Also, we wished to look at thegender-typed nature of the audience separately from both the genderand gender role orientation and the writing task; consequently, weposed a third research question related to this issue. Thus, ourresearch questions were as follows:

Research Question 1: Will there be a residual difference in writers’ gender-typed linguistic markers related to their biological gender after controllingfor the effects of their expressive or instrumental gender role orienta-tions, or do effects of psychological gender role schemata eliminate anyresidual variance due to biological sex?

Research Question 2: After controlling for the effects of expressive and in-strumental gender role identities, will there be a difference in writers’gender-typed linguistic markers as a function of the writing task alone(self-description vs. responding to a specific audience)?

Research Question 3: After controlling for the effects of expressive or instru-mental gender role identities, will there be a difference in writers’ gender-typed linguistic markers as a function of the gender role bid (instrumentalvs. expressive) issued by a personal ad?

METHOD

PARTICIPANTS

A total of 84 undergraduate college students attending a large,southeastern U.S. university participated in the study. Participantswere recruited from introductory communication courses and receivedcourse credit for their involvement. Participants’ ages ranged between18 and 34, with a mean age of 21.4 years. The final sample includedonly participants who reported that they were heterosexual and nativeEnglish speakers. Thus, the responses of 4 students were not used.Twenty men and 24 women were randomly assigned to the instrumen-tal ad condition, and 21 men and 20 women were assigned to theexpressive ad condition. About half of the participants (46%) reportedthat they had never read a personal ad, whereas 54% reported thatthey had read newspaper personal ads previously. None reported thathe or she had ever responded to or placed a personal ad, and mostreported that they were not likely to respond to (95%) or to place (95%)a personal ad in the future.


STIMULUS MATERIALS

Two personal ads were constructed for each of the two study condi-tions. The first ad was intended as a bid for an expressive gender roleschema companion (enjoys walking in the woods, attending art shows,and seeking a meaningful relationship). The other ad bid for aprototypical instrumental gender role schema companion (enjoys fit-ness training and NASCAR races and seeks a mate to “walk on the wildside”). The ad texts were otherwise comparable (equal length, physicallayout, etc.). The ad texts were constructed from portions of personalads appearing in the local newspaper. They were presented to partici-pants in a format designed to simulate an actual newspaper ad.

Participants also completed a 20-item adaptation (Wheeless &Dierks-Stewart, 1981) of the Bem Sex Role Inventory (BSRI). Thisscale asks participants to rate themselves according to whether theytend to exhibit 10 expressive traits (e.g., caring) and 10 instrumentaltraits (e.g., assertiveness). Responses were rated on a Likert-type scaleranging from 1 (never true) to 7 (always true).Finally,participants com-pleted a brief questionnaire inquiring about their familiarity with per-sonal ads and also about the stability and types of intimate relation-ships in which they were currently involved.

PROCEDURE

Participants engaged in two writing tasks. For the first, they wroteself-descriptions that could hypothetically be submitted to a datingservice. Participants were instructed to represent themselves as hon-estly as possible so that the dating service could match them with com-panions who would be truly compatible. This first writing sample wasintended to assess each participant’s “baseline” gender self-presentationbefore exposure to the stimulus ads. Participants were instructed towrite letters that described themselves realistically and yet that wouldbe most likely to attract someone to whom they would be compatible.

Following this first writing episode, participants read one of the twosimulated personal ads (either instrumental or expressive). Partici-pants were instructed to produce a second writing sample, intended tobe a letter of response to the personal ad. For this second writing task,participants were asked to assume that they had chosen the particularad and were interested in meeting this person. They were also told to“try to present yourself in a way that this person would find appealing”to maximize the chance that they would write letters designed toattract (rather than repel) the hypothetical other.

Following this, participants completed an adaptation (Wheeless &Dierks-Stewart, 1981) of the BSRI. They also completed a brief ques-tionnaire inquiring about their familiarity with personal ads, their


own sexual orientation, their native language status, and other infor-mation not used in the current study.

Language Coding

The scheme for coding features of gender-linked language closelyapproximated that developed by Rubin and Greene (1992; see alsoRubin & Greene,1995).These original coding features were selected onthe basis of their use in prior research both in oral (e.g., Rubin & Nel-son, 1983) and written (e.g., Hiatt, 1977) gender-linked language. Thecoding scheme is detailed rather than impressionistic. Thus, this cod-ing strategy attempts to mitigate charges that various studies oflanguage and gender cannot be compared because the composition oftheir categories is unspecified (Aries, 1996). However, individuallycoded features were logically clustered into a smaller number ofdependent variables to make explication and statistical analyses moremanageable.

The language features chosen for use in the current study wereselected on the basis of results found by Rubin and Greene (1992). Spe-cifically, those categories that revealed significant differences withrespect to either the biological gender or gender role orientation of thewriters were selected. These included language features traditionallyassociated with female writing: markers of nonessential information(dashes, parentheses), which may signal digressiveness and/or embel-lishment of the text; markers of excitability (exclamation points,underlining), which may signal high emotionality and also embellishthe basic text; and hedges, including intensifiers, deintensifiers,proximals, modal adjuncts, and auxiliaries of possibility. Hedges maysignal tentativeness or lack of assertiveness on the part of the writer.Finally, first-person pronouns (the use of “I” or “me”) were considered.The use of first-person pronouns in conjunction with verbs of percep-tion, affect, or cognition (e.g., “I guess . . .” or “I feel . . .”) signals tenta-tiveness, whereas their use in other contexts (e.g., “I went . . .” or “. . . tome”) signals subjectivity. Both of these traits are associated with femi-nine style, though empirical findings are mixed (see Rubin & Greene,1992).

Two coding categories that were more traditionally associated withmen’s writing were also included. The first of these was connectives,including adversatives, illustrators, additives, temporals, and logicalconnectors. Connectives signal explicit organization and hierarchicalrelations. Also, syntactic complexity was created by calculating wordsper sentence. Sentence length was related to an instrumental genderrole orientation in Rubin and Greene’s (1992) study.

Finally, markers of audience acknowledgements (the use of second-person pronouns and questions) were also included as a kind of manip-ulation check. The first writing task was intended to be written to a


https://www.researchgate.net/publication/229068236_Men_And_Women_in_Interaction_Reconsidering_the_Differences?el=1_x_8&enrichId=rgreq-9622fc65-9eca-48a9-bff3-eb1a700c82a8&enrichSource=Y292ZXJQYWdlOzI0MDI4MTI5MDtBUzoyNjU5MjUyMzI0MjcwMDlAMTQ0MDQxMjkyNDQzMQ==

generalized and nondisclosed audience. On the other hand, for the sec-ond letter, a specific audience was identified for the writers. If the writ-ers indeed perceived a difference between the two tasks, then it waslikely that they would acknowledge a specifically identified audiencemore frequently than they would a generalized one.

Both writing samples (self-description and response to ad) producedby each participant were coded using this coding scheme. One coderanalyzed all of the texts, and a second coder independently analyzedone third of the texts. Reliabilities for the 19 specific language featurescoded were estimated by Pearson correlation coefficients and were typ-ically .89 or higher, with only proximals falling below this (r = .82).

Data Analysis

To avoid effects due to sheer verbosity, each of the seven languageindices listed above was converted to a relative frequency, with totalnumber of words serving as the denominator (syntactic complexity wasalready denominated by number of words). Each of the eight relativefrequencies was subjected to a 2 (Participant Gender) × 2 (Instrumen-tal or Expressive Ad Bid) × 2 (Self-Description or Response to Ad-Writing Task) repeated measures ANCOVA conducted with the GLMprocedure in SAS (Version 6.1).

Covariates were measured instrumental and measured expressivegender role orientations, as measured by the BSRI. A hierarchicalmodel was employed, with covariate effects entered first, then maineffects, then interactions. Participants were nested in combinations ofgender by ad bid and crossed with the repeated measure (writing task).Where interactions proved statistically significant, Bonferonni’s t cri-terion for means comparisons was used to test for preplanned butnonorthogonal post hoc contrasts. Because only the univariate resultswere of interest to the research questions proposed by the currentstudy, multivariate results are not reported (see Huberty & Morris,1989).

Covariates

The mean score on the BSRI was 5.24 for instrumental traits and5.79 for expressive gender role traits. Scores for instrumental traitsranged from 3.1 to 7.0, with a slight negative skew (–.12). Similarly,scores for expressive traits ranged from 4.2 to 7.0,with a negative skew(–.48), indicating a tendency for more people to report having thesetraits than not.For instrumental gender role schema scores, the meanswere 5.13 for women and 5.36 for men. No significant differencebetween these means was indicated by t-test analysis. For expressivegender role schema scores, the means were 5.94 for women and 5.62 for


men. This constituted a statistically significant difference, t(1, 83) =–2.29, p < .025.

Manipulation Check 1

To address the more audience-centered nature of the letter-writingmanipulation, audience acknowledgements (second-person addressesand questions) were coded for both letters. As might be expected, writ-ing task was the sole statistically significant factor for relative fre-quency of audience acknowledgements, which are considered to bemarkers of interpersonal sensitivity, F(1, 81) = 142.18, p < .0001, η2 =0.80. This indicates that participants indeed produced more responsesacknowledging their hypothetical audience when they responded tothe personal ad (M = 0.02749) than when they wrote their initialself-descriptors (M = 0.00242).

Manipulation Check 2

To address the gender-typed nature of the stimulus ads, a post hocmanipulation check was conducted with a separate group of partici-pants who received course credit for their participation. Twenty-fourstudents read and responded to questions regarding the expressive ad,and 24 read the instrumental ad. All were asked to complete theadapted BSRI (Wheeless & Dierks-Stewart, 1981) with regard to theirperceptions of the gender schema of the person who had written the ad.A two-tailed t test, t(1, 46) = 7.06, p < .000, revealed that participantsindeed rated the ad intended to be instrumental as more instrumental(M = 5.36) than the ad intended to be expressively oriented (M = 3.16).Similarly, a two-tailed t test, t(1, 46) = –5.10, p < .000, revealed thatrespondents perceived the expressive ad (M = 4.90) as more expressivethan the instrumental ad (M = 3.57).

RESULTS

Results are presented below for each of the six gender-related lan-guage features. The cell means for all seven coded language features(including audience acknowledgements) appear in Table 1. Results ofthe univariate ANCOVAs for all seven of the coded language featuresare summarized in Table 2.

Female/Expressive-Typed Language Features

Nonessentials. The univariate effect for gender of participants wassignificant for nonessential information, F(1, 79) = 12.70, p < .001, η2 =


Table 1Cell Means for the Seven Coded Language Features

Male Participants Female Participants

Instrumental Ad Bid Expressive Ad Bid Instrumental Ad Bid Expressive Ad Bid

Self- Self- Self- Self-Writing Task Description Response Description Response Description Response Description Response

Nonessentials 0.00181 0.00106 0.00216 0.00172 0.00368 0.00434 0.00686 0.00510Excitability 0.00094 0.00423 0.00208 0.00056 0.00309 0.01037 0.00457 0.00407Hedges 0.02210 0.03146 0.02332 0.02879 0.04925 0.04925 0.04504 0.05811First person 0.10324 0.10247 0.10773 0.08788 0.02826 0.02535 0.02309 0.04118Connectives 0.04619 0.05024 0.04521 0.04521 0.10881 0.10449 0.11266 0.09433Syntactic complexity 12.9387 14.3715 13.2152 14.9645 13.8150 14.4815 13.0785 14.6470Audience acknowledgment 0.00172 0.02673 0.00348 0.03062 0.00135 0.02814 0.00331 0.02418

407

Table 2Summary of Biological Gender by Ad Type by Writing Task ANCOVAs for the Seven Dependent Variables

NE EX HG FP CN SC AU

Source df MS F MS F MS F MS F MS F MS F MS F

Gender 1 .00048 12.70*** .00056 7.05** .00012 0.47 .00094 1.31 .00209 5.01* 0.73028 0.05 .00008 0.33Ad type 1 .00005 1.44 .00011 1.40 .00024 0.94 .00109 1.52 .00001 0.03 0.23883 0.02 .00005 0.22G × A 1 .00003 0.86 .00001 0.07 .00014 0.56 .00004 0.06 .00145 3.47 5.09079 0.37 .00015 0.63Errorbetween 79 .00004 — .00008 — .00025 — .00072 — .00042 — 13.9318 — .00023 —Task 1 .00001 0.49 .00019 5.24* .00237 12.51*** .00494 10.77** .00041 1.46 77.4700 14.37*** .02634 142.18***G × T 1 .00000 0.00 .00019 1.80 .00000 0.00 .00001 0.02 .00049 1.73 2.37367 0.44 .00005 0.29A × T 1 .00001 0.42 .00042 11.35*** .00078 4.10* .00289 6.30 .00005 0.19 3.92723 0.73 .00004 0.21G × A × T 1 .00002 0.68 .00002 0.64 .00164 8.65** .00007 0.15 .00125 4.42* 0.90556 0.17 .00017 0.92Errorwithin 81 .00003 — .00004 — .00019 — .00046 — .00028 — 5.39017 — .00019 —

Note. NE = nonessentials, EX = excitability markers, HG = hedges, FP = first-person pronouns, CN = connectives, SC = syntactic complexity, AU = audienceacknowledgements, G = gender, A = ad type, T = task.*p < .05. **p < .01. ***p < .001.

408

0.37. The means indicated that women (M = 0.00490) in this studywere about 3 times more likely to produce markers of nonessentialinformation (e.g., dashes and parentheses) than men (M = 0.00169).

Excitability. The gender of participants also exhibited a significantmain effect for markers of excitability, F(1, 79) = 7.05, p < .01, η2 = 0.29,indicating a tendency for women (M = 0.00563) to use about 3 timesmore of these excitability markers (e.g., exclamation points and under-lining) than men (M = 0.00194). In addition, there was a significantmain effect for writing task, F(1, 81) = 5.24, p < .02, η2 = 0.25, indicatinga tendency for respondents to use more markers of excitability whenresponding to a specific ad bid (M = 0.00502) than when they wrotetheir initial self-descriptors to a generalized audience (M = 0.00268).

A significant interaction between writing task and ad type alsoemerged for excitability, F(1, 81) = 11.52, p < .001, η2 = 0.35. Pairwisecontrasts of interest between cell means indicated that for the instru-mental ad condition, respondents tended to use more excitabilitymarkers when they responded to the specific ad bid (M = 0.00758) thanwhen they wrote their initial self-descriptors to a general audience (M =0.00211). In turn, those responding to an instrumental genderrole–specific ad bid tended to produce more excitability markers thandid those responding to an expressive gender role ad bid (M = 0.00227).

Hedges. There was a significant covariate effect for expressive gen-der role scheme, βExpr = –.01373, F(1, 78) = 5.13, p < .03, on hedges. Thenegative regression weight for expressive gender role traits indicates anegative relationship between expressivity and the relative frequencyof hedges. There was also a significant main effect for writing task onparticipants’ use of hedges, F(1, 81) = 12.51, p < .0007, η2 = 0.37. Here,participants writing their initial self-descriptors to a general audience(M = 0.02437) tended to use fewer hedges than when they responded tothe specific audience personal ad (M = 0.03136).

There was a significant two-way interaction between the writingtask and the type of gender role ad bid for hedges, F(1, 81) = 4.10, p <.05, η2 = 0.22. However, there was also a significant three-way inter-action between writing task,gender role ad bid,and participant genderfor hedges, F(1, 81) = 8.65, p < .004, η2 = 0.31. Pairwise contrastsbetween cells of interest revealed that women produced more hedgeswhen responding to an expressive ad bid (M = 0.04118) than they didbefore exposure to the ad (M = 0.02309). They also produced signifi-cantly more hedges than men when responding to an expressive ad (M =0.02879).

First-person pronouns. There was a significant difference by writingtask, F(1, 79) = 10.77, p < .002, η2 = 0.35, for the use of first-person pro-nouns. Specifically, when responding to an ad, writers tended to use


fewer first-person pronouns (M = 0.09752) than when writing their ini-tial self-descriptors to a general audience (M = 0.10812). There wasalso a significant interaction between gender role ad bid and writingtask, F(1, 79) = 6.30, p < .01, η2 = 0.27. Pairwise contrasts between cellsof interest indicated that participants responding to the expressive adtended to use significantly fewer first-person pronouns (M = 0.09103)than they did when they wrote their initial self-descriptors to a generalaudience (M = 0.11013).

Male/Instrumental-Typed Language Features

Connectives. The instrumental covariate exerted a significant effect,βInstr = –.00768, F(1, 81) = 5.39, p < .023, on connectives. The negativeregression weight for instrumental gender role traits indicates a nega-tive relationship between instrumentality and relative frequency ofconnectives. There was also a significant main effect by gender, F(1, 79) =5.01, p < .028, η2 = 0.24, with women (M = 0.05031) using more connec-tive phrases than men (M = 0.04550).

In addition, there was a significant three-way interaction betweenwriting task, gender role ad bid, and gender for connectives, F(1, 81) =4.42, p < .04, η2 = 0.23. A comparison between means of interestrevealed that after exposure to the expressive ad bid, women producedsignificantly more connectives (M = 0.05811) than men (M = 0.04063).There were no other significant contrasts of interest.

Syntactic complexity. There was also a significant difference by writ-ing task for syntactic complexity, F(1, 81) = 14.37, p < .0003, η2 = 0.39.This variable was composed of the ratio of words to sentences. Here,when writing in response to personal ads, writers were likely to pro-duce longer sentences (M = 14.614) than when writing self-descriptors(M = 13.287).

DISCUSSION

Our goal was to consider the joint effects of two contextual factors,(a) audience-related gender role bidding and (b) type of writing task, onthe written language of men and women in light of communicationadaptation (Burgoon et al., 1995) and communication accommodation(Giles et al., 1991) theories. In addition to considering the biological gen-der of the language users, we also took into account their instrumentaland expressive gender role schemata (Bem, 1974). For some languagefeatures associated with gendered language (e.g., nonessentials,hedges, connectives), either biological gender or psychological genderschemata exerted independent effects. In general, however, the results


support previous notions (e.g., Rubin & Greene, 1992) that the writingcontext plays at least as much of a role in shaping gender-typed lan-guage use as the gender of the writer. For a number of such features(markers of excitability, use of first-person pronouns), writersappeared to be adopting a complementary or interpersonal compensa-tion strategy (Burgoon et al., 1995) rather than converging with theperceived gender role orientation of their audience.

GENDER EFFECTS

Our first research question focused on effects of the biological gen-der of the writer, in contrast with psychological gender role orientation,on the production of gender-typed written language. Given that lin-guistic style is socioculturally rather than biologically conditioned,andgiven previous findings that biological sex is not the dominant deter-minant of female-typed language (e.g., O’Barr & Atkins, 1980), we sur-mised that we would find few effects of biological sex after covaryingfor gender role schemata. The results, however, were mixed on thisscore. Two language features, syntactic complexity and first-personpronouns, showed no effects for either the writer’s own biological gen-der or gender role orientation.

Simple main effects for biological sex, but no covariate effects forgender role schema, did emerge for two variables. Replicating previousfindings (Rubin & Greene, 1992, 1995), women as compared with menin this study did produce a higher relative frequency of nonessentials(e.g., parenthetical expressions), postulated as indices of digressionand nonlinearity in women’s discourse (Hiatt, 1977). Women alsoexpressed more markers of excitability (e.g., exclamations), postulatedas tokens of emotionality. Consequently, for these two language fea-tures at least, the results show some support for the prevalent ideathat women tend to use more female-typed language features in theirwriting. These findings contradict our supposition in the present studythat gender role schemata are more predictive of gender-typed lan-guage than is biological sex.

Connectives (e.g., “because,” “unless,” “then,” “and”) were associatednot only with biological sex but also with psychological gender roles.For connectives,women and low instrumentals tended to produce moreconnectives than men and high instrumentals. This finding runs con-trary to previous suppositions, which attribute more hierarchicallyorganized discourse rich in explicit connections to men (see Hiatt,1977). However, these results do again replicate those found in Rubinand Green’s (1992) study.

A number of gender-related effects also emerged for the category ofhedges (e.g., “kinda,” “I think,” “probably,” “might have”). People whoreported high levels of expressive gender role orientation used hedgesinfrequently. Although hedges have been stereotypically associated


with women’s speech (e.g., Lakoff, 1975), empirical research has gener-ally not borne out any such association (Crawford, 1995). In the pres-ent study, the negative slope for expressive gender role schemata onthis variable suggests that personality traits such as nurturing andcooperativeness can actually depress expressions of tentativeness.

Nor was there any main effect of biological sex on use of hedges oncethe effects of psychological gender had been partialed out. On the otherhand, the analysis of interaction effects suggested that women did usehedges more adaptively than men, for example, deploying a higher fre-quency of hedges when responding to an expressive gender role bidcompared with an instrumental gender role bid. Indeed, a similarresult emerged for connectives. For these two language features, then,it appears that women were more responsive to the perceived nature ofthe audience than men. Similarly, Hogg (1985) found that womenengaged in more adaptive use of gender-typed features than men.

CONTEXT EFFECTS

Our second research question pertained to the situational variableof the writing task:describing oneself to a generalized other as opposedto responding to specific ad poster.Writing task exerted main effects onseveral of the linguistic indices, thus confirming that most participantswere indeed sensitive to the differing rhetorical demands posed by thetwo situations. When responding to a more defined audience, writersused more markers of excitability, more hedges, and greater syntacticcomplexity than they did when responding to a more amorphouslydefined audience. Writers were also less likely to use first-person pro-nouns when addressing the more specific audience. In the case inwhich participants were composing for the more generalized audience,they chose to talk about themselves as opposed to asking questionsabout the other or suggesting activities that they could do together.

Responding to a specific bid for relationship thus resulted in a moreinterpersonal style of writing. This more interpersonal style includedgreater affective expression, seen in the increased use of excitabilitymarkers, and this stylistic shift held equally for men and for women.By the same token, in the explicitly interpersonal context of addressinga particular ad poster, writers were more tentative than when simplydescribing themselves. Hedging is typical of face-saving discourse strat-egies writers sometimes use to establish or maintain a potentially ten-uous written interaction (Scollon & Scollon, 1995, see especially p. 88).Finally, when responding to the specific ad poster, writing style wasmarked by relative syntactic complexity. This is consistent with earlierfindings that mature writers reduce complexity when writing to a gen-eralized other with whom no prior interpersonal bond exists (Rubin,1982).Also,descriptive tasks, including apparently the self-descriptionelicited in this study, tend to engender relatively simple syntax


compared with discourse engaged to influence an other (Crowhurst &Piché, 1979). That is because purely descriptive prose requires fewhypotactic constructions indicating causality or condition.

AUDIENCE EFFECTS

The third and most central research question concerned the effectsof the gender role bid (instrumental or expressive) issued by the hypo-thetical target audience’s personal ads. No main effects for this factoremerged, but we were especially interested in any interactions thatmay have emerged between writing task and gender role bid of theadvertisement. It is within those interactions that one would find themost direct evidence of communication adaptation.

Consider, for example a writer whose “baseline” style as evinced inthe description written to an indeterminate audience is gender typedas feminine (e.g., many markers of excitability and nonessentials). Ifthis individual is asked to respond to an instrumental gender role per-sonal ad, he or she may adapt by reducing the frequency of these lan-guage features, in which case we would conclude that adaptation byconvergence had taken place (Giles et al., 1991). Alternatively, thiswriter might have actually increased his or her level of excitability andnonessential markers; this would constitute a form of divergence.

Linguistic divergence, according to CAT, often serves as a strategyfor conveying a hostile or alienated posture vis-à-vis one’s interlocutor,in other words, for increasing psychosocial distance. On the otherhand, CAT also acknowledges that divergent linguistic style can alsobe a means of expressing affiliative affect, as when one member of adyad complements the seemingly opposite language pattern of his orher partner (Burgoon et al., 1995). Thus, for example, when one part-ner responds to his or her mate’s dominance display by expressingsubmissiveness, that divergence is surely a form of complementarycommunication adaptation. By the same token, when men and womenenacting heterosexual courtship interactions, as in the elicitation pro-cedure for this study, diverge in their use of gender-typed language fea-tures, it seems most reasonable to interpret such patterns as instancesof complementary language use.

Two language variables in these data did evince just such a comple-mentary pattern. First, there was indeed a statistically significantinteraction between writing task and gender role bid of the personal adfor markers of excitability. In responding to the instrumental genderrole bid (traditional masculine values), participants increased theiruse of this female-typed feature. The instrumental ad thus appeared toelicit a complementary adaptation from writers (an increase in excit-ability markers relative to baseline self-descriptions). In a parallelfashion, the interaction for first-person pronouns showed similar evi-dence of complementary adaptation. In the face of an expressive gender


role bid (traditionally feminine), writers reduced their output of thesefemale-typed markers from their baseline output.

Thus, this study provides at least some evidence of a complementarytype of communication adaptation in the context of responding to per-sonal ads. If the person who places the ad expresses an instrumentalgender role identity himself or herself, the linguistic response is toemphasize the responder’s expressive gender role identity by increas-ing the output of subjectivity markers. Conversely, if the ad appears toexpress an instrumental gender role identity, the responder may useexcitability markers to emphasize his or her expressive traits. It mustbe noted, of course, that this linguistic complementary patternemerged for only two of the six gender-typed language features. Fur-thermore, it must be acknowledged that this interaction between writ-ing task and gender bid from which we infer this complementary typeof adaptation is only one of several significant sources of variation forthese language variables.

Limitations and Directions for Future Research

Several methodological limitations to the current study must alsobe noted. First, although the stimulus personal ads were fabricated byusing real models, they expressed very uniformly expressive andinstrumental traits. This arrangement was necessary for the potencyof the experimental manipulation. In most naturalistic personal ads,however, gender role traits of one type are mitigated by at least sometraits from the alternative gender role position. Thus, the resultinggender role bids may have been more caricatured and possibly lessattractive than those appearing in natural texts. Second, the extent towhich participants were invested in the experimental task remainsuncertain. No doubt, the intensity of their language adaptations dif-fers in genuine interpersonal contexts compared with a laboratorywriting context. Nonetheless, there is no reason to expect that thedirection of these language shifts would differ were we to sample froma corpus of actual responses to personal ads.

The present study was also limited to an examination of written lan-guage production. Future research, as suggested by Rubin and Greene(1992), may profitably explore the perceptions of readers regardingthese language features. Also, it would be interesting to connect writ-ers’ specific relational goals (e.g., identity protection vs. affinity seeking)with their language use in more realistic contexts in which relationalconsequences might be observed. In other words, what are the rela-tional sequelae of shaping one’s presentation of self to fit the audience?Is such adaptation (whether divergent or convergent) more or lesslikely to attract another, as predicted by CAT?

Although the language features investigated in the current studyhave been associated with particular writer traits (i.e., they are gender


typed), it is unknown to what extent participants themselves feel thatthey are altering such features to their language use. An experiment inwhich writers were cued to deliberately express greater tentativeness,excitability, or connectedness, for example, would shed further light onthis issue. On the other hand, in cases in which language users areunaware of their linguistic adaptations, one cannot necessarily con-clude that they lack associated interpersonal goals. Much communica-tion adaptation in writing as well as in speech may be subliminal(Rubin, 1998).

Future studies might also examine more macroscopic discoursevariables to examine similar questions of adaptation in gender iden-tity by considering, for example, the strategic content of interpersonalmessages. One might ask, for example, whether the content of a per-sonal ad narrative evinces adaptation to audience, and if so, do adapta-tions in message content merely mirror linguistic adaptation or func-tion differently? Thus, a comparison of gender-related content withgender-linked language would further understanding of the issue ofgender as it relates to self-presentation within the heterosexual datingcontext.

In sum, we conclude that contextual influences need to be accountedfor when analyzing gender and language use in written discourse. Thefindings of the present study, though modest, do offer some concreteillustration of ways in which people enact varying gender identities inthe face of an interlocutor’s bid for one or another gender role. Theseidentities are enacted by modulating the use of socially meaningfullanguage features, for example, in a manner that complements theinterlocutor’s apparent gender identity.

REFERENCES

Ahuvia,A.C.,& Adelman,M.B. (1992).Formal intermediaries in the marriage market:Atypology and review. Journal of Marriage and the Family, 54, 452-463.

Aries, E. (1996). Men and women in interaction: Reconsidering the differences. New York:Oxford University Press.

Ayres, J. (1992). Personal ads: An exploratory investigation into the relationships amongad characteristics, communication apprehension and contact. CommunicationReports, 5, 67-72.

Balester, V. M. (1993). Cultural divide: A study of African-American college-level writers.Portsmouth, NH: Boynton/Cook.

Bell, R. A., & Daly, J. A. (1984). The affinity-seeking function of communication. Commu-nication Monographs, 51, 91-113.

Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consultingand Clinical Psychology, 42, 155-162.

Bem, S. L. (1981). Gender role schema: A cognitive account of sex typing. PsychologicalReview, 88, 354-364.

Bem, S. L. (1984). Androgyny and gender schema theory: A conceptual and empiricalintegration. Nebraska Symposium on Motivation, 32, 179-226.


Bem, S. L. (1993). The lenses of gender: Transforming the debate on sexual inequality.New Haven, CT: Yale University Press.

Blom, J., & Gumperz, J. J. (1972). Social meaning and linguistic structure: Code-switchingin Norway. In J. J. Gumperz & D. Hymes (Eds.), New directions in sociolinguistics(pp. 407-434). New York: Holt, Rinehart & Winston.

Bohan, J. S. (1993). Regarding gender: Essentialism, constructionism and feminist psy-chology. Psychology of Women Quarterly, 17, 5-21.

Bolig, R., Stein, P. J., & McKenry, P. C. (1984). The self-advertisement approach to dating:Male-female differences. Family Relations, 33, 587-592.

Burgoon, J. K., Stern, L. A., & Dillman, L. (1995). Interpersonal adaptation: Dyadic inter-action patterns. New York: Cambridge University Press.

Chafe, W. (1986). Writing in the perspective of speaking. In C. Cooper & S. Greenbaum(Eds.),Studying writing:Linguistic approaches (pp.12-39).Beverly Hills,CA:Sage.

Collier,M.J.,& Thomas,M. (1988).Cultural identity:An interpretive perspective. In Y.Y.Kim & W. B. Gudykunst (Eds.), Theories in intercultural communication (pp. 99-122).Newbury Park, CA: Sage.

Coupland, J. (1996). Dating advertisements: Discourses of the commodified self. Dis-course and Society, 7, 187-207.

Crawford, M. (1995). Talking difference: On gender and language. London: Sage Ltd.Crowhurst, M., & Piché, G. L. (1979). Audience and mode of discourse effects on syntactic

complexity at two grade levels. Research in the Teaching of English, 13, 101-109.Deaux, K., & Hanna, R. (1984). Courtship in the personals column: The influence of gen-

der and sexual orientation. Sex Roles, 11, 363-375.Douglas,W. (1987).Affinity-testing in initial interactions.Journal of Social and Personal

Relationships, 4, 3-15.Edelsky,C. (1976).The acquisition of communication competence:Recognition of linguis-

tic correlates of sex roles. Merrill-Palmer Quarterly of Behavior and Development, 22,47-59.

Freimuth, M. J., & Hornstein, G. A. (1982). A critical examination of the concept of gen-der. Sex Roles, 8, 515-532.

Giles, H., Coupland, N., & Coupland, J. (1991). Accommodation theory: Communication,context, and consequence. In H. Giles, J. Coupland, & N. Coupland (Eds.), Contexts ofaccomodation: Developments in applied sociolinguistics (pp. 1-68). Cambridge, UK:Cambridge University Press.

Greenlees, I. A., & McGrew, W. C. (1994). Sex and age differences in preferences and tac-tics of mate attraction: Analysis of published advertisements. Ethology andSociobiology, 15, 59-72.

Grob, L. M., Meyers, R. A., & Schuh, R. (1997). Powerful/powerless language use in groupinteractions:Sex differences or similarities? Communication Quarterly,45, 282-303.

Harrison, A. A., & Saeed, L. (1977). Let’s make a deal: An analysis of revelations and stip-ulations in lonely hearts advertisements. Journal of Personality and Social Psychol-ogy, 35, 257-264.

Hiatt, M. (1977). The way women write. New York: Teachers College Press.Hogg, M. A. (1985). Masculine and feminine speech in dyads and groups: A study of

speech style and gender salience. Journal of Language and Social Psychology, 4,99-112.

Hogg, M. A., & Rigoli, N. (1996). Effects of ethnolinguistic vitality, ethnic identification,and linguistic contacts on minority language use. Journal of Language and SocialPsychology, 15, 76-89.

Holmes, J. (1986). Functions of you know in women’s and men’s speech. Language inSociety, 15(1), 1-21.

House, A., Dallinger, J. M., & Kilgallen, D. (1998). Androgyny and rhetorical sensitivity:The connection of gender and communicator style. Communication Reports, 11,11-21.


Huberty, C. V., & Morris, J. D. (1989). Multivariate analysis versus multiple univariateanalyses. Psychological Bulletin, 105, 302-308.

Jason, L. A., Moritsugu, J. N., & DePalma, D. M. (1982). Advertising as a strategy formeeting people. Psychological Reports, 71, 1311-1314.

Johnson, D., & Roen, D. (1992). Complimenting and involvement in peer reviews. Lan-guage in Society, 21, 27-57.

Koestner, R., & Wheeler, L. (1988). Self-presentation in personal advertisements: Theinfluence of implicit notions of attraction and role expectations. Journal of Social andPersonal Relationships, 5, 149-160.

Kramarae,C. (1990).Changing the complexion of gender in language research.In H.Giles &W. P. Robinson (Eds.), Handbook of language and social psychology (pp. 345-360).New York: John Wiley.

Kramer, C., Thorne, B., & Henley, N. (1978). Review essay: Perspectives on language andcommunication. Signs: A Journal of Women in Culture and Society, 3, 638-651.

Lakoff, R. (1975). Language and women’s place. New York: Harper & Row.Meyerhoff, M. (1998). Accommodating your data: The use and misuse of accommodation

theory in sociolinguistics. Language & Communication, 18, 205-225.Montini, T., & Ovrebro, B. (1990). Personal relationship ads: An informal balancing act.

Psychological Perspectives, 33, 327-339.Mulac, A., Lundell, T. L., & Bradac, J. J. (1986). Male/female language differences and

attributional consequences in a public speaking situation: Toward an explanation ofthe gender-linked language effect. Communication Monographs, 53, 115-129.

Nichols, P. C. (1983). Linguistic options and choices for Black women in the rural South.In B. Thorne, C. Kramarae, & N. Henley (Eds.), Language, gender and society (pp. 54-68).Rowley, MA: Newbury House.

O’Barr,W.,& Atkins,B.K.(1980). “Women’s language”or “powerless language.”In S.McConnell-Ginet, R. Borker, & N. Furman (Eds.), Women and language in literature and society(pp. 93-110). New York: Praeger.

Parrott, R., Lemieux, R., Travillion, L., & Harris, T. (1997). Interfacing interpersonal andmediated communication: Strategic self-disclosure in personal ads. Southern Com-munication Journal, 64(2), 318-331.

Redd, T. M. (1995). Untapped resources: “Styling” in Black students’ writing. In D. Rubin(Ed.), Composing social identity in written language (pp. 221-240). Hillsdale, NJ:Lawrence Erlbaum.

Roulis, E. (1995). Gendered voice in composing, gendered voice in evaluating: Genderand the assessment of writing quality. In D. Rubin (Ed.), Composing social identity inwritten language (pp. 151-186). Hillsdale, NJ: Lawrence Erlbaum.

Rubin, D. L. (1982). Adapting syntax in writing to varying audiences as a function of ageand social cognitive ability. Journal of Child Language, 9, 497-510.

Rubin, D. L. (1995). Introduction: Composing social identity. In D. Rubin (Ed.), Com-posing social identity in written language (pp. 1-30). Hillsdale, NJ: LawrenceErlbaum.

Rubin,D.L. (1998).Writing for readers:The primacy of audience in composing. In N.Nel-son & R. C. Calfee (Eds.), The reading-writing connection (97th yearbook of theNational Society for the Study of Education,part 2) (pp.53-73).Chicago:University ofChicago Press.

Rubin, D. L., & Greene, K. (1992). Gender-typical style in written language. Research inthe Teaching of English, 26, 7-40.

Rubin, D. L., & Greene, K. (1995). The suppressed voice hypothesis in women’s writing:Effects of revision of gender-typical style. In D.Rubin (Ed.), Composing social identityin written language (pp. 133-150). Hillsdale, NJ: Lawrence Erlbaum.

Rubin, D. L., & Nelson, M. W. (1983). Multiple determinants of stigmatized speech style:Women’s language, powerless language, or everyone’s language? Language andSpeech, 26, 273-290.


Sachdev, I., & Bourhis, R. Y. (1990). Language and social identification. In D. Abrams &M. A. Hogg (Eds.), Social identity theory (pp. 211-229). New York: Springer-Verlag.

Scollon, R., & Scollon, S. W. (1995). Intercultural communication: A discourse approach.Oxford, UK: Basil Blackwell.

Sitton, S., & Rippee, E. T. (1986). Women still want marriage: Sex differences in lonelyhearts advertisements. Psychological Reports, 58, 257-258.

Sterkel, K. (1988). The relationship between gender and writing style in business com-munication. Journal of Business Communication, 25, 17-38.

Tannen, D. (1982). The oral/literate continuum in discourse. In D. Tannen (Ed.), Spokenand written language:Exploring orality and literacy (pp.1-16).Norwood,NJ:Ablex.

Thiessen, D., Young, R. K., & Burroughs, R. (1993). Lonely hearts advertisements reflectsexually dimorphic mating strategies. Ethology and Sociobiology, 14, 209-229.

Waters, M. C. (1990). Ethnic options: Choosing identities in America. Berkeley: Univer-sity of California Press.

West, C., & Zimmerman, D. (1987). Doing gender. Gender and Society, 1, 125-151.Wheeless, V. E., & Dierks-Stewart, K. (1981). The psychometric properties of the Bem

Sex-Role Inventory. Communication Quarterly, 29, 173-186.Willis, F. N., & Carlson, R. A. (1993). Singles ads: Gender, social class and time. Sex Roles,

29, 387-404.Yaeger-Dror, M. (1998). Factors influencing the contrast between men’s and women’s

speech. Women and Language, 21(1), 40-45.Zeni, J., & Thomas, J. K. (1990). Suburban African American basic writers: A text analy-

sis. Journal of Basic Writing, 9(2), 15-39.


JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001Beaumont et al. / SIMILARITIES AND DIFFERENCES

SIMILARITIES AND DIFFERENCESIN MOTHER-DAUGHTER AND

MOTHER-SON CONVERSATIONSDURING PREADOLESCENCE

AND ADOLESCENCE

SHERRY L. BEAUMONTVIVIANE C. B. VASCONCELOS

MONICA RUGGERIUniversity of Northern British Columbia

Temporal and functional conversation styles were examined in discussions betweenmothers and their preadolescent or adolescent sons and daughters. Conversations wereaudiotaped and coded for speakers’ rates of overlaps between speaking turns, simulta-neous speech, and successful interruptions. Results indicated that boys and girls of bothages used a high-involvement conversational style including frequent overlaps, simulta-neous speech, and interruptions. Mothers used a high-considerateness style characterizedby significantly lower rates of overlaps, simultaneous speech,and interruptions than theirchildren. Secondary analyses examining the functions of speakers’ simultaneous speechand successful interruptions indicated that adolescents produced less confirming simul-taneous speech than preadolescents, and more confirming simultaneous speech was pro-duced in the preadolescent boy dyads than in either the adolescent boy dyads or thepreadolescent girl dyads. Finally, adolescent boys produced significantly more rejectingsuccessful interruptions than their mothers.

Over the past 30 years, a large body of research has been focused ondetermining whether parents and children experience an increase incommunication difficulties during the period of adolescence. Althoughthere is still some controversy over whether parent-adolescent verbalconflict is inevitable, it has been firmly established that conflict

419

AUTHORS’ NOTE: This study was supported by a research grant awarded to the first au-thor from the Social Sciences and Humanities Research Council of Canada. Portions ofthis article were presented at the 1999 Biennial Meeting of the Society for Research inChild Development. Appreciation is extended to the families who participated in thestudy and to Paula Beaumont, Jacqui Boonstra, Jeneen Clark, Laura McMullen, KristenMix, Greg Pope, Liz Rocha, Heather Sapergia, and Michelle Smith for their assistance invarious phases of this research. Correspondence concerning this article should be ad-dressed to Sherry L.Beaumont,Department of Psychology,University of Northern BritishColumbia,3333 University Way,Prince George,BC,Canada V2L 4Z9.E-mail may be sentto [email protected].


between parents and adolescents is common in many families (forrecent reviews, see Arnett, 1999; Laursen, Coy, & Collins, 1998), andparent-adolescent conflict has been implicated in many negative out-comes for adolescents (e.g., low self-esteem, depression, drug abuse,delinquency) (Montemayor, 1983) as well as parents’ experiences ofmidlife stress (Julian, McKenry, & Arnold, 1990).

It has been suggested that parent-adolescent discord is related toqualitative differences in parents’ and adolescents’ communicationstyles (i.e., different patterns of speech behaviors) (Robin & Foster,1984). Researchers who have examined the quality of parent-adolescentcommunication have relied primarily on the use of self-report mea-sures. Although self-reports provide insights into adolescents’ andtheir parents’ perceptions of their interactions, they do not providemore objective information about overt communication behaviors thatmight shed light on the question of whether parents and adolescentsdisplay different communication styles. Nevertheless, relative to thepreponderance of self-report studies measuring parent-adolescentperceptions, relatively few researchers have investigated the quality ofparent-adolescent communication using observational methods.Those who have used observational methods have focused primarily onchanges in the content of parent-child communication from preadoles-cence through middle adolescence (e.g., Montemayor, Eberly, &Flannery, 1993; Powers, Hauser, Schwartz, Noam, & Jacobson, 1983)rather than comparing possible differences in the structural or stylis-tic features of parents’ and adolescents’ communication styles (e.g.,turn-taking habits). It is conceivable that parent-adolescent verbalconflict might be influenced or perhaps exacerbated by differences inparents’ and adolescents’ structural use of language (i.e., conversa-tional styles). The goal of the present study was to extend previousresearch on parent-adolescent communication by systematicallyexamining similarities and differences in the conversational styles ofmothers and their preadolescent or adolescent sons and daughters.

Many previous studies have examined individual or group differ-ences in communication styles as a function of such factors as gender,age, and interpersonal familiarity. Those studies have focused primar-ily on identifying differences between speakers in the use of certaincommunicative behaviors and determining the intentions of thosespeakers when using those behaviors. For example, many researchershave argued that speakers who interrupt their partners frequently doso in an attempt to dominate conversations (e.g., Zimmerman & West,1975). Mishler and Waxler (1968) suggested that this pattern of behav-ior is evident in middle-class, Western families. That is, they claimedthat parents tend to interrupt their children more than children inter-rupt their parents, reflecting the parents’ greater power in the family.However, when children become adolescents, the balance of power inthe family is shifted, and normal interaction patterns are disrupted


(Mishler & Waxler,1968). In support of that hypothesis, several studieshave found that adolescents tend to interrupt their mothers more oftenthan when they were younger and more often than their mothers inter-rupt them (this is not the case for fathers, who tend to interrupt theirchildren more regardless of the children’s ages; e.g., Hill, 1988; Papini,Datan, & McCluskey-Fawcett, 1988; Steinberg, 1981). Those research-ers have interpreted this result as evidence of adolescents’ attempts todominate their mothers.

The interpretation of interruptions as attempts to dominate isfound not only in research on family interaction but also in the litera-ture addressing the issue of gender differences in communicationstyles. Several studies on gender differences in communication haveshown that men tend to interrupt women more than women interruptmen. For example, in an often-cited example of this type of research,Zimmerman and West (1975) recorded cross-gender conversationsbetween previously unacquainted adults and found that the men inter-rupted 3 times more than the women. However, a more recent review ofinterruption studies by James and Clarke (1993) showed that out of 21studies, only 6 conclusively found that men interrupted at a higherrate than women (the other studies found no differences between menand women in the rate of interruptions or that women interruptedmore than men). Many of those past studies on interruption behaviorwere criticized by James and Clarke on methodological grounds.Some ofthe researchers collected their conversation samples in uncontrolled,naturalistic environments in which factors such as the topic of conver-sation or the familiarity of partners could influence rates of interrup-tions. In contrast, conversations recorded in controlled laboratoriescan lack ecological validity. In addition to problems with setting, inves-tigators have defined interruptions in many different ways, and ratesof interruptions are typically measured in the absence of other relatedconversational behaviors (James & Clarke, 1993).

In addition to methodological concerns, sociolinguists have criti-cized the dominance hypothesis for the meaning of interruptions ontheoretical or interpretive grounds (e.g., Beattie, 1981; Scollon, 1985;Tannen, 1983, 1984, 1989). Specifically, they have claimed that focus-ing on what one speaker is doing to another speaker dismisses thedyadic nature of interruptions. In addition, they have tended to focuson the fact that interruptions may occur in conversations simplybecause the speakers use different conversational styles. Gumperz(1976) and Tannen (1983, 1984, 1989) have argued that individualshave particular conversational styles that include variations in pacing,pausing, and the use of simultaneous speech (i.e., talking at the sametime as other speakers), and interruptions take place when speakershave different habits with regard to these behaviors. For example,Tannen (1983, 1984) identified two broad categories of conversationalstyle, which she called the high-involvement and high-considerateness

Beaumont et al. / SIMILARITIES AND DIFFERENCES 421

styles. The high-involvement style is characterized by a faster rate ofspeech, faster turn taking, an avoidance of interturn pauses, and fre-quent initiations of simultaneous speech. The high-consideratenessstyle, on the other hand, consists of slower speech, slower turn taking,longer pauses between turns, and an avoidance of simultaneousspeech. According to Tannen, high-involvement speakers use simulta-neous speech to build rapport and signal involvement, whereashigh-considerateness speakers avoid simultaneous speech to honorthe principle not to impose. In other words, it is the intentions of thespeaker to signal considerateness or involvement that give rise to thelinguistic devices, either silence or simultaneous speech, that are usedto signal how the utterance is meant (Tannen, 1984).

Tannen (1983, 1984) and others (e.g., Gumperz, 1976; Scollon, 1985)have claimed that the most successful conversations occur betweenspeakers with the same style (whichever it might be) because theyshare expectations for the pace of turn taking and the use of simulta-neous speech. On the other hand, when speakers with opposing stylesinteract, the person who uses faster turn-taking habits and moresimultaneous speech will always interrupt his or her partner fre-quently. For example, high-involvement speakers tend to interrupt high-considerateness speakers. That is, whenever a high-consideratenessspeaker pauses within his or her turn, a high-involvement speaker willstart talking because he or she believes that silence signals a lack ofrapport, and because the high-considerateness speaker believes thatoverlapping speech is imposing, he or she will stop talking. Therefore,whether or not the first speaker (the high-considerateness speaker inthis example) stops talking is what determines whether the initiationof simultaneous speech becomes a successful interruption. “If the firstspeaker does not stop, there is no interruption” (Tannen, 1989, p. 270).In other words, conversations that include many one-sided interrup-tions may be due to a “clash” in conversational styles as opposed todominance attempts (Tannen, 1983, 1989). High rates of mutual inter-ruptions might simply indicate that both partners are using a similarhigh-involvement style that produces a good deal of simultaneousspeech (when the first speaker does not stop talking) and successfulinterruptions (when the first speaker does stop talking) and does notnecessarily reflect a power struggle.

This alternative explanation for the meaning of interruptions wasinvestigated in a series of studies (Beaumont, 1995, 2000; Beaumont &Cheyne, 1998) that examined the possibility that adolescents inter-rupt their mothers unintentionally simply because they have devel-oped different habits or styles for pacing and pausing and the use ofinterruptions and simultaneous speech. Beaumont (1995) examinedvariations in the conversations of adolescent girls and their mothersand friends across two age groups (11 to 12 and 15 to 16). The resultsindicated that girls of both ages and their friends exhibited a high-


involvement conversational style that included a fast rate of turn tak-ing (overlaps between turns), frequent interruptions, and simulta-neous speech, and the use of this style had a tendency to increase fromearly to middle adolescence. Furthermore, girls were found to use thesame style in conversations with their mothers. In contrast, mothersexhibited a high-considerateness style with fewer overlaps, interrup-tions, and simultaneous speech. Therefore, mothers and daughtersdemonstrated a conflict in styles, with daughters interrupting morethan their mothers.

In a follow-up study, Beaumont (2000) proposed that becausewomen tend to use a fast-paced style when talking with their friends(e.g., Coates, 1989), it is likely that the more controlled, slow-pacedstyle that was exhibited by mothers interacting with their daughters isa style that mothers reserve particularly for conversations with theirchildren. That hypothesis was investigated by Beaumont by comparingthe conversational styles that mothers used with their preadolescentand middle-adolescent daughters with those used in conversationswith their female friends. The results indicated that in conversationswith their daughters, mothers exhibited a high-considerateness con-versational style with few interruptions and little simultaneousspeech, but in conversations with friends, mothers used a fast-paced,high-involvement conversational style with frequent interruptionsand a lot of simultaneous speech. Furthermore, there was little differ-ence in the styles used by mothers of preadolescents and mothers ofmiddle adolescents. Beaumont claimed that those results confirmedthat mothers tend to use a high-considerateness conversational style,even when their daughters reach adolescence, and as a result, mothersand daughters experience a clash in conversational styles.

In contrast to researchers who have concluded that adolescents’interruptions of their mothers reflect dominance attempts (e.g., Hill,1988; Steinberg, 1981), previous research by Beaumont (1995, 2000)suggests that adolescent girls may use frequent interruptions simplyas part of a high-involvement conversational style, a style that con-flicts with the style used by their mothers. Adolescent girls’ use of ahigh-involvement conversational style might develop as a result of thedemands associated with increased participation in intimate conver-sations with their peers. That is, previous researchers have found notonly that adolescents spend more time talking with their peers than doyounger children (e.g., Csikszentmihalyi & Larson, 1984) but also thatadolescent friendship conversations are characterized primarily byintimate self-disclosure and self-affirmations, and this is particularlytrue of girls’ conversations (e.g., Mettetal, 1983; Youniss & Smollar,1985).

Therefore, it may be that adolescent girls learn to use a conversa-tional style that includes frequent overlapping and simultaneousspeech (some of which might be interruptive, depending on whether


their partners stop talking) for the purpose of showing involvementand support for what their partners are saying (i.e., a high-involvementstyle) (Tannen, 1984). In fact, many researchers have suggested thatinterruptive behavior is not intended to usurp the speaking rights ofthe current speaker but is supportive or cooperative in nature (e.g.,Bennett, 1981; Kennedy & Camden, 1983; Tannen, 1983, 1984, 1989).In support of the hypothesis that frequent interruptions can be part ofa particular conversational style, investigators have found that inter-ruptions rarely occur in isolation. For example, research by Feldsteinand Welkowitz (1987) demonstrates that interruptions are part of aparticular conversational style that includes other nonverbal behav-iors. That is, in conversations between adult female peers, the fre-quency of interruptions, as well as instances of noninterruptive simul-taneous speech, was related to the duration of pauses within andbetween turns. In other words, women who tended to interrupt alsoused a fast rate of speech with short pauses between turns and a lot ofnoninterruptive simultaneous speech. This style has been found to beexhibited most often by women who perceive themselves to be outgo-ing, talkative, adventurous, and socially bold (Feldstein, Alberti, &BenDebba, 1979), terms that are all synonymous with the concept ofdominance. Nevertheless, although the use of this style of speechmight be interpreted as reflecting dominance, it does not appear to berelated to personality measures of dominance (e.g., Dindia, 1987;French & Local, 1983; Murray, 1987).

The preceding review suggests at least three theoretical argumentsfor interpreting the meaning of interruptions in dyadic conversations,two of which provide arguments for the functions of interruptions.Interruptions may serve either (a) dominance or (b) affiliative func-tions, or they may (c) simply reflect conflicting conversational styles.Several authors (e.g., James & Clarke, 1993; Walker, 1991) have notedthat it is not an easy matter to determine the functions of interruptionsby examining simple frequencies of interruptions. It is possible thatinterruptions reflect different functions in different contexts. For exam-ple, interruptions in conversations with friends where interpersonalpower is more equal may reflect “engagement,” whereas interruptionsin conversations with mothers, where power is not distributed equally,may indeed reflect dominance attempts (Beattie, 1981; Greenwood,1989). To investigate that possibility, Beaumont and Cheyne (1998)examined the functions of adolescent girls’ interruptions and simulta-neous speech in conversations with their mothers or friends. Theresults indicated that adolescent girls’ interruptions of mothers andfriends were found to serve the same functions. That is, adolescentgirls used confirming, disconfirming, and rejecting interruptionsequally with mothers and friends. Those findings indicate that inter-ruptions in adolescents’ conversations serve both agreement and dis-agreement functions. Therefore, Beaumont and Cheyne concluded that


those results are not consistent with the notion that interruptionsreflect only dominance attempts but are more consistent with thehypothesis that the use of frequent interruptions likely reflects the useof a particular temporal conversation style.

The purpose of the present study was to extend previous research(Beaumont, 1995, 2000; Beaumont & Cheyne, 1998) by examining dif-ferences in the temporal and functional conversation styles of mothersand their children as a function of age and gender. The use of successfulinterruptions and overlapping and simultaneous speech was exam-ined in mother-son and mother-daughter conversations in two agegroups (preadolescence and adolescence). It was hypothesized that theresults would parallel those of Beaumont (1995,2000),which indicatedthat both preadolescents and adolescents use more overlaps, simulta-neous speech, and successful interruptions than their mothers, withthe tendency for more high-involvement features to be produced inadolescent versus preadolescent dyads. That is, it was expected thatboth preadolescent and adolescent girls and boys would use higherrates of interruptions and overlapping and simultaneous speech thantheir mothers, with more high-involvement features in the adolescentthan the preadolescent conversations but with little differencebetween the styles of boys and girls.

This latter hypothesis regarding no differences in the conversa-tional behaviors of boys and girls can be supported by previous find-ings. Other researchers have found more similarities than differencesbetween boys’ and girls’ temporal communication behaviors with par-ents and peers. For example, Greenwood (1989) and Marche and Peter-son (1993) found that both boys and girls interrupted their peers atsimilar high rates, with slightly more interruptions used by adoles-cents compared with preadolescents, and Walker (1991) found a simi-lar result for interruptions of parents. Nevertheless, given thatprevious researchers also have found that adolescent sons tend to bemore withdrawn than daughters in conversations with their parents(Kahlbaugh & Haviland, 1994; Whalen, Henker, Hollingshead, & Bur-gess, 1996), it was expected that mothers would use a more high-considerateness style with their sons than with their daughters in anattempt to provide their sons with a nonimposing conversational set-ting. This hypothesis is supported by a recent meta-analyses on gendereffects on parents’ talk to their children, which showed that mothersuse fewer directives in conversations with their sons than in conversa-tions with their daughters (Leaper, Anderson, & Sanders, 1998). Ifmothers of sons do use more high-considerateness features (i.e., fewerinterruptions and less overlapping and simultaneous speech) thanmothers of daughters, then a greater difference would be seen in thestyles of mothers and sons than the styles of mothers and daughters.This latter hypothesis can be supported by other research, which hasshown, for example, that there are greater differences between


mothers and sons than between mothers and daughters in the rates ofsupportive, directive, and emotional communications (e.g., Whalen et al.,1996).

Secondary analyses also were conducted to determine the functionsof mothers’ and their children’s simultaneous speech and successfulinterruptions. Using a scheme developed by Kennedy and Camden(1983), all successful interruptions and instances of simultaneousspeech (i.e., unsuccessful interruptions) were coded for five possiblefunctions: clarification, agreement, disagreement, tangentialization(e.g., mocking), and subject change. These five subtypes of interrup-tions were designed to tap three theoretical constructs introduced byWatzlawick, Beavin, and Jackson (1967). Those theorists claimed thateach communicated message reveals some aspect of a speaker’sself-definition. As a result, any response to a previously communicatedmessage can serve to confirm, reject, or disconfirm that person’s (A’s)definition of self. Specifically,

confirmations are responses which express approval, understanding oracceptance of A’s position. Rejections are responses from B which some-how indicate disagreement or disapproval of A. No matter how painful,rejection at least recognizes what is being rejected and, therefore, doesnot necessarily deny the reality of A’s view of self. In contrast,disconfirmations are the most dysfunctional types of communicationbecause they negate the reality of A as a source of self-definition. (Ken-nedy & Camden, 1983, p. 48)

Kennedy and Camden (1983) argued that clarification and agree-ment interruptions reflect confirmation, tangentialization and subjectchanges represent disconfirmation, and disagreement indicates rejec-tion.Those authors further suggested that if interruptions reflect dom-inance attempts on the part of the speaker, then one would expect tosee most interruptions serving rejection or disconfirmation functions.The results of their research on cross-gender communication do notsupport that explanation. Specifically, Kennedy and Camden foundthat both men and women used more confirmation than disconfirma-tion and rejection interruptions. For the present study, this secondarycoding will provide information about similarities or differences in thefunctions of simultaneous speech and successful interruptions used bymothers and their children. Following the results of Beaumont andCheyne (1998), it was hypothesized that both boys and girls would usesimilar rates of all three speech act functions (confirming, rejecting,and disconfirming). However, following from previously cited researchon differences in mothers’ speech behavior with their sons versus theirdaughters, it was also hypothesized that mothers of sons would useless rejecting and disconfirming and more confirming simultaneousspeech and successful interruptions than mothers of daughters.


In summary, the conversations of mothers and their preadolescentor adolescent sons and daughters were audiotaped and coded for fea-tures of temporal and functional conversation styles.Speakers’ rates ofsuccessful interruptions, simultaneous speech, and overlaps betweenspeaking turns were analyzed to find evidence of possible similaritiesor differences in mothers’ and sons’ or daughters’ conversationalstyles. It was hypothesized that both preadolescents and adolescents(both boys and girls) would use more overlaps, simultaneous speech,and successful interruptions than their mothers, with more frequentsimultaneous speech and successful interruptions produced in adoles-cent versus preadolescent dyads. For analyses regarding the functionsof interruptions, each previously identified successful interruption orinstance of simultaneous speech was coded for one of five functions(clarification, agreement, disagreement, tangentialization, and sub-ject change). Analyses were subsequently conducted on composite datareflecting the three concepts of confirmation (clarification + agree-ment), disconfirmation (tangentialization + subject change), and rejec-tion (disagreement). It was hypothesized that both boys and girlswould use similar rates of all three speech act functions but that moth-ers of sons would use less rejecting and disconfirming and more con-firming simultaneous speech and successful interruptions than mothersof daughters.

METHOD

PARTICIPANTS

The participants included 24 Caucasian boys and 24 Caucasiangirls and their mothers living in Prince George,British Columbia,Can-ada. The participants were informed of the study through visits by aresearcher to school classrooms or through advertisements in localbusinesses. Following the classroom visits, students were either askedto take an information letter home to their mothers or were mailed aninformation letter. Only those students who received parent permis-sion and who themselves agreed to participate were included in thestudy. All participants gave oral consent to participate, and mothersgave written consent for themselves and their sons or daughters toparticipate. The rate of consent for participation was somewhat low atapproximately 11%. This low consent rate was likely due to the factthat recruitment took place in a small city (population 80,000) with anew university, and requests for participation in research were quitenovel.

Half of the children who participated were in Grades 4 and 5(preadolescents; age range = 8.25 to 10.83 years, M = 9.81 years; n = 12


girls and 12 boys), and the other half of the children were in Grades 8through 11 (adolescents; age range = 13.00 to 16.83 years, M = 15.72years; n = 12 girls and 12 boys). Mothers of preadolescents ranged inage from 30 to 45 years (M = 36.71 years), and mothers of adolescentsranged in age from 29 to 50 years (M = 40.57 years).There was a signifi-cant difference in the mean ages of the two groups of mothers, t(46) =–3.12, p = .003, with mothers of adolescents being significantly older onaverage than mothers of preadolescents.

Demographic information indicated that the majority of the chil-dren lived with both of their biological parents (8 preadolescent boys,7 preadolescent girls, 8 adolescent boys, and 10 adolescent girls). Onepreadolescent boy and girl and 2 adolescent boys and girls lived alonewith their mothers, whereas 3 preadolescent girls and boys and 1 ado-lescent boy lived with their biological mothers and stepfathers. In addi-tion,1 preadolescent girl and 1 adolescent boy lived with their adoptiveparents. Boys and girls from both age groups also were comparable interms of birth order: Four preadolescent boys, 3 preadolescent girls,5 adolescent boys, and 2 adolescent girls were the oldest children intheir families; 1 preadolescent boy, 3 preadolescent girls, 1 adolescentboy, and 3 adolescent girls were middle children; and 6 boys and 6 girlsin both age groups were youngest children.Finally,1 preadolescent boyand 1 adolescent girl were only children.

On the basis of information about women’s and their spouses’ occu-pations (where applicable), 77% of the participating families were mid-dle class, 16.7% were lower class, and 6.3% were unemployed. Thisratio was similar across both age and gender groups (according to theindex developed by Blishen, Carroll, & Moore, 1987). Therefore, thecultural and socioeconomic profile of this sample could be described aspredominantly middle class and White (European background), whichis characteristic of the majority of school-age children attending publicschools in Prince George (according to 1996 census information).

PROCEDURE

Each boy and girl participated in discussions with their mothers. Allobservations were conducted at the homes of the participants. Prior tothe observations, participants independently completed demographicquestionnaires and the Revealed Differences Questionnaire (RDQ)(Mishler & Waxler, 1968). The RDQ consists of 38 hypothetical issuesabout interpersonal problems and values, for which participantsselected one of two possible answers for each item. Various versions ofthe RDQ have been used extensively in previous studies of familyinteractions for the purpose of generating discussion topics. Disagree-ment topics were used for the purpose of providing a direct comparisonwith previous studies on family interruption patterns (e.g., Hill, 1988;Steinberg, 1981) and for the purpose of simulating naturally occurring


differences of opinion that can occur in parent-adolescent verbal con-flicts while controlling for the topic of those disagreements.

The researchers compared the mothers’ and children’s answers tothe RDQ, noting all item numbers for which the dyad members dis-agreed on the answer. Five of these items were then selected for thedyad to discuss. To ensure that dyads from both age and gender groupstalked about a similar set of RDQ items, the discussion items werematched, where possible, across dyad types. The procedure for match-ing the discussion items consisted of (a) calculating the proportion ofall dyads who disagreed on each item, (b) calculating the differencebetween the proportion of mother-son dyads and the proportion ofmother-daughter dyads for each age group who disagreed on each item,and (c) rank ordering the item numbers according to the size of the dif-ference (from smallest to largest) between Steps 1 and 2. This proce-dure was updated as the completed questionnaires were received fromparticipants. The item numbers that each particular dyad discussedwere selected according to the rank orderings. Across all dyads, 31 ofthe total 38 items of the RDQ were discussed; however, 5 particularitems were discussed most often. Forty-four of the 48 dyads discussedat least 1 of those 5 items (23 discussed 2 or more), with consistentnumbers across the four dyad types for each of the 5 items. These itemshad to do with social issues such as parental influence on adolescents’behaviors and ethical questions about lying, stealing, or the care of theelderly. The dyad members were given 1 item at a time to discuss forapproximately 4 minutes each. The researchers left the room whiledyad members conversed, and all discussions were audiotaped via twolapel microphones feeding into separate channels of a stereo taperecorder.

CODING FOR TEMPORAL CONVERSATION STYLE

The audiotapes were transcribed verbatim, on a turn-by-turn basis,by several trained research assistants and checked for accuracy by theauthors. All instances when 2 speakers talked at the same time werepreserved in the transcripts by typing overlapping words withinslashes and ensuring that lines of text that overlapped were adjacentto one another. Coding was completed on a turn-by-turn basis by twotrained observers using the transcripts and the audiotapes to ensureaccuracy. Reliability was ensured by calculating agreement on 15% ofthe taped conversations. The coders were also free to change the tran-scripts whenever there was disagreement with the original transcrip-tions. These transcription disagreements were included in the agree-ment matrix. Overall agreement was found to be high, with a Cohen’skappa of .92.

The following categories of speech were coded for each speaker onthe basis of definitions used in previous studies (e.g., Beaumont, 1995;


Walker, 1991). In the following definitions, the first speaker is the per-son who held the conversational floor, and the second speaker is theperson who intruded on the first speaker’s speaking turn.

Overlaps between turns (O) were defined as instances when the sec-ond speaker cut off only one word (or less) of the first speaker’s com-plete utterance or when the two speakers began speaking at the sametime after a pause. An overlap was credited to the speaker who initi-ated it (i.e., the speaker who was not currently holding the floor). Over-laps were included as a measure of speakers’ pace of turn taking. Thatis, one would expect a faster paced (high-involvement) speaker to useoverlaps more frequently than a slower paced (high-considerateness)speaker. Agreement (agreements / [agreements + disagreements]) foroverlaps was 81.5%.

Simultaneous speech (SS) was defined as an instance in which thesecond speaker began talking before the first speaker had finished hisor her utterance and both speakers continued talking and completedtheir utterances. Simultaneous speech, then, demonstrates a type ofunsuccessful interruption (i.e., the second speaker is not successful ingetting the first speaker to stop talking). An instance of simultaneousspeech was credited to the speaker who initiated it (i.e., the “inter-rupter”). Agreement for simultaneous speech was 82.5%.

Successful interruptions (SI) were defined as instances when thesecond speaker cut the first speaker off before he or she had finished acomplete utterance (i.e., more than the last word of the utterance). Suc-cess was determined by examining whether the first speaker abruptlystopped talking before his or her idea was completed, in contrast to con-tinuing to speak simultaneously with the interrupter’s speech. A suc-cessful interruption was credited to the person who initiated it (i.e., theinterrupter). Agreement for successful interruptions was 82.8%.

Listener responses (short remarks that encourage the speaker tocontinue, e.g., “mhmm”) and unsuccessful interruptions (attempts tointerrupt in which the first speaker continues to talk and the inter-rupter stops talking) also were coded to ensure that the scheme wasmutually exclusive and exhaustive in coding all possible violations ofthe turn-taking rule (agreement = 80.3% for listener responses and76.9% for unsuccessful interruptions). Nevertheless, there were nohypotheses with regard to these two behaviors; therefore, no analyseswere conducted for those variables.

The first three speech acts listed earlier correspond to the criticalfeatures that discriminate the high-involvement and high-consideratenessstyles introduced by Tannen (1983, 1984). Specifically, the high-involvement conversational style is characterized by a fast rate of turntaking (and, therefore, frequent overlaps between speaking turns), fre-quent initiations of noninterruptive simultaneous speech, and possi-bly interruptions (depending on whether the person being interrupted


stops talking, as measured in the current coding scheme by O, SS, andSI, respectively).

CODING FOR FUNCTIONAL CONVERSATION STYLE

In addition to coding for temporal conversational style, data on thefunctions of speakers’ conversational behaviors were obtained byusing the coding scheme developed by Kennedy and Camden (1983). Atrained research assistant coded each previously identified SI and SSfor one of six possible functions. The definitions for each coding cate-gory were as follows (Kennedy & Camden, 1983, p. 51; examples arefrom Beaumont & Cheyne, 1998, pp. 279-280):

Clarification. A speech that attempts to understand the interruptedperson’s message. For example:

A: He should be home for dinner at least two or three times a week,and if hecan’t

M: Two or three?

Agreement. A speech that demonstrates agreement, support, con-currence, compliance, or understanding and can be demonstratedthrough further development or elaboration of the first speaker’s idea.For example (speech that occurs within slashes occurred simultaneously):

M: I’d hope that my life would still be full enough /that/A: /Yeah./ You’d live by yourself or you’d get married again.

Disagreement. A speech that demonstrates rejection, disagreement,challenge, or contradiction of the first speaker’s communication. Forexample:

A: It’s not worth saying in /the first place/.M: /But don’t you/ think he’d feel better if she told him.

Tangentialization. A speech that (a) reflects awareness of the firstspeaker’s statement; and (b) in some way minimizes, makes light of, ormocks the first speaker’s message. For example:

M: I guess you’re right, but what /I said is true too./A: /So I win. I win./ I win. I win.

Subject change. A speech that (a) reflects no awareness of the firstspeaker’s statement and (b) has no theme in common with the firstmessage and/or is a substantial change of topic. For example:


M: I would never wait until he was 20 years old then try to /deal/A: /The/ phone. The phone is ringing.

Other. This code was given whenever an instance of simultaneousspeech or a successful interruption could not be coded in one of theother five categories. This category typically was given when either (a) aspeech had little to do with what the first speaker was saying but didnot reflect a clear subject change; or (b) a speech reflected awareness ofthe first speaker’s statement but was not clearly an agreement, dis-agreement, clarification, or tangentialization. For example:

A: Sunday, when I got home from work I read it. I was finished reading it. Iwrote Matt a five page letter and the

M: You didn’t write me one.

Nine of the 48 tapes (15%) were coded by a second observer for agree-ment, which was found to be high, with an overall Cohen’s kappa of .93.Percentage agreement figures for individual categories were 86.6 forclarification, 87.8 for agreement, 90.5 for disagreement, 75.0 fortangentialization, 77.8 for subject change, and 83.3 for other.

TREATMENT OF THE DATA

The frequencies of each speech type (O, SS, SI, clarification, agree-ment, disagreement, tangentialization, subject change, and other) foreach speaker were summed across the four RDQ topics. Because thenumber of speech acts a person initiated was constrained by theamount of time that person talked, raw frequencies for each speakerwere transformed into rates using the individual’s talking time as thedenominator, a strategy that is typical in previous research in this area(e.g., Hill, 1988). The majority of previous studies on interruption pat-terns have measured talking time in seconds; however, Hill (1988)found that counting the number of words (from the transcript) thateach person spoke was a more efficient and reliable measure of talkingtime and was highly correlated with seconds spent talking. Therefore,in the present study, the number of words rather than seconds wasused as an index of overall talking time.The rates that were calculated,however, were extremely small because the denominators were solarge. To provide more meaningful data, each rate was multiplied by aconstant of the approximate average number of words spoken by allparticipants (i.e., 1,092) over the entire 20-minute session (followingthe precedent set by Kollock, Blumstein, & Schwartz, 1985). This for-mula yielded a figure indicating the number of interruptions and so onthat occurred for every 1,092 words spoken (which is equivalent to thenumber of words the average speaker spoke in a 20-minute conversa-tion). This calculation was used to provide data that are easier to inter-


pret and generalize because they are presented in the form of an indi-vidual’s rate of interruptions and so on per 20 minutes of conversation(corrected for the number of words that person spoke) rather than afractional rate of interruption and so on produced as a function of thenumber of words that person spoke. For example, if a person produced15 interruptions and spoke 1,500 words in 20 minutes, the trans-formed rate without multiplying by a constant would be .01; however, ifthis rate is multiplied by a constant of 1,092 words, the transformedrate is 10.92. Therefore, one could then say that this speaker produceda corrected rate of 10.92 interruptions in approximately 20 minutes ofconversation (if he or she had spoken the average number of words)rather than saying that he or she produced a rate of .01 interruptionsin 20 minutes of conversation. Thus, this calculation results in moremeaningful data by providing comparable rates of each speech act per20-minute conversation rather than rates of each speech act per num-ber of words spoken. What are examined in this study, therefore, arespeakers’ rates of each speech act per 1,092 words, which is roughlyequivalent to a 20-minute conversation.

RESULTS

ANALYSES OF TEMPORAL CONVERSATION STYLE

Before proceeding with the analyses, the temporal conversationstyle dependent variables (O, SI, and SS) were checked for skewness,and data for all three variables were found to be positively skewed.Therefore, as suggested by Tabachnick and Fidell (2001), the rateswere subjected to a square root transformation, and these transformedrates were used in the analyses (however, means and standard devia-tions for the untransformed rates are presented). Preliminary analy-ses were conducted to determine whether speakers’ rates of O, SS, andSI were intercorrelated. Correlations computed for all speakers (N =96) were found to be significant for all combinations of the dependentvariables (r = .44 and p < .001 for SI and SS, r = .36 and p = .001 for SIand O, and r = .47 and p < .001 for O and SS). Therefore, speakers’ ratesof O, SS, and SI were analyzed by a 2 (Age) × 2 (Gender) × 2 (Speaker)× 3 (Speech Type) MANOVA with age (preadolescent, adolescent) andgender (girl, boy) as between-dyad variables and speaker (mother,child) and speech type (O, SS, SI) as within-dyad variables. Univariateanalyses were examined where appropriate, and significant F ratiosfor interactions among variables were followed by Tukey’s honestlysignificant difference (HSD) tests of differences between means. Analpha level of .05 was used for all statistical tests.

MANOVA results revealed significant multivariate main effects forspeaker, F(1, 44) = 68.03, p < .001, η2 = 0.61, and speech type, F(2, 43) =


61.08, p < .001, η2 = 0.74. These main effects were qualified by a signifi-cant interaction of speaker and speech type, F(2, 43) = 7.74, p = .001, η2 =0.27, and a significant interaction of age, speaker, and speech type, F(2,43) = 4.43, p = .018, η2 = 0.17.

Subsequent univariate analyses indicated that the significantmultivariate main effect of speaker was significant for all threedependent variables: for O, F(1, 44) = 45.22, p < .001, η2 = 0.51; for SS,F(1,44) = 84.38,p < .001,η2 = 0.66;and for SI,F(1,44) = 15.71,p < .001,η2 =0.26. Examination of the relevant means for these main effects ofspeaker revealed that children produced significantly higher rates ofoverlaps (M = 16.28, SD = 7.11), simultaneous speech (M = 10.61, SD =5.98), and successful interruptions (M = 7.29, SD = 5.67) than theirmothers (M = 10.07, SD = 5.40; M = 3.23, SD = 2.29; and M = 3.70, SD =3.46, respectively).

The significant multivariate three-way interaction between age,speaker, and speech type was explored further by examining the age byspeaker interaction at the univariate level for each of the threedependent variables. The univariate results indicated that this inter-action was significant for the dependent variable SS,F(1,44) = 4.14,p =.048, η2 = 0.09. As shown in Figure 1, examination of the relevantmeans for the age by speaker interaction indicated that mothers ofboth preadolescent (M = 2.59, SD = 1.58) and adolescent (M = 3.86, SD =2.71) children produced significantly less simultaneous speech thantheir children (M = 11.73 and SD = 6.72 for preadolescents, M = 9.48and SD = 5.03 for adolescents). Furthermore, there was no significantdifference in the mean rates of simultaneous speech produced bypreadolescent and adolescent children; however, mothers of adoles-cents tended to produce more simultaneous speech than mothers ofpreadolescents (p = .06).

ANALYSES OF FUNCTIONAL CONVERSATION STYLE

Examination of the mean rates for each functional coding category(clarification, agreement, disagreement, subject change, tangential-ization, and other) indicated that some types of SI and SS occurred atrelatively low rates (e.g., subject change, tangentialization, other). As aresult, following Beaumont and Cheyne (1998), analyses were con-ducted using the three conceptually driven categories of confirmation,disconfirmation, and rejection, as proposed by Kennedy and Camden(1983). Specifically, the rates for agreement and clarification weresummed to indicate rates of confirmation. The rates for subject changeand tangentialization were summed to indicate rates of disconfirm-ation, and the category of disagreement alone was used as an index ofrejection. The category of other was not included in analyses because itdid not clearly reflect any of the three conceptually driven categories.


Before proceeding with the analyses, each dependent variable waschecked for skewness, and data for all three variables were found to bepositively skewed. Therefore, the rates were subjected to a square roottransformation,and these transformed rates were used in the analyses(however, means and standard deviations for the untransformed ratesare presented). Speakers’ rates of the three types of SI and SS wereanalyzed separately using 2 (Age) × 2 (Gender) × 2 (Speaker)mixed-model ANOVAs with age (preadolescent, adolescent) and gen-der (girl, boy) as the between-dyad variables and speaker (mother,child) as the within-dyad variable. Significant F ratios for interactionsamong variables were followed by Tukey’s HSD tests of differencesbetween means.

Confirmation. The ANOVA results for speakers’ rates of confirmingSS revealed a significant main effect of speaker, F(1, 44) = 56.72, p <.001, η2 = 0.56; a significant interaction of age and speaker, F(1, 44) =8.50, p = .006, η2 = 0.16; and a significant interaction of age and gender,F(1, 44) = 16.18, p < .001, η2 = 0.27. As shown in Figure 2a, examinationof the relevant means for the age by speaker interaction indicated thatboth preadolescents (M = 7.43, SD = 4.97) and adolescents (M = 3.70,SD = 2.55) produced significantly more confirming SS than their moth-ers (M = 1.32, SD = 1.20 and M = 1.58, SD = 1.16, respectively). In addi-tion, adolescents produced significantly less confirming SS thanpreadolescents, although there was no significant difference in therates of confirming SS produced by mothers of preadolescents andmothers of adolescents.


rate

of

SS

0

2

4

6

8

10

12

preadol adol

mother child

Figure 1. Age by Speaker Interaction for Speakers’ Rates of SimultaneousSpeech.

Note. SS = simultaneous speech, preadol = preadolescent, adol = adolescent.

a. Age by Speaker interaction for confirming SS

rate

of

con

firm

ing

SS

0

1

2

3

4

5

6

7

8

preadol adol

mother child

b. Age by Gender interaction for confirming SS

rate

of

con

firm

ing

SS

0

1

2

3

4

5

6

girl boy

preadolescent adolescent

c. Gender by Speaker interaction for rejecting SI in adolescent dyads

rate

of

reje

ctin

g S

I

0

1

2

3

4

5

6

7

8

girl boy

mother child

Figure 2. Significant Two-Way Interactions for Features of Functional Conversation Style.Note. SS = simultaneous speech, SI = successful interruptions, preadol = preadolescent, adol = adolescent.

436

As shown in Figure 2b, examination of the relevant means for theage by gender interaction for confirming SS revealed that althoughthere was no significant difference in the rates of confirmation pro-duced in preadolescent and adolescent girl dyads (M = 2.98, SD = 2.07and M = 3.30, SD = 1.62, respectively), there was significantly moreconfirming SS in the preadolescent boy dyads than in the adolescentboy dyads (M = 5.77, SD = 2.63 and M = 1.97, SD = 1.11, respectively).In addition, although there was no significant difference in the rates ofconfirming SS produced by adolescent girls and boys, preadolescentgirls produced significantly less confirming SS than preadolescentboys.

The ANOVA results for speakers’ rates of confirming SI revealed nosignificant main effects or interactions.

Disconfirmation. The ANOVA results for speakers’ rates ofdisconfirming SS revealed a main effect only of speaker, F(1, 44) =14.35, p < .001, η2 = 0.25, with children producing more disconfirmingSS than their mothers (M = 0.94, SD = 1.74 and M = 0.17, SD = 0.34,respectively).

The ANOVA results for speakers’ rates of disconfirming SI revealedno significant main effects or interactions.

Rejection. The ANOVA results for rejecting SS revealed significantmain effects of speaker, F(1, 44) = 9.42, p = .004, η2 = 0.18, and age, F(1,44) = 4.67, p = .036, η2 = 0.10. Examination of the relevant means forthe main effect of speaker revealed that children (M = 2.35, SD = 2.39)produced significantly more rejecting SS than their mothers (M = 0.98,SD = 1.18). The relevant means for the age main effect indicated thatthe mother-adolescent dyads produced significantly more rejecting SSthan the mother-preadolescent dyads (M = 3.22, SD = 2.55 and M =1.68, SD = 1.41, respectively).

The ANOVA results for rejecting SI revealed a significant maineffect of speaker and a significant gender by speaker interaction thatwere qualified by a significant three-way interaction of speaker, age,and gender, F(1, 44) = 4.14, p = .048, η2 = 0.09. This significantthree-way interaction was examined further by conducting separatespeaker by gender ANOVAs for preadolescent and adolescent dyads.This interaction was not significant for the preadolescent dyads; how-ever, it was significant for the adolescent dyads,F(1,44) = 8.18,p = .009,η2 = 0.27. As shown in Figure 2c, examination of the relevant means forthe speaker by gender interaction for the adolescent dyads revealed nosignificant difference in the rates of rejecting SI produced by daughtersand their mothers (M = 4.04, SD = 2.78 and M = 3.78, SD = 3.67, respec-tively); however, sons produced significantly more rejecting SI thantheir mothers (M = 7.28, SD = 6.41 and M = 1.37, SD = 1.36, respec-tively). This significant difference in rates of rejecting SI between


mothers and sons appeared to be due to boys’ producing significantlymore rejecting SI than girls and mothers of sons’ producing signifi-cantly less rejecting SI than mothers of daughters.

DISCUSSION

The purpose of this study was to further examine the conversationalstyles of mothers and their children by considering two factors thatmight influence differences in conversational styles: age and gender ofthe child. It was hypothesized that both girls and boys would usehigher rates of O, SS, and SI than their mothers; there would be subtledifferences in the mother-child and mother-adolescent conversations,with more high-involvement features used in the adolescent dyads;and mothers’ behaviors would vary as a function of gender, but chil-dren’s would not. These results were generally confirmed. Specifically,mothers used a high-considerateness conversational style, whereasboth preadolescent and adolescent boys and girls demonstrated theuse of a high-involvement conversational style, as evidenced by signifi-cantly higher rates of simultaneous and overlapping speech and inter-ruptions (with no differences in the temporal conversation styles pro-duced by boys and girls). In addition, although no differences inmothers’ or children’s temporal conversational styles emerged as afunction of child gender, subtle differences in the functional use ofinterruptions and SS emerged in mother-daughter versus mother-sonconversations.

With regard to differences as a function of age, mothers of adoles-cents tended to produce more SS than mothers of preadolescents. Thisfinding is consistent with previous research by Beaumont (2000), whofound that in mother-daughter conversations, there was significantlymore SS produced in adolescent than preadolescent dyads. Althoughthis finding suggests that mothers use a more high-involvement con-versational style with adolescents than preadolescents, it is importantto note that both adolescents and preadolescents used significantlymore SS than their mothers. In other words, there is evidence that bothpreadolescents and adolescents experienced a clash in conversationalstyles with their mothers. This lack of difference in the styles ofpreadolescents and adolescents is surprising given that previousresearchers have found that adolescents tend to interrupt their moth-ers more often than when they were younger (e.g., Hill, 1988). Further-more, this lack of difference in children’s styles as a function of age isnot consistent with the hypothesis put forth by Beaumont (1995) thatincreasing differences in mothers’ and children’s conversational stylesacross age might account for increased mother-child conflict duringadolescence. Nevertheless, even though preadolescent-mother dyadsdemonstrated just as much difference in styles as adolescent-mother


dyads, it may be that these differences in styles have a greater impacton adolescents’ compared with preadolescents’ perceptions of conflictwith their mothers. This hypothesis is supported by research that hasshown that during playback of taped conversations with their mothersand friends, adolescent girls rated themselves and their mothers dif-ferently in terms of levels of friendliness, but they rated themselvesand their friends similarly in levels of friendliness (Beaumont, 1992).Furthermore, the clash in conversational styles between mothers anddaughters appeared to be perceived more negatively by middle-adolescent girls (15- to 16-year-olds) than preadolescent girls (11- to12-year-olds) (Beaumont, 1996). The results of previous studies sug-gest that whereas both preadolescent and adolescent daughters andtheir mothers tend to experience a clash in styles, it may be the oldergirls who react most negatively to differences in their own and theirmothers’ conversational styles. Although boys’ perceptions of audiotapedconversations with their mothers have not yet been examined, morenegative perceptions of adolescents compared with preadolescents (asdemonstrated by Beaumont, 1996) might account for (or at least berelated to) the current finding that there was more rejecting SS inadolescent-mother than in preadolescent-mother conversations andthat adolescents produced less confirming SS than preadolescents.These findings are consistent with those of previous researchers whofound less supportive communication behaviors in parent-adolescentthan parent-preadolescent conversations (e.g., Papini et al., 1988).

In addition to differences as a function of age, differences betweenmother-daughter and mother-son conversations also emerged. Spe-cifically, although there were no differences in mothers’ or children’stemporal conversation styles as a function of child gender, subtle dif-ferences in the functional use of interruptions emerged inmother-daughter versus mother-son conversations. For the adolescentage group, there was no difference in the rates of rejection interrup-tions produced by mothers and daughters; however, boys produced sig-nificantly more interruptions that served a rejection function thantheir mothers. This significant difference between mothers and adoles-cent sons appeared to be due to what both mothers and their sons weredoing. That is, adolescent boys produced significantly more rejectioninterruptions than adolescent girls, and mothers of adolescent sonsproduced significantly more rejection interruptions than mothers ofadolescent girls. As a result, there was a divergence in the interruptionbehavior of mothers and adolescent sons. This finding suggests thatalthough both girls and boys experience a clash in conversationalstyles with their mothers, the greatest difference in styles (at least inthe functional use of interruptions) might be between mothers andadolescent sons.

Why would mothers and their children use different conversationalstyles, and why might mothers and adolescent sons use more dis-


similar styles than mothers and adolescent daughters? A possibleexplanation can be found in a combination of sociolinguistic theory andfindings from studies in social and developmental psychology.Gumperz (1976) claimed that speakers’ styles vary with age and com-municative experiences (e.g., cultural or subcultural experiences) suchthat speakers of relatively the same age who have had similar commu-nicative experiences should exhibit similar habits with regard to voicequality and temporal patterning. Asymmetrical conversation, or con-versation between speakers with different habits or styles, occurswhen speakers “do not find some common base of experience on whichto build the interaction” (Gumperz, 1976, p. 276). Empirical researchby social psychologists confirms that the temporal speech styles ofsame-sex peers become similar after engaging in only a few conversa-tions. For example, Jaffe and Feldstein (1970) found significant simi-larity between dyad members’ rates of speech and turn taking afteronly three interactions. It is important to note, though, that this simi-larity in speakers’ temporal styles was most likely to occur betweenspeakers who perceived themselves to have similar personalities(Welkowitz & Feldstein, 1969).

These findings offer some explanation for the differences found inmothers’ and children’s conversational styles. Specifically, mothers’and children’s differing styles may be attributable to the asymmetricnature of their relationships. In other words, mothers and theirpreadolescent and adolescent children may not have similar communi-cative experiences on which to base their conversations. Bothpreadolescents and adolescents, but particularly adolescents, mayhave increased experience using a fast-paced, high-involvement stylefrom interacting with their peers. In contrast, mothers’ experiencesusing a more considerate pace of turn taking is likely based on years ofexperience interacting with children who slowly develop conversa-tional skills. Results from studies of interactions between mothers andyounger children offer some support for this hypothesis. Specifically,researchers have demonstrated that from the time their children arefirst learning to speak, Western mothers tend to structure face-to-faceinteractions with their children using a strict turn-taking pattern withvery little SS (e.g., Kaye & Charney, 1981). Furthermore, mother-daughter and mother-son conversations appear to differ in quality. Forexample, mothers spend more time talking to their daughters than totheir sons about emotions (e.g., Dunn, Bretherton, & Munn, 1987), andsons appear to be more withdrawn in conversations with their mothersthan are daughters (e.g., Whalen et al., 1996). Therefore, the currentfinding that mothers of adolescent sons produced fewer rejecting inter-ruptions than mothers of adolescent daughters might be because con-versations with sons are somewhat novel situations compared withconversations with daughters, and mothers may have been trying to be


even more considerate with sons than with daughters given that boystend to be more withdrawn.

Some limitations of the methods used in this study must beacknowledged. First, as mentioned in the Method section, the rate offamilies who agreed to participate in this study was somewhat low(11%). Other researchers who have studied parent-adolescent inter-actions in families from larger cities have tended to obtain consentrates of approximately 20% to 50% (e.g., Holmbeck & Hill, 1991;Montemayor et al., 1993; Papini et al., 1988). However, in the presentstudy, considerably different methodological techniques were usedthan in other previous studies, and these techniques might have influ-enced the consent rate. Specifically, most other studies have tended toobserve parent-adolescent interactions in laboratories, and the fami-lies are paid for their participation. In contrast, for the present study,the observations took place in the participants’ homes, and partici-pants were not paid for their participation. Therefore, the require-ments of the present study may have been more intimidating and lessrewarding than previous studies that obtained higher consent rates.Because no other previous studies have used similar methods, thepresent findings will need to be replicated using a larger sample(obtained with a higher participation consent rate) with participantsfrom varied socioeconomic and cultural backgrounds. Nevertheless,the present results for middle-class, Caucasian participants are cer-tainly comparable to those obtained in Beaumont’s (1995) study, whichconsisted of only a slightly larger sample size (and participation con-sent rate) obtained in a much larger, tricity area.

It is also important to note that there was high variability amongindividual speakers with regard to the use of O, SI, and SS, as evi-denced in the large standard deviations found in this study. This vari-ability is not surprising given that sociolinguists also have found sig-nificant individual differences with regard to pacing and pausing andthe use of SS (e.g., Tannen, 1984). In fact, Tannen (1984) has acknowl-edged that it may be helpful to think of the high-involvement andhigh-considerateness conversational styles as varying along a contin-uum from slow paced to fast paced rather than two totally distinct cate-gories. Within this conceptualization, an individual speaker can beconsidered a high-involvement speaker only in relation to the contrast-ing style of any given partner. The results of the present study confirmthat there are large individual differences with regard to temporalspeech patterns and suggest that as a group, preadolescents and ado-lescents demonstrate speech patterns that can be considered morehigh involvement than those of their mothers. That is, individualmother-child dyads may vary in the extent to which they experience aclash in styles depending on the magnitude of the differences in thestylistic features of their conversational styles.


In conclusion, the results of this study add to the literature onmother-child communication by providing information about mothers’speech styles with both boys and girls who are older than preschoolage. The results support the hypothesis presented in previous researchby Beaumont (1995, 2000) that mothers and children experience aclash in conversational styles, with some indication that mothers andadolescent sons may experience a greater difference in styles thanmothers and adolescent daughters. Other researchers have found thatspeakers with differing speech styles form negative perceptions of oneanother and of their conversations (e.g., Giles, 1979; Ryan, 1979).Therefore, it is possible that a greater clash in mothers’ and sons’ con-versational styles may lead to more negative perceptions and misunder-standings than in mother-daughter relationships (for clarification, seeBeaumont, 1996). Examining these findings within the framework ofsociolinguistics suggests that differences in mothers’ and children’sstyles may have implications for the quality of mother-child relation-ships and suggests that more positive communication might comeabout from mothers’ and children’s modifying their conversationalbehaviors to create more convergence in styles. These results also haveimplications for future research on parent-child relationships. Spe-cifically, future longitudinal or sequential research should be directedtoward identifying developmental changes in children’s conversa-tional styles and possible changes in parents’ conversational styles astheir children enter early adulthood. In addition, it would be useful forresearchers to focus on isolating the attitudinal, personality, and emo-tional characteristics that are associated with parents’ and children’sconversational styles and the influence of differing styles on parent-child relationships.

REFERENCES

Arnett, J. J. (1999). Adolescent storm and stress, reconsidered. American Psychologist,54, 317-326.

Beattie, G. W. (1981). Interruption in conversational interactions and its relation to sexand status of the interactants. Linguistics, 19, 15-35.

Beaumont, S. L. (1992, June). Mothers’ and daughters’ interpretations of their conversa-tions. Paper presented at the annual meeting of the Canadian Psychological Associa-tion, Quebec City, Canada.

Beaumont, S. L. (1995). Adolescent girls’ conversations with mothers and friends: A mat-ter of style. Discourse Processes, 20, 109-132.

Beaumont, S. L. (1996). Adolescent girls’ perceptions of conversations with mothers andfriends. Journal of Adolescent Research, 11, 325-346.

Beaumont, S. L. (2000). Conversational styles of mothers and their preadolescent andmiddle adolescent daughters. Merrill-Palmer Quarterly, 46, 119-139.

Beaumont, S. L., & Cheyne, J. A. (1998). Interruptions in adolescent girls’ conversations:Comparing mothers and friends. Journal of Adolescent Research, 13, 272-292.


Bennett, A. (1981). Interruption and the interpretation of conversation. Discourse Pro-cesses, 4, 171-188.

Blishen, B. R., Carroll, W. K., & Moore, C. (1987). The 1981 socioeconomic index for occu-pations in Canada. Canadian Review of Sociology and Anthropology, 24, 465-488.

Coates,J. (1989).Gossip revisited:Language in all-female groups.In J.Coates & D.Cameron(Eds.), Women in their speech communities (pp. 94-122). New York: Longman.

Csikszentmihalyi, M., & Larson, R. (1984). Being adolescent: Conflict and growth in theteenage years. New York: Basic Books.

Dindia, K. (1987). The effects of sex of subject and sex of partner on interruptions.Human Communication Research, 13, 345-371.

Dunn, J., Bretherton, I., & Munn, P. (1987). Conversations about feeling states betweenmothers and their young children. Developmental Psychology, 23, 132-139.

Feldstein, S., Alberti, L., & BenDebba, M. (1979). Self-attributed personality characteris-tics and the pacing of conversational interaction. In A. W. Siegman & S. Feldstein(Eds.), Of speech and time: Temporal speech patterns in interpersonal contexts (pp. 73-87). New York: John Wiley.

Feldstein, S., & Welkowitz, J. (1987). A chronography of conversation: In defense of anobjective approach. In A. W. Siegman & S. Feldstein (Eds.), Nonverbal behavior andcommunication (2nd ed., pp. 435-499). Hillsdale, NJ: Lawrence Erlbaum.

French, P., & Local, J. (1983). Turn-competitive incomings. Journal of Pragmatics, 7,17-38.

Giles, H. (1979). Ethnicity markers in speech. In K. R. Scherer & H. Giles (Eds.), Socialmarkers in speech (pp. 251-289). Cambridge, UK: Cambridge University Press.

Greenwood, A. (1989). Discourse variation and social comfort: A study of topic initiationand interruption patterns in the dinner conversations of preadolescent children.Unpublished doctoral dissertation, City University of New York, New York.

Gumperz, J. J. (1976). Language, communication and public negotiation. In P. Sanday(Ed.), Anthropology and the public interest (pp.273-292).New York:Academic Press.

Hill, J.P. (1988).Adapting to menarche:Familial control and conflict. In M.Gunnar (Ed.),21st Minnesota symposium on child psychology (pp. 43-77). Hillsdale, NJ: LawrenceErlbaum.

Holmbeck, G. N., & Hill, J. P. (1991). Conflictive engagement, positive affect, and menar-che in families with seventh-grade girls. Child Development, 62, 1030-1048.

Jaffe, J., & Feldstein, S. (1970). Rhythms of dialogue. New York: Academic Press.James,D.,& Clarke,S. (1993).Women,men and interruptions:A critical review. In D.Tannen

(Ed.), Gender and conversational interaction (pp. 231-280). New York: Oxford Univer-sity Press.

Julian, T. W., McKenry, P. C., & Arnold, K. (1990). Psychosocial predictors of stress associ-ated with the male midlife transition. Sex Roles, 22, 707-722.

Kahlbaugh, P. E., & Haviland, J. M. (1994). Nonverbal communication between parentsand adolescents: A study of approach and avoidance behaviors. Journal of NonverbalBehavior, 18, 91-113.

Kaye, K., & Charney, R. (1981). Conversational asymmetry between mothers and chil-dren. Journal of Child Language, 8, 35-49.

Kennedy, C. W., & Camden, C. T. (1983). A new look at interruptions. Western Journal ofSpeech Communication, 47, 45-58.

Kollock, P., Blumstein, P., & Schwartz, P. (1985). Sex and power in interaction: Conversa-tional privileges and duties. American Sociological Review, 50, 34-46.

Laursen, B., Coy, K. C., & Collins, W. A. (1998). Reconsidering changes in parent-childconflict across adolescence: A meta-analysis. Child Development, 69, 817-832.

Leaper, C., Anderson, K. J., & Sanders, P. (1998). Moderators of gender effects on parents’talk to their children: A meta-analysis. Developmental Psychology, 34, 3-27.

Marche,T.A.,& Peterson,C. (1993).The development and sex-related use of interruptionbehavior. Human Communication Research, 19, 388-408.


Mettetal, G. (1983). Fantasy, gossip, and self-disclosure: Children’s conversations withfriends. In R. N. Bostrom (Ed.), Communication yearbook (Vol. 7, pp. 717-735).Beverly Hills, CA: Sage.

Mishler, E. G., & Waxler, N. E. (1968). Interaction in families: An experimental study offamily process and schizophrenia. New York: John Wiley.

Montemayor, R. (1983). Parents and adolescents in conflict: All families some of the timeand some families most of the time. Journal of Early Adolescence, 3, 83-103.

Montemayor, R., Eberly, M., & Flannery, D. J. (1993). Effects of pubertal status and con-versation topic on parent and adolescent affective expression. Journal of Early Ado-lescence, 13, 431-447.

Murray, S. O. (1987). Power and solidarity in “interruption”: A critique of the SantaBarbara School conception and its application by Orcutt and Harvey (1985). Sym-bolic Interaction, 10, 101-110.

Papini, D. R., Datan, N., & McCluskey-Fawcett, K. A. (1988). An observational study ofaffective and assertive family interactions during adolescence. Journal of Youth andAdolescence, 17, 477-492.

Powers, S. I., Hauser, S. T., Schwartz, J. M., Noam, G. G., & Jacobson, A. M. (1983). Adoles-cent ego development and family interaction: A structural-developmental perspec-tive. New Directions for Child Development, 22, 5-25.

Robin,A.L.,& Foster,S.L. (1984).Problem-solving communication training:A behavioral-family systems approach to parent-adolescent conflict. In P. Karoly & J. J. Steffen (Eds.),Adolescent behavior disorders: Foundations and contemporary concerns (pp. 195-240).Lexington, MA: D. C. Heath.

Ryan, E. B. (1979). Why do low prestige language varieties persist? In H. Giles & R. N. St.Clair (Eds.), Language and social psychology (pp. 53-69). Baltimore: University ParkPress.

Scollon,R. (1985).The machine stops:Silence in the metaphor of malfunction. In D.Tannen &M. Saville-Troike (Eds.), Perspectives on silence (pp. 21-30). Norwood, NJ: Ablex.

Steinberg, L. D. (1981). Transformations in family relations at puberty. DevelopmentalPsychology, 17, 833-840.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Boston:Allyn & Bacon.

Tannen, D. (1983). When is an overlap not an interruption? One component of conversa-tional style. In R. DiPietro, W. Frawley, & A. Wedel (Eds.), The 1st Delaware sympo-sium on language studies (pp. 119-129). Newark: University of Delaware Press.

Tannen, D. (1984). Conversational style: Analyzing talk among friends. Norwood, NJ:Ablex.

Tannen, D. (1989). Interpreting interruption in conversation. In B. Music, R. Graczyk, &C. Wiltshire (Eds.), Papers from the 25th annual regional meeting of the Chicago Lin-guistic Society (pp. 266-287). Chicago: Chicago Linguistic Society.

Walker, L. J. (1991). Verbal interactions within the family context. Canadian Journal ofBehavioural Science, 23, 441-454.

Watzlawick, P., Beavin, J., & Jackson, D. (1967). Pragmatics of human communication.New York: Norton.

Welkowitz, J., & Feldstein, S. (1969). Dyadic interaction and induced differences in per-ceived similarity. Proceedings of the 77th Annual Convention of the American Psycho-logical Association, 4, 343-344.

Whalen, C. K., Henker, B., Hollingshead, J., & Burgess, S. (1996). Parent-adolescent dia-logues about AIDS. Journal of Family Psychology, 10, 343-357.

Youniss, J., & Smollar, J. (1985). Adolescent relationships with mothers, fathers, andfriends. Chicago: University of Chicago Press.

Zimmerman, D. H., & West, C. (1975). Sex roles: Interruptions and silences in conversa-tion. In B. Thorne & N. Henley (Eds.), Language and sex: Differences and dominance(pp. 105-126). Rowley, MA: Newbury House.


JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001Wiggins / FOOD EVALUATION IN CONVERSATION

CONSTRUCTION AND ACTIONIN FOOD EVALUATION

Conversational Data

SALLY WIGGINSLoughborough University

This study engages both social psychological research on “attitudes” and discursive workon “evaluative practices.” Methodological constraints in both of these fields have resultedin a relative lack of attention to everyday interaction. By using conversational data, thecurrent study extends discursive research and highlights the constructive and con-structed nature of food evaluations. Family mealtimes were audiotaped, transcribed, andanalyzed using discursive and conversational analytic procedures. Direct evaluativeexpressions such as “like” and “nice” were examined in terms of their construction andplacement in the talk. The rhetorical organization of these expressions highlighted theextent to which food evaluations are oriented to actions such as accounts, compliments,and offers of food. Examples of these activities are discussed in relation to theinteractional construction of evaluations. Implications of the study for the fields of foodpreferences and health promotion are also addressed.

Food and eating practices are now widely recognized to be richlysocial as well as biological phenomena (Beardsworth & Keil, 1997;Lupton, 1996; Scapp & Seitz, 1998). Research within the psychology offood, for example, now includes explorations into the social influencesthat may determine an individual’s eating behavior (e.g., Birch, 1990;Clendenen, Herman, & Polivy, 1994; Lyman, 1989). Sociological andanthropological research has also developed in this area, although itsemphasis is on the cultural and global significance of food at a societallevel (e.g., Counihan & Van Esterik, 1998; Murcott, 1983).

Despite these developments, the social significance of food is stillrelatively underexplored, and only a handful of studies have empha-sized the relationship between food, talk, and interaction (e.g.,Pontecorvo & Fasulo, 1999; Wiggins, Potter, & Wildsmith, 2001). Thisrelationship involves the way in which food is managed, requested,evaluated, and organized on a practical level. It also includes theexpression of food and eating in the media, cultural artifacts, and

445

AUTHOR’S NOTE: I would like to thank Howard Giles, Jonathan Potter, and two anony-mous reviewers for extremely helpful remarks on an earlier draft of this article. An earlierversion of this article was presented at the 7th International Conference on Language andSocial Psychology, Cardiff, United Kingdom, July 2000.


consumer advertising.This interactional focus is distinct from theoriessuch as Barthes’s (1979) structural account of the signification of foodsystems. What is highlighted in this study is what can be seen as thebroadly practical and communicative aspects of food—how it allows usto express and engage with others—as a critical component of our dailylives. Food is not just social in the sense of sharing meals with others; itbecomes social as soon as we orient to it as “food” and in the terms thatwe use to describe it.

ATTITUDES AND FOOD PREFERENCES

To contextualize and introduce the concerns of the current article, itwill be useful to examine existing social psychological research on food.Food has entered psychology under two broad areas of interest. Thefirst of these is characterized by work on bodily matters such as diet-ing, body image, and eating disorders. The second has been a focus onfood preferences and consumption behavior. Within this latter field,the topics have been linked together using the framework of “attitude”models, in which an attitude (or preference) toward a food may be mea-sured to predict consumptive behavior (Ajzen, 1988, 1991). For exam-ple, a study by Cantin and Dubé (1999) distinguished between theaffective and cognitive bases for preferred beverages and the degree towhich this influenced an individual’s consumption of a drink. Otherresearch has used variations in product information to determinewhether these affect attitudes or behaviors toward different foods (e.g.,Aaron, Mela, & Evans, 1994; Berg, Jonsson, & Conner, 2000; Bonhamet al., 1995).

Food preference research is predominantly based on the assump-tions of attitudinal research. The major contributors to such work areFishbein and Ajzen’s (1975) theory of reasoned action and, later, thetheory of planned behavior (TPB) (Ajzen, 1988, 1991). The TPB exam-ines how factors such as attitudes, subjective norms, and perceivedbehavioral control can be used as means to predict behavioral inten-tions (Bagozzi & Kimmel, 1995; Eagly & Chaiken, 1993; Netemeyer,Burton, & Johnston, 1991). The model is based on a cognitive constructof individual attitudes—beliefs or opinions toward specific objects—and draws heavily on questionnaire and rating scale measures (e.g.,Armitage & Conner, 1999; Conner, Martin, Silverdale, & Grogan, 1996;Schifter & Ajzen, 1985; Sparks, Hedderley, & Shepherd, 1992; Sparks &Shepherd, 1992). As such, the key research questions are concernedwith measuring and determining the components of attitudes towardfood, examining how these could be altered, and determining how theyare linked to consumption behavior.

The implications of attitudinal research can be traced throughadvertising and marketing fields, consumer research, and health


promotion work (e.g., Dubé, Chattopadhyay, & Letarte, 1996). If eatingpractices can be causally linked to definable attitudes toward food,then by changing perceptions, one could potentially alter an individ-ual’s behavior. This is of considerable importance at the levels of bothsocial policy and commerce because it could modify health-related eat-ing patterns and alter consumer spending. In line with this, contempo-rary health and nutritional campaigns often rely on the notion of theindividual as information processor (a cognitive model). Hence, provid-ing “knowledge” of food nutrition and basic biological needs is used tochange people’s attitudes and thus their behavior.

The assumptions that underlie the TPB model are broadly sharedby the social psychology of food, even where the model is not directlyapplied (Conner, Povey, Sparks, James, & Shepherd, 1998). The stan-dard methodology within this approach is to use questionnaires withevaluative terms placed against rating scales. For example, attitudestoward food are often assessed using semantic differential scales suchas “enjoyable-unenjoyable”or “good-bad.”A participant is asked to ratethe degree to which he or she finds a food “good” or “enjoyable” (Sparkset al., 1992; noted as the usual method within TPB models). This restson the assumption that there are underlying food attitudes (or “prefer-ences”) that can be distinguished from evaluations of particular foods.That is, a “taste evaluation” is a single, immediate judgment of a food,whereas a food attitude is an “evaluative summary”of these judgments(Lozano, Crites, & Aikman, 1999, p. 208). This makes sense in theabstract, as a theoretical distinction, but it is harder to reconcile inpractical terms. The distinction becomes somewhat blurred when onetries to define the boundaries between what constitutes an evaluationrather than an attitude at an interactional level.

A further problem with the attitudinal approach arises with thegeneric use of evaluative expressions such as good or like within ques-tionnaire designs. On the basis of a researcher’s notion of the meaningof these terms, the attitudinal approach fails to take into account how aparticipant may interpret what is meant by a good or enjoyable food.Moreover, it imposes categories of food evaluations such as taste,health value, and sensory appeal onto participants’ responses (e.g.,Geiselman et al., 1998; Lindeman & Stark, 1999; Martins, Pelchat, &Pliner, 1997). The problem with relying on analysts’ (rather than par-ticipants’) categories is that the meaning of a response is often alteredin the process (Hutchby & Wooffitt, 1998). The particular form of anevaluation has an indexical meaning that is lost when it is taken out ofcontext. The assumption, therefore, is that not only will participantshave the same notions of the meanings of expressive terms, but also,this meaning is extractable from the context within which it wasproduced.

Wiggins / FOOD EVALUATION IN CONVERSATION 447

EVALUATIVE PRACTICES AND INTERACTION

Discursive psychological approaches have challenged attitudinalresearch in a number of ways (e.g., Billig, 1989, 1992; Burningham,1995; Potter & Wetherell, 1987, 1988; Puchta & Potter, 1999;Verkuyten, 1998). Most notably, the notion of a stable attitude concepthas been problematized through an examination of the variability ofattitude expression. For example, interviews with New Zealandersrevealed a number of conflicting standpoints within the same stretchof talk (Potter & Wetherell, 1987). Taking a conventional approach, itwould be difficult, if not impossible, to isolate the single attitude thatunderpinned a response. A further challenge has been concerned withthe link between attitudes and behavior. Problems identifying the pre-cise relationship between these concepts continue to engage attitudi-nal researchers (Eagly & Chaiken, 1998). However, discursiveapproaches propose that expressing an evaluation is itself a behavior.An attitude becomes the topic of study rather than an entity to be iso-lated and measured. Therefore, it becomes less appropriate to searchfor links between what may be distinct and separate practices.

Discursive approaches are concerned with the construction of evalu-ations as interactional practices rather than with assumptions aboutcognitive constructs or processes. For instance, an evaluation is exam-ined within the context of an interaction. What talk precedes and fol-lows an evaluation? What are the features of the interaction that mayhave required such an assessment? In other words, evaluative prac-tices are regarded as interindividual rather than intraindividual activ-ities. The practical uses of evaluations are also emphasized in terms ofwhat else is achieved when giving an evaluation. In this respect, theapproach differs from an examination of the functions of holding a par-ticular attitude (e.g., Herek, 1987; Maio & Olson, 1994). The focus is onthe action-orientation of these constructions and how this is achievedinteractionally (Edwards, 1995, 1997; Potter, 1996, 1998a). For exam-ple,Marshall and Raabe (1993) illustrated how attitudes toward politi-cal ideologies varied within participants’ interview responses.Whether they were “for” or “against” privatization was dependent onthe question they were being asked. Indeed, the meaning of the termprivatization was constructed in their response and so was itself opento negotiation.

Up to the present time, discursive work on evaluations has drawnprimarily on interactions within interviews (e.g., Billig, 1992;Burningham, 1995; Marshall & Raabe, 1993). Although these studieshave seriously challenged social psychological work on attitudes, theyare at risk for falling into a similar methodological trap. That is, theyoften abstract evaluations from their everyday locations and placethem within a research agenda. The problems with this type of databecome apparent when one considers evaluations to be contextualized,


indexical practices. That is, they perform particular actions within aconversation and are thus expressed for that specific moment in time.If we then look at an evaluation within an interview, we are looking atsomeone “doing” an interview. Although this can allow us to examinesome of the uses of evaluations, it does not tell us anything about howpeople use evaluations in their everyday lives.

The current article focuses on the conversational use of food evalua-tions. By using this form of data, both the construction and occurrenceof food evaluations are a consequence of the interaction rather than theresearch agenda (Potter, 1998b). Drawing on developments within thediscursive framework enables a move toward more “naturalistic”materials. From this, we can examine how evaluations are used to con-struct and make sense of daily realities. More important, there is littlediscursive research on conversations around or about food. Notableexceptions are the work of Blum-Kulka (1994, 1997) and of Fasulo andcolleagues (Ochs, Pontecorvo, & Fasulo, 1996; Pontecorvo & Fasulo,1999), though these are more concerned with the construction of cul-tural (e.g., Italian, American, Jewish) identities through talk aboutfood.

More concisely, the aims of the study are to (a) examine food evalua-tions in terms of their placement in conversation, (b) consider the rhe-torical design of food evaluations, and (c) examine the extent to whichfood evaluations are bound up with other activities in interaction.Thus, the goal is to engage with discursive and attitudinal researchand with the study of food preferences in social psychology.

METHOD

PARTICIPANTS

Four families were recruited via personal contacts according to thefollowing criteria:The family included at least one son or daughter whowould be present at meals, and eating together was a usual routine forthe family. There were practical reasons for selecting these criteria,because the pilot research suggested that families with children aremore likely to eat together regularly than, say, groups of young adultssharing accommodations. Thus, it was considered more appropriate torecruit those who participate in meals together as a matter of course.

PROCEDURE

Each family was provided with a small tape recorder and audiotapesand was asked to record its mealtimes as regularly as possible. It wasemphasized that the aim of the study was to collect a “corpus of


naturalistic conversation amongst family members,” and as such, noalteration to daily household activities was required. The data collec-tion resulted in over 15 hours of recorded conversations. The tapeswere all transcribed to first pass (words only); then, smaller sectionsthat involved talk about food or eating practices were fully transcribed.The detailed sections were transcribed using conventional Jeffersonnotation (see the appendix; see also Hutchby & Wooffitt, 1998). Thiswas done to highlight characteristics of the talk that would assist inthe analysis.

ANALYTIC PROCEDURE

The transcribed tapes were analyzed using discursive psychologicaland conversation analysis techniques (Hutchby & Wooffitt, 1998; Pot-ter, 1996, 1998a; Potter & Edwards, 2001). This involved making notesduring the transcription process and on repeated readings of the tran-scripts and listening to the tapes. The analysis focused on participants’uses of explicit evaluative terms such as good or like and how these dif-fered from the generic use of such terms in the dominant literature.The extracts presented are taken from the larger data corpus and wereselected for their ability to illustrate and develop the main analyticaltopics.

ANALYSIS

The analysis highlights two topics, which are contrastive with stan-dard “attitudinal” and food preference research: (a) the discursive con-struction of food evaluations as an interactional achievement in con-versation; and (b) how food evaluations are bound up with otheractivities in conversation such as complaining, complimenting, andaccounting for behavior.

REEVALUATING CONVERSATION

Using conversational data immediately highlights a very simple butimportant point: Food evaluations are expressed in interaction. It isimportant because evaluating a food as representative of an attitude isoften treated as an individual response. Using the example from theintroduction, participants may be asked to respond (in written form)on a good-bad rating scale in terms of how they perceive a type of food(Sparks et al., 1992). Yet, the interactional expression of evaluationsreveals them to be contested, negotiable activities (Billig, 1996;Edwards & Potter, 1992; Wetherell & Potter, 1992). The point here is


that the notion of evaluative practices has not been applied to foodevaluations nor, to some extent, to conversational data. As was dis-cussed in the introduction, food preference research still adopts abroadly cognitive framework.

The extract below is used to demonstrate the construction of foodevaluations as part of conversation. It is taken from near the end of afamily meal at which the conversation turns to the food that is beingeaten. Lesley and Paul are the parents.

(1) SKW/ J1b-M4 (410-422)1. Lesley: >d’you like your< ↑sausage bits2. (1.0)3. Paul: >pardon<4. Lesley: d’you like ↓°them°5. Paul: → mmm: ↑they’re alright >they’re a bit< ↑spi:cy6. Lesley: >h↓mm<7. (1.0)8. Paul: what are: ↓they9. Lesley: >°well° just< slice:s with- (0.4) °with:°10. (0.6)11. Lesley: spices °rather than:: n:ormal ones (.) >(thought12. they’d)< be: (0.2) >nice for a ↑change°<

The first point to note about this extract is that Paul’s evaluation(line 5) is the second part of a question-answer pair, with Lesley’s turnin line 1 or line 4 serving as the first part (Pomerantz, 1984a; Sacks,1987). In some ways, then, this could be comparable to a questionnaireitem in which a participant is asked for his or her attitude toward afood. However, differences become immediately apparent when oneconsiders the rest of the extract.One noticeable feature is the structureof Paul’s response (line 5), consisting of an evaluation (“mmm they’realright”) and a description (“they’re a bit spicy”). Each part can beheard to support the other: The “spiciness” accounts for the food beingjust “alright,” which itself projects a probable hearing of “a bit spicy.” Inother words, both description and evaluation are constructed throughreference to the other.Not only does Paul provide an evaluation,he alsogives a reason, or account, for it.

Another point of difference from questionnaire designs occurs whenPaul asks “what are they” (line 8) in reference to the food that he hasjust assessed. Attitudinal research tends to assume that participantsknow the identity of the “attitude object” (e.g., Eagly & Chaiken, 1993,1998), yet here, Paul has been able to evaluate the food without (appar-ently) knowing exactly what the food is.1 Questioning the food in thisway constructs it as being unidentifiable on sight alone and thus ofpotentially dubious quality. Lesley must then account for the sausages,and she does so in terms of familiar foods: “just slices with, with spicesrather than normal ones” (lines 9 to 11). Evaluating a food, then, is also


implicated within the identification and description of the food itself;the nature of the attitude object cannot be taken at face value (Potter &Wetherell, 1988).

What has been demonstrated here is that food evaluations are con-structed interactionally and are bound up with conversational normsand practices. The situated, contextual nature of discourse prevents asimple way of abstracting an underlying attitudinal construct. More-over, because other speakers are also involved in their construction,the notion of evaluations as reflecting individual, mental attitudesbecomes problematic.

VARIATIONS ON AN EVALUATIVE THEME

Conversations are overwhelmingly sites of action, for example,arguing, requesting, and praising. As part of conversation, food evalua-tions are similarly involved in these activities. The following examplesare by no means an exhaustive list but provide an illustration of thevariety and depth of this involvement. Speakers here are doing morethan just evaluating food.

Compliments and Praises

The first example is based on what may seem to be an obviouspoint—that one can compliment a cook by praising his or herfood—though it is exactly its simplicity that is of interest here. The fol-lowing two extracts explicate how this may be realized conversation-ally using food evaluations. In the first, Laura and her 11-year-olddaughter Beth are hosts to their (adult) relatives, Doris and Bill. Atthis point, the family has just finished their main course, and Bill is inthe midst of a previous discussion.

(2) SKW/ G2a-M8 (315-319)1. Doris: → that was lovely Lau:↓ra [thank ↓yo:u2. Bill: [because eh-3. Beth: it is [love:↓ly4. Laura: [>did you enjoy that< there is5. some >d’you ↑want< (0.2) there’s a6. bit mo:re >if you want<

The evaluation on line 1 (“that was lovely”) works as a complimentby providing a reason for the thanking. By referring to the food itselfrather than, say, Doris’s personal taste preferences, the complimentdisplays an “objective” quality. This is also qualified by Beth’s repeat ofthe expression, providing corroboration with the initial praise (line 3).However, it is the explicitness of the evaluation that highlights its use


as a compliment. In saying “that was lovely Laura,” Doris brings atten-tion to the food despite the conversation’s focusing on a different topic(as indicated by Bill, line 2). The direct reference marks it as being anobservable event,as something to be noted (Sacks,1992).Confirmationof this is given by Laura, who by offering more food (see the Fishing forFood section) displays an orientation to the evaluation as being morethan just an evaluation.

The next example illustrates how the use of evaluations as compli-ments may be an issue for the speakers themselves. Here, Simon andAnna are the parents at the meal. Their daughter, Jenny (23 years old),has just introduced a cake for dessert.

(3) SKW/ D2a-M3 (1024-1031)1. Simon: → ve:ry well iced like- (0.2) uhm:: nice2. (0.2) nicely uh:: (0.2) °iced°3. (1.0)4. Jenny: you don’t have to say ↓that Dad5. >cos I didn’t make ↑it<6. Anna: o::h its terrible isn’t it7. Jenny: ye[a::h8. Anna: [Jenny could have done better than this:

Characterized by frequent pauses and “uhm”s, Simon’s initiallystrong evaluation (“very well iced,” line 1) soon appears disjointed andrepetitive. By referring only to the icing, the turn seems to be exagger-ated, and Jenny orients to this as being uttered merely out of an obliga-tion to praise the chef (lines 4 and 5). “Having” to say something com-plimentary about the food suggests that the evaluation is insincere.This construction of “real” versus contrived assessments is taken onestep further in Anna’s turn, in which she makes reference to the cake’sbeing “terrible” (line 6). Situated directly after Jenny’s admittance tonot making the cake, it is hearable as highlighting the potential use offood evaluations as complimentary devices.

Fishing for Food

Analysis of the conversational data further revealed that an offer offood would often follow an evaluation of the item. Because offers gener-ally follow requests (Sacks, 1992), this suggests that the secondspeaker is orienting to the evaluation as if it were an indirect requestfor more food. Returning to the second extract, the focus here is onLaura’s response to the preceding compliments.

(2) SKW/ G2a-M8 (315-319)1. Doris: → that was lovely Lau:↓ra [thank ↓yo:u2. Bill: [because eh-3. Beth: it is [love:↓ly


4. Laura: [>did you enjoy that< there is5. some >d’you ↑want< (0.2) there’s a6. bit mo:re >if you want<

The main point to note here is the construction of both the offers offood and the expression “did you enjoy that,” in personal terms. Byalluding to the praise as being indicative of a personal pleasure, theoffer of food is made relevant by orienting to the individual. This can becontrasted with, for example, talk about qualities of the food itself. It isthrough the interactional features of the conversation that the evalua-tion is subsequently treated as a potential request for more food.

The next extract also highlights this orientation to “fishing” formore food, although there are no explicit compliments in this inter-action. Lesley and Paul are the parents at this family meal, and thissection of talk occurs at the end of the main course.

(4) SKW/ J1a-M2 (515-521)1. (2.0)2. Lesley: °mmm that was a ↑nic:e quiche (.) I ↑liked that°3. (0.6)4. Lesley: °mm-↑hm°5. Paul: → some left6. (1.0)7. Lesley: °mmm (0.2) no not for ↓me (thanks —-)°

The expression “some left” (line 5) is particularly significant here inthe way it makes the extra food relevant to the current conversation.Paul displays an orientation to the evaluation as indicating a potentialdesire for more food. As such, it works as an indirect offer, retrospec-tively constructed by Lesley’s refusal on line 7. In other words, shetreats Paul’s turn as an offer to take more food. As with the compli-ments, it is the explicit nature of the food evaluation that marks it assomething to be attended to.Speakers have the opportunity to orient tofood evaluations because they are part of the conversational order.Exactly how this is done is managed as an interactional concern.

Obligations to Eat

The use of mealtime conversations as data raises issues of eatingand not eating, particularly within a family setting. It is generally thecase that parents are the providers of the food and have a responsibil-ity to make sure that their children eat it. Similarly, children have anobligation to eat the food that their parents have given them (see DeBourdeaudhuij, 1997). What shall be considered here, then, is howevaluations may be used to justify having not eaten one’s food and howthis also highlights potential obligations. The first example of this istaken from near the end of a family meal in which Lesley (the mother)


makes a comment about food left on her son Chris’s plate. Chris is 10years old.

(5) SKW/ J1b-M3 (441-448)1. (6.0)2. Lesley: you ↓struggling Chris?3. Chris: → °don’t like carrots:°4. (1.0)5. Lesley: you don’t like the ↓carrot:s6. Chris: °no°7. Lesley: but you didn’t ↑have man::y8. (4.0)

The relevance of Chris’s evaluation occurs here as an account for noteating (or “struggling” with, line 2) his food. As such, he displays notonly an expression of a “dislike” for carrots but also that there is a needfor him to account for his behavior. Lesley’s subsequent orientation tothe evaluation confirms this, in which she specifies the source of thedislike as being “the carrots” (line 5, emphasis added). In doing so, apotential complaint of the food is constructed. With a particular focuson the evaluation, it appears as if it is the food rather than the evalua-tor that is to blame. Lesley then counters this (line 7) with a referenceto quantity,highlighting again Chris’s obligation to eat all of his meal.

What the above extract shows, then, is that evaluations can be usedto justify not eating one’s food and that this highlights potential obliga-tions of the speakers. To make this point a little clearer, the followingextract shows a rather ironic instance of how an evaluation may beused to negotiate this. Here, the parent-child roles have been reversed,as 23-year-old Jenny teases her father,Simon,about not eating his food(Michael is Jenny’s partner).

(6) SKW/ D5a-M8 (1801-1812)1. Jenny: >have you< finished Dad?2. Simon: yes: (.) thank you very ↓much pet3. Jenny: °oh:: and you haven’t <finished your ↓sprouts>°4. Simon: I’ve- (.) I know [I hav(hh)en’t finished all=5. Jenny: [heh heh heh6. Simon: =m(h)y spro(hh)uts d(hh)e-7. Jenny: well they’re the best ↑bit8. Simon: I know9. Michael: hhhheh10. Simon: → they were ↑actually very good11. Jenny: °hheh heh°12. (2.0)

Despite the element of teasing in this sequence (lines 3 and 5), anaccount is still provided by Simon for not finishing his sprouts. Heachieves this both by displaying an agreement with Jenny (“I know,”


line 8) and by his own evaluation of them as being “actually very good”(line 10). The evaluation here rhetorically defends against possiblecounterarguments (Billig, 1996) by suggesting that it is not distastethat prevents Simon from finishing the food. In particular, the termactually suggests that the evaluation works against expectations. Inagreeing that they are “very good” or “the best bit,” Simon heads off thepotential claim that one may have an obligation to eat one’s food, ir-respective of taste preferences. His account confirms the need for anexplanation, despite the reversal of “roles” in this interaction.

Experience Claims

Another discursive activity associated with evaluations is that ofexpressing knowledge or experience of a particular food or drink(Pomerantz, 1984b). The examples used here are concerned with alco-holic drinks, which are perhaps more suited to such an activity, espe-cially if children or teenagers are involved. Experience with alcohol is aparticularly adult activity associated with certain category-bound fea-tures (Sacks, 1992) such as maturity, responsibility, and independence.Implying that one has drunk alcohol on a number of occasions suggeststhat these qualities can be attributed to the speaker.

The example below demonstrates this role of food evaluations. Inthis case, they have also been used to defend against a challenge to theknowledge claim. Here, 11-year-old Beth asks her mother, Laura, if shecan try some of the red wine that the others are drinking. Also presentat this meal are Beth’s aunt (Doris) and uncle (Bill); Beth is the onlychild.

(7) SKW/ G2a-M8 (740-749)1. Beth: can I try some ↑wi:ne2. Laura: °oh::: (.) (↑mm-hm)°3. (1.0)4. Beth: → don’t [↑like red really5. Laura: [its very nice:6. (1.0)7. Laura: ↑well=8. Bill: =how d’you know (0.6) have you ↑ever tried it9. Beth: I’ve tried it about a ↑million times10. → I hate all ↑red (.) it’s too strong

Beth’s request to “try some wine” (line 1) portrays her here as aminor needing permission to drink alcohol. Her subsequent evaluationon line 4 is therefore a possible anomaly, contrasting sharply with theinitial display of inexperience. The reference to “red” is particularlynoticeable by virtue of its being a familiar and abbreviated expression.By using an evaluation in this way, Beth reconstructs her identity in


terms of an experienced wine drinker, enough, at least, to distinguishbetween red and white varieties.

The evaluation is then taken up by Bill, who directly challenges theexperiential basis for the claim (“how do you know, have you ever triedit,” line 8). Beth’s response to this is to upgrade her evaluation usingboth stronger terms and an objective judgment (“it’s too strong”). Theextreme case formulations here are rhetorical, working to qualify herevaluation and demonstrate a commitment to the claim (Edwards,2000). Having tried it a “million times” and hating “all red” wine areextremely unlikely considering Beth’s age. What these achieve, then, isto defend the identity being constructed by invoking “factual” evidence(Potter, 1996). Therefore, the evaluation has been used here to bothproclaim and support Beth’s status as being familiar with alcohol.

The next example similarly works on the experience-evaluationlink, in which an older daughter expresses knowledge about types ofalcoholic beverages. This family meal comprises Sandra (the mother),her daughters Julie (18 years old) and Amy, and her son Darren.

(8) SKW/ K1a-M2 (1695-1705)1. Julie: you don’t [drink vodka an-2. Amy: [°(I don’t want water)°3. Julie: >↓straight< you drink it neat4. (1.0)5. Darren: >do you<6. Julie: → but its horri↓ble7. (1.2)8. Sandra: °a[h:°9. Amy: [I h:ope [you’ve not trie:d it ↓young lady10. Julie: [↑not on its own:11. (2.0)

A reference to the way that “you drink” vodka (line 3) presents Julieas one who has either drunk it herself or who is particularly knowl-edgeable on the subject. Either way, she may have to account for thebasis of her knowing this, and indeed, Darren suggests that clarifica-tion may be necessary here (line 5). It is at this point that the evalua-tion (“but its horrible,” line 6) constructs and confirms the experience.Amy highlights the implications of this in a style reminiscent of Bill inExtract 7. This, then, becomes the issue for the participants, not theevaluation per se.

CONCLUSION

The use of conversational data raises some key issues for researchconcerned with food evaluations. One of the points that this study has


highlighted is that such evaluations are not freestanding represen-tations but are embedded in and part of the business of discursiveactivities. The extent of this involvement can only be fully appreciatedwhen one considers the rhetorical, interactional nature of discourse,something that is often obscured in more traditional methods.Although recent attitude theories have grown in complexity (Eagly &Chaiken, 1993, 1998), they have yet to capture the sorts of constructiveand action-oriented features of evaluations documented in this study.

The study has also highlighted limitations of questionnaires thatdraw on generic evaluative terms such as good or like. One the featuresthat such questionnaires miss is the way evaluative terms can be con-structed to produce a particular evaluation. For example, there aresubtle but crucial differences in saying that one dislikes “the carrots”(i.e., these particular carrots here) or “carrots” (i.e., the generic foodtype), and these could have important implications for the interactionin terms of whether a criticism or complaint is being made. Also seenwas the way speakers were able to move between evaluations of differ-ent strengths—for example, from “don’t like” to “hate”—when manag-ing a particular identity claim.

The flexible, variable, and constructed nature of evaluative termsmakes it difficult to sustain the notion that a speaker has a particular,enduring, and unitary attitude toward food. There is no “neutral” set-ting within which to establish the nature of the food preference andtherefore to assess whether an assessment is representative or not.The suggestion offered here is that it is more heuristic and closer to thelogic of the materials to move away from the idea of an underlying atti-tude. In evaluating a food discursively, one constructs the notion of“taste” by reference to the food itself. Having a food preference, then,becomes something to be worked up in talk rather than being measur-able through rating scales. Evaluating a food is achieved locally andconversationally in collaboration with other speakers as part of partic-ular tasks. In essence, the evaluation becomes part of the interactionand that from which the meaning is constructed.

The picture of food evaluations as flexible and locally organized hasimplications for the application of this study. The first thing to note isthat there is no expectation that the pattern found here will generalizeto other settings. Quite the contrary, what has been highlighted is theway evaluations are bound up with other actions. In a different setting—dining out with a close friend, say, or eating in a staff canteen—thesorts of actions in which food is embedded might be quite different. Thenormative backdrop may be different (is the food eaten as a one-offactivity or as a daily routine), allowing for different contrastive evalua-tions; is the food to be gotten out of the way or topicalized as a notablepart of an evening? The point of this study is not to show a standardpattern to eating evaluations (although certain patterns might be


identified); rather, it is to start to show how evaluations may have dif-ferent roles.

Discursive work on evaluations has a very different focus from tra-ditional attitudinal research, which has often been concerned withcausal determinants of behavior. The reasons why people eat are notthe concern here. Taking a social, discursive stance, I am interested inhow constructions of food preferences are used in conversation andhow this can shed light on conventional notions of a preference towarda food. Trying to express into words the many varied sensations experi-enced when eating food seems to me to be missing the point. What ismore interesting is how constructions of food (and preferences) canalso construct our relations with others and with the experience of eat-ing itself.

There are two broad sets of implications from this work for healthpromotion and social policy. First, the problems that have been high-lighted for current social psychological approaches to food attitudesraise issues about the sorts of policy suggestions that are made on thebasis of attitude research (see also Eldridge & Murcott, 2000). The sec-ond point is more positive. The identification of food evaluations asbeing bound up with other activities opens up new possibilities forapproaching health policy. For example, research could examine howhealthily eating is managed in practice on a daily basis; how it is man-aged and made accountable as healthy and how it causes interactionaltrouble or becomes a source of dispute. How do people incorporate dif-ferent foods into their diets, and how are these related to interactionalactivities? Could eating “healthy” become a game for children in whichfood is not constructed as “good for you” but as “enjoyable?” Eating isalready a part of our relationships with other people; what could bechanged is how these relationships are constructed.

The nature of food itself is part of this constructive relationship. Thecurrent focus of health research is to promote “healthy eating” habits,but this assumes that what is healthy is relatively unproblematic(although there are different arguments on this matter).What needs tobe addressed is how food is constructed as healthy, locally within set-tings such as family meals or evenings out with friends, and theassumptions these definitions draw on. It also highlights a differentapproach to advertising material, media representations, and educa-tional material. For example, the notion that there are food “groups”(i.e., carbohydrates, protein, fats, etc.) is often used in home economicsclasses in schools and colleges. This constructs a particular abstractversion of “food as nutrition,” which then places limits on what peopleshould or ought to eat. Opening up and examining these types of con-structions and considering the way they are built, ironized, resisted,and ignored in interactions during eating could provide an importantnew basis for considering the value and development of food policy.


APPENDIXTranscription Notation Used in the

Analysis of the Conversations

Notation Description

Don’t Underlining indicates stress or emphasis in the speech.Che::ese Colons are used to represent extended, drawn-out speech.(2.0) Numbers in brackets refer to pauses in tenths of a second. Pauses less

than two tenths of a second long are indicated by (.).(Mine’s) Words in parentheses indicate the transcriber’s best estimate of an un-

clear section of speech.(h) Indicates laughter within speech.[] Square brackets indicate the beginning and end of overlapping talk.= Equal signs indicate continuous talk between speakers.° Degree signs enclose talk that is lower in volume than the surrounding

talk.↑↓ Pointed arrows indicate a marked rising or falling in speech intonation.> < Greater-than and less-than signs enclose speech that is noticeably faster

than the surrounding talk. When the order is reversed (< >), this in-dicates slower speech.

→ Indicates a specific line of the extract discussed in the text.

NOTE

1. Regardless of whether Paul “knows” what the food is, the line “what are they” dis-plays a lack of knowledge and works here as a request for clarification.The concern, then,is not with what participants may “think” about the food but how this is displayed in, orsuggested by, their talk.

REFERENCES

Aaron, J. I., Mela, D. J., & Evans, R. E. (1994). The influences of attitudes, beliefs andlabel information on perceptions of reduced fat spread. Appetite, 22, 25-37.

Ajzen, I. (1988). Attitudes, personality and behavior. Milton Keynes, UK: Open Univer-sity Press.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and HumanDecision Processes, 50, 179-211.

Armitage, C. J., & Conner, M. (1999). The theory of planned behavior: Assessment of pre-dictive validity and “perceived control.” British Journal of Social Psychology, 38,35-54.

Bagozzi, R. P., & Kimmel, S. K. (1995). A comparison of leading theories for the predictionof goal-directed behaviors. British Journal of Social Psychology, 34, 437-461.

Barthes, R. (1979). Toward a psycho-sociology of contemporary food consumption. InR. Forster & O. Ranum (Eds.), Food and drink in history (pp. 166-173). Baltimore:Johns Hopkins University Press.

Beardsworth, A., & Keil, T. (1997). Sociology on the menu. London: Routledge.


Berg, C., Jonsson, I., & Conner, M. (2000). Understanding choice of milk and bread forbreakfast among Swedish children aged 11-15 years: An application of the theory ofplanned behavior. Appetite, 34, 5-19.

Billig, M. (1989). The argumentative nature of holding strong views: A case study. Euro-pean Journal of Social Psychology, 19, 203-223.

Billig, M. (1992). Talking of the royal family. London: Routledge.Billig, M. (1996). Arguing and thinking: A rhetorical approach to social psychology (2nd ed.).

Cambridge, UK: Cambridge University Press.Birch, L. L. (1990). The control of food intake by young children: The role of learning. In

E. D. Capaldi & T. L. Powley (Eds.), Taste, experience and feeding (pp. 116-135). Wash-ington, DC: American Psychological Association.

Blum-Kulka, S. (1994). The dynamics of family dinner talk: Cultural contexts for chil-dren’s passages to adult discourse. Research on Language and Social Interaction, 27,1-50.

Blum-Kulka, S. (1997). Dinner talk: Cultural patterns of sociability and socialisation infamily discourse. London: Lawrence Erlbaum.

Bonham, P., Greenlee, D., Herbert, C. S., Hruidi, L., Kirby, C., Perkins, A., Salkind, N. J., &Wilfong, R. (1995). Knowledge of brand and preference. Psychological Reports, 76,1297-1298.

Burningham, K. (1995). Attitudes, accounts and impact assessment. SociologicalReview, 43, 100-122.

Cantin, I., & Dubé, L. (1999). Attitudinal moderation of correlation between food likingand consumption. Appetite, 32, 367-381.

Clendenen, V. I., Herman, C. P., & Polivy, J. (1994). Social facilitation of eating amongfriends and strangers. Appetite, 23, 1-13.

Conner, M., Martin, E., Silverdale, N., & Grogan, S. (1996). Dieting in adolescence: Anapplication of the theory of planned behavior. British Journal of Health Psychology, 1,315-325.

Conner, M., Povey, R., Sparks, P., James, R., & Shepherd, R. (1998). Understandingdietary change: Contributions from social psychology. In A. Murcott (Ed.), “Thenation’s diet”: The social science of food choice (pp. 43-56). London: Longman.

Counihan, C., & Van Esterik, P. (Eds.). (1998). Food and culture: A reader. London:Routledge.

De Bourdeaudhuij, I. (1997). Family food rules and healthy eating in adolescents. Jour-nal of Health Psychology, 2, 45-56.

Dubé, L., Chattopadhyay, A., & Letarte, A. (1996). Should advertising appeals match thebasis of consumers’ attitudes? Journal of Advertising Research, 36, 82-89.

Eagly,A.H.,& Chaiken,S. (1993).The psychology of attitudes.New York:Harcourt Brace.Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T.

Fiske, & G. Lindzey (Eds.), The handbook of social psychology (Vol. 1, 4th ed., pp. 269-322). New York: McGraw-Hill.

Edwards, D. (1995). Two to tango: Script formulations, dispositions and rhetorical sym-metry in relationship troubles talk. Research on Language and Social Interaction, 28,319-350.

Edwards, D. (1997). Discourse and cognition. London: Sage Ltd.Edwards, D. (2000). Extreme case formulations: Softeners, investment and doing

nonliteral. Research on Language and Social Interaction, 33(4), 347-373.Edwards, D., & Potter, J. (1992). Discursive psychology. London: Sage Ltd.Eldridge, J., & Murcott, A. (2000). Adolescents’ dietary habits and attitudes: Unpacking

the “problem of (parental) influence.” Health, 4, 25-49.Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction

to theory and research. Reading, MA: Addison-Wesley.Geiselman, P. J., Smith, C. F., Williamson, D. A., Champagne, C. M., Bray, G. A., & Ryan,

D. H. (1998). Perception of sweetness intensity determines women’s hedonic andother perceptual responsiveness to chocolate food. Appetite, 31, 37-48.


Herek, G. M. (1987). Can functions be measured? A new perspective on the functionalapproach to attitudes. Social Psychology Quarterly, 50, 285-303.

Hutchby, I., & Wooffitt, R. (1998). Conversation analysis: Principles, practices and appli-cations. Cambridge, MA: Polity.

Lindeman, M., & Stark, K. (1999). Pleasure, pursuit of health or negotiation of identity?Personality correlates of food choice motives among young and middle-aged women.Appetite, 33, 141-161.

Lozano, D. I., Crites, S. L., & Aikman, S. N. (1999). Changes in food attitudes as a functionof hunger. Appetite, 32, 207-218.

Lupton, D. (1996). Food, the body and the self. London: Sage Ltd.Lyman, B. (1989). A psychology of food: More than a matter of taste. New York: Van

Nostrand Reinhold.Maio,G.R.,& Olson,J.M. (1994).Value-attitude-behavior relations:The moderating role

of attitude functions. British Journal of Social Psychology, 33, 301-312.Marshall, H., & Raabe, B. (1993). Political discourse: Talking about nationalization and

privatization. In E. Burman & I. Parker (Eds.), Discourse analytic research: Reper-toires and readings of texts in action (pp. 35-51). London: Routledge.

Martins, Y., Pelchat, M. L., & Pliner, P. (1997). “Try it, it’s good for you”: Effects of tasteand nutrition information on willingness to try novel foods. Appetite, 28, 89-102.

Murcott, A. (Ed.). (1983). The sociology of food and eating. Aldershot, UK: Gower.Netemeyer, R. G., Burton, S., & Johnston, M. (1991). A comparison of two models for the

prediction of volitional and goal-directed behaviors: A confirmatory analysisapproach. Social Psychology Quarterly, 54, 87-100.

Ochs, E., Pontecorvo, C., & Fasulo, A. (1996). Socializing taste. Ethnos, 61, 7-46.Pomerantz, A. M. (1984a). Agreeing and disagreeing with assessments: Some features of

preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structuresof social action: Studies in conversation analysis (pp. 57-101). Cambridge, UK: Cam-bridge University Press.

Pomerantz, A. M. (1984b). Giving a source a basis: The practice in conversation of telling“how I know.” Journal of Pragmatics, 8, 607-625.

Pontecorvo,C.,& Fasulo,A. (1999).Planning a typical Italian meal:A family reflection onculture. Culture & Psychology, 5, 313-336.

Potter, J. (1996). Representing reality: Discourse, rhetoric and social construction. Lon-don: Sage Ltd.

Potter, J. (1998a). Discursive social psychology: From attitudes to evaluative practices.European Review of Social Psychology, 9, 233-266.

Potter, J. (1998b). Discourse analysis as a way of analyzing naturally occurring talk. InD. Silverman (Ed.), Qualitative research: Theory, method and practice (pp. 144-160).London: Sage Ltd.

Potter,J.,& Edwards,D. (2001).Discursive social psychology. In W.P.Robinson & H.Giles(Eds.), The new handbook of language and social psychology (pp. 103-118).Chichester, UK: Wiley.

Potter, J., & Wetherell, M. (1987). Discourse and social psychology: Beyond attitudes andbehavior. London: Sage Ltd.

Potter, J., & Wetherell, M. (1988). Accomplishing attitudes: Fact and evaluation in racistdiscourse. Text, 8, 51-68.

Puchta,C.,& Potter,J. (1999).Asking elaborate questions:Focus groups and the manage-ment of spontaneity. Journal of Sociolinguistics, 3, 314-335.

Sacks, H. (1987). On the preferences for agreement and contiguity in sequences in con-versation. In G. Button & J.R.E. Lee (Eds.), Talk and social organization (pp. 54-69).Clevedon, UK: Multilingual Matters.

Sacks, H. (1992). Lectures on conversation (G. Jefferson, Ed.) (Vols. 1 & 2). Oxford, UK:Basil Blackwell.

Scapp, R., & Seitz, B. (Eds.). (1998). Eating culture. Albany: State University of New YorkPress.


Schifter, D. E., & Ajzen, I. (1985). Intention, perceived control and weight loss: An appli-cation of the theory of planned behavior. Journal of Personality and Social Psychol-ogy, 49, 843-851.

Sparks, P., Hedderley, D., & Shepherd, R. (1992). An investigation into the relationshipbetween perceived control, attitude variability and the consumption of two commonfoods. European Journal of Social Psychology, 22, 55-71.

Sparks, P., & Shepherd, R. (1992). Self-identity and the theory of planned behavior:Assessing the role of identification with “green consumerism.” Social PsychologyQuarterly, 55, 388-399.

Verkuyten, M. (1998). Attitudes in public discourse: Speakers’ own orientations. Journalof Language and Social Psychology, 17, 302-322.

Wetherell, M., & Potter, J. (1992). Mapping the language of racism: Discourse and thelegitimation of exploitation. London: Harvester Wheatsheaf.

Wiggins, S., Potter, J., & Wildsmith, A. (2001). Eating your words: Discursive psychologyand the reconstruction of eating practices. Journal of Health Psychology, 6, 5-15.


JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001Stahlberg et al. / EFFECTS OF GENERICS

NAME YOUR FAVORITE MUSICIANEffects of Masculine Generics and

of Their Alternatives in German

DAGMAR STAHLBERGSABINE SCZESNYUniversity of Mannheim

FRIEDERIKE BRAUNUniversity of Kiel

This article reports on two experiments with native speakers of German that were con-ducted to determine the influence of different types of German generics on the cognitiveinclusion of women. The results of these studies show that masculine versus other types ofgenerics influence the retrieval of male and female exemplars from memory. This is thefirst piece of empirical evidence for this kind of effect with regard to the German language.

Masculine generics have been the central issue in the debate about“sexist” and “nonsexist” language that was raised by feminist languagecritics in the 1970s (e.g., Miller & Swift, 1977). When general state-ments are formulated in the masculine, critics argue, women are lin-guistically ignored and become invisible (e.g., “a typical doctor . . . he”).An impressive body of empirical research indicates that in English,masculine generics evoke predominantly “male” associations and inconsequence put women at a disadvantage (e.g., Hamilton, Hunter, &Stuart-Smith, 1992; Moulton, Robinson, & Elias, 1978). Little empiri-cal research, however, has been done on other languages. It is doubtfulwhether the findings for English are valid for languages that are struc-turally rather different.

The present contribution reports on empirical research on mascu-line generics in German. The linguistic situation in German is very dif-ferent from that in English. Every German noun has a specific gender(feminine, masculine, or neuter) that is marked on articles, adjectivalsuffixes, and various types of pronouns. As a consequence, where mas-culine generics occur, they concern much more linguistic forms, andmasculine markings are much more frequent in a text. The frequency

464

AUTHORS’ NOTE: Thanks are extended to Constantin Dony, Sabine Kindsvater, JensKowalski, Manuela Langkath, Frederik Pohl, and André Siegel for their assistance withdata collection and to Geoff Haig for his advice on the English translation.


of masculine markers in German might thus intensify male associa-tions and produce a stronger male bias than in English. On the otherhand, because every German noun has a gender—even nouns such asZahnbürste (feminine), “toothbrush,” or Funken (masculine),“spark”—speakers of German might perceive the semantic tie betweengender and sex as weaker.1

In German-speaking countries, as elsewhere, feminists (e.g.,Grabrucker, 1993) demand that masculine generics be replaced withnonsexist alternatives, for which two different strategies are proposed:neutralization and feminization.

Neutralization relies on linguistic forms that do not express sex.These are nouns of neutral gender (e.g., das Individuum [neuter], “theindividual”), nondifferentiating forms (e.g., die Angestellten, “theemployees” [plural of the feminine die Angestellte, “the employee,” aswell as the masculine der Angestellte, “the employee”]), forms withfixed gender that refer to both women and men (e.g., die Person [femi-nine], “the person,” or der Mensch [masculine], “the human being”), orcollectives (e.g., das Personal [neuter], “the staff”).

The term feminization, on the other hand, is used for expressionsthat make the inclusion of women explicit. Feminization thus involvesusing feminine-masculine word pairs, for example, Lehrerinnen undLehrer, “female and male teachers,” or the so-called capital I forms. Thecapital I form was invented in feminist circles to substitute masculinegeneric plurals of the type Leser (masculine), “readers,” at least in writ-ing. The new form is created by using a feminine plural as the baseform (e.g., Leserinnen [feminine], “readers”) and capitalizing the initiali of the suffix to highlight the generic function. The resultingLeserInnen, “readers,” closely resembles the feminine pluralLeserinnen but is meant to stand for Leserinnen und/oder Leser,“female and/or male readers.” Nowadays, capital I forms are used incertain newspapers and magazines, in unofficial messages or letters,and in a few scientific publications, but they are generally not acceptedfor official usage.

Until now, there have been very few studies on the effects of the dif-ferent types of generics in German (Braun, Gottburgsen, Sczesny, &Stahlberg, 1998; Irmen & Kaczmarek, 2000; Irmen & Köhncke, 1996;Scheele & Gauler, 1993). These studies indicate that different mentalrepresentations are triggered by different generic terms for person ref-erence depending on their grammatical gender. For example, the pro-portion of women and men in a given group is estimated differently,depending on the generic forms used to describe that group (Braunet al., 1998). Also, longer time spans were needed to read sentenceswith feminine references that referred back to masculine generic sen-tences than to complement sentences without “gender switches”(Irmen & Kaczmarek, 2000).

Stahlberg et al. / EFFECTS OF GENERICS 465

The present studies were planned to investigate the influence ofmasculine versus alternative types of generics on the retrieval of maleand female exemplars from memory. Our general hypothesis was thatmasculine generics would facilitate the retrieval of male exemplars,compared with alternative generics (such as neutralizing forms orforms that refer to women explicitly), but would impede the retrieval offemale exemplars. Furthermore, we expected that different alterna-tive generics would trigger significantly different responses dependingon how explicitly women were referred to.

In Experiment 1, respondents were asked to name their favoriteheroes, favorite musicians, and so on. The hypothesis was that mascu-line generics would trigger fewer “female” responses than alternativeforms. Moreover, neutralizing forms were expected to trigger fewerfemale responses than feminine-masculine word pairs. The first factorwas this type of generic. The second factor, sex of participant, wasincluded to analyze whether the expected effects of type of genericwould be independent of the sex of participants.

Participants (50 female and 46 male native speakers of Germanbetween 17 and 58 years of age) filled out a questionnaire that wasmodeled on a list of questions used by a national German newspaper tointerview prominent persons. The questionnaire contained 16 ques-tions such as “What would you personally consider a severe tragedy?”The six critical questions targeted the participants’ favorite heroes innovels, real life, and history and their favorite painters, musicians, andathletes. The questionnaire was presented in three different genericlanguage versions: masculine (e.g., Romanheld [masculine], “hero in anovel”) versus neutralizing (forms not differentiated for gender, e.g.,heldenhafte Romanfigur, “heroic character in a novel”)2 versus feminine-masculine word pairs (e.g., Romanheldin [feminine] oder Romanheld[masculine], “heroine or hero in a novel”). The number of womenreported in response to the critical questions (summarized over all sixquestions) was the dependent variable.

Data were analyzed with a 3 (Type of Generic) × 2 (Sex of Partici-pant) factorial ANOVA with the number of women reported as thedependent variable. The ANOVA showed a significant main effect fortype of generic, F(2, 90) = 4.93, p < .01. The contrast indicated that mas-culine generics (M = 0.67) triggered fewer female responses than alter-native formulations (neutralizing generics or feminine-masculineword pairs, M = 1.67, p < .05). No significant difference was foundbetween the two conditions, neutralizing and feminine-masculinepairs (p = ns). The main effect for the sex of participant variable wasalso significant, F(1, 90) = 12.53, p < .01, with female participants men-tioning more women than male participants (M = 1.81 vs. M = 0.83).There was no interaction of the type of generic and sex of participantvariables, F(2, 90) = 0.98, p = .38.


Experiment 2 was planned as a replication of Experiment 1 withslightly differing material. In Experiment 1,we basically asked for per-sonal preferences, whereas in Experiment 2, we were interested in thecognitive availability of female and male exemplars depending on thegenerics used in the questions. Therefore, participants were asked toname several athletes, singers, and other people described with differ-ent types of generics. Masculine generics were tested against twoforms of the feminizing type: feminine-masculine word pairs and capi-tal I forms. It was hypothesized that the alternative generics wouldraise the number of women reported compared with masculinegenerics. In addition, we expected that capital I forms would triggerthe highest number of women reported, for their near identity withfeminine generics combined with their relative novelty and theirorthographic exceptionality made their nonsexist intention especiallysalient.

Participants (45 female and 45 male native speakers of Germanbetween 18 and 45 years of age) filled out a questionnaire that claimedto investigate the effects of media consumption on the recall of promi-nent people. The questionnaire contained questions about the partici-pants’ interest in and occupation with the media. The critical itemswere embedded in a section that asked participants to name the threefamous people in a given category who first entered their minds:“Please name three athletes (or singers, politicians, hosts of televisionshows).” Three types of generics were used in the three different ver-sions of the questionnaire: (a) masculine generics (e.g., Politiker [mas-culine], “politicians”), (b) feminine-masculine word pairs (e.g.,Politikerinnen [feminine] und Politiker [masculine], “female and malepoliticians”), and (c) capital I forms (e.g., PolitikerInnen, “politicians”).Participants were randomly assigned to one of these language condi-tions. The number of women reported in response to the four criticalquestions (summarized over all four questions) was the dependentvariable.

Data were analyzed with a 3 (Type of Generic) × 2 (Sex of Partici-pant) factorial ANOVA with the number of women reported as thedependent variable. The ANOVA revealed a main effect for type ofgeneric, F(2, 84) = 9.97, p < .001. Masculine generics triggered the few-est female responses, whereas feminine-masculine word pairs andespecially capital I forms made participants respond with more femalenames (M = 2.37, M = 2.67, and M = 4.60, respectively). The contrastbetween masculine generics on one hand and the two alternativegenerics on the other was statistically significant (p < .01).3 Furtherpost hoc comparisons, however, showed that capital I forms and wordpairs apparently had markedly different effects. Whereas capital Iforms made participants respond with significantly higher numbers ofwomen than masculine generics (Scheffé’s post hoc test; p < .01), the


difference between word pairs and masculine generics did not reachsignificance (Scheffé’s post hoc test; p > .10). Another main effect wasfound for the sex of participant variable, F(1, 84) = 19.9, p < .001.Female respondents mentioned more women than male respondents(M = 4.2 vs. M = 2.2). There was no interaction between sex of partici-pant and type of generic, F(2, 84) = 0.61, p > .5.

Taken together, these results indicate that different linguistic formsindeed make a difference: Wherever generics produced significanteffects, masculine generics caused the lowest number of female exem-plars to be retrieved from memory, whereas alternative forms facili-tated the retrieval of female exemplars. Our results thus confirm theassumption of feminist language critics that masculine generics havedetrimental effects on the cognitive inclusion of women and that alter-native forms are better suited to make hearers or readers think offemale people. The masculine gender of traditional German genericsapparently has a semantic component of “maleness” that makes theseforms less compatible with the idea of female reference. Thus, theinsight we could gain from the present studies is that masculine versusother types of generics influence the retrieval of male and femaleexemplars from memory.This is the first piece of empirical evidence forthis kind of effect with regard to the German language.

As expected, the various nonsexist alternative forms did not alwaysyield the same results. Contrary to our hypothesis, though, neutraliz-ing forms were as effective retrieval cues for female exemplars asfeminizing ones in Experiment 1. The findings in Experiment 2 sup-port the hypothesis that capital I forms enhance the retrieval of femaleexemplars relative to masculine generics, whereas no comparableenhancement was found for feminine-masculine word pairs.

In both experiments, sex of participant was included as a second fac-tor to analyze whether the hypothesized effects of type of generic wouldbe moderated by participants’ sex. In both experiments, we did not finda Type of Generic × Sex of Participant interaction. However, bothexperiments revealed a main effect of sex of participant.Female partic-ipants named more female exemplars than males. This higher salienceof one’s own sex replicates previous findings for the English language(see, e.g., Moulton et al., 1978; Prentice, 1994).

NOTES

1. The investigations by Zubin and Köpcke (1984) and Konishi (1993), however, sug-gest that even with nonhuman nouns, gender is in some way associated with gender orgender stereotypes.

2. Although the grammatical gender of Figur, “character,” is feminine in German, theword can be used for male as well as female referents. It is not differentiated for gender.


3. Four participants responded to capital I forms with female names exclusively. Be-cause these participants might have mistaken capital I forms for feminine generics (i.e.,sex-specific forms), statistical analyses were repeated without the respective partici-pants. The exclusion of these doubtful cases did not, however, lead to significant changesin the results.

REFERENCES

Braun, F., Gottburgsen, A., Sczesny, S., & Stahlberg, D. (1998). Können GeophysikerFrauen sein? Generische Personenbezeichnungen im Deutschen. Zeitschrift fürGermanistische Linguistik, 26, 265-283.

Grabrucker, M. (1993). Vater Staat hat keine Muttersprache. Frankfurt, Germany:Fischer.

Hamilton, M. C., Hunter, B., & Stuart-Smith, S. (1992). Jury instructions worded in themasculine generic: Can a woman claim self-defense when “he” is threatened? In J. C.Chrisler & D. Howard (Eds.), New directions in feminist psychology: Practice, theory,and research (pp. 169-178). New York: Springer.

Irmen,L.,& Kaczmarek,N. (2000,September).Beeinflusst das grammatische Geschlechtdie Repräsentation von Personen in einem mentalen Modell? Ein Vergleich zwischeneiner englischsprachigen und einer deutschsprachigen Stichprobe. Poster presentedat the 42nd Congress of the German Psychological Association, University of Jena.

Irmen, L., & Köhncke, A. (1996). Zur Psychologie des “generischen” Maskulinums.Sprache & Kognition, 15, 152-166.

Konishi, T. (1993). The semantics of grammatical gender: A cross-cultural study. Journalof Psycholinguistic Research, 22, 519-534.

Miller,C.,& Swift,K. (1977).Words and women:New language in new times.Garden City,NY: Anchor.

Moulton, J., Robinson, G. M., & Elias, C. (1978). Sex bias in language use: “Neutral pro-nouns that aren’t.” American Psychologist, 33, 1032-1036.

Prentice, D. A. (1994). Do language reforms change our way of thinking? Journal of Lan-guage and Social Psychology, 13, 3-19.

Scheele, B., & Gauler, E. (1993). Wählen Wissenschaftler ihre Probleme anders aus alsWissenschaftlerInnen? Das Genus-Sexus-Problem als paradigmatischer Fall derlinguistischen Relativitätsthese. Sprache & Kognition, 12, 59-72.

Zubin, D. A., & Köpcke, K. M. (1984). Affect classification in the German gender system.Lingua, 63, 41-96.


JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001BOOK REVIEWS

Book Reviews

Pour Une Écologie Des Langues Du Monde. Louis-Jean Calvet. Paris:Plon, 1999. 304 pp. ISBN 2-259-18975-X (paper).

Calvet’s latest book,Pour une écologie des langues du monde (For an Ecologyof the Languages of the World), as the title suggests, is a defense of what he callsan ecological approach to the analysis of the world’s languages. The book issubdivided into six main chapters in addition to an introduction and a conclu-sion. The latter is followed by a selected bibliography. Given that Calvet alsopresents in his book a large number of terms,many of which are familiar to spe-cialists of sociolinguistics (e.g., écologie, pidgin, créole, vernaculaire) but someof which are less familiar (e.g., acclimatement, a process of removal andreadaptation for a species, and acclimatation, a process of readaptation andreproduction; pp. 142, 213), an index at the end of the book listing the terminol-ogy would have been a useful adjunct for both students and specialists. How-ever,Calvet does indicate many of the terms in italics in the body of the text.

The content and the order of the chapters are built on four structuralsubmodels (one model for each of chapters 2 through 5), which Calvet presentsas his general ecological model of the languages of the world. The submodelsare first listed and defined in the introduction (p. 16):

1. the gravitational or macro submodel,based on a “galaxies and constella-tions” analogy of languages in the world language situation;

2. the homeostatic or regulatory submodel, based on an autoregulation ofoutside stimuli, which results in a tendency to neutralize, change, ormaintain languages and language situations in place;

3. the representational or what could be called the social psychologysubmodel, which the author terms “épilinguistique” (p. 158). These rep-resentations refer to “la façon dont les locuteurs pensent des pratiques,comment ils se situent par rapport aux autres locuteurs, comment ils sesituent par rapport aux autres langues en présence [the way in whichspeakers think of their everyday usage, how they situate themselveswith regard to other speakers, how they situate themselves with regardto other languages present]” (p. 158); and

4. the transmission submodel, an explanatory model that accounts forchange and continuity (evolution) in tandem with the two previoussubmodels. All three in combination allow us to understand change interms of the overarching or macro (gravitational) submodel but on ascale extending from the macro to the micro spheres.

The final chapter covers five case studies, which further emphasize, largelythrough a descriptive, anecdotal approach, some of the conceptual and theoret-ical questions dealt with in earlier chapters:

470


1. the problem of polycentric languages such as Arabic and the fixing oftheir inner boundaries, including the proper identification of each form,both spoken and written;

2. the problem of nomenclature or the use of several glossonyms for thesame or seemingly the same speech form;

3. the problem of bicentricity, whereby it is difficult to determine clearlywhether one or two languages are involved (e.g., Serbo-Croatian,Hindi-Urdu, etc.);

4. the introduction of normative French into a microlinguistic contact situ-ation (niche linguistique) in which the regular languages are Englishand Louisiana Creole and the effects of this introduction on both theforms and functions of the languages present; and

5. the introduction into the island of Saint Barthélemy of a French patois(along with normal French). Because of the geographical isolation notonly of the island but of its internal geography (its windward and lee-ward parts), only the windward side of the island developed a creolizedvariety of French, because of the arrival of a significant number of Blackimmigrants.

PRATIQUES ET REPRÉSENTATIONS(PRACTICE AND REPRESENTATION)

Chapter 1, called “L’écologie des langues” (“The Ecology of Languages”;pp. 34-74), presents both a definition of the term ecology and a defense ofCalvet’s application of the ecological model to the science of language. Hisobjective, Calvet argues, is scientific rather than political. “La référence àl’écologie n’est pas synonyme de défence d’espèces menacées, mais implique larecherche d’un modèle explicatif [The use of the term ecology is not synony-mous with the defense of a threatened species but implies the search for anexplanatory model]” (p. 34). The above model, then, constitutes the central aimof the book and is also the main object of this review.

The importance of the model is further emphasized in the introduction,called“Pratiques et représentations” (“Practice and Representation”), which pres-ents a binary framework (cadre), with various degrees of emphasis on oneaspect or the other, in all four submodels. For example, pratiques (languagepractices or usage) is language use, and représentations is an “interpretationor representation” on the part of the speakers and listeners of these practicesor usages. Because we cannot deal with languages out of context, especially inan ecological model, the chapter would perhaps have better been entitled“Situations-pratiques-représentations.”Calvet does include all these elementsin his equation, which he says amounts to a model of social communication.“L’articulation ent re les situations, les pratiques et les représentations relèved’un modèle de la complexité de la communication sociale qui reste à élaboreret auquel est consacré ce livre [The interrelating of situations, practices, andinterpretations suggests a complex model of social communication, which stillrequires elaboration and to which this book is dedicated]” (p. 17). To be produc-tive, such a model would first have to describe language form (structure) aswell as language function (usage) and changes in both. It would then have toidentify and measure the sources of the changes. Are both forms and functionsinfluenced by situations and representations “de savoir si le milieu a une

BOOK REVIEWS 471

influence sur les pratiques, sur la forme des langues [to know if the milieu hasan influence on usage, on the form of languages]” (p. 281)?

The author further explains some, but not all, of these elements in his defi-nition of linguistic ecology:

L’écologie linguistique étudie les rapports entre les langues et leurmilieu, c’est-à-dire d’abord les rapports entre les langues elles-mêmes,puis entre ces langues et la société, mais il s’agit de construire un modèlethéorique dépassant l’opposition articifielle entre linguistique etsociolinguistique, il s’agit d’intégrer les langues dans leur contexte social[Linguistic ecology studies the relationship between languages and theirmilieu, that is to say, first the rapport among languages themselves, thenbetween these languages and the society, but it is a question of construct-ing a theoretical model which goes beyond the artificial oppositionbetween linguistics and sociolinguistics, it is a question of integratinglanguages into their social context]. (p. 17)

First, in the language-in-society research, we have now gotten well beyondthe dichotomy of linguistics and sociolinguistics as mentioned by Calvet andprogressed to what I would term a language in society and social psychological(sociolinguistic) dichotomy. As two separate traditions, these have already mademuch progress in explaining language and societal change, but unfortunately,each has remained tradition bound and has not taken the next obvious step inuniting the two traditions in a combined research effort, to be able to accountfor change not only on each of the macro and micro levels but also on change asinfluenced by both situations (language in society) and representations (socialpsychology). This Calvet may be darkly hinting at when he suggests first astudy of “languages among themselves”; otherwise it makes little sense in themodel of social communication that he proposes here. Second, when we speakof “studying language in society,” it is useful to distinguish (as does the literaturein great detail) whether we are dealing with situational- or representational-type research. This Calvet does in his four-tiered general model: His firstsubmodel deals with situations, whereas his third submodel deals clearly withrepresentations.

Calvet is not blind to the fact that in transplanting the ecological modelfrom the biological sciences to the language sciences, he (and others) is makinga transition that only partly “fits” or that can fit only in a metaphorical sense:“Les langues n’ont pas de vie, de famille, ne sont ni des instruments, ni desorganismes extérieurs à ceux qui les utilisent [Languages have no life, or fam-ily, nor are they instruments, nor organisms outside of those who use them]”(p. 16). This theme of the nonobjectivity of languages, “la langue est une pra-tique sociale au sein de la vie sociale, une pratique parmi d’autres, inséparablede son environnement [language is a social practice at the center of social life,one among many practices, inseparable from its environment]” (p. 33), is com-pounded by Calvet’s denial of their reality: “Les langues n’existent pas, lanotion de langue est une abstraction reposant sur la régularité d’un certainnombre de faits, de traits, dans les productions des locuteurs, dans leur pra-tiques [Languages do not exist, the notion of language is an abstraction basedon the regularity of a certain number of facts and traits in the speech of thespeakers and in their practices]” (p. 281).


Although language is not a three-dimensional object (nobody has seen onewalking on the streets), philosophically, one cannot deny the reality of thingssimply because they are not three-dimensional objects. For example, it wouldbe folly to deny the existence of hate or love simply because we have not“touched” them. The manifestations of both hate and love are indeed every-where around for us to “see.” Language does, however, have a halfway realitybetween three-dimensional objects and abstract concepts whose manifesta-tions only we can see. Language we can definitely hear and in its most conven-tional form see in writing. These manifestations of language are less concretethan any three-dimensional object but more concrete than completely abstractconcepts.The fact that language is an abstraction of sorts with boundaries thatare flexible cannot permit us to say it does not exist.

But, Calvet is correct in remarking that languages, however they aredenoted or named, are a relative construct and that their limits can be difficultto define because different sets of criteria may be applied according to variousobjectives to operate a separation. For example, some linguists and languageteachers may advance the criterion of comprehension as a most important ped-agogical criterion for their pupils. Ethnographers may see things differentlyand encourage a specific group representation of what their language is orshould be. If language variants are relatively undisturbed by mass populationmovements, one would get a rainbow effect, with one color shading intoanother almost imperceptibly. It has been said that if a traveler in India wereto walk across the nation from east to west, he would find no point where com-munication would be broken; that is, wherever the traveler was situated,speakers would be able to communicate with those immediately to the east andwest of them. So, where can the language boundaries be drawn if the criterionof comprehension is applied? The language census of India (which is now over100 years old) has used a rather wise approach in the gathering of languagedata. First, the name of the language (whether mother tongue or second lan-guage) is solicited directly from the speaker; that is, no language name is sug-gested by the census official. These names, which are called mother tonguesand so on, are later classified or grouped by linguists into “languages.” Hence,mother tongue figures are based on a grouping of freely solicited names basedon sociopolitical identities established within communities. In the 1961 cen-sus, this resulted in over 1,600 mother tongue names, which were later classi-fied by the census department into 200 languages. Each set of figures (mothertongue/language) can of course produce its own set of boundaries,and both setscan be justified or rationalized according to different objectives.

THE THEORETICAL MODEL

As Calvet stresses at the outset the importance of his conceptual model,“Encore une fois, la linguistique ne peut pas faire l’économie d’une théorieglobale de la communication sociale [To repeat, linguistics cannot do without aglobal theory of social communication]” (pp. 23-24). It is important to examinethis ecolinguistic model in terms of its capacity not just to describe ecologicallythe language situation of the world but especially to demonstrate the

BOOK REVIEWS 473

mechanism of language change. My approach here is to examine each ofCalvet’s four submodels with the above in mind.

THE GRAVITATIONAL SUBMODEL

In my opinion, it makes good sense to start, as Calvet does, with thesubmodel that is the most encompassing both spatially and linguistically.Here, we are analyzing languages in broad spatial terms. Geographically, weare dealing with continents and countries and linguistically with broadlydefined language units, that is, ones that the Indian census would term lan-guage units. According to Calvet, this gravitational submodel is “destiné àrendre compte de la situation linguistique mondiale, des rapports macro-sociolinguistiques entre les différentes langues [destined to take into accountthe world language situation, the macrosociolinguistic relationships betweendifferent languages]” (p. 16).

This submodel is further developed in chapter 2, “La galaxie des langues”(pp. 75-99), in terms of an analogy (galaxies and constellations) used by deSwann (1993) and others and based on a principle of attraction between cen-ters and peripheries; that is, languages in peripheral positions are attracted tothose in central positions. The glue that holds this all together is the gravita-tional pull by the larger bodies on the smaller ones in the case of constellationsand, by analogy, in the case of languages, the desire or need of speakers insmaller language networks to communicate with speakers in larger languagenetworks. Calvet notes that bilingual speakers are usually bilingual in eitheran equal (horizontal bilingualism) or larger (vertical bilingualism) languagenetwork (p. 78).

From this constellational model, Calvet develops a typology of languages onfour levels, which he denotes as (a) hypercentral, (b) supercentral, (c) central,and (d) peripheral. Visually, this hierarchy is pyramidal in shape, with one lan-guage (English) at the summit and more and more languages on each consecu-tive descending level. It is easy enough to classify English at the summit and4,000 to 5,000 languages at the base of the pyramid (i.e., at both extremes), butif one is to do a careful ranking of many languages, it would be difficult to deter-mine where to place many specific cases, especially those on the supercentraland central (or middle) levels. For example, Calvet places Hindi and Malay onthe second or supercentral level and German on the third or central level alongwith another European language, Czech. One could easily argue that Germanis just as important as Hindi or Malay and much more important than Czechand that German should rather be placed on the second or supercentral level.Why it is not, in spite of its number of speakers, is not clearly explained byCalvet. To do a serious rather than ad hoc classification, what one reallyrequires is a good system of ranking based on quantitative measurements.This Calvet unfortunately does not provide, apart from the comment that“mais le nombre important de locuteurs ne suffit pas à conférer le statut delangue super-centrale [but the number of speakers is not sufficient to confersuper-central status on a language]” (p. 79). I would agree with Calvet that it isnot only the number of speakers that confers status,but if not speakers,what isit exactly? What other criteria are involved? Unfortunately,Calvet does not tellus, but he tries perhaps to grapple with this problem in the same chapter byexplaining “le problème de la dynamique de ce système [the problem of the


dynamics of this system]” (p. 81). He points to demography, conquest, theexpansion of religion, and conversions and commerce as possible triggers ofchange. Calvet speaks of an internal dynamic that rewards already-powerfullanguages: “Plus une langue a de la valeur et plus elle en acquiert [the morevalue a language has, the more value it acquires]” (p. 81). Finally, he states thatthis value is acquired through the number of users: “Plus elle (la langue) ad’utilisateurs et plus elle augmente sa valeur [the more users a language has,the more its value increases]” (p.81).First, it is difficult to follow this argumentbecause of Calvet’s use of alternate terms such as status and value (statut andvaleur) and speakers and users (locuteurs and utilisateurs), which are notclearly defined,although one could presume they have different meanings sim-ply because of the choice of alternate terms. For example, if speakers and usersboth refer to almost the same thing, namely, speakers and locuteurs, then Iwould not agree with Calvet’s statement “plus elle (la langue) a d’utilisateurset plus elle augmente sa valeur” (p. 81). Second, this explanation of the dynam-ics of the situation does not really help shed light on the above problem of theranking of the world’s languages, and it is too schematic to help us betterunderstand the whole process of language change in the macro sphere.

Under the subheading “La constellation hindie” (“The Hindi Constellation”;p. 85), and in keeping with the overall galactic model, Calvet introduces thethree-dimensional figures of my unipolar and multipolar language constella-tions, showing two states in India (pp. 86-87; there also exists a third constella-tion without poles, which he also mentions in the text but does not show as afigure). In my article, I demonstrated that constellations are constructed noton the basis of the number of speakers but on the basis of the uses and func-tions of a language and that a shift in the composition of the constellation (e.g.,from unipolar to multipolar) may have a bearing not only on the distribution ofthe functions but also on the number and types of significant situational vari-ables that come into play. The total number of speakers or the number ofbilinguals may be a highly significant variable in one constellation but less soin another. What is important in my view is that these uses and functions arequantified in terms of vitality ratings (dependent variable),which are based onan analysis of various situational variables (independent variables) accordingto constellation type. In my article, these variables were demographic in type,but other types of variables could also be added. It is unfortunate that Calvet,in reproducing my table and figures, has removed the vitality ratings, whichare basic to the overall exposé. My quantified approach could be helpful notonly in understanding the dynamics of the constellation model but also in cre-ating a quantified ranking of languages through the use of vitality ratings. Myapproach also helps clarify such terms as speakers and users on one hand andfunctions and uses on the other. In Calvet’s (situations)-pratiques-représentations, the above quantificational and analytical approach could shedlight on the situations-pratiques part of his equation.

Later in chapter 2, Calvet devotes a section to writing systems, “La galaxiedes écritures” (“The Galaxy of Writing Systems”), and in doing so continues touse the same analogy of galaxies and hierarchies that he uses for languages.He correctly notes that any number of writing systems can depict the sounds ofa single language and that the same writing system,when used for several lan-guages, modifies its sounds accordingly for representing the same grapheme.Just as English dominates as the hypercentral language in Calvet’s galaxy of

BOOK REVIEWS 475

languages, so is Latin the hypercentral script in his galaxy of scripts. However,there seems to be no supercentral level (as there is for languages) but ratheronly a three-level hierarchy for scripts, namely, (a) hypercentral, (b) central,and (c) peripheral, and these are shown in his figure on page 96. Calvet’s repre-sentations of Tamil and Hindi are misleading in this figure. Because Tamil hasa written history of over 2,000 years, its script cannot be considered a directdevelopment of the Arabic script, as is the case for Urdu. Although I am noexpert on scripts, it is probable that both Arabic and Tamil have an earlier com-mon origin in one of the Semitic alphabets, perhaps Kharoshthi, which was aderivation of Aramaic and used throughout northwest India during the days ofPersian rule.As for Hindi,being an offshoot of Sanskrit, it should appear underSanskrit rather than on the same line.Putting them together,as Calvet does, islike placing Latin alongside Portuguese. Finally, the supercentral languagethat is denoted by Devanagari is not Hindi, as claimed by Calvet (“ledévanagari note le hindi [Devanagari denotes Hindi]”; p. 97), but is, of course,Sanskrit.

THE HOMEOSTATIC SUBMODEL

Conceptually, it is somewhat difficult to locate this submodel (i.e., in termsof a macro-micro scale), but in mentioning “la régulation des situationslinguistiques et des langues [the regulation of language situations and lan-guages]” in his definition (p. 16), Calvet might appear to be conceptually in theabove situations-pratiques part of the equation. A careful reading of the chap-ter, however, suggests that the représentations part of the equation is also in-cluded, in that we are not just including social functions but also the forming ofthese social functions according to social needs: “Ce modèle implique donc quel’on distingue entre les besoins linguistiques de la société et les fonctionssociales de la langue [this model implies then that one distinguishes betweenthe linguistic needs of the society and the social functions of language]” (p.101).

This submodel is depicted as a rheostat that controls change as part of aninner dynamic, especially from outside forces that are wont to create imbalance.It is “destiné à rendre compte de la régulation des situations linguistiques etdes langues [destined to take into consideration the regulation of language sit-uations and languages]” (p. 16). Calvet further explains,

Le modele homéostatique ou modele d’autorégulation implique unprocessus de réponse à des stimuli extérieurs qui peut engendrer lechangement: la régulation est une réactionà un stimulus extérieur parun changement interne qui tend à neutraliser les effets de ce stimulus[The homeostatic or self-regulated model implies a response process tooutside stimuli that can create change: regulation is a reaction to an out-side stimulus by an internal change,which tends to neutralize the effectsof the stimulus]. (p. 100)

Calvet then gives several examples of how particular language functions(e.g., langue véhiculaire) are the result of a specific language form coming outof a milieu where there is no common language but are also a response to a so-cial need for unification.Although Calvet gives several such examples that fur-ther highlight change of language form in response to a social need, what he of-


fers is basically a general descriptive mechanism with supporting examplesrather than a detailed description of a rheostatic mechanism or even a quanti-tative model. Although I cannot claim to have the last word on this issue, thehomeostatic model (the general power model of contextual and vital forces andtheir resolution) that I proposed (McConnell, 1991) does give detailed featuresof such a mechanism, which is also supported by the possibility of a quantifica-tion of the forces at play. The aim of my model is to explain functional change(or maintenance) of languages in a constellation, and it hypothesizes that thischange (or lack thereof) is based first on an inner resolution of forces that canbe measured at any given moment by the vitality rates of each of the languagesin place. It is also based on an outer concordance of forces that can influence (ornot influence) the inner forces already in place.

According to the present model, internal forces and their resolution maylargely account for vitality levels with either no influence from externalforces or with external forces intervening only periodically and dramati-cally, or as an additional sustaining factor to internal forces already inplace. (McConnell, 1991, p. 227)

My model is conceptually simpler than Calvet’s in that it aims to account forchange in language use and function (e.g., functional spread), whereas Calvetfocuses both on a change in function and on a change in form: “Et il nous fautdonc rendre compte en meme temps de la régulation interne de la langue et dela régulation sociale des situations linguistiques [One has to take into accountat the same time the internal regulation of a language and the social regula-tion of social situations]” (p. 104). However, to really advance things theoreti-cally, what we require here is a more detailed descriptive or quantifiable ap-proach to the homeostatic model rather than what Calvet has proposed.

THE REPRESENTATIONAL SUBMODEL

According to Calvet’s definition, this submodel is “destiné à rendre comptede la façon dont les locuteurs, individuellement et collectivement, percoiventleurs pratiques et celle des autres [destined to give an account of the way inwhich speakers, individually and collectively, perceive their practices and thatof others]” (p. 16). It would be difficult to contest the need for such a submodel,which nicely fits under the représentations part of the equation that Calvetproposes. In spite of the reference to Labov,whose early work identified specificlanguage forms with specific social classes, including individual opinions onthe forms and norms used, and to Lambert’s matched guise test, which demon-strated stereotypes related to language group identification (the characteris-tics and traits of the same person were rated differently depending on the lan-guage used), Calvet only skims over the vast literature available in this area,which may be variously labeled sociolinguistics or social psychology and whichis the central concern of this journal.

I noted earlier that the theoretical advancement of the whole field of lan-guage and communication was partly stymied by the lack of coordinationbetween specialists of social psychology and those working in the sociology oflanguage (i.e., the micro and macro spheres). That Calvet has brought thesetwo important aspects of social communication together in his equation

BOOK REVIEWS 477

certainly again serves as a reminder that both are required if we are to have anoverall theory of communication regarding language and social change. Calvetfirmly states that “la social donne naissance à du linguistique, que les condi-tions écolinguistiques produisent par le biais des représentations, des pra-tiques et donc de la forme [the social gives birth to the linguistique, thatecolinguistic conditions produce through the bias of representations, usagesand therefore form]” (p. 179). This is to say that the concerns of social psychol-ogy, which are largely subjective and specific to individuals and groups, havean important bearing on usage (both as to choice of function and choice of form)as well as on the form of language or language variety per se. Calvet backs thisup with his Malinké-Bambara example (pp. 177-180), which shows a strongpreference for the language of the city (Bambara) as opposed to the language ofthe village (Malinké) and a corresponding adaptation to city phonological fea-tures, especially by those migrating to the urban area.

Calvet dedicates a large part of chapter 4 to the dichotomy of security andinsecurity (sécurité and insécurité). “Mais les représentations portent le plussouvent sur l’ensemble de la langue et peuvent entre autres choses révéler unesécurité ou une insécurité dans différents domains . . . ayant un effet derétroaction sur les usages, les modifiant [But representations usually encom-pass the whole language and can among other things reveal a security or aninsecurity in different domains . . . having a retroactive effect on usage by mod-ifying it]” (p. 158). As for representations, Calvet identifies what he calls threedomains: (a) language form (forme des langues), that is, how people speak orhow they should speak (e.g., patois vs. standard); (b) language status (statutdes langues), that is, what people think they should speak (e.g., ParisianFrench);and (c) identifying function (fonction identitaire), that is,what charac-terizes the community (e.g., local French). These three domains are com-pounded in terms of security and insecurity (the view of the speaker) and againin terms of securance and insecurance (sécurisation and insécurisation; theview of the speaker’s observer) to produce numerous representations that pur-portedly yield a multiplication of language practices and forms. All this seemsneedlessly complicated and epistemologically does not help us measure theinfluence that representation does have on both choice and change of languagefunctions and forms as opposed to situational and contextual factors that alsoact on these same functions and forms. This is where a combination of theabove two traditions, the sociology of language and social psychology, would besuch a great asset and allow us to determine whether the thrust of changecomes mainly from the constellational matrix or from the representationalone. Frequently, language forms and functions are givens in any specific con-text, but forms and functions are also manipulable to some degree to allow forchange and evolution of both in communication networks.

THE TRANSMISSION SUBMODEL

According to Calvet, this submodel is “destiné à rendre compte de la façondont les situations et les langues évoluent [destined to take into account themanner in which situations and languages evolve]” (p. 16). Although Calvetclaims to be dealing with “un autre facteur de régulation et d’évolution[another factor of regulation and evolution]” (p. 183), what he is really dealingwith is the transmission and spread of languages either intergenerationally or


simply over time, that is, the old question of the loss, spread,or gain of languagespeakers either as mother tongue or second language speakers: “Nous avonsvu que le facteur limitant principal des langues était le nombre de leurlocuteurs [We have seen that the main limiting factor of languages was thenumber of their speakers]” (p. 185).

Calvet then moves on to a lively discussion of the creation of new languageforms and particularly of the genesis of creole languages, which are unique inthat they can give us an example of a relatively recent and rapid change in lan-guage form and function (p. 213). “Mon hypothèse est donc que ‘‘la genèse” descréoles met en jeu deux processus corrélatifs, la non-transmission des languesdes esclaves d’une part, l’appropriation “sauvage”, informelle, d’une autrelangue présente clans la niche écolinguistique d’autre part [My hypothesisthen is that the genesis of creoles brings into play two correlated processes, thenontransmission of languages of the slaves on the one hand and the informal or“savage” appropriation of another language present in the ecological niche onthe other hand]” (p. 213). This is interesting in itself because it sheds light on a“short”evolutionary process of language formation and change but finally addsnot a great deal to the general theoretical explanation of language spread,maintenance, and loss. Other than making further references to the naturaltransmission of languages and to migration (p. 222) or to the functions of lan-guages (p. 223) or an appeal for a need to measure such language changes,there is very little in this chapter that advances the overall theoretical planand that could not have just as well been placed in chapter 2 under thehomeostatic model or be partly dealt with by my general power model of con-textual and vital forces.

CONCLUSION

Calvet’s aim for this book was an ambitious one: to build a conceptual frame-work for language in society and, by the use of an ecological model, to accountfor change and direction of change in language form and function throughappropriate representations and situations. “Ce livre s’attachera à proposer lecadre conceptuel dans lequel une telle reconstruction (langue et société) estpossible [This book will try to propose a conceptual framework in which such areconstruction (language in society) is possible]” (p.22).To this end,Calvet pro-poses a four-tiered interactive model that would “tenter de rendre compte de lacommunication sociale clans toute sa complexité [attempt to give an account ofsocial communication in all its complexity]” (p. 16). Although Calvet in hismodel-building process does propose some useful constructs and typologies, inparticular that of his situations-pratiques- représentations, the reconstructionis based on a descriptive and anecdotal modus vivendi. This can be useful toclarify a point but, especially in an ecological-type model, never gets to whatwill finally be required: a quantitative modus vivendi. This would allow us, forexample, to measure things such as the functional strength of a language orlanguages or the relative importance of specific external forces (independentvariables) at play according to the type of constellation. It would then allow usto hypothesize as to whether certain types of language configurations wouldprivilege certain types of variables. It would also call for a much more inte-grated effort on the part of language sociologists and psychologists so that

BOOK REVIEWS 479

language change could be detected and measured both from “above” and from“below”. Calvet’s model is structurally not detailed enough, nor is it quantita-tively organized to do this job.

—Grant D. McConnellUniversity of Laval

REFERENCES

McConnell, G. D. (1991). A macro-sociolinguistic analysis of language vitality:Geolinguistic profiles and scenarios of language contact in India. Sainte-Foy, Canada:Presses de l’Université Laval.

McConnell, G. D. (1997). Analyses et comparaisons des situations de contact en Inde. InN. Labrie (Ed.), Plurilingua XX, études récentes en linguistique de contact (pp. 294-317). Bonn Germany: Duemmler.

Swann, Abraham de. (1993). The emergent world language system: An introduction.International Political Science Review, 14(3), 219-226.

Folk Linguistics. Nancy A. Niedzielski and Dennis R. Preston. Berlin,Germany: Mouton De Gruyter, 1999. 375 pp. ISBN 3-11-016251-2(hardcover).

A book of “stankos”? A new word for me, and one that I guess will be un-familiar to many scholars of the post-Chomsky generation.Stankos was appar-ently coined by Leonard Bloomfield’s family to describe the language beliefs ofnonlinguists.Bloomfield often collected what he considered “ignorant or stupidremarks about language” made by people outside of academia. Presumably, hedid so for reasons of amusement rather than as potential data that could war-rant serious investigation. Despite Bloomfield’s misgivings and those of manyprofessional linguists subsequently, here is a book dedicated to stankos andone moreover that sets its store by arguing that nonlinguists’ beliefs about lan-guage not only deserve careful consideration but also should be systematicallycollected and interpreted.

Received linguistic science has tended to refute the idea that nontrainedprofessionals have any direct role to play in the discipline.Although these ordi-nary language users, or “folk,” as Niedzielski and Preston prefer to refer tothem, have been readily used as informants, consultants, and respondents indescriptive language studies as well as in the analytic and experimental meth-odologies developed in language attitude studies and sociolinguistics, theircontributions to linguistic argument and theory in general have not been muchexplored. A tradition for the discussion of “folk linguistics” is accorded veryrecent history by Niedzielski and Preston. It is a tradition that is directlyrelated to Hoenigswald’s (1966) call of just over 35 years ago to take profes-sional cognisance of “not only what goes on (language), but also in . . . how peo-ple react to what goes on (they are persuaded, they are put off, etc.) andin . . . what people say goes on (talk concerning language)” (p. 20).


One scholar who has responded positively to this call in the interveningyears is Dennis Preston (1989, 1993, 1994, 1996). He has tended to be some-thing of a lone voice, however, because relatively few scholars have pursuedHoenisgwald’s (1966) suggestions in any systematic way. Indeed, this is thefirst monograph-length treatment of folk linguistic beliefs of which I am aware.It is welcomed accordingly because if native language users speak about a lin-guistic phenomenon, then it should surely merit professional concern.Whether the concerns raised by Niedzielski and Preston will have a majorimpact on the wider linguistic community remains to be seen. My guess is thatenthusiasm for the book will tend to be limited to those scholars who arealready empathetic to the Hymesian ethos in which it is framed.

The data of folk linguistic beliefs are taken as the object of enquiry.The aimsare to demonstrate how these data can be “more systematically collected andinterpreted” and “to know the organizing principles behind belief.” The argu-ment and findings are acknowledged to be “complex.”Folk Linguistics is a chal-lenging text in terms of both its content and its readability. It is clearly estab-lished that the folk have much to say about language and that what is saidshould be an integral part of the ethnography of a speech community. A role forfolk linguistics is also ascribed in the field of applied linguistics, with particu-lar respect to issues of first- and second-language acquisition and language useand norms in education. Folk notions with respect to grammar are more brieflyexplored.

Rather than its substantive arguments, it was the data of Folk Linguisticsthat captured my attention, for therein is the essence of its scholarly contribu-tion and also the potential for the kind of methodological and empirical criti-cism that I think it might attract. The study on which the monograph is basedwas undertaken more than 10 years ago by a graduate sociolinguistics class atEastern Michigan University. It is not easy to disentangle the methodologyand fieldwork practices of the class or to establish the nature of the corpus thatis used by Niedzielski and Preston in their subsequent discussion. Indeed, thesection that describes this study (Section 1.2) is oddly brief, given the kind ofrigour and detail that is presented in chapter 2, for example.

As I understand it, 68 respondents from southeastern Michigan were iden-tified on the basis of an adaptation of social network theory. The networksappear to be of the loose-knit rather than dense variety. Respondent details interms of age, gender, ethnicity, and educational background are described ingeneral terms, and the onus is put on readers to “form whatever conclusionsthey may like about the nature of the folk linguistic data reported in conjunc-tion with the respondents’ social status, ethnicity and the like” (p. 325).

The folk thus identified were then made the subjects of “interview-like” and“conversation-like” interactions. It is difficult to determine precisely what tookplace in these interactions, though I understand that talk about language wasubiquitous enough to provide a corpus of recordings. Details about the processof data collection, transcription, and coding that took place are unfortunatelycommitted to an appendix (p. 325 ff) and a footnote on page 339. I remainunclear as to why “the nonnative fieldworkers were given a list of sentencesdrawn from contemporary theoretical work in syntax and asked to get nativespeakers’ reactions to them” (p. 40). Indeed, more explicit instruction on the

BOOK REVIEWS 481

anticipated uses of the data would help a great deal. The reader is simplyinformed that

the remainder of what went on with whom and why will be discussed inplace as we provide a more detailed account of the topics which preoccu-pied our folk linguists and of our approaches to and interpretations ofwhat they said. (p. 40)

My intention in raising these issues is not to be niggardly but rather to suggestthat a case for the folk might have been made even more compelling if the rela-tionship between methods and theory had been made more explicit (see Milroy,1987).

The many topics and issues that the folk talk data cover are considered inchapters about regional variation (chapter 2); ethnicity, status, speech style,and gender (chapter 3); language acquisition and applied linguistics (chapter 4);and general and descriptive linguistics (chapter 5). I particularly enjoyed theerudition and detail of Preston’s earlier work concerning folk dialect percep-tions and the mapping of regional speech zones in the United States (chapter 2).However, the “conversational evidence” presented at the end of this chaptercould perhaps have been intercalated in the body of the chapter, as in theremaining chapters of the book. The question that underpins them is, “Is therea link between what linguists have found and folk linguists believe?”

The data and discussion presented in Folk Linguistics provide strong evi-dence for a claim in the affirmative,notwithstanding Niedzielski and Preston’smodestly expressed hope “to have provided at least an opening description ofthis rich territory.” In my opinion, it will also help redress the widely held andunflattering view of folk talk as imperfect science (McGregor, 1998). Whatnonlinguists have to say about language is worth listening to and can only com-plement linguistic description in ways that deserve more than cursory consid-eration or the pejorative status of stanko in the Bloomfieldian sense. Ordinarylanguage knowledge is an intrinsic part of Gumperz’s interactional socio-linguistic theory (Figueroa, 1994). Perhaps this can provide an avenue forexploring the cognitive models that folk use in reasoning about language, asraised in discussion by Niedzielski and Preston in their final chapter.Althoughtheir wish not to “frighten readers by making them believe that we intend to goand on” (p. 322) may be a little tongue in cheek, it does indicate that in the caseof folk linguistics, much remains to be both written and said.

—Graham McGregorOtago Polytechnic

REFERENCES

Figueroa, E. (1994). Sociolinguistic metatheory. Oxford, UK: Pergamon.Hoenigswald, H. M. (1966). A proposal for the study of folk-linguistics. In W. Bright (Ed.),

Sociolinguistics (pp. 16-26). The Hague, the Netherlands: Mouton.McGregor,G. (1998).Whaddaweknow? Language awareness and non-linguists’ accounts

of everyday speech activities. Language Awareness, 7, 32-51.


Milroy, L. (1987). Observing and analysing natural language: A critical account ofsociolinguistic method. Oxford, UK: Basil Blackwell.

Preston, D. (1989). Perceptual dialectology. Dordrecht, the Netherlands: Foris.Preston, D. (1993). The uses of folklinguistics. International Journal of Applied Linguis-

tics, 3, 181-259.Preston, D. (1994). Content-oriented discourse analysis and folk linguistics. Language

Sciences, 16, 285-331.Preston, D. R. (1996). Whaddayaknow? The modes of folklinguistic awareness. Language

Awareness, 5, 41-74.

The Social Context of Nonverbal Behavior. Pierre Philippot, Robert S.Feldman, and Eric J. Coates (Eds.). Cambridge, UK: Cambridge Uni-versity Press, 1999. ISBN 0-521-58371-3 (hardcover); 0-521-58666-6(paperback).

Traditionally, the study of nonverbal behavior has followed what some callthe “channel”approach,whereby behaviors such as gaze, touch,proxemics,andso on are studied individually, without regard for other behaviors that theyaccompany. More recently, scholars have taken a “functional” approach to thestudy of nonverbal behavior (e.g., Andersen, 2000; Patterson, 1983), examiningvarious social functions (e.g., intimacy, emotional expression, etc.) served bycombinations of nonverbal behaviors. The Social Context of Nonverbal Behav-ior takes yet a different position by analyzing nonverbal behaviors as a func-tion of the social contexts in which they are embedded and that they create andshape. In the preface, the editors appropriately point out that the volume pre-sents some of the most contemporary perspectives on nonverbal behavior, withan eye toward the controversy over biological versus social forces behind theproduction of nonverbal behavior.These chapters are written by a mix of estab-lished scholars and young scientists in the area of nonverbal communication.This book is appropriate for advanced undergraduate- and graduate-levelcourses on nonverbal behavior and/or emotion. Researchers working in thearea of nonverbal communication and emotion will also find considerable util-ity in many of the book’s chapters.

The book is divided into four major sections. The first section, entitled“Social Norms and Nonverbal Behavior,” starts with a chapter by Kupperbuschet al. that addresses cultural influences on nonverbal behaviors. This chaptercontains an excellent overview of the classic studies that supposedly estab-lished the facial expression universality position. In their analysis, the authorsoutline the components of a theory of culture and emotional expression andprovide some useful suggestions for future research. However, one can see theauthors struggle with the idea of “culture,” rightfully noting that there is greatheterogeneity within a given culture. Their admonition to assess culture at theindividual level ultimately raises questions about what culture really means.The next chapter in this section,by LaFrance and Hecht, is an interesting anal-ysis of gender and role norms in smiling. There is a strong analysis of displayrules and how they often prescribe smiling for those in low-status roles, regard-less of whether they really feel like smiling. The authors present the results ofa study in which they manipulated power and found, consistent with several

BOOK REVIEWS 483

other lines of research on nonverbal behavior,greater variability in the smilingbehavior of high-powered actors compared to those with less power. The finalchapter in this section on social norms is by Saari and Weber, examining thecomplex function of emotional displays in children. Saari and Weber provide astimulating discussion of different types of expressive manipulations in chil-dren along with an equally compelling treatment of individual differences inemotional displays. The authors argue that emotion management plays amajor role in children’s coping. Saari and Weber try to cover a lot of ground inthis chapter, resulting in a sometimes unfocused collection of findings withinwhich a number of engaging topics are discussed.

The second part of the book, “Transmission of Social Norms RegulatingNonverbal Behavior,” picks up where the first part left off. Halberstad et al.’schapter presents a synthesis of the literature on family expressiveness and itsrelationship to various child outcomes such as social competence, aggression,popularity, self-esteem, and temperament. The authors summarize a fairlylarge body of literature in this chapter with a quasi-meta-analytic techniquebased on counting studies with significant associations but absent any calcula-tion or presentation of effect sizes. Unfortunately, readers will be left wonder-ing how strong or weak these associations are.Also, for certain dependent vari-ables, the number of studies available is often less than five, again makingconclusions somewhat tenuous. Nevertheless, in typical form, Halberstad andher colleagues offer a fine methodological critique of studies on family expres-siveness and demonstrate an unparalleled command of this literature. But, forthose who are genuinely curious about nonverbal behavior, the “family expres-siveness” construct does not readily translate to any particular profile or mani-festation of nonverbal behaviors. Coats et al. follow with a chapter about theinfluence of television on children’s nonverbal behavior. Social learning theo-rists are sure to gravitate toward their thesis that television is a socializingagent for children’s emotional expressiveness. They argue convincingly thattelevision can teach display rules to young viewers. Their data indicate thatfrequent viewers of television are good encoders of the same sorts of emotionsthat are often depicted on television and that they are more emotionallyexpressive, as well as better decoders, than less frequent viewers. Finally, achapter by Kirouac and Hess on group membership and nonverbal decodingfinishes off this section. Whereas earlier chapters argue that norms for non-verbal behavior come from family and media, Kirouac and Hess indicate thatour group membership also influences both the encoding and decoding of emo-tional displays. Their analysis suggests that people use knowledge of a target’sgroup membership when making inferences about that person’s nonverbalbehavior. Factors such as culture, gender, and status are all offered as groupvariables that affect the production and interpretation of nonverbal behavior.Like several other chapters in the book, however, Kirouac and Hess make sucha good case for the many factors that can affect encoding and decoding of non-verbal behavior that clear predictions and simple summaries become a nearimpossibility.

The third section of this book, entitled “Immediate Social Factors DuringInteraction,” explores social contexts associated mostly with facial expression.This section’s first chapter,by Hess et al., is an intriguing analysis of mimicry,abehavioral phenomenon as mysterious as it is ubiquitous. Potential answers tosome vital questions about the functions of nonverbal mimicry and a candid


assessment of the baselessness of clinical lore on mimicry are presented. Thesecond chapter in this section, authored by Fernandez-Dols, presents asituationist analysis of facial expression and emotion that is more of a critiqueof traditional assumptions linking emotion and facial expression than a pre-sentation of an alternative perspective.Fernandez-Dols’s analysis rests in parton the poor connection between the etymology of the term expression (“tosqueeze out”) and the phenomenology of facial behavior. One has to wonder ifperhaps this approach pays sufficient attention to both the lay and profes-sional usage of the term expression and whether the issue at hand is moresemantic than empirical. The following chapter by Wagner and Lee exploresfacial behavior as it is affected by the presence of others. These authors alsoquestion the necessary link between emotion and facial expression,noting thatexpressions can be facilitated or inhibited and rendered qualitatively differentby the presence of others and the relationship between an encoder and his orher audience.Wagner and Lee’s analysis,however, is propelled by some curiousassumptions, such as Fridlund’s position that people often behave as if they arewith a mental companion, so there is no such state as “alone,” and Wagner andLee’s own assertion that nonverbal behavior cannot be understood without ref-erence to the verbal behavior that it accompanies. These arguments are for themost part extraneous to the overall thrust of the chapter but will raise suspi-cion in some readers. In the final chapter of this section, Manstead et al. reviewtheir program of work on social and emotional functions of facial displays. Theauthors strike a near ideal balance between a wholesale demolition of the emo-tional expression view (that facial expression is indicative of underlying emo-tion) and blind acceptance of a direct and invariant path from emotion to facialexpression. Although they appear mildly sympathetic to Fridlund’s claim thatthere is no causal connection between emotion and facial behavior, Mansteadet al. present a good case for at least some relationship between internal emo-tional experience and facial expression. The authors are also refreshingly can-did about the untestability of the hypothesis that people exhibit facial expres-sions pursuant to various social motives.

The chapters in this third section of the book, aside from Hess et al.’s, pre-sent a fairly thorough critique and analysis of many traditional assumptionsabout emotion and facial behavior. However, as a whole, they argue that therelationship between emotion and facial expression is affected by factors a, b, c,d, e, f, and g; useful for explanation but a nightmare for prediction.

The fourth and final section of this book, “The Role of Nonverbal Behavior inthe Facilitation of Social Interaction,” for the most part explores nonverbalbehavior in relational contexts. This section starts with a chapter by Pattersonthat provides an excellent, readable discussion of the evolution of his parallelprocess model. Patterson provides a clear historical context for this theory,adeptly reviewing its predecessors, such as arousal labeling theory and equi-librium theory, refreshingly garnished with a frank assessment of the scopeand weaknesses of the parallel process model. In the past decade, it has becomefashionable in the field of nonverbal behavior research to articulate theoriesthat sometimes call on dozens of factors to explain the action and function ofnonverbal behaviors in social contexts. Although Patterson’s model does not goquite that far, the number of variables required to provide an explanation orprediction of nonverbal behavior and the social judgments it can produce canbe downright burdensome. Despite that, Patterson’s model is sensible and

BOOK REVIEWS 485

understandable. This chapter is sure to find its way onto the required readinglist of many courses on nonverbal behavior. Feeney et al.’s chapter on non-verbal behavior and conflict in close relationships follows. This is an overviewof a program of research conducted with Noller, with particular attention tothree of their recent studies.The third of these is particularly novel, examiningconflict patterns in families with twins. Most notably, they found that fathers’conflict styles not only affect their relationships with the twins but that theyalso affect the relationship between the twins. This line of research clearlyshows how nonverbal behavior is a manifestation of satisfaction with or with-drawal from close relationships and some of the collateral consequences it canhave. This fourth section of the book concludes with a strong chapter by Ander-son et al. on deception in close relationships. Anderson and his associatesexplore the truth bias and social context of deception. They provide a compel-ling account of why the process of deception may vary in close relationships. Asit turns out, there is not a great deal of attention granted to particular non-verbal behaviors (something that this chapter has in common with many oth-ers in this book) and not a great deal of data available for review. Anderson et al.’schapter raises several hypotheses about deception in close relationships, andwhen there are existing studies that speak to these hypotheses,Anderson et al.do a good job of extracting useful evidence.

As a whole, The Social Context of Nonverbal Behavior avoids a mere rehashof channel or functional approaches to nonverbal behavior. However, does itsucceed in situating the analysis of nonverbal behavior in a social context?Most of the contributing authors make a good case for considering such contex-tual factors as culture, gender, race, family, close relationships, and even televi-sion for understanding how and why nonverbal behaviors function. With thatsaid, at times, the social context analysis can be dissatisfying, taking the formthat “to understand the meaning and function of X, you need to consider these15 social factors.” The search for parsimonious social regularities sometimeslooks both dismal and futile after reading certain sections of the book. One alsohas to wonder if some of the authors stayed a bit too far away from the channelapproach.There are virtually no mentions of certain behaviors such as gesture,proxemics, paralanguage, clothing use, posture, and touch to be found any-where in this book. Readers who want to know about these behaviors will haveto make inferences based on discussions of family expressiveness, “emotionalexpression,” and “conflict style,” for example. However, this book is sure toappeal to those interested in emotional expression.At least half of the chaptersin this book are squarely focused on emotional expression. In this sense, thebook is unbalanced, dominated by an analysis of this one function of nonverbalbehavior, its most commonly associated channel (facial expression), and theattendant debate over the universality versus cultural specificity of facialexpressions of emotion. This debate is nothing new, and the chapters thataddress it summarize more than advance the controversy. Despite its unevencoverage of nonverbal functions and behaviors, this book represents an impor-tant step toward the understanding of nonverbal behavior as a function of thesocial context in which it is embedded, particularly in the area of emotionalexpression.

—Chris SegrinUniversity of Arizona


REFERENCES

Andersen, P. A. (1999). Nonverbal communication: Forms and functions. Mountain View,CA: Mayfield.

Patterson, M. L. (1983). Nonverbal behavior: A functional perspective. New York:Springer-Verlag.

Discourse in Late Modernity: Rethinking Critical Discourse Analysis.Lillie Chouliaraki and Norman Fairclough. Edinburgh, UK: Edin-burgh University Press, 1999. 168 pp. ISBN 0-7486-1082-0 (paper).

Applied Discourse Analysis: Social and Psychological Interventions.Carla Willig (Ed.). Buckingham, UK: Open University Press, 1999.166 pp. ISBN 0-335-20226-8 (paper).

Critical Textwork: An Introduction to Varieties of Discourse and Analy-sis. Ian Parker and the Bolton Discourse Network. Buckingham, UK:Open University Press, 1999. 226 pp. ISBN 0-335-20204-7 (paper).

All three of these books proclaim a concern with expanding the scope of criti-cal analysis of discourse. In this sense, they seek to positively counter, respec-tively, the absence of a developed analytic framework for grasping contempo-rary historical and ideological shifts, a reluctance to explore the practicalusability of findings, and the exclusive concern with written or spoken dis-course. Although all three deal with issues of relevance to those interested inlanguage and social psychology, only the collection of articles edited by CarlaWillig explicitly engages with psychological topics.Broadly speaking, there is ashared preference of analytic style, that is, toward a focus on the use of histori-cally situated discursive resources and the attendant positionings as opposedto a more ethnomethodologically informed excavation of the fine grain of local,situated practical activity.

Chouliaraki and Fairclough set out to formulate a research agenda for criti-cal discourse analysis (CDA) that offers the possibility of critically illuminat-ing what are seen as the large-scale societal and economic changes associatedwith late modernity. Overall, it represents a substantial work of theoreticalsynthesis—incorporating clearly written chapters on, for example, the theoryof critical social science, the theory of discourse,and sociological theories of latemodernity—and ambitiously brings forth a complex amalgamation of a widerange of analytically relevant themes.

Part of the book’s aim is to ensure that CDA is equipped to demonstrate howsocietal transformations, as analysed by theorists such as Anthony Giddens,David Harvey, and Jürgen Habermas, are instantiated in the domain of discur-sive practice. It is the discursive aspects of contemporary change that are to beempirically explored by CDA. In the process, it is expected that the sociologicalanalyses, which themselves broadly recognise that the societal transform-ations to a significant degree consist of transformations in language, will inturn be enriched. The different forms of change foregrounded by the theorists

BOOK REVIEWS 487

of societal transformation—forms that are said to have corresponding effectsin the discursive domain—include the growth of reflexivity, globalisation, theunsettling of identities,and the colonisation of one social practice by discoursesor genres originating in another realm. The most important and overridingprocess, however, in the light of which the above themes take on a particularanalytic significance, is that of hybridisation. On one hand, hybridity is recog-nised here as an inherent feature of discourse. The coexistence of different his-torically constituted themes within an unstable mix means that there hasalways been the potential for different discursive elements to be articulatedtogether as historical circumstances shift. On the other hand, Chouliaraki andFairclough wish to suggest that late modernity is characterised by an intensifi-cation of hybridity: “In late modernity, boundaries between social fields andtherefore between language practices have been pervasively weakened andredrawn, so that the potential [for articulating practices together in new ways]seems to be immense” (p. 13). Examples of such a process raised in the bookinclude the presence of the informal language of everyday life within publicmedia discourse, the extension of the commodified discourse of advertising intotraditional spheres of political discourse, and the local incorporation of ele-ments of globally dominant cultural material by more marginal culturalpractices.

It is this theme of hybridity, or “interdiscursivity,” that suggests an analyticfocus for CDA’s attempt to explore changing relations of power. The task is oneof investigating the (practically realised) relationship between conflicting dis-cursive elements, for each of the changes mentioned above is said to exist in acomplex relationship with its opposite: colonisation coexists with appropria-tion, globalisation with localization, and reflexive consciousness with ideologi-cal naturalisation. Such a route involves both an acceptance of the grand socio-logical narratives of late modernity and a determination of their limits,especially those that display an overoptimism about the universalisation ofreflexivity and the openness of the social.

Mapping the semiotic diversity of the hybrid text involves identifying“orders of discourse,” defined as “the socially ordered set of genres and dis-courses associated with a particular social field, characterised in terms of theshifting boundaries and flows between them” (p. 58). In thus tracking how dif-ferent discursive resources are drawn on and articulated in interaction, itstudies something that mediates “between communicative interaction andtext on one hand, and social structures and processes on the other” (p. 116).

Most of the book consists of theorising, bringing a diverse selection of con-cepts into association with one another, in the process attempting to instigateinterdisciplinary bridges between CDA and other areas such as the field the-ory of Pierre Bourdieu. This strong emphasis on theory is in fact unapologetic-ally projected by Chouliaraki and Fairclough at the start of the book, wherethey state that although viewing the project of CDA as bringing theory andpractice together, they have consciously set out to produce a theoretical book.However, there are still a few developed empirical examples of CDA in actiongiven at various points in the book, and it is here that we are able to someextent to judge the strength of the CDA project in general. The main examplesinclude an analysis of an advertisement by a charity for the homeless, focusingon the theme of the commodification of language; an examination of a publicexchange of letters regarding alleged acts of unwarranted police violence (here


it is argued that highlighting the hybridity of the text can enhance the analysisoriginally made of these data by Dorothy Smith); and an analysis of part of aworkplace meeting that, Chouliaraki and Fairclough claim, shows howemployees integrate a new discourse of “quality practice”—relating to team-based work organisation—imposed from outside their company. Althoughthese topics of analysis are interesting, they do suggest that despiteChouliaraki and Fairclough’s expressed support for a genuine dialoguebetween theory and practice, there is a risk of theory dominating or drowningout practice. On one hand, this is perhaps more obvious when CDA is used toanalyse spoken interaction. CDA’s emphasis on linguistic features—reflected,for instance, in the book’s effort to build a close relationship between CDA andsystemic functional linguistics—would seem to make it more suitable for ana-lysing written texts. Despite the intention that CDA be sensitive enough topick up the subtleties of “the dialogical interactions of everyday life” (p. 118), itis doubtful whether it succeeds in demonstrating with any detail that, at a par-ticular moment, with its particular rhetorical concerns, a particular discourseis being drawn on and thus realised as part of practical sense making in action.But, it is also the operation of studies of written texts that appears quite mech-anistic, such that the specificity of local textual relations is not fully explored.It seems to be treated as sufficient to impute the co-occurrence of different dis-courses in a piece of text at the cost of ignoring the full social-argumentativedimensions, including the various significant absences and the paradoxicalmanoeuvrings associated with the management of dilemmas. It follows thatalthough Chouliaraki and Fairclough’s version of CDA is still able to provide avaluable picture of broad discursive change, certain features that couldsharpen the critical edge of the analysis and increase the likelihood of discover-ing “unsuspected structural connections . . . across practices” (p. 34) areunderattended to.

In sum, there would have been much to gain by tempering the extendedcommitment in this book to theoretical synthesis in favour of fuller empiricalengagement with practical complexities and ideological subtleties. There is arisk otherwise of being seen to promise a large amount but to deliver little. Onefinds oneself unconvinced of the real necessity of so much of the theorising inthe book. Still, it remains the case that Chouliaraki and Fairclough deservepraise for their attempt to put large-scale historical change under the analyticspotlight. Their wish to take seriously contemporary processes of market-isation and the changing power relations within the workplace is especiallyadmirable. It is in this sense that their work challenges those interested in lan-guage and social psychology to wake up to topics that, despite affecting such alarge portion of contemporary social life, so often go unstudied.

In contrast to the theoretical synthesising of Discourse in Late Modernity,the other two books under review consist primarily of empirical studies.But, inthe case of Willig’s edited collection of research entitled Applied DiscourseAnalysis: Social and Psychological Interventions, there is not to be the usualapproach to presenting discussion of empirical findings. Willig claims that thebook is “motivated by the desire to move beyond critical commentary andtoward an active engagement with social and political practice” (p. 1). It ispointed out that critical analyses typically fail to spell out how the potentialimplications of the analyses, such as increasing awareness of and resistance totaken-for-granted ideological habits, may be put into practice. That is to say,

BOOK REVIEWS 489

they fail to draw out specific recommendations for developing strategic inter-ventions and alternative discursive constructions. Accordingly, each chapter ofthe book consists of a section of analysis (in most cases of spoken interaction)followed by a section evaluating the implications for potential intervention.The types of data analysed by the six chapters include self-help guides forstress; police interviews with suspected criminals; accounts given by working-class female speakers about their smoking; talk about human reproductivetechnologies; women talking about sexual safety, condom use, and trust; and aprofessional psychiatrist’s accounting for the failure of medication.

Arguably, three of the chapters come closest to satisfying the book’s aims.Two of these show the potential value for health promotion policy of listeningto individuals who constitute potential targets of such campaigns. In Gillies’schapter, these individuals are working-class female smokers; in Willig’s chap-ter, they are sexually active women within long-term relationships. Attendingto such accounts, it is suggested, can increase understanding of the practicaleffects of existing health messages that although designed to help, may, iftaken seriously, end up disempowering people. Gillies argues that this is thecase with warnings against smoking that construct the latter in terms of physi-ological addiction and consequently risk discouraging attempts to give up. Anempirical basis for these conclusions is the readiness of all her respondents todraw on a discourse of addiction in their accounts, routinely portraying theirsmoking as resulting from deep-rooted physiological dependence. It is alsohighlighted how the theme of addiction exists in the interviewees’ talk within adilemmatic relationship with the opposite theme of self-control. This observa-tion leads into a discussion of the policy maker’s own parallel dilemmas, espe-cially concerning the structural constraints on working-class women forwhom, Gillies argues, an unmitigated campaign emphasis on self-control islikely to foster self-blaming and guilt.

More generally, there is also guidance given for future health campaignsthrough analysis of potentially dangerous representations. Relevant here arethe various rhetorical strategies employed to discount the health hazards ofsmoking and to establish contrasts between the acceptable (because it is con-siderate) smoking of self and the unacceptable smoking of others.

In the case of Willig’s study, potentially dangerous representations includethose that legitimate sexual risk taking, such as the notion that protected sex-ual intercourse could threaten trust within a marriage or other long-term rela-tionship. One policy recommendation that is said to follow from this is to breakthe exclusive association of condom use with casual relationships. Also dis-cussed are the notions of sex as male preserve and women’s passivity more gen-erally, which are portrayed as making women less resistant in the face of thethreat of HIV infection.

In contrast to the two studies just mentioned, the chapter by Harper focusesnot on lay accounts but on that of a professional psychiatrist. What is beingaccounted for here is the failure of prescribed psychiatric medication to havethe intended effects on one user of mental health services. Harper criticallyexamines the speaker’s rhetorical appeal to the chronicity of the user’s condi-tion. It is suggested that through such an appeal, responsibility for treatmentfailure can be pushed away from the medication and professionals and ontousers and the illness. In the process, the interests of the pharmaceutical indus-try are served. The main value of the findings, it is claimed, is their potential in


training mental health professionals and also in encouraging the provision ofbetter information for users of mental health services as well as for their care-givers and relatives.

A notable weakness of this collection is what is frequently an overly basiclevel of analysis. This can contribute to a sense that some of the implicationsfor interventions that are subsequently drawn out largely resemble wishfulthinking. One instance of such an analysis is to be found in Auburn, Lea, andDrake’s study of police interviews with suspects of crimes.The focus is on how apolice interviewer expresses disbelief in aspects of a suspect’s account and war-rants this in different ways, the aim being to obtain the suspect’s amendmentof his or her original account and acceptance of the institutionally preferredversion of events. Subsequently, it is suggested that such observations couldempower certain potentially vulnerable individuals by enabling them to resistpositioning as offenders and to be aware of interactional means used to con-struct the institutionally preferred version. However, with the level of analysishere remaining very basic, such that no serious critique of interviewing prac-tices is actually developed, it is difficult to see how the study could performsuch an empowering role. To the extent that the book as a whole shows suchtendencies, it reads at times as a hypothetical exercise, merely indicating pos-sible directions for when applied discourse analysis is seriously implementedin depth. A related observation here is the fact that to inject some iron into theoverall argument of the book, it would have been useful to include attempts tofollow through an actual attempt to employ discourse analytic findings in anintervention.

Finally, it is worth noting that Willig’s introduction and conclusion repre-sent careful attempts to think through some of the ethical and political dilem-mas surrounding intervention and to try to theorise its relationship with prac-tice. She ends by signalling the need to differentiate between discourseanalysis as a method of intervention and discourse analysis as a source ofapplicable “findings.” The former, it is argued, has greater radical potential,allowing challenge to the conceptual opposition between theory and practice.

Critical Textwork: An Introduction to Varieties of Discourse and Analysisopens with the stated aim of extending discourse analysis beyond the usualfocus on interview transcripts or other written texts such as newspaper arti-cles. There is an aspiration to demonstrate how a wider range of texts “can beopened up and read using innovative methods” (p.1).Text in this case is definedin the broadest possible terms as “any tissue of meaning which is symbolicallysignificant for the reader” (p. 4), that is, wherever meaning resides.

The book consists of no fewer than 16 short chapters written by people froma wide variety of disciplinary backgrounds and including nonacademics, all ofwhom are associated with the Bolton Discourse Network. The latter isdescribed at the beginning of the book as “an interdisciplinary forum for inter-disciplinary research into different forms text, based at Bolton Institute” (p. vii).Each chapter follows the same format: a general introduction to the form oftext to be considered and to previous approaches to it, an account of the specifictext of interest, presentation of this text within a “text box,” presentation of ananalytic reading of the text,and consideration of the disadvantages and advan-tages of the analytic approach. The first set of chapters looks at spoken andwritten texts, including interviews, love letters, children’s fiction, and class-room interaction. The second section takes visual texts as its topic, namely,

BOOK REVIEWS 491

comics, advertising, television, and film. This is followed by an examination of“physical texts” such as cities, organisations, gardens, and sign language. Thefinal section is entitled “Subjectivity in Research” and is presented as shifting“attention from the ‘written’—whether we are looking at actual written, atvisual or at physical material as kinds of writing—to the researcher as readerof the texts” (p. 9). In this case, the topics include the body, ethnography, the sig-nificance of silence in particular interactive settings, and self-advocacy groupsin the lives of people with learning difficulties.

The quality of the analyses in this collection is very mixed: Some areextremely basic and unilluminating, and a small minority are of more interest.A few of the latter category can be mentioned here. Rudd’s chapter on chil-dren’s literature attempts to engage with conventional critical evaluations ofthe fiction of Enid Blyton, including the suggestion that implicit within theFamous Five series of books is a middle-class and sexist worldview. To this end,Rudd focuses on the contradiction between these dismissive accounts by adultcritics and the fact that Blyton remains the most popular writer ever with chil-dren. On reanalysing certain controversial passages from the Famous Fivebooks, Rudd claims to find a more complex discourse in which traditional sex-ual relations are continuously being explored and questioned as well as main-tained. Second, Rudd’s efforts to listen to children talk about the books shedlight on some unexpected interpretations of the different fictional charactersthat help challenge some of the adults’ assumptions and indicate a potentialfor a more radical reading.

Concerning the more exotic forms of text analysed in Critical Textwork, twochapters in the section on physical texts stand out. First, Ford provides a stim-ulating discussion of the way the design of some celebrated Victorian suburbangardens reflected wider ideological themes of early Victorian society. She isaided in this task by a copy of an original intricate diagrammatic representa-tion of one such garden. It is demonstrated how the garden was imbued with amultitude of subtle and not-so-subtle indicators of taste and distinction. Assuch, it could be employed within the struggle by the new middle class to“achieve the appearance which would reflect their upwardly mobile aspira-tions” (p. 149). A central theme of the study concerns the way in which an ideol-ogy of sexual difference was realised through associated cultural representa-tions, such as that which took for granted the intimate association of womenwith the flower garden.

Another quite innovative piece of research, by Allbutt, Gray, and Schofield,takes as its text a photographic “transcript” of a section of an interview con-ducted in sign language with a deaf participant. The discussion of the photo-graphs of the different signs provides some interesting insights, includingthose relating to identity construction, into an area of interaction of which lit-tle is known to those outside the communities directly affected. In so doing, itcontributes to the positive notion of deaf people as constituting a distinctivecultural group.

Finally, Nightingale engages in a thought-provoking argument for the limitto social constructionist approaches to the body—the specific case discussedconcerns the topic of bodily disability—and the benefits of complementing itwith appreciation of the physical body and personal experiences developed inrelation to the latter.


A key criticism of this collection is that there is no obvious reason why somany different analyses needed to be included together within the same collec-tion. Some of the analyses appear to be provided not because of their substan-tive analytic insights but for the sake of demonstrating the range of texts thatcan be read. Arguably, there may have been more value in including fewer,more detailed studies that had more restricted and coherent overall purposesor rationales as opposed to something as all inclusive as the analysis of mean-ingful texts.

—David WeltmanLoughborough University

BOOK REVIEWS 493

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001ABOUT THE AUTHORS

About the Authors

Sherry L. Beaumont (Ph.D., University of Waterloo) is an associate professor ofpsychology at the University of Northern British Columbia. Her research interestsfocus on speech styles, parent-child communication, and socioemotional and per-sonality development across the life span. Recent publications include articles inMerrill-Palmer Quarterly, the Journal of Adolescent Research, and DiscourseProcesses.

Friederike Braun is a professor of general linguistics presently at the Universityof Passau, Germany. Her fields of interest are language and gender, feminist lan-guage critique, and the linguistic treatment of gender in grammatically genderlesslanguages such as Turkish.

Donald L. Rubin (Ph.D., University of Minnesota) is a professor in the Depart-ment of Speech Communication and Language Education and in the Program inLinguistics at the University of Georgia. His research specializations include gen-der and language, comparisons between oral and written discourse, and socialevaluation of language varieties. He has most recently published in Research in theTeaching of English,Communication Education,and the Journal of Communication.

Monica Ruggeri obtained her B.Sc. in psychology from the University of North-ern British Columbia in 1997.

Sabine Sczesny is an associate professor of social psychology at the University ofMannheim, Germany (Faculty of Social Sciences). Her research interests includehealth psychology, aggression and violence, gender stereotypes, and language andgender.

Dagmar Stahlberg is a full professor of social psychology at the University ofMannheim, Germany (Faculty of Social Sciences). Her current research interestsfocus on hindsight bias and decision making, gender stereotypes and leadership,and nonsexist language.Her publications include articles in the Journal of Person-ality and Social Psychology, the European Journal of Social Psychology, and Orga-nizational Behavior and Human Decision Processes.

Viviane C. B. Vasconcelos obtained her B.Sc. in psychology from the Universityof Northern British Columbia in 2000. She now works in Kamloops, BritishColumbia.

Sally Wiggins is a doctoral student with the Discourse and Rhetoric Group atLoughborough University, Great Britain. Her research focuses on the constructionof food evaluations and the development of discursive psychology. Earlier work in-cludes a coauthored article on the discursive construction of eating practices.

Laura L. Winn (Ph.D., University of Georgia–Athens) is an assistant professor ofinterpersonal communication at Wayne State University in Detroit. Her researchand teaching interests focus on social class, race and gender, and their role withinclose relationships. More specific interests include family and dating relationshipsand perceptions of verbal and nonverbal behaviors.

494

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY,Vol. 20 No. 4, December 2001 494 2001 Sage Publications

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY / December 2001INDEX

INDEX

to

JOURNAL OF LANGUAGEAND SOCIAL PSYCHOLOGY

Volume 20

Number 1 & 2 (March/June 2001) pp. 1-256Number 3 (September 2001) pp. 257-388Number 4 (December 2001) pp. 389-500

Authors:

ANTAKI, CHARLES, “ ‘D’you Like a Drink Then Do You?’: Dissembling Language andthe Construction of an Impoverished Life,” 196.

BALDWIN, JOHN R., “Gender and Discourse, edited by Ruth Wodak” [Book Review],358.

BARNES, BRENDON, INGRID PALMARY, and KEVIN DURRHEIM, “The Denial ofRacism: The Role of Humor, Personal Experience, and Self-Censorship,” 321.

BEAUMONT, SHERRY L., VIVIANE C. B. VASCONCELOS, and MONICA RUGGERI,“Similarities and Differences in Mother-Daughter and Mother-Son ConversationsDuring Preadolescence and Adolescence,” 419.

BRADAC, JAMES J., see Winn, L. L.BRAUN, FRIEDERIKE, see Stahlberg, D.CORRIGAN, ROBERTA, “Implicit Causality in Language: Event Participants and Their

Interactions,” 285.COWAN, JODIE, see McClure, J.DURRHEIM, KEVIN, see Barnes, B.GILES, HOWARD, see Winn, L. L.GIRBAU, DOLORS, “Children’s Referential Communication Failure: The Ambiguity

and Abbreviation of Messages,” 81.HILTON, DENIS J., see McClure, J.HOWE, CHRISTINE, and DONNA MCWILLIAM, “Peer Argument in Educational Set-

tings: Variations Due to Socioeconomic Status, Gender, and Activity Context,” 61.HUMMERT, MARY LEE, and DEBRA C. MAZLOFF, “Older Adults’ Responses to

Patronizing Advice: Balancing Politeness and Identity in Context,” 167.ISHIDA, LUCYNA, see McClure, J.KALBFLEISCH,PAMELA J., “Deceptive Message Intent and Relational Quality,”214.LI, HAN Z., “Co-Operative and Intrusive Interruptions in Inter- and Intra-Cultural

Dyadic Discourse,” 259.

495


LOUIS, WINNIFRED, and DONALD M. TAYLOR, “When the Survival of a Language Isat Stake: The Future of Inuttitut in Arctic Québec,” 111.

MAZLOFF, DEBRA C., see Hummert, M. L.MCCLURE, JOHN, DENIS J. HILTON, JODIE COWAN, LUCYNA ISHIDA, and MARC

WILSON, “When People Explain Difficult Actions, Is the Causal Question How orWhy?,” 339.

MCCONNELL, GRANT D., “Linguistic Cultures of the World: A Statistical Reference, byPhilip M. Parker” [Book Review], 367.

MCCONNELL, GRANT D., “Pour Une Écologie Des Langues Du Monde, by Louis-JeanCalvet” [Book Review], 470.

MCGREGOR, GRAHAM, “Folk Linguistics, by Nancy A. Niedzielski and Dennis R. Pres-ton” [Book Review], 480.

MCWILLIAM, DONNA, see Howe, C.NESDALE,DREW,“Language and the Development of Children’s Ethnic Prejudice,”90.OHBUCHI, KEN-ICHI, see Takaku, S.PALMARY, INGRID, see Barnes, B.PEDLOW, ROBERT, ROGER WALES, and ANN SANSON, “Children’s Production and

Comprehension of Politeness in Requests: Relationships to Behavioral Adjustmentin Middle Childhood,” 23.

REISSLAND, NADJA, “The Emergence of the Speech Capacity, by D. Kimbrough Oller”[Book Review], 382.

ROBINSON, W. PETER, “Epilogue,” 248.ROBINSON, W. PETER, “A Tale of Two Histories: Language Use and Education in Rela-

tion to Social Class and Gender,” 231.ROSSMAN, LILIANA CASTAÑEDA, “Moral Conflict: When Social Worlds Collide, by

W. Barnett Pearce and Stephen W. Littlejohn” [Book Review], 375.RUBIN, DONALD L., see Winn, L. L.RUGGERI, MONICA, see Beaumont, S. L.SANSON, ANN, see Pedlow, R.SCZESNY, SABINE, see Stahlberg, D.SEGRIN, CHRIS, “The Social Context of Nonverbal Behavior, edited by Pierre Philippot,

Robert S. Feldman, and Eric J. Coates” [Book Review], 483.STAHLBERG, DAGMAR, SABINE SCZESNY, and FRIEDERIKE BRAUN, “Name Your

Favorite Musician: Effects of Masculine Generics and of Their Alternatives in Ger-man,” 464.

TAKAKU, SEIJI, BERNARD WEINER, and KEN-ICHI OHBUCHI, “A Cross-CulturalExamination of the Effects of Apology and Perspective Taking on Forgiveness,”144.

TAYLOR, DONALD M., see Louis, W.VASCONCELOS, VIVIANE C. B., see Beaumont, S. L.WALES, ROGER, see Pedlow, R.WEINER, BERNARD, see Takaku, S.WELTMAN, DAVID, “Discourse in Late Modernity: Rethinking Critical Discourse Analy-

sis, by Lillie Chouliaraki and Norman Fairclough, Applied Discourse Analysis: Socialand Psychological Interventions, edited by Carla Willig, and Critical Textwork: AnIntroduction to Varieties of Discourse and Analysis, by Ian Parker and the Bolton Dis-course Network” [Book Reviews], 487.

WIGGINS, SALLY, “Construction and Action in Food Evaluation: Conversational Data,”445.

WILSON, MARC, see McClure, J.WINN, LAURA L., and DONALD L. RUBIN, “Enacting Gender Identity in Written Dis-

course: Responding to Gender Role Bidding in Personal Ads,” 393.WINN, LAURA L., HOWARD GILES, and JAMES J. BRADAC, “Language, Truth, and

Social Reality in the Work of W. Peter Robinson: A Prologue,” 5.


Articles:

“Children’s Production and Comprehension of Politeness in Requests: Relationships toBehavioral Adjustment in Middle Childhood,” Pedlow et al., 23.

“Children’s Referential Communication Failure: The Ambiguity and Abbreviation ofMessages,” Girbau, 81.

“Construction and Action in Food Evaluation: Conversational Data,” Wiggins, 445.“Co-Operative and Intrusive Interruptions in Inter- and Intra-Cultural Dyadic Dis-

course,” Li, 259.“A Cross-Cultural Examination of the Effects of Apology and Perspective Taking on For-

giveness,” Takaku et al., 144.“Deceptive Message Intent and Relational Quality,” Kalbfleisch, 214.“The Denial of Racism: The Role of Humor, Personal Experience, and Self-Censorship,”

Barnes et al., 321.“ ‘D’you Like a Drink Then Do You?’: Dissembling Language and the Construction of an

Impoverished Life,” Antaki, 196.“Enacting Gender Identity in Written Discourse: Responding to Gender Role Bidding in

Personal Ads,” Winn and Rubin, 393.“Epilogue,” Robinson, 248.“Implicit Causality in Language: Event Participants and Their Interactions,” Corrigan,

285.“Language and the Development of Children’s Ethnic Prejudice,” Nesdale, 90.“Language, Truth, and Social Reality in the Work of W. Peter Robinson: A Prologue,”

Winn et al., 5.“Name Your Favorite Musician: Effects of Masculine Generics and of Their Alternatives

in German,” Stahlberg et al., 464.“Older Adults’ Responses to Patronizing Advice: Balancing Politeness and Identity in

Context,” Hummert and Mazloff, 167.“Peer Argument in Educational Settings: Variations Due to Socioeconomic Status, Gen-

der, and Activity Context,” Howe and McWilliam, 61.“Similarities and Differences in Mother-Daughter and Mother-Son Conversations Dur-

ing Preadolescence and Adolescence,” Beaumont et al., 419.“A Tale of Two Histories: Language Use and Education in Relation to Social Class and

Gender,” Robinson, 231.“When People Explain Difficult Actions, Is the Causal Question How or Why?” McClure

et al., 339.“When the Survival of a Language Is at Stake: The Future of Inuttitut in Arctic Québec,”

Louis and Taylor, 111.

Book Reviews:

“Discourse in Late Modernity: Rethinking Critical Discourse Analysis, by LillieChouliaraki and Norman Fairclough, Applied Discourse Analysis: Social and Psycho-logical Interventions, edited by Carla Willig, and Critical Textwork: An Introductionto Varieties of Discourse and Analysis, by Ian Parker and the Bolton Discourse Net-work,” Weltman, 487.

“The Emergence of the Speech Capacity, by D. Kimbrough Oller,” Reissland, 382.“Folk Linguistics, by Nancy A. Niedzielski and Dennis R. Preston,” McGregor, 480.“Gender and Discourse, edited by Ruth Wodak,” Baldwin, 358.“Linguistic Cultures of the World: A Statistical Reference, by Philip M. Parker,”

McConnell, 367.“Moral Conflict: When Social Worlds Collide, by W. Barnett Pearce and Stephen W.

Littlejohn,” Rossman, 375.“Pour Une Écologie Des Langues Du Monde, by Louis-Jean Calvet,” McConnell, 470.“The Social Context of Nonverbal Behavior, edited by Pierre Philippot, Robert S.

Feldman, and Eric J. Coates,” Segrin, 483.

INDEX 497

Date post:	13-May-2023
Category:	Documents
Upload:	unibe-ch
View:	0 times
Download:	0 times

Name Your Favorite Musician: Effects of Masculine Generics and of their Alternatives in German

Documents