LanguageVariationandChange,18(2006),317–330.PrintedintheU.S.A. ©2006CambridgeUniversityPress0954-3945006$9.50 DOI:10.10170S0954394506060157 Nonresponsive performance in radio broadcasting: Acase study Juan Antonio Cutillas-Espinosa and Juan Manuel Hernández-Campoy UniversidaddeMurcia ABSTRACT The present study analyzes the speech of a radio presenter in a local station in Murciaandcomparesittotheaudience’slinguisticbehaviorasshowninthephone callsreceivedduringtheprogram.Wealsoanalyzethedataobtainedinaninterview withtheradiopresenter.Ourresults,whichshowaradicaldivergencebetweenthe presenter’sspeechandthatofhisaudience,arecontrastedwithbothAudienceDesign andSpeakerDesigntheoreticaltenets,usingtheexplicitknowledgeofthepresent- er’s attitudes and opinions to contrast theory and fact. We conclude that neither modeloffersacompletelysatisfactoryexplanationofthepatternsfound.Finally, wereflectontheneedtoconsidernotonlyperformance,butalsothescript(inthe formofaprofessionalvoiceusedfollowingaparticularlinguisticpolicybasedon sociolinguisticnormsandattitudestolanguage)thatconditiontheindividuallin- guisticbehavior,thussuggestingtheneedtoconsidercommunity-specificfactors intheexplanationofstylisticvariation. Styleenjoysapivotalpositioninsociolinguisticvariation,withstylistic(orintra- speaker) variation constituting a principal component together with linguistic variationandsocial(orinterspeaker)variation(seeEckert&Rickford,2001:1). Historicallyspeaking,thetraditionaldelimitationofstyleinvariationiststudies isbasedonthespeechstylescontinuumestablishedbythepioneeringstudiesby Labov(1966)andTrudgill(1972,1974)forthesociolinguisticinterview—style asareflectionofthespeaker’sattentiontohisorherownspeech(AS:Attention toSpeechmodel)andusingtheinterlocutorand0orthetopicand0orthecontext ofconversationasfactorsindeterminingthelinguisticvarietytobeemployedin a given situation. On the other hand, based on Giles’Speech Accommodation Theory(SAT;seeCoupland1980,1985,1988;Giles&Coupland,1991;Giles& Powesland, 1975; Giles & Smith, 1979) in the Audience Design model (AD), “intraspeakervariationisaresponsetointerspeakervariation,chieflyasmani- festedinone’sinterlocutors”(Bell,1984:158).Inordertosatisfytheaudience,to persuade them and to identify with them, we design our speech as if it were a We are very grateful for Peter Trudgill, David Britain, Natalie Schilling-Estes, Dennis Preston, DagmarScheu,RafaelMonroy,andJoséJiménez-Canoforcommentsonearlierversionsofthis paper. 317 318 J. A. CUTILLAS-ESPINOSA & J. M. HERNÁNDEZ-CAMPOY producttobesoldtoouraudience,adoptingthosefeaturesthataremoreaccept- ableforthem(seealsoBiber&Finegan,1994foradifferentperspective). Butstyleingeneralisamultidimensionalphenomenonthatcannotbemod- elledonasingleunidimensionaltheory,andthusstylisticstudieshavetoprogress, asEckert&Rickford(2001:2)suggest,inunderstandingasmorepermeablethe boundariesbetweenthethreemaincomponentsofsociolinguisticvariation:sty- listic(orintraspeaker)variation,linguisticvariation,andsocial(orinterspeaker) variation.Inthissense,someoftheaforementionedapproachesfollowaunidi- mensionalframeworkinthattheyareeitherderivativeofattentiontospeech(AS) or reactive to audience-related concerns (SAT andAD). In contrast, the much morerecentapproachofSpeakerDesigntheory(SD)viewsstylisticvariationas aresourceintheperforming(activecreation,presentation,andevenrecreation) ofspeakerindividualandinterpersonalidentity,thatis,stylisticvariationviewed as a resource for creating as well as projecting one’s persona (see Coupland, 1985,2001a;Eckert&Rickford,2001;Schilling-Estes,1999,2002;Traugott& Romaine,1985). OBJECTIVES Thisarticleisacasestudythatexaminesthepeculiaritiesofthespeechofaradio presenterinalocalstationinMurciaandcomparesittotheaudience’slinguistic behaviorinthecontextofavernacularspeechcommunity.Ourmainpurposeisto find out whether the Audience Design (AD) model or Speaker Design (SD) approachcanaccountforthepatternsobserved.Ifneitherofthemhappenstobe totallysatisfactory,weshalltrytoofferanalternativeaccount,basedonorderived fromADorSDapproaches,butwiththenecessarymodificationstoaccountfor thedataadequately. THE SOCIOLINGUISTIC SITUATION OF THE LOCAL DIALECT MurcianSpanish,likeAndalusian,canbestbeconsideredasasouthwardexten- sion of varieties originating in the central-north areas of the Iberian Peninsula (see Alvar, 1996; Hernández-Campoy, 2003; Hernández-Campoy & Jiménez- Cano,2003;Hernández-Campoy&Trudgill,2002;Lapesa,1988;Zamora-Vicente, 1989). Sociolinguistically speaking, this variety is stereotyped as ‘the orchard pro- nunciation’(“el habla de la huerta”), with connotations of ruralness, and even relativestigmatizationas‘badspeech’forMurcianspeakers(seeSánchez-López, 1999,2004;andCutillas-Espinosa,2001,2004),whoconsidereditunaesthetic, incorrect,inadequate,andsubstandard.Buttheparticularlycontradictorysitua- tion is that this overt view of their local variety as ‘bad speech’is for public consumptiononly.DespiteMurcianspeakers’negativevaluejudgmentsoftheir own speech, they do not abandon it entirely. In many ways, the local accent clearly has covert prestige (see Trudgill, 1972), combined with what could be NONRESPONSIVE PERFORMANCE IN THE MEDIA 319 calledalinguisticinferioritycomplex.InLabov’s(1966)terms,thisisanareaof linguisticinsecurityamongstlocalspeakerswithastrongdoubleconsciousness situation. METHODOLOGY Datagathering Linguisticdatawereobtainedbyanalyzingrecordingsoftheinteractionsbetween aradiopresenterandmembersofhisaudienceinadailymusicprogrambroadcast live by the MQM (Más Que Música) local radio station from 6 to 8 p.m. This program is broadcast from Santomera, a small town some 12 kilometers from Murcia.Itsaudienceparticipatesveryactively,byphoningtoaskforasongtobe played,ortogivetheirnumberinordertomeetotherpeople,orjusttotalktothe radio presenter.The average duration of calls ranges from 3 to 5 minutes. It is necessary to distinguish between the general audience of the program (all its listenersandtheiraveragesocialcharacterization)andthelimitedgroupofpeo- plewhoactuallyphonedtheradiostation,whosespeechistheobjectofstudyin thisarticle.Weshallcallthelattertheaudienceinterlocutorsfromnowon. Additionaldatawereobtainedfromaninterviewwiththeradiopresenter,both quantitativeandqualitative.Asfarasthequantitativeanalysisisconcerned,we lookedatthesamevariablesstudiedinthepresenter’sspeechontheradio.We wantedtocheckwhetherhishighuseofstandardformswasexclusivetobroad- castingorextensiveinhiseverydayuseoflanguage,thusmakingsurethatlin- guisticdivergencefromtheaudienceinterlocutorswasaphenomenonrelatedto themassmediasetting.Theinterviewwasconductedbytheauthorsthemselves inaninformalatmosphere,threemonthsaftertherecordingswerecarriedoutand oncethedatahadbeenanalyzed.Wetriedtoemphasizethisinformalitybyusing thenonstandardMurcianaccent.Bothofusarenativespeakersofthisaccent,and weuseitinoureverydaylife.Thechoiceoftheaccentwasconscious,aswehave alreadyremarked,butitisworthpointingoutthatitwasalso‘natural’insofaras theradiopresenteraddressedususingthissameaccentfromtheverybeginning of the interview. We tried to disguise the fact that we were questioning him, transformingthequestionsintosubject-raisingremarks,sothatbytheendofthe interviewwewereengagedinalively,relaxedconversation.1 As far as the qualitative analysis is concerned, we tried to elicit from him opinions that might provide evidence supporting some account of his stylistic behavior.Thequestionnairewasdevisedtoobtaininformationaboutthefollow- ingaspects,movingfromafirststageof‘subtle’askingtoamoredirecttypeof questionattheend:(a)presenter’sawarenessofthesocialandlinguisticfeatures of his audience (questions 1–7); (b) opinions about nationwide radio stations, theirpresenters,andtheirstatusasmodels(questions8–11);(c)attitudestoward Murcianspeech(questions12–16);(d)opinionsaboutmediacommunicationand languageuse(question17);and(e)directquestionsaboutthepresenter’sdiver- gencefromhisaudienceinterlocutor’sspeech(questions18–22). 320 J. A. CUTILLAS-ESPINOSA & J. M. HERNÁNDEZ-CAMPOY TABLE1.Typologyofinformants Numberof Groups TypeofInformants Informants NumberofInstances Group1 MQMlisteners 20 1,009 Group2 MQMradiopresenter 1 437(broadcasting) 602(interviewed) Total 21 2,048 Informants Twenty-one Murcian informants (20 listeners and the radio presenter) partici- pated,fromwhom1,446occurrencesofthelinguisticvariableswerequantified forthestudybasedontheradioprogram(Table1).Wealsoquantified602occur- rences from the presenter’s speech in the interview. The informants who were recordedandanalyzedweremostlymaleandfemaleadolescents(secondaryschool students)andmaleworking-classadults(skilledandnonskilledmanuallaborers: electricians, plumbers, wood workers, metal workers, etc.). Their phone calls usually took place while studying, in the case of the adolescents, or at work (sometimesontheroad)inthecaseofworkers.Thisinformationaboutthegeo- graphicalprovenanceandoccupationoftheaudienceinterlocutorswasregularly partofthecontentofthephoneexchangesbetweentheradiopresenterandthe listeners:Whoareyou?Whatdoyoudo?Whereareyoufrom?Whereareyou now? or What are you doing at the moment? and What do you want? or What wouldyouliketosay? Linguisticvariables Forthepurposeofourstudy,fourvariableswereselected:(r),(l),(s),andcon- sonant reduction, which are all salient features in this accent. The deletion of certainconsonantsimpliesvowelchangesthatareoutsidethescopeofourstudy (seeHernández-Campoy&Trudgill,2002). Variable (r). ‘r’ in syllabic coda position, as in comer (‘eat’) and carne (‘meat’),hastwodifferentrealizations:(1)themaintenanceof[r]asinStandard CastilianSpanishinallpost-vocalicpositions,and(2)r-deletion(word-finally) orassimilation(word-medially),whicharethenonstandardrealizations. realization (r) Variant1(standard): [r] ['karne],[ko'mer] Variant2(nonstandard) ø [ko'mE:] regressiveassimilation ['kænne] Variable (l). ‘l’in syllabic coda position, as in el (‘the’) and alto (‘high’) has three different realizations (one standard and two nonstandard): (1) the NONRESPONSIVE PERFORMANCE IN THE MEDIA 321 maintenance of [l] as in Standard Castilian Spanish in all post-vocalic posi- tions,(2)l-deletioninword-finalposition,and(3)regressiveassimilationwhen followed by another consonant (normally in word-internal position), or even liquid permutation: realization (l) Variant1: [l](standard) [el],['alto] Variant2: ø(nonstandard) [E:] Variant3: [r](nonstandard) [er] (liquidpermutation) ['arto] Variable (s). ‘s’in syllabic coda position, as in casas (‘houses’), has two differentrealizations:(1)themaintenanceof[s]asinStandardCastilianSpanish inallpost-vocalicpositions,and(2)s-deletionasthenonstandardform—unlike Andalusian, the ‘intermediate’aspirated form is not usual in Murcian Spanish (seePenny,1991,2000). realization (s) Variant1: [s](standard) ['kasas] Variant2: ø(nonstandard) ['kæsæ:] Consonantreduction. Thisincludescasesofconsonantloss,normallyafter a stressed syllable, as in the -ado0-ido verbal endings (cansado, ‘tired’) or the prepositionpara(‘to0for’),andthesimplificationofstresseddiphthongs(muy, ‘very’)byomittingthesecondvowelelement(foundinsomeelderlypeopleor uneducatedspeechingeneral): cansado: MurcianSpanish: [kan'sao] StandardCastilianSpanish: [kan'saDo] para: MurcianSpanish: [pa] StandardCastilianSpanish: [para] muy: MurcianSpanish: [mu] StandardCastilianSpanish: [mwi] RESULTS Tables2and3showthedifferencesinuseofstandardformsbetweentheaudi- ence interlocutors and the radio presenter (Figure 1). There is a radical diver- genceinlinguisticbehaviorbetweenthepresenterandtheaudienceinterlocutors, whichremainsmoreorlessstablethroughoutthefourdifferentvariablesunder study.2 Table4showstheresultsofthequantificationoftheinterviewdata,inwhich thepresenterdisplaysadrasticallydifferentlinguisticbehaviorfromthatshown inhisradioprogramspeech,adifferencethatisstatisticallysignificant:beingH o (cid:1)nullhypothesis,x2(H (cid:1)10.82;p,.001)(cid:1)395.026.Nonstandardformsare 0 322 J. A. CUTILLAS-ESPINOSA & J. M. HERNÁNDEZ-CAMPOY TABLE2.MQMaudiencemembers’scores Variable Variants TotalNo. Total% (r) [r] 370100 37% Ø0Assimilation 630100 63% (l) [l] 23065 35.4% [r] 19065 29.2% Ø 23065 35.4% (s) [s] 00359 0% Ø 3590359 100% Reduction No 17048 36% Yes 31048 64% Total Standard 770572 13.4% Nonstandard 4950572 86.6% TABLE3.MQMradiopresenter’sscores (broadcasting) Variable Variants TotalNo. Total% (r) [r] 64065 98.5% Ø0Assimilation 1065 1.5% (l) [l] 56056 100% [r] 0056 0% Ø 0056 0% (s) [s] 2630295 89% Ø 320295 11% Reduction No 18021 86% Yes 3021 14% Total Standard 4010437 92% Nonstandard 360437 8% TABLE4.MQMradiopresenter’sscore(interview) Variable Variants TotalNo. Total% (r) [r] 1120154 73% Ø 420154 27% (l) [l] 48057 84% [r] 0057 0% Ø 9057 16% (s) [s] 40351 1 Ø 3470351 99% Reduction No 15040 37.5% Yes 25040 62.5% Total Standard 1790602 30% Nonstandard 4230602 70% NONRESPONSIVE PERFORMANCE IN THE MEDIA 323 figure1. Percentageofuseofstandardformsbypresenterandaudienceinterlocutorsin thefourdifferentvariablesunderstudy. figure2. Presenter’suseofstandardforms(%)inbroadcastingandintheinterview. nowpredominantonthewhole,andthereweresignificantincreasesofnonstan- dardformsforallvariables(seeFigure2).Inspiteofthis,hisspeechisstillfar morestandardizedthanthatofhisaudienceinterlocutorsforvariables(r)and(l). Ontheotherhand,thepresenterisasnonstandardashisaudienceinterlocutors forvariables(s)and‘reduction’.Heusesthezerovariantof(s)inpracticallyall cases(99%)andaproportionofreductionthatapproximatelyequalsthatofhis audienceinterlocutors(seeFigure3).Thisreinforcesthestatusof(s)and‘reduc- tion’asvernacularlandmarks.Inaddition,itshowsthatthelinguisticbehaviorof thepresenteriscausedbythebroadcastingcontext,andconsequently,itcannot beregardedasageneralfeatureofhiseverydayspeech. 324 J. A. CUTILLAS-ESPINOSA & J. M. HERNÁNDEZ-CAMPOY figure3. Percentageofuseofstandardformsbythepresenter(intheinterview)andthe audienceinterlocutorsoftheradioprogram. ANALYSIS OF RESULTS AND DISCUSSION Two observable patterns arise: (1) radical divergence between presenter’s and audience’sspeechinbroadcasting,and(2)muchhigheruseofnonstandardforms inthecontextoftheinterviewwiththepresenter.Weshallnowhavealookathow thesepatternscouldbeexplained. TheAudienceDesignandSpeakerDesignmodels It seems that the predictions of Bell’s model are contradicted by the linguistic behaviorofthepresenterinourstudy.Hedoesnotseemtobeinterestedinexpress- ing‘sharedidentity’withtheaudienceinterlocutors,atleastbylinguisticmeans. However,theinterviewrevealedthatheisfullyawareofthesocialandlinguistic features of his potential audience. In his answer to question 1, he said that it consistedofworking-classpeople,teenagestudents,andhousewives,inthatorder. In spite of this, he designs a type of speech that does not fit their social and linguisticcharacteristics.Wecannotconfirmhereaudience-basedstyleshifting in the sense of Selting (1983) and Bell (1982, 1991, 2001). This conclusion is furthersupportedbythequantitativedataobtainedintheinterview.Itprovidesus withanotherparadox,becausetheaudienceinthiscase(tworesearchersfromthe University),probablybecauseoftheirstatusandeducationalbackground,should also have caused a high level of standardization—though we consciously used thelocalMurcianaccent.However,whatwegetisjusttheopposite(seeTable4). Forsomereason,thepresenterregardstheinterviewsituationaslessdeservingof hismoststandardformsthanthebroadcastingone.Someadditionalkindofcon- sideration(notaudience)isneededtoexplainthispattern. As discussed earlier, more recently, speaker design approaches suggest that stylistic variation is not a response to an existing situation, but rather a way to NONRESPONSIVE PERFORMANCE IN THE MEDIA 325 shape communicative situations and project a self-image. In other words, lan- guagedoesnotmerelyreflectcategoriessuchasgenderorstyle;rather,itisan essential component in the building of such categories (Bell, 1999; Cameron, 2000;Coupland,1985,2001a;Kiesling,1998;Wong&Zhang,2000;foranover- view, see Eckert, 2000 or Schilling-Estes, 2002).The speaker design approach wouldsuggestthatourpresentercouldbeperforminganidentityandbuildinga persona.But,sinceheadmitsinquestion13thathedoesnotparticularlylikethe accentheusesatwork(thestandard),whatarethemotivations?Inotherwords, thefactthattheannouncerperformsaparticularroledoesnotexplainthenature ofthatrole.Ourspeakeracknowledgesthatheisawareofsocialattitudestowards thevernacularaccentofMurcia,thestandardquestion,languageandthemedia, andsoforth.Itisobviousthatalloftheseinfluencesareessentialindetermining hisfinallinguisticperformance.Butinthisparticularcase,theperformanceitself doesnotprovideanexplanation;rather,itisjustasymptomofasocialpressure inaparticulardirection—morespecifically,awholesetofbeliefsaboutthestan- dardaccentinSpain,theaccentusedintheRegionofMurcia,andtheappropriate varietytobeusedinbroadcasting.Thus,thescriptofourpresenter’slinguistic behaviorusingaprofessionalvoiceisnotacompletelyfreechoice,giventhatit hasprobablybeenwrittenordecidedbysomeoneelseasalinguisticpolicyfor theradiostation,whichisbasedonlinguisticattitudes,norms,andconceptsof appropriatenessandcorrectnessthatgointhedirectionofthestandard.Tounder- standourpresenter’sbehavior,weinevitablyneedtogofurtherthantheperfor- mance,andanalyzethescript.Indoingthis,weinsistonthenotionthatindividual creativityisrestrictedbyrulesorstructuralconstraints(seeBell,1999;Cameron etal.,1992;Eckert,2000)andsuggestthattheseconstraintsactuallycomprisea unitmadeupoflinguisticattitudes,norms,andconceptsofappropriatenessand correctness.Thisunitiswhatwecallscript. Predetermined performance: Script as radio talk norms behindaprofessionalvoice Inthissectionwewillanalyzethedifferentaspectsthatmayhaveinfluencedthe linguistic behavior of our presenter. In other words, we shall try to analyze the ‘script’,understoodasakindoflinguistic(languageoraccent)policyorinstruc- tionstobecompliedwith,abstractedfromasetofattitudes,norms,andbeliefs about ‘appropriate’and ‘correct’speech in a given situation, within a specific communityunderstudy,andwhichnormally—butnotalways—gointhedirec- tion of the standard prestige variety.This script may take the form of a profes- sional voice used as a result of this linguistic policy, which may or may not coincidewiththeuser’slinguisticpreferences.Webaseouranalysisonsomeof thepresenter’sopinionsintheinterview. Allofthesefactorscontributetoamentalscriptinwhichthestandardassuch is hardly under discussion. Nonstandard varieties are accepted to be ‘wrong’ versions of some ‘right’linguistic variety, a fact that is reinforced by spelling conventionsandprescriptivepressures.Thisstateofaffairsisalsomaintainedby allowingaplacefornonstandardness.Someregionalaccentshavetheirownsta- 326 J. A. CUTILLAS-ESPINOSA & J. M. HERNÁNDEZ-CAMPOY tusandrespectabilityandthisminimizesthepossibilityofa‘revolt’againstthe standard.Thus,partialtolerancebecomestheperfectalibiforgeneralizedprej- udiceagainsttheregionalvarietiesandpraiseofthestandard. Our presenter answered quite clearly to the question of why he used such a standardaccenttoaddresshisaudienceinterlocutors.Hesaidthat,inbroadcast- ing, there are many people listening and it is inappropriate to use vernacular forms.Therearecontradictoryfeelingsanditseemsthatourpresenterisnotan exception,butratheraconfirmationofageneralpatternintheRegionofMurcia. Asstatedearlier,itiswhatLabov(1966)calledacaseofdoubleconsciousness. Ontheonehand,theMurcianaccentis‘coarse’,rural,even‘ugly’;ontheother hand, it is our accent, and it is as respectable as any other, and we should all defend it. In his answer to question 13, our presenter moves from the negative appreciationoftheaccenttoitspraisewithoutapause.Thisapparentcontradic- tion does not seem to be a problem. On the contrary, it helps to establish two well-differentiated spheres of influence—the Murcian accent is acceptable in familyandeverydaylife,butitshouldnotbeusedinthemedia. This pattern is further reinforced by broadcasting managers. Presenters are instructedtoavoidvernacularforms,whichareregardedasasignofdisrespect totheaudience.Pronunciationhastobe‘correct’(again,prescriptivismandthe popular beliefs about spelling and pronunciation), clear, and intelligible.There doesnotseemtobeanythingintrinsicallywrongabouttheMurcianaccent,but from the viewpoint of broadcasting, it is seen as inappropriate. These princi- ples are deeply rooted in the mental script of sociolinguistic behavior of the whole community, in such a way that the audience interprets linguistic diver- gence as a sign of respect, and not contempt. The broadcasting context is a sufficient justification. Additionally, Bell is right in stating that “communicators need to persuade theiraudiences,”butthispersuasiondoesnotalwaysnecessarilyimplytheuseof linguistic means, as Bell himself admitted in later work (see Bell & Johnson, 1997).Ourpresentershowsidentitywithhisaudience,buthedoesnotneedto expressitbyusingvernacularforms.Heaccommodatestothisaudience’stastes intopicsandmusic.Fromtheconversationalviewpoint,thetopicofshorttele- phone interactions always depends on the characteristics of the audience inter- locutor.Mostconversationsdealwiththecaller’sjob—whattheyaredoingwhile theytalktohim,wheretheyaregoingiftheyphonefromavehicle—oraboutthe caller’sstudies,withtheinevitablecomplaintsaboutteachersandexams;orelse they are concerned with interpersonal relationships, such as: Have you got a boyfriend? or Would you like to have one? These are the topics that all of his audiencewillbeabletounderstandandidentifywith. Inshort,MQMpresentersprovidetheiraudiencewiththemusicandthecon- versation they want. They do it by using an accent that, in most cases, is very distinct from their own. However, this does not seem to affect their audience’s interactionwiththepresenter.Itisassumedtobeapartofthenormalcommuni- cative conventions used in broadcasting. This reflects what Jiménez-Cano & Hernández-Campoy(2004:83)describedasbeingadiglossicsituationinMurcia: