Introducing Functional Grammar by Geoff Thompson
The purposes of linguisticanalysis
1.1Startingpoints
Amanisdrivingthroughapartofthecountryhedoesn’tknow,andhegetslostin whatlookstohimlikethemiddleofnowhere,completelydeserted.Finally,hesees anoldmanworkinginafield,andhestopsthecarandcallsouttohim,‘Excuseme, howdoIgetfromhereto...?’(thetowndependsonwhichcountryyouhearthe storyin).Theoldmanthinksforawhile,andthenhesays,‘Well,ifIwereyouI wouldn’t start fromhere.’
WhatIwantthisstorytohighlightisthefactthatwhereyoucangetto–in languagedescriptionasinanythingelse–dependsagreatdealonwhereyoustart from;andthatstartingfromthewrongplacemaymakeitmuchmoredifficulttoget tothedesiredkindofdestination.Inthesecondhalfofthelastcentury,therebuilt upanimmenselyinfluentialviewofwhatthestudyoflanguageshouldinvolve whichinsiststhatthereisonlyoneproperplacetostart–fromaviewoflanguageas anabstractsetofgeneralizedrulesdetachedfromanyparticularcontextofuse.It wouldbepossibletoignorethisviewandsimplystartwiththeapproachthatIwill besettingoutinthebook–basedonaviewofhowlanguagefunctionsasasystem ofhumancommunication.However,acomparisonofdifferentpossibleapproaches willhelpustounderstandbetternotonlythedestinationsthateachapproachallows ustoheadforbutalsothereasonswhywemightchooseoneoftheapproachesin preferencetoanother.Therefore,inthischapterIwillbrieflyoutlinetheapproach thatwasdominant,attemptingtoshowwhyitwassoattractivebutalsoshowingwhy anincreasingnumberoflinguistshavecometofeelthatitdoesnotmakeiteasyfor ustotalkaboutmanyofthemostcentralfeaturesoflanguage.Iwillthengoonto introduce an alternative approach which takes full account of those features, and whichoffersamoreappropriateplacetostartfromifweareinterestedinlanguage inuse.
Wecanbeginbylookinginformallyatabitoflanguage,selectedmoreorlessat random.Thiscomesfromanadvertisementaimedatattractingpeopletotakeup nursingasacareer.Beforereadingon,canyoudecidewhataspectsofthesentence youmightwanttoconsiderinprovidingalinguisticdescriptionofit?
Of course, you’re unlikely to be attracted to nursing because of the money.
WhenIhaveaskedstudentstodothiskindofpreliminaryanalysis,some(oftenthose whohavelearntEnglishasaforeignlanguageandthereforehavemorebackgroundin traditionalgrammaticalparsing)breakitupintoitscomponentsasfarastheycan(this isinfacttrickierthanitmightlook).Theylabelthepartsofthesentenceusingterms likeSubjectandVerb,ornon-finiteverbandprepositionalphrase.Theymaycomment onthefactthat‘tobeattracted’isapassiveform,andthattheunderstoodSubjectis ‘you’,carriedoverfromtheSubjectoftheprecedingverb‘(a)re’.Somementionthat thestructure‘beunlikelytobeattracted’isnotpossibleintheirownlanguageandthat, inaway,itisanillogicalstructure(sinceitisnot‘you’whoare‘unlikely’,but‘you beingattractedtonursing’).Whattheyareessentiallyfocusingoniswhatthedifferent partsofthesentenceareandhowtheyfittogether–inotherwords,theform.
MoststudentsforwhomEnglishistheirmothertongue,ontheotherhand,focus onissuessuchaswhoexactly‘you’is(sincethewriterisnotaddressinganyonefaceto face),andwhythewriterassumesthisabout‘you’soconfidently(‘Ofcourse’).Some pickupon‘you’reunlikelyto’,whichsoftensthepossiblearroganceofthewriter telling‘you’about‘your’ownfeelings;otherscommentontheimplicationthat‘you’ arelikelytobeattractedtonursingforotherreasonsapartfrommoney;andafew wonderwhythewriterdecidednottosay‘nursingisunlikelytoattractyou’.Whatall thesepointshaveincommonisthattheyareconcernedwiththefunctionofthe sentence,whatthewriter’spurposeisinwritingthesentence–inotherwords,with themeaning.Underlyingthepoints,thoughnotusuallymadeexplicit,isalsotheidea ofchoice:thattherearepotentiallyidentifiablereasonswhythewriterisexpressingthe messageinthisparticularwayratherthaninotherpossibleways.
Bothofthesewaysoflookingatthesentencetellussomethingusefulaboutit, and,intheinformaldescriptionsgivenhereatleast,thereisagooddealofpotential overlap.Anyfullanalysisofthesentencewillinevitablyneedtotakeaccountofboth themeaningandtheform(andofthelinksbetweenthem).However,inordertomaketheanalysisfairlyrigorousratherthanjustanunorderedlistofpointsaboutthe sentence,weneedtodecideonareasonablysystematicmethod;andinpracticethis involveschoosingbetweenformandmeaningasourstartingpoint.Thismayatfirst seemsimplyadifferenceinemphasis,but,ifcarriedthroughconsistently,each approachinfactendsupwithastrikinglydifferentkindofdescriptionoflanguage.
1.1.1Going in throughform
Themostfullydevelopedandinfluentialversionoftheapproachthroughformis thatproposedbyNoamChomskyandhisfollowers,originallyknownastheTG (Transformational–Generative) approach, although a number of variations have
developedfromthatstartingpoint.Chomskyinsistedthatlinguisticsshouldgo beyondmerelydescribingsyntacticstructures,andaimtoexplainwhylanguageis structuredinthewayitis–whichincludesexplainingwhyotherkindsofstructures arenotfound.Hearguedthat,inordertodothisadequately,itwasessentialtomake languagedescriptionabsolutelyexplicit.AlthoughtheaimofTGwasnottoproduce acomputerprogramthatcouldgeneratelanguage,itwascomputersthatprovided thedrivingmetaphorbehindtheapproach.Acomputeriswonderfullyliteral:it cannotinterpretwhatyoumean,andwilldoexactly–andonly–whatyoutellitto do.Thereforeinstructionstothecomputerhavetobeexplicitandunambiguous:this includesgivingtheminexactlytherightorder,sothateachstepinanoperationhas the required input from preceding steps, and formulating them so as to avoid triggeringanyunwantedoperationsbymistake.TGsetouttoproviderulesofthis kindfortheformationofgrammaticallycorrectsentences.(Notethatthefollowing outlinedescribesTGinitsearlyform.Thetheoryhaschangedradicallysincethe 1960s,becomingmoreabstractandmorepowerfulinitsexplanatoryforce;butthe basicconcerns,andthekindoffactsaboutlanguagethatitattemptstoexplain,have remained essentially thesame.)
Insettingupitsrules,TGstartedfromanotherdeceptivelysimpleinsight:that everyverbhasaSubject,andthatunderstandingasentencemeansaboveallidentifying theSubjectforeachverb.InEnglish,Subjectsnormallyappearinfrontoftheverb, soitmightbethoughtthatidentifyingthemwouldbetooeasytobeinteresting. However,therearemanycaseswheretheSubjectdoesnotappearinthe‘right’ position–ordoesnotappearatall(wehavealreadyseenthattheSubjectof‘tobe attracted’hastobecarriedoverfromadifferentverb).Wearesoskilledat understandingwhodoeswhatinasentencethatwetypicallydonotevennoticethat insuchcaseswehavetointerpretsomethingthatisnotexplicitlysaid.Onewell- knownexampleusedbyChomskywasthepairofsentences:
John is eager to please. John is easy to please.
Theseappear,onthesurface,tohavethesamestructure;butinfactweunderstand thatinthefirstcaseitisJohnwhodoesthepleasing(i.e.istheunderstoodSubjectof ‘toplease’),whileintheseconditisanunnamedpersonorthing(and‘John’is understoodastheObjectof‘toplease’).Thisgameof‘hunttheSubject’canbecome evenmorecomplexandexciting–thekindof(invented)sentencethatmadeTG linguistssalivatewithdelightisthefollowing:
Which burglar did the policeman say Mary thought had shot himself?
Here,weunderstandthattheSubjectof‘hadshot’is‘whichburglar’–eventhough therearetwootherpossiblenounsthatarecandidatesfortheSubjectrole(‘the policeman’and‘Mary’)inbetween.Addingtotheexcitementisthefactthatwealso understandthat‘himself’referstotheburglar,eventhough‘thepoliceman’iscloser inthesentence;whereas,ifwereplaceditwith‘him’,itmightrefertothepoliceman oranothermaleperson,butitcouldnotrefertotheburglar.
Buthowdoweunderstandallthis?Andhowcanthelinguistshow,inanexplicit way,whatitisthatweactuallyunderstand?Oneproblemisthat,inordertolabel partofthesentenceas‘Subject’,wehavefirsthadtoidentifythatpartashavinga particularrelationtotheverb(the‘doer’oftheverbratherthantheObjector‘done- to’):inotherwords,wehaveactuallyjumpedovertheinitialstage.Thatmeansthat ourdescriptionisnotinfactfullyexplicit.Weneedtoworkwithlabelsthattellus whateachconstituentisinitself,notwhatitdoesinthesentence.Atthesametime, wealsoneedtoshowwhereeachconstituentfitsinthebasicstructure.Chomsky’s famous first rule capturedthis:
S→NPVP
Thisisanon-verbal(andthusapparentlylessambiguous)wayofsayingthatevery sentenceinalanguageconsistsofanounphrasefollowedbyaverbphrase–ifitdoes notshowthesefeaturesitisnotagrammaticallyacceptable‘sentence’.Ithastobe borneinmindthatSactuallyreferstoaclauseratherthanwhatistraditionallycalled a sentence (in some later versions of the approach, the label ‘IP’, standing for inflectionalphrase,wasusedinstead);andVPhereincludeseverythingintheclause apartfromthefirstNP.Translatedintoover-simplefunctionalterms,itmeansin effectthateveryclausemusthaveaverbandeveryverbmusthaveaSubject.Using thisrule,theunderlyingmeaningsofour‘burglar’examplecanbesetoutasfollows, witheachofthethreeclausesinthesentencelabelledasanS(theinvertedcommas roundthewordssignalthatwearedealingwiththeabstractconceptsthatthewords refertoratherthanthewordsthemselves):
S1→NPVP
[‘the policeman’] [‘did say’ (something)]
S2→NPVP
[‘Mary’][‘thought’(something)]
S3→NPVP
[‘which burglar’] [‘had shot himself’]
Notethatthisanalysisalsobeginstoelucidatewhy‘himself’referstotheburglar. When the Object of a verb refers to the same entity as the Subject, a reflexive pronounisnormallyused:compare‘Marywashedher’and‘Marywashedherself’.
AsthefinalSabovesuggests,theVPelementdoesnotonlyincludetheverbbut anyotherelementsthatdependontheverb.Wecanthereforegoonsplittingthe clauseelementsintotheircomponentpartsuntilwereachthebasicconstituents (essentially words, though with some exceptions). This splitting up must,however, bedoneinthecorrectsequenceinordertoshowthedependenciesbetweendifferent partsoftheclausecorrectly.Forexample,two(simplified)furtherrulesare:
VP→VNP
NP→DetN
Thefirstruleallowsustoshowthatsomeverbphrasesconsistofaverbandanoun phrase(anounphraseinthispositionistraditionallycalledtheObject).Thisaccounts fortheVPinS3above:
VP→VNP
[‘hadshot’][‘himself’]
Thesecondruleallowsustoanalysewithinthenounphrase,andtoshowthatitmay consistofadeterminer(e.g.‘the’)andanoun(e.g.‘policeman’).
However,wehavenotyetdealtwiththeVPinS1orS2.Thiswillallowusto showhowS1–3combineintothesentenceasweactuallyseeit.Althoughtheoperation isimmenselycomplexinpractice,itissimpleintheory:itturnsoutthatwecan identifynotonlyafinitesetofexplicitrulesgoverningthepossiblecombinations(the complexitycomesespeciallyfromtheinteractionbetweentherules),but,more crucially,anevenmorerestrictedsetofunderlyingregularitiesinthetypeofrules thatarepossible.Thecrucialrulethatweneedtoaddis:
VP→VS
Thisrulemeansthatverbphrasesmayincludenotonlyaverb(V)butalsoanotherS (thisistechnicallyknownasrecursion:aclauseappearswheretheObjectmightbe). Thismaybeeasiertograspifwerevisetheanalysisofourexampletotakethesenew rulesintoaccount:
S1→NPVP→[VS]
[‘thepoliceman’][[‘didsay’][‘S2’]] S2→NP VP→ [V S]
[‘Mary’][[‘thought’][‘S3’]]
S3→ / NP / VP→ / [V / NP][‘which burglar’] / [‘had shot’] / [‘himself']
IhaveconcentratedsofarontheSubjectintheclauses,butexactlythesamekindof analysiscanbedoneforObjectsandotherclauseconstituentsthatappearinthe ‘wrong’placeorthatgoverntheformandinterpretationofotherconstituents(as ‘whichburglar’governstheinterpretationof‘himself’).WhataretheS1–3underlying thisversionoftheexample?
Which burglar did the policeman say Mary told him she had shot?
Itisperhapssurprisingthat,usingsuchapparentlymarginalexamples,theapproach shouldhavethrownsomuchlightonhowsentencesarestructured;andyetthe insightsgainedhavebeenextensiveandinsomewaysrevolutionary.Forourpresent purposes,however,itislessimportanttolookatthesediscoveriesinanydetailthan toconsiderwheretheapproachleadsus.Thefirstthingtosayisthatthisapproachis almostexclusivelyinterestedinwhatwecancall‘propositionalmeaning’–the
‘content’ofthesentence(notethat,fromthispoint,boldtypefacewillbeusedwhen animportanttechnicaltermisintroduced).Thefollowingtwosentenceshaveexactly thesamepropositionalcontentandthereforethesameanalysisintermsofSs:
Theburglarhadshothimself.Hadtheburglarshothimself?
S1→ / NP / VP → / [V / NP][‘the burglar’] / [‘had shot’] / [‘himself’]
Thedifferenceinsurfaceform(‘Theburglarhad’vs.‘Hadtheburglar’)resultsfrom rulesthatallowtheauxiliary‘did’toappearinfrontoftheNPastheStransforms intothesentences.Ontheotherhand,thefactthatastatementandaquestionserve entirely different functions in communication is regarded as irrelevant in the grammaticalanalysis–itistakenintoaccountinadifferentpartofthelinguistic description(thoughtherewasrelativelylittleinterestindevelopingthatpartwithin theapproach).Chomskymadeaprincipleddecisiontoexcludehowweusesentences incommunication(e.g.asstatementsorquestions):themodelisnotdesignedto show,forexample,thatonesentencefunctionsastheanswertoaprecedingquestion. Theaimistodiscovertherulesthatgovernhowconstituentscanbeputtogetherto formgrammaticallycorrectsentences,andtoformulatetheserulesinasgeneralaway aspossible(ideally,sothattheyapplytoallhumanlanguageratherthanjustindividual languages);thereforeeachsentenceisanalysedincompleteisolation,bothfromother sentencesandfromthesituationsinwhichitmightbeused.Thislimitationisself- imposedbecausegenerativelinguistsfeelthatitisonlyworthdescribingthoseaspects oflanguagethatcanbedescribed‘scientifically’(i.e.withabsoluteexplicitness).The waysinwhichlanguageisusedarethoughttobe,unfortunately,toomessyandare thereforeignored,atleastuntilsomeonecanfindawayofdescribingthemaccording toscientificgenerallaws.
Butiftheroadtowardsanexaminationofuseisblockedoff,whereelsecanwego fromthisstartingpoint?Theanswerisinwards,intothebrain.Thefactthatweas languageuserscanhandlethecomplexrelationsbetweenSsandclauses/sentences–
i.e.wecanidentifytheseparateconstituentsinthesentenceandassignthemtotheir correctplaceinthestructureoftheappropriateS–tellsus,itisargued,agreatdeal abouthowourbrainsmustwork.Atthesametime,thefactthatwedonotneedto beexplicitlytaughthowtodothismeansthatwemustinsomewaybebornwith therequiredmentalcapacities.Thusarigorouslyformalapproachtothedescription oflanguageleadsustowardsneurologyandgenetics.Clearly,thesearefascinating andworthwhileareas,buttheydoinvolvegivingupanyideaoflookingatlanguage inuse.InfactthelogicofChomsky’sapproachleadshimtoargueinOnNatureand Language (2002: 76) that ‘language is not properly regarded as a system of communication.Itisasystemforexpressingthought,somethingquitedifferent.’
1.1.2Going in throughmeaning
It may well be possible, and intellectually productive, to view language, asthe generativeapproachdoes,asasystemofabstractrulesthatareappliedinordertoend
upwithagrammaticallyacceptablesentence;buttherearegravedoubtsabout whetherthisviewcapturestoanyusefulextentwhatgoesonwhenusersactually produceorunderstandlanguage.Moreimportantly,thereislittledoubtthatitdoes notreflecthowtheusersthemselvesviewlanguage.Theyrespondabovealltothe meaningsthatareexpressedandthewaysinwhichthosemeaningsareexpressed.For theuser,despitetheclearsimilaritiesintermsofpropositionalcontent,thefollowing sentenceshaveverydifferentmeaningsbecausetheyaredesignedtoelicitdifferent responsesfromtheaddressee(acknowledging,agreeing/confirmingorinforming):
Colds last seven days on average.
Coldslastsevendaysonaverage,don’tthey? Docoldslastsevendaysonaverage?
Similarly,thereareimportantdifferencesbetweenthefollowingsentencesbecauseof thespeaker’schoiceofaformalorcolloquialwording:
Wouldyoumindhelpingmewiththis? Canyougissahand[=givemeahand]?
Thesyntacticunderpinningintheexamplesaboveisofcourseessentialinexpressing thedifferentmeanings,butonlyasatoolthatenableswhatmostpeopleseeasthe primaryfunctionoflanguage–communicatingmeaningsinparticularcontexts–to becarriedout.Asalways,theexactnatureofthetooluseddependsonthetaskin hand.Inlinguisticterms,wecanexpressthisastheassumptionthat,ifwestartfrom thepremisethatlanguagehasevolvedforthefunctionofcommunication,thismust haveadirectandcontrollingeffectonitsdesignfeatures–inotherwords,theform oflanguagecanbesubstantiallyexplainedbyexaminingitsfunctions.Ofcourse,we needtotakeintoaccounttheconstraintsofthe‘rawmaterials’:thepre-determined (genetic)characteristicsofthehumanbrainthatalloworencouragecertainkindsof languageforms,anddisallowordiscourageotherkinds.Generativeapproaches provideapossiblewayofinvestigatingthosecharacteristics(thoughtheirvalidityhas beenincreasinglyquestioned).Buttheyclearlyrepresentonlyhalfthestory:westill needtoexaminetheformativeinfluencesoftheusestowhichlanguageisput.(We canseethecontrastbetweenthetwoapproachesasareflectionoftheolddichotomy ofnaturevs.nurture–and,asalways,theanswerismostlikelytolieinacombination ofboth.)
Whathappens,then,ifweheadintheotherdirectionand(likelanguageusers) startfrommeaning?Themeaningsthatwemaywanttoexpress,ortheusestowhich wemaywanttoputlanguage,areclearly‘messy’:theyappearsovariedandso dependentontheinfiniterangeofdifferentcontextsthatitisdifficultatfirsttosee howwemightimposesomeorderonthem.However,ifwelookatthegrammatical optionsopentous,wecaninfactrelatethoseoptionsfairlysystematicallytodifferent kindsofmeanings.Letustakejusttwoexamplesofareasthatwewillexaminein moredetaillater.Wecanrelatethepresenceofmodalverbsto(amongstother things) expressing the speaker’s feeling that what they are saying needs tobe
negotiatedwiththeaddressee.Inthefollowingexample,thespeakerevaluates‘this seemingstrangeatfirst’asonlypotentiallyvalid(‘may’)toshowawarenessofthefact thats/hecannotbesurewhetheritdoesseemstrangetotheaddressee:
This may seem strange at first.
Andwecanrelatetheorderingofpartsoftheclausetothespeaker’sdesiretosignal howthismessagefitsinwiththeprecedingmessage(s).Comparewhatcomesfirstin thesecondsentencesineachofthesepairs(andthinkaboutwhytheorderisdifferent, andwhetherthesecondsentencescouldbeswapped):
Whatisaplatelet?Aplateletisadisc-shapedelementinthebloodthatisinvolved inbloodclotting.
Onekindofbloodcellisadisc-shapedelementthatisinvolvedinbloodclotting.
This is called a platelet.
Itmayseemodd(notemyuseof‘may’toavoidimposingthisopiniononyou!)to saythatorderingintheclausehas‘meaning’;butitisonlyoddifwerestrictmeaning to‘propositionalmeaning’–which,asIhavesuggested,isanarrowerdefinitionthan wewant.Ifwetakemeaningasbeingthesumofwhatthespeakerwantsthehearer tounderstand–inotherwords,ifweequatethemeaningofasentencewithits function–thenunderstandinghowthepresentmessagefitsinitscontextisclearly partofthemeaning,justasthedifferencebetweenastatementandaquestionispart of themeaning.
Indescribingthevariouskindsofmeaningsinthisfairlygeneralway(e.g.‘signalling howthismessagefitsinwiththeprecedingmessage(s)’),wearealreadybeginningto setupcategoriesoffunctionsthatweperformthroughlanguage;andwecanthengo backtotextstoseeifthereareothergrammaticalfeaturesthatseemtobeperforming thesamekindoffunction.Butwearestillindangerofendingupwithafairlyrandom- seeminglistoffunctions.Isthereanywayofarrivingatanevenmoregeneralized groupingofmeaningtypes,sothatwecanstarttoexplainwhywefindtheparticular kindsoffunctionsthatwedo?Forthis,weneedtostepbackand,ratherthanlooking atlanguagestructures,thinkaboutwhatwedowithlanguage.Inthebroadestterms, weuselanguagetotalkaboutthingsandevents(‘It’sraining’)andtogetthingsdone (‘Sitdown’).Asweshallsee,thesearenotmutuallyexclusive(thecommand‘Sitdown’ involvesreferencetotheparticulareventofsittingratherthananyother;andtelling someonethatit’sraininghastheeffectofchangingtheirknowledge):indeed,thebasic principleisthateverytimeweuselanguagewearedoingbothsimultaneously.Wewill alsoseethatweneedtoaddathirdmajorfunction,akindoflanguage-internal‘service function’;but,havingsimplyestablishedherethatitispossibletoidentifyaverysmall numberofbroadfunctions,wecanleavefurtherspecificationuntil,inChapter3,we startexploringhowthesemajorfunctionscanbeusedtoilluminateandexplainthe choicesthatareavailableinlanguage.
Ihaveatseveralpointsusedtheterm‘choice’indiscussingmeanings.Ifwewant toexaminewhatapieceoflanguageisintendedtodo(i.e.itsfunction),wecannot
avoidthinkingintermsofchoice.Clearly,speakersdonotgoroundproducing de-contextualized grammatically correct sentences: they have reasons for saying something,andforsayingitinthewaytheydo.Totakeasimpleexample,ifyou wanttofindoutsomeinformationyouaremostlikelytoaskaquestionratherthan makeastatement;and,atamoredetailedlevel,youaremorelikelytouseaninformal wordingifyouaretalkingtoafriendratherthanaformalone:
What the hell was that noise?
Butnotethat,indescribingtheexampleinthisway,wehaveinfactsetuptwosets ofcontext-dependentchoices:questionvs.statement,informalvs.formal.Ifyou havereasonsfordoing(saying)onething,theimplicationisthatyoucouldhave done(said)somethingelseifthereasons(thecontext)hadbeendifferent.
FunctionalGrammarsetsouttoinvestigatewhattherangeofrelevantchoicesare, bothinthekindsofmeaningsthatwemightwanttoexpress(orfunctionsthatwe mightwanttoperform)andinthekindsofwordingsthatwecanusetoexpressthese meanings;andtomatchthesetwosetsofchoices.Inordertoidentifymeaning choices,wehavetolookoutwardsatthecontext:what,inthekindofsocietywe
livein,dowetypicallyneedorwanttosay?Whatarethecontextualfactorsthat
makeonesetofmeaningsmoreappropriateorlikelytobeexpressedthananother? Butatthesametimeweneedtoidentifythelinguisticoptions(i.e.thelexicaland structural possibilities that the language system offers for use), and to explore the meaningsthateachoptionexpresses.Thesearecomplementaryperspectivesonthe samephenomenon:one,asitwere,fromthebottomup–fromwordingtocontext
–andtheotherfromthetopdown–fromcontexttowording.Lookingfromthe bottomup,theuseofthe‘thehell’inthequestionabovemeans–i.e.hasthe functionofexpressing–informality(amongstotherthings):inotherwords,one thingthatourgrammaticaldescriptionmustaccountforisthelexicalandstructural meansbywhichdifferentdegreesofformalityareexpressed.Lookingfromthetop down,thefactthatthespeakeristalkingtoafriendmakesappropriatetheuseof informalwordings:inotherwords,weneedadescriptionofthesocialcontextwhich includesdegreesoffamiliaritybetweenpeopleinteractingwitheachotherasa relevantfactorinfluencingtheirlanguagechoices.
Notethattheuseoftheterm‘choice’doesnotnecessarilyimplyaconscious processofselectionbythespeaker:whatweaimtouncoverthroughafunctional analysisarethemeaning-wordingoptionsthatareavailableinthelanguagesystem andthefactorsthatleadthespeakertoproduceaparticularwordingratherthanany otherinaparticularcontext(insomeways,itwouldalmostbeastruetotalkofthe wordingchoosingthespeaker).Inwritingthisbook,therearecertainchoicesthatI amveryawareofmaking–e.g.Ihaveconsciouslysetouttosound‘interactive’in thisbook,andsoIsometimesaddress‘you’directlyratherthanalwaysavoidingthis byusingpassives,etc.(bothoptionsarepossibleinatextbook,whereasinacademic journalarticles,forexample,directaddresstothereaderas‘you’isveryrareindeed). Buttherearemany‘choices’thatIamconstrainedtomakebythekindofcontextin whichIamusinglanguage:forexample,itisveryunlikelythatIwillusethestructures
associatedwithswearing,exceptperhapsinquotes.Itisonlyinconsciouslytryingto imaginethe‘wrong’choicesthatsuchchoicesevenpresentthemselvesaspossible: butthechoicenottoswearhasneverthelessbeenmade(or,rather,madeforme). Thesearedeliberatelycrudeexamples;buttheprincipleappliesineverydetailofthe wordingsthatI‘choose’.
Oneimportantimplicationofthefunctionalviewoflanguageisthatcontextand languageareinterdependent.Thismightseemtoostrongawayofputtingit:itlooks asthoughlanguagecouldbeseenasdependentoncontext.Forexample,ateacher mayask‘display’questionstowhichs/healreadyknowstheanswer,andtoevaluate theanswergivenbyapupilascorrectornot:
Teacher:Whatisthewomanwearingonherhead?Student:Ahat?
Teacher:Ahat,yes.
Onecouldassumethatthisis‘allowed’becauseoftheclassroomcontext,wherethe teacherhasaparticularkindofauthority;butitisequallytruetosaythat,byspeaking inthisway,theteacherandstudentarecontributingtocreatingthecontextasbeing thatofaclassroominteraction.Ifthesameteacherbehavedlikethiswiththesame studentwhentheyhappenedtomeetinthestreet,itwouldalmostcertainlybe inappropriate because it would project the context as if it were theclassroom. Similarly,ifaTVjournalistinterviewingagovernmentministeraskedadisplay questionandevaluatedtheminister’sanswerascorrect,itwouldsoundoddprecisely becauseitwouldconjureupthewrongcontext,withthewrongrelationshipbetween thetwospeakers.Wecanusetheterm‘construe’totalkaboutthiskindofreflexivity. Thequestionandevaluationoftheresponseconstrueaclassroomcontext:thatis, theysimultaneouslyreflectandconstructthatcontext.Totakeadifferentexample, ‘theglassbroke’construesaslightlydifferentviewofeventsfrom‘Ibroketheglass’ (hingingonthequestionofagency–seeChapter5).
Atabroaderlevel,ourexperiencesintheworldclearlyinfluencewhatwenormally talkaboutandthewaywetalkaboutit.Forexample,weconstantlyadjusttheway wetalktothepersonwearespeakingtosoastotakeintoaccountwhatwethink theyalreadyknow,andtonegotiateourmoment-by-momentrelationshipwith them(asIamdoingwithyou–notehowIhavechosentousethemoreinteractive ‘we’hereratherthan,say,‘speakers’);andthelexicalandgrammaticalresourcesof thelanguagethereforeofferwaysofconductingthisnegotiation.Atthesametime, thewaywenormallytalkabouttheseexperiences(andthewaywehearotherpeople talkaboutthem)influencesthewayweseethem:forexample,wegenerallyaccept withoutconsciousquerythefactthatadvertiserstalkabouttheirproductsassolutions toourproblems(asopposedtotalkingaboutourwillingnesstopayfortheproducts asthesolutiontotheadvertisers’problems,whichisatleastequallyvalidaview).
Byformulatingourapproachtolinguisticdescriptioninthekindoftermsused above–choicesamongstrelevantoptionsincontext–wearedeliberatelyopening upthepathtowardsgrammaticallybasedtextanalysis(where‘text’meansany
instanceoflanguageinuse):ateachstage,wecanaskwhythewriterorspeakeris
expressing this particular meaning in this particular way at this particular point.I mentioned earlier that generative approaches take linguistics towardsbiology; functionalgrammartakesittowardssociology:thesystematicstudyofrelevant featuresinthecultureandsocietythatformthecontextinwhichlanguageisused, andwhichareatthesametimeconstructedbythewayinwhichlanguageisused. Bothapproaches,throughformandmeaning,askessentiallythesamequestionabout language:howcanweexplainwhylanguagehasthemainfeaturesthatitdoes?But whereas the form-based approach finds the answer in the way our brains are structured,themeaning-basedapproachfindsitinthewayoursocialcontextis structured.(Ofcourse,thedifferentanswersdependverylargelyonthefactthateach approachtakesadifferentviewofthe‘mainfeatures’thatneedtobeexplained.) Althoughourfocusintherestofthebookwillbeonchoiceswithinthegrammatical systems,weshallberegularlylookingoutwardstowardsthewidercontextualfactors thatareconstruedbythesechoices.
1.2 Language, context and function: a preliminary exploration
Ifitistruethatlanguageandcontextareinextricablylinked,anynaturallyoccurring stretchoflanguageshould,toagreaterorlesserextent,cometrailingcloudsof contextwithit:weshouldbeabletodeduceagreatdealaboutthecontextinwhich thelanguagewasproduced,thepurposeforwhichitwasproduced,andthereasons whyitwasexpressedinthewayitwas.(Thisiswhyformallinguistsgenerallyprefer inventedexamples:apseudo-sentenceliketheburglarexampleaboveisdesignedto givenocluesabout‘distracting’elementssuchaswhomighthaveutteredthese words,inwhatcircumstancesorwhy.)Wecancheckthiscontext-embeddednessof reallanguageinapreliminarywaybylookingatasimpleexample.Ihavedeliberately chosenonethatconjuresupaveryclearcontext;butcanyougofromthattoexplain asmuchaspossibleaboutthelanguagechoicesintermsofwhotheinteractantsare andwhatthespeaker’spurposesare?Mycommentaryfollows,butyouwillfindit usefultotryyourownanalysisbeforereadingit.
Once upon a time, there was a big, bad bear.
Thecontextisobviouslyafairystory,probablytoldbyanadulttoayoungchild. Thisismostclearlysignalledby‘Onceuponatime’,whichisusedalmostonlyin fairystories(somuchsothat,ifusedinanothercontext,itconjuresupthevery specificfairy-talecontext,howeverfleetingly).Theindividualstorytellerhardly needsto‘choose’thisopening:heknowsthatthisishowfairystoriesstart.However, itisworthconsideringwhythistypeofnarrativeshouldhavesuchanimmediately recognizable opening. One important factor is the addressee: a relatively unsophisticatedlanguageuser,forwhomveryclearsignalsofpurposearenecessary. Theconventionalopeningsignalssomethinglike:‘I’mnotgoingtotellyoutodo anything;I’mnotgoingtoscoldyou;allyouneedtodoistositbackandenjoythe storythatiscomingup.’Inaddition,althoughtheexpressionbelongsgrammatically tothegroupofadverbialsthatspecifytime(‘Once’,‘Yesterday’,‘Threeyearsago’,
etc.),itclearlydoesnotinfactspecifyarealtime.Itthussignalsthatthenarrativeis afictionaloneratherthan,say,anaccountofwhatthetellerdidlastyear.
Theclausestructure(‘therewas...’)isanexistentialone(see5.2.5).Itintroduces oneofthemaincharacterswithoutsayingthatthebearwasinvolvedinanyparticular action–theactionwillpresumablystartinthenextclause.Thusitstagesthe information, building up the story in increments that are manageable to the inexperiencedlanguageprocessortowhomthestoryisaddressed.Whatwearetold aboutthebearapartfromitsexistenceisthatitisbigandbad.Thealliterationis obviouslystrikinghere:itappealstochildren’spleasureinincidentalpatterningsof sound,ratherlikewordplayatamoresophisticatedlevel(inmanyadulttextsweare morelikelytorewritesomethingtoremovealliterationifithappenstooccur).Atthe sametime,itservestoreinforcethenon-real,poeticnatureofthestory,perhaps reducingthepotentialscarinessoftheanimal(cf.theeffectof‘anenormous,savage bear’).Itisalsoworthcommentingonthefactthatthespeakerevaluatesthecharacter asheintroducesit.Insophisticatednarrativessuchasnovels,weexpecttobeskilfully guidedtowardsanevaluationofcharacterswithouthavingtheauthor’sevaluation thrustuponus;butherethechildistoldinadvancethatthebearisbad.Theadult takesontheresponsibilityofsettingouttherequiredsetofvaluesforthechild,partly nodoubtasareflectionofhisassessmentofthechild’srestrictedabilitytodothe necessaryinferencingforhimself.Inaddition,theevaluationopensupgeneric expectationsofhowthestorywillunfold:thebearwillsomehowcauseproblemsfor thegoodcharacterswhowillappearinamoment,butwillintheendbedefeated. Childrenlearnveryrapidlytorecognizeconventionalstorylines,aslongasthe signals are clearenough.
Theseareonlysomeofthemainpointsthatcanbemadeabouthowthispieceof languageworksinitscontext–Ihavenot,forexample,touchedonthebroader issuesoftheroleofstory-tellinginthesocializationofchildren.Ihavedeliberately outlinedthepointsasinformallyasIcan;butwhatIhopethediscussionshowsisthe kindoffeaturesthatwewanttobeabletodiscussinamoreformalizedway.The grammaticalsystemthatwesetupshouldprovidecategoriesthatrelatetothe communicativepurposesandchoicesthatwehaveidentified.Intherestofthebook, IshallbesettingoutafunctionalapproachbasedcloselyonMichaelHalliday’swork, whichallowsustodothisinasystematicandsatisfyingway.