7

UniversalGrammar1Slidesonthenetat:http://http-server.carleton.ca/~ecorazza/http://http-server.carleton.ca/~ecorazza/online_courses/Mind_World_and_knowledge/2Chomsky’sEpiphenomenalismaboutLanguageLanguagevs.Grammar “Grammar”isaprecisedefinitetermwhile“l(fā)anguage”isavagueandderivativetermwhichwecouldwelldispenseof,withoutmuchloss. Thegrammarinsomeonemind/brainisrealwhilelanguageisnot.3Theaimoflinguisticscanbesummarizedbyfourquestions.

1. Whatconstitutesknowledgeoflanguage?

2. Howissuchknowledgeacquired?

3. Howissuchknowledgeputtouse?

4. Whatarethephysicalmechanismsthatserveasthe materialbasis?4Deepvs.SurfaceStructurePortRoyalGrammar(1660)

ItisheavilyinfluencedbyDescartes. Itaimstoproposethegeneralformofanypossiblegrammar. Insodoingitelaboratestheuniversalstructureunderlyingthe“naturalmannerinwhichweexpressourthoughts”.5Theinner/outeraspectoflanguage AccordingtoPortRoyalgrammarianswemustdistinguishbetweenlanguagehavinganinnerandanouteraspect. Hencewedistinguishbetweenasentencequa

expressionofathoughtandthephysicalshapeofasentence(i.e.anutterance). Toshowthestructureofthemindthegrammarshouldreflectpropertiesofallminds,itshouldbeuniversal.6MentalGrammar Thedeepstructureisoftenonlyimplicitanddoesnotgetexpressed.Itisonlyrepresentedinthemind. Thesamedeepstructurecanberealizeddifferentlyindifferentlanguages(e.g.:“Videocanemcurrentum”and“Jevoisunchienquicourt”). Therulesofthisgrammararenotrepresentedinthelanguageuser:theyaresimplythere.Yettheymustbe“l(fā)earned”.Butseepovertyofthestimulusargument.7TransformationRules Therearetransformationrulesoperatingfromdeeptosurfacestructure.Itisthelinguist’sjobtofigureouttheserules. ThegrammariansofPortRoyalarethefirsttorecognizethetwosystemsofrules: 1. Abasesystemgeneratingdeepstructure.

2. Atransformationalsystemmappingthesedeep structuresintosurfacestructure.8UG

UGcorrespondstothedeepstructure.Sinceitistheexpressionofthough,itiscommontoalllanguages. Itisthusuniversal.HenceUniversalGrammar,UG. Thetransformationrulesconvertingthedeepstructureintosurfacestructuremaydifferfromlanguagetolanguage. Differentoutputscancorrespondtothesameinnerstructure.9PortRoyal WithintheCartesiantraditionexemplifiedbythegrammariansofPortRoyal,thedeepstructureiswhatconstitutesthemeaning(sense)inthemind. Itcanbetransmittedindifferentway(e.g.:active/passive). E.g.:differentlanguagesordifferentsurfacestructurestransmitthesamemeaning/sensewhichisamentalentity.10NowadaysUGmeanstheinitialstateofalanguagelearner. Itisthe“innate”(geneticallytransmitted)aspectofgrammaticalrules;thelanguageinstinct(Pinker). Itisthataspectofthehumanmindthatcausesonetolearnthelanguage. UGquainitialstateisbiologicallydetermined. Assuch,itdoesnotbelongtoaspecificlanguage.11UGneednotbesupposedtobewhatisuniversalamonglanguages(seeJackendoff2002:72ff.). Itismerelythehumancapacity,i.e.,theinitialstate,allowingonetolearnalanguage. Theaspectsoftheinitialsateoneendsupusinginone’slearningperiodsdependsonthestimuli/input. Languages(inputs)affectthedevelopmentoftheinitialstateandthustheoutputsoneendsupproducing(cf.switchesmetaphorexplainingthelearningofphonetics).12Innate Wedonotnecessarilymeanthatitispresentatbirthorinanembryo. Itrathermeansthatitautomaticallyappearsduringthedevelopment,regardlessonwhetheritispresentatbirthornot. Itdoesnotmeanthatitisfreefromtheinputoftheenvironment.E.g.visioncapacity.13Deepvs.SurfaceStructure,andCreativity Thedeep/surfacestructuredistinctioniswhathelpsexplaininglinguisticcreativity. ThePortRoyal’sdistinctionbetweendeepandsurfacestructureimplicitlycontainsrecursivedevicesallowingforinfiniteusesofthefinitemeansthatitdisposes. Thedeepstructureiswhatgetsrepresentedinthemindwhenasentenceisproduced/heard(seeLF).14Linguisticcreativityandtheargumentformentalgrammar Theexpressivevarietyoflanguageuseimpliesthatthebrainofalinguisticallycompetentusercontainsasetofunconsciousgrammaticalprinciples. (cf.JackendoffR.1994.PatternsintheMind.BasicBooksHarperCollins,NewYork:6).15LOT Inadoptingthelanguageofthoughthypothesis,LOT(orMentalese)theargumentformentalgrammarcanbestatedalongthecompositionalprincipleforthoughts,orwhatFodorcharacterizesastheproductivityofthought.

Theclassicalargumentthatmentalstatesarecomplexadvertstotheproductivityoftheattitudes…TheLOTstoryis,ofcourse,aparadigmofthissortofexplanation,sinceittakesbelievingtoinvolvearelationtoasyntacticallystructuredobjectforwhichacompositionalsemanticsisassumed.(FodorJ.1987.Psychosemantics.MIT:147-8)16Logicalvs.GrammaticalForm Arnauld&Nicole(inPort

RoyalLogic1662:160)highlightthedifferencebetweendeep(logical)structureandsurface(grammatical)structure. In:

(1) Nowfewpastorsatthepresenttimeare readytogivetheirlifefortheirflocks thegrammatical(surface)structureisaffirmative,whileitsunderlyingstructure(LF)isnegative.17 (1)containstheimplicitnegativesentence(“itcontainsthisnegationinitsmeaning”): (1a) Severalpastorsatthepresenttimearenotreadyto givetheirlivesfortheirflocks Thesamewith: (2) Comeseeme whosedeepstructureis: (2a) Iorder/begyoutocomeseeme18AccordingtothePortRoyalgrammariansthereisatransformationenablingtogofrom(1a/2a)(deepstructure)to(1/2)(surfacestructure). Wehavehiddenunderlyingstructureandagrammaticaltransformationsoperatingbetweenthedeepstructure(LF)andthesurface(orgrammatical)structure. E.g.:thesurfacestructure“OnlythefriendsofGodarehappy”isatransformationofthedeepstructure“ThefriendsofGodarehappy”and“allotherwhoarenotfriendofGodarenothappy”.19Understanding Tounderstandasentenceonemustgraspthesense,i.e.themeaning(“naturalorder”)thespeakerhasinmind. Onegraspsitinreconstructingitsmeaning,i.e.incomingtoentertainitsunderlyingstructure(LF)andthemeaningsofthesinglewords. Thefundamentalprinciplesatworkarereorderingandellipsiswhichenablethehearertorecoverinhermindthemeaningthespeakerhasinher.20LinguisticexplanationanddescriptionGrammaireGénérale(PortRoyal) Cartesianlinguisticsdidnotconfinetoameredescriptionofalanguageanditsgrammar. Itaimedtocapturetheuniversal(mental)structureunderlyinglanguages.

21PortRoyalgrammar,likemodern(Chomsky’sinspired)linguisticscanbeviewedasabranchofpsychologyorcognitivesciences. Thegeneralgrammarisakindofuniversalgrammar. Assuch,itdiffersfromthespecialgrammarwhichislanguagespecific.ItdiffersfromthegrammarofEnglish,Chinese,etc.22Linguistics/GeneralGrammarasaScience

GeneralGrammaris…therationalscienceoftheimmutableandgeneralprincipleofspokenandwrittenlanguage,whateverlanguagethismaybe…GeneralGrammarisascience,becauseitsobjectisrationalspeculationontheimmutableandgeneralprincipleoflanguage…Thescienceofgrammarisanteriortoalllanguagesinsofarasitsobjectspresupposeonlythepossibilityoflanguagesandarethesameasthosewhichguidehumanreasoninitsintellectualoperations…becausetheyareeternallytrue(Bauzé1767).23ShortcomingsofCartesianLinguistics(1600-1700)Theunderlyingassumption UG(theabstractstructureunderlyinganaturallanguagesentence)isakindofsentenceitself. Itisgenerallyassumedthatdeepstructureconsistsofactualsentencesinasimplerormorenaturalorganization.

24Theunderlyingassumptionisgratuitousandcanbedismissed. ItrestsontheCartesianideathatthegeneralprinciplesunderlyinganddeterminingourthoughtsandperceptionsmustbeaccessible

to

introspectionandcanbebroughttoconsciousnesswithcareandattention. IfweassumethatUGisunconsciouswedon’thavetoassumethatthegeneralprinciplearesentence-likeentities.25LanguageacquisitionUniversalConditions Theyarenotlearnedandmustexistforlanguageknowledgetobeexplained. Theyarethepre-requisiteleadingtoknowledge: principlesornotionsimplantedinthemind…adirectgiftofNature,aperceptofnaturalinstinct…[they]remainlatentwhentheircorrespondingobjectsarenotpresent,andevendisappearandgivenosignoftheirexistence(Herder1624).26Thiscontrastswiththeempiricistviewthat

ourmindisacleansheet,asthoughweobtainedourcapacityfordealingwithobjectsfromobjectsthemselves(Herder1624).

Themindisnotatabula

rasa.27Nativism/InnatismTheuniversalprinciplesareinnateandimplicit. Yet,wemayrequireexternalstimulustoactivatethemandmakethemavailabletointrospection. ThisisoneofthemainprinciplesunderlyingthepsychologyofCartesianlinguisticsandrationalismingeneral(seee.g.Leibniz).

[I]tistruethatitispurelyarbitrarytoconnectacertainideatooneparticularsoundratherthananother.Butideas—atleastthosethatareclearanddistinct—arenotatallarbitrarythingsdependingonourfancy.(Arnauld&Nicole1662:28)28Plato’sProblem NativismprovidesasolutiontoPlato’sproblem(cf.Plato’sMenoandTheaetetus). Foritprovidesascienceoflanguagethatshowshowaninternalbiologicalmechanismcan,withlittleinputfromtheexternalenvironment(povertyofthestimuliargument)develop(almostautomatically)ineachindividualtherichcompetenceknownas“knowingalanguage”.29SolvingPlato’sproblemforlanguageacquisition Itinvolvessayingbothwhatisknownwhenoneknowsalanguageandhowonecomestoknowit. Weshoulddothiswithascienceofthemind,notphilosophicalspeculations.

30Chomskyvs.Plato Platoappealstomyth,invokingthepre-existenceoftheboy’ssoulwithothersoulsintheworldofForms(ideas)andingoingtroughaprocessofreminiscence. Chomskysolvesitinproposinganaturalistictheoryofabiologicalsystemthatmakeslanguageacquisitionvirtuallyautomatic.31GeneralPresuppositionsofCartesianLinguisticsTheprincipleoflanguageandnaturallogicareknownunconsciouslyandtheyareinlargepartapreconditionforlanguageacquisitionratherthanamatterofinstitutionortraining. Linguisticsasasciencetryingtobringtolighttheseunderlyingprinciplesbecomesabranchofpsychology. Thusthisart[logicorartofthinking]doesnotconsistinfindingthemeantoperformtheseoperations,sincenaturealonefurnishedthemingivingusreason,butinreflectingonwhatnaturemakesusdo.(Arnauld&Nicole1662:23)32ThePovertyoftheStimulusArgumentGenerallanguage-acquisitionschema Input

LAD

Output (primary (Grammarconsistingof linguisticdata) principles,parameters andlexicon)33Languageacquisitionisamatterofgrowthandmaturationofrelativelyfixedprinciplesunderappropriateexternalconditionsandtraining. Cf.growthandmaturationofbones:thestructureofthebonesisgeneticallyprogrammed,yetitneedsexercisetodevelop.34Creolization Onelearnsalanguagebecauseoneisprogrammedtolearnalanguage,i.e.becauseofone’sinitialstate,UG. Theprocessofcreolizationunderlieswhathappenswhenachildlearnshermothertongueinnormalsituations.

Thesamekindoflinguisticgeniusisinvolvedeverytimeachildlearnshisorhermothertongue....letusdoawaywiththefolklorethatparentsteachtheirchildrenlanguage.(S.Pinker.1994.TheLanguageInstinct:39)35

Thecruxoftheargumentisthatcomplexlanguageisuniversalbecausechildrenactuallyreinventit,generationaftergeneration—notbecausetheyaretaught,notbecausetheyaregenerallysmart,notbecauseitisusefultothem,butbecausetheyjustcan’thelpit.(Pinker1994:32)Theargumentofinnateknowledge

Itrestsontheactualwaychildrenacquiretheirmothertongue. Itisanempiricalhypothesiswhichpositsthatourbrainisgeneticallyprogrammedtoinventalanguage.36

Chomsky

Large-scalesensorydeficitseemstohavelimitedeffectonlanguageacquisition.Blindchildrenacquirelanguageasthesighteddo,evencolortermsandwordsforvisualexperiencelike“see”and“l(fā)ook.”Therearepeoplewhohaveachievedclosetonormallinguisticcompetencewithnosensoryinputbeyondthatcanbegainedbyplacingone’shandonanotherperson’sfaceandthroat.

Theanalyticmechanismofthelanguagefacultyseemtobetriggeredinmuchthesamewaywhethertheinputisauditory,visual,eventactual,andseemtobelocalizedinthesamebrainareas,somewhatsurprisingly.

37

Theseexamplesofimpoverishedinputindicatetherichnessofinnateendowment—thoughnormallanguageacquisitionisremarkableenough,asevenlexicalaccessshows,notonlybecauseofitsrapidityandtheintricacyofresult.Thusveryyoungchildrencandeterminethemeaningofanonsensewordfromsyntacticinformationinasentencefarmorecomplexthattheycanproduce. Aplausibleassumptiontodayisthattheprinciplesoflanguagearefixedandinnate,andthatvariationsisrestrictedinthemannerindicated.Eachlanguage,then,is(virtually)determinedbyachoiceofvaluesforlexicalparameters:withthearrayofchoices,weshouldbeabletodeduceHungarian;withanother,Yoruba.…Theconditionsoflanguageacquisitionmakeitplainthattheprocessmustbelargelyinner-directed,asinotheraspectsofgrowth,whichmeansthatalllanguagesmustbeclosetoidentical,largelyfixedbyinitialstate.(Chomsky2000.NewHorizons…:121-2)38Theparadoxoflanguageacquisition

[A]nentirecommunityofhighlytrainedprofessionals,bringingtobearyearsofconsciousattentionandsharingofinformation,hasbeenunabletoduplicatethefeatthateverynormalchildaccomplishesbytheageoftenorso,unconsciouslyandunaided.(Jackendoff1994:26)39LanguagePerceptionandUnderstanding Perceptionofspeechrestsoninnatediscriminatorycapacities. Thereisafundamentaldifferencebetweentheperceptionofspeechandtheperceptionofunarticulatedsounds. Speechperception,unlikevisualperceptionforinstance,requirestheactivationofthegenerativerulesplayingtheroleintheproductionofspeech.40Boththeperceptualmechanismandthemechanismofspeechproductionmakeuseofthesameunderlyingsystemofgenerativerules. Itisbecausetheseunderlyingsystemsarethesameamongusthatcommunicationcanoccur.Itisbecauseofthisuniformityofhumannaturethatwetalkthewaywedoandsucceedinunderstandingeachothers(cf.Humboldt1836).41Everyyoungchild(raisedinanEnglishspeakingcommunity)wouldknowthatinEnglish‘blug’isphoneticallypossiblewhile‘bkr’is