https://ebookmass.com/product/language-change-variation-and-
Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...
Lexical Variation and Change: A Distributional Semantic Approach Geeraerts
https://ebookmass.com/product/lexical-variation-and-change-adistributional-semantic-approach-geeraerts/
ebookmass.com
(Re)Creating Language Identities in Animated Films: Dubbing Linguistic Variation 1st Edition Vincenza Minutella
https://ebookmass.com/product/recreating-language-identities-inanimated-films-dubbing-linguistic-variation-1st-edition-vincenzaminutella/
ebookmass.com
Expressing the Self: Cultural Diversity and Cognitive Universals Minyao Huang (Editor)
https://ebookmass.com/product/expressing-the-self-cultural-diversityand-cognitive-universals-minyao-huang-editor/
ebookmass.com
The One for You Roni Loren
https://ebookmass.com/product/the-one-for-you-roni-loren-2/
ebookmass.com
Psychology for VCE Units 3 and 4, 8e LearnON and Print 8th Edition John Grivas
https://ebookmass.com/product/psychology-for-vceunits-3-and-4-8e-learnon-and-print-8th-edition-john-grivas/
ebookmass.com
Advanced Concrete Technology, 2nd Edition Zongjin Li
https://ebookmass.com/product/advanced-concrete-technology-2ndedition-zongjin-li/
ebookmass.com
Smart Polymers and Their Applications 2 ed. 2nd Edition Aguilar Maria Rosa. (Ed.)
https://ebookmass.com/product/smart-polymers-and-theirapplications-2-ed-2nd-edition-aguilar-maria-rosa-ed/
ebookmass.com
Purchasing & Supply Chain Management 7th Edition Robert M. Monczka
https://ebookmass.com/product/purchasing-supply-chain-management-7thedition-robert-m-monczka/
ebookmass.com
Management Leading Collaborating in a Competitive World 13th Edition Thomas Bateman
https://ebookmass.com/product/management-leading-collaborating-in-acompetitive-world-13th-edition-thomas-bateman/
ebookmass.com
https://ebookmass.com/product/leading-managing-occupational-therapyservices-an-evidence-based-approach-2nd-edition-ebook-pdf/
ebookmass.com
LanguageChange,Variation,andUniversals
LanguageChange, Variation,andUniversals
AConstructionalApproach
PETERW.CULICOVER
3
GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom
OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries
©PeterW.Culicover2021
Themoralrightsoftheauthorhavebeenasserted
FirstEditionpublishedin2021
Impression:1
Allrightsreserved.Nopartofthispublicationmaybereproduced,storedin aretrievalsystem,ortransmitted,inanyformorbyanymeans,withoutthe priorpermissioninwritingofOxfordUniversityPress,orasexpresslypermitted bylaw,bylicenceorundertermsagreedwiththeappropriatereprographics rightsorganization.Enquiriesconcerningreproductionoutsidethescopeofthe aboveshouldbesenttotheRightsDepartment,OxfordUniversityPress,atthe addressabove
Youmustnotcirculatethisworkinanyotherform andyoumustimposethissameconditiononanyacquirer PublishedintheUnitedStatesofAmericabyOxfordUniversityPress 198MadisonAvenue,NewYork,NY10016,UnitedStatesofAmerica
BritishLibraryCataloguinginPublicationData Dataavailable
LibraryofCongressControlNumber:2021931013
ISBN978–0–19–886539–1
DOI:10.1093/oso/9780198865391.001.0001
PrintedandboundinGreatBritainby ClaysLtd,ElcografS.p.A.
LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.
4.2Constructionalinnovation
4.3.1Multiplegrammarsvs.multipleconstructions
4.3.2Definingcompetition
4.4.1Representationalcomplexity
4.4.2Computationalcomplexity
4.4.3Interpretivecomplexity
PARTII.VARIATION
5.Argumentstructure
5.3.1Differentialsubjectmarking
5.3.2Differentialobjectmarking
5.4Modelingdifferentialmarking
5.4.1AcquisitionofASCs
5.4.2Simulation
6.Grammaticalfunctions
6.3.1PlainsCreeargumentstructure
7.2.1Gapsandchains
7.2.2Relatives
7.2.3Topicalization
7.3Scopeinsitu
7.3.1Wh-in-situ
7.3.2Insituinpolysynthesis
7.3.3Otherinsitu
7.3.4Cryptoconstructionalinsitu
7.4ExtensionsofA constructions
7.5TowardanA
7.6Summary
PARTIII.CHANGE
8.ConstructionalchangeinGermanic
8.1Introduction
8.2BasicclausalconstructionsofModernGerman
8.2.1Initialpositionintheclause
8.2.2Positionofthefiniteverbinthemainclause
8.2.3Positionoftheverbinasubordinateclause
8.2.4Positionoftheverbinquestions
8.3ThedevelopmentofEnglish
8.3.1Thepositionoftheverb
8.3.2The‘loss’ofV2inEnglish
8.3.3Thelossofcasemarking
8.4ThedevelopmentofModernGermanfromOldHighGerman215
8.5Verbclusters
9.ChangesoutsideoftheCCore
9.1.1Reflexivityinconstructions
9.1.2Variationandchangeinreflexiveconstructions
9.2Auxiliary do
9.2.1Theemergenceof do
9.2.2Thespreadof do
9.3Prepositionstranding
9.3.1Whyp-stranding?
9.3.2P-passive
9.3.3Coercion
10.Constructionaleconomyandanalogy
Acknowledgments
Thisbookhasbeenalongtimeinthemaking,andhasbeenprofoundly influencedbymanypeople.Asalways,thankstoRayJackendoffandSusanne Winklerfortheirfriendship,collegiality,advice,andsupport.JackHawkins readseveralearlyversionsofthemanuscriptandgenerouslysharedhis insights—theycanbeseenthroughout.IamverygratefultoJeffersonBarlew fortheformaldescriptionoftheconstructionalframeworkdevelopedas partofourcollaborationonminimalconstructions,whichappearsinthe AppendixtoChapter2andisbasedonBarlew&Culicover(2015).AndIowe atremendousdebttoGiuseppeVaraschin,whoreadtheentiremanuscript invariousincarnationsandmadecountlessdetailedandconstructive suggestions,virtuallyallofwhichhavebeenincorporatedintothecurrent version.
ThanksalsotoBrianJosephandNoahDiewald,fromwhomIlearnedso muchinthecourseofourdiscussionsinourCreeReadingGroup,toGreg Carlson,AshwiniDeo,AdeleGoldberg,BjörnKöhnlein,AndrewMcInnerney, RafaelaMiliorini,LorenaSainz-MazaLecanda,RichardSamuels,YourdanisSedaris,AndreaSims,ShaneSteinert-Threlkeld,ElenaVaiksnoraite,and JoshuaWamplerforstimulatingdiscussionsonarangeoftopics,toPhilip MillerforhelpfulcommentsonthematerialinChapter3andforhisgeneral perspectiveonconstructionalapproachestogrammar,toAfraAlishahifor ourcollaborationonthesimulationofchangeinargumentstructureconstructions,toAndrzejNowakforourcollaborationonlanguagechange,and toMarianneMithunforhelpfulinsightintoactive-stativelanguages.Thanks toZoeEdmiston,whoseinterestinPlainsCreestimulatedmyown,andto MortenChristiansenandNickChaterforgivingmetheopportunitytowrite theforewordtotheirrecentbookandtothinkfreshlyaboutthefoundations oflinguistictheory.
Iamespeciallyindebtedtoseveralanonymousreviewers,whoseconstructivesuggestionshavepointedtoarethinkingofthisbookinwaysthathave ledtosignificantimprovements.Ofcourse,anyerrorsanddeficienciesthat remainaremyresponsibilityalone.
TheDepartmentofLinguisticsandTheOhioStateUniversityawardedme aSpecialAssignmentintheAutumnof2016,whichmadeitpossibleformeto
xacknowledgments makesignificantprogressonthefirstdraft.MythankstoShariSpeer,forher constantencouragementandsupport.
IoweadeepdebttotheworkofIvanSagandParthaNiyogi,andIwish Icouldthankthemnow—sadly,theyleftusfartoosoon,withsomuchleft forustodowithouttheirguidanceandinsights.Iamgratefulaswellto thelateJohnDaveyofOxfordUniversityPress,whoseunflaggingsupport andencouragementovermanyyearshelpedmethroughthepublication offivesubstantialvolumes,eachofwhichcontributessignificantlytothe foundationsofthepresentwork.Anditiswithgreatregretandsadnessthat Iamunabletosharethisworkwithmydearfriendandcolleague,thelate MichaelRochemont.
Finally,deepestthanksandmuchloveasalwaystoDiane,Daniel,andJen foralwaysbeingthere.
Preface
Thisbookbeganwithanaggingworry.Byandlarge,thegrammaticalliterature assumesthatgrammaticalfunctionssuchassubjectandobjectareuniversal andindependentlyrepresentedinsyntax,andplayanintegralroleinthe descriptionoftheform/meaningcorrespondencesthatcomprisewhatwecall ‘language’.Butthesesupposeduniversals,likemanyothers,poseseveralmysteries.Wheredotheycomefrom?Aretheypartofthebiologicalendowment forlanguagethatisencodedinourgenes?IsthereadevicecalledUniversal Grammarinourbrainsthatincorporatessuchnotionsassubjectandobject,or theequivalent?Didbiologicalevolutionselectforlanguageswhosegrammars makeuseofthesegrammaticalfunctions?
AsIlookedmoreintotheliteratureongrammaticalfunctions,itbecame clearthattheyarenotuniversal—notalllanguagesappeartomakeuseof them,andwhereitappearsthattheyareused,theyarenotthesamecrosslinguistically.Ofcourse,wefindtheterms‘subject’and‘object’usedallthe timetodistinguishtheparticipantsinarelationsuchas‘ThetigerbitSandy’in agivenlanguage.Anditispossibletostipulatethat‘thetiger’insomelanguage hasthesyntacticrepresentationofwhatwecall‘subject’inalanguagelike English.Butoncloserinvestigation,oftenwhatisbeingdistinguishedarethe phrasesthatdenoteentitieswiththematicroleslikeagentandpatient.
Itisoftensaidthat“Ajourneyofathousandmilesbeginswithasingle step”,andeachjourneystartsfromadifferentplace.Thisisonesuchjourney. Nomatterwherewestartfrominsyntactictheory,theinterconnectionstake ustoplacesthatwedidnotenvisionattheoutset.Inthiscase,asIthought aboutgrammaticalfunctionsandhowargumentsaredistinguishedcrosslinguistically,Ifoundmyselfengagedinsomethingmuchmorefar-reaching andambitious:theexplanationoflanguagechange,variation,andtypology. Whydoeschangeproceedincertaindirections,whydowegetthevariation thatwedo,whydowegetcertainvariantsandnotothers?Howdogrammars carryoutthetaskofencodingtheexpressivefunctionsoflanguage,andwhat, ifany,arethelimitsonhowthisisdone?Whyarecertainpatternsubiquitous, acrosslanguagesandinasinglelanguage,whileothersarerareornonexistent?
Itisofcoursepossibletoformulatedescriptionsofchange,variation,and typologicalpatternsinanyreasonablyexplicitdescriptiveframework.Butin ordertoproperlyexplainthem,weneedtherightdescriptiveframework.Ihad beenworkingonconstructionalphenomenaforsometime,beginninginfact withmydissertation(Culicover1971),morerecentlywiththepublicationof mybooks SyntacticNuts (Culicover1999), GrammarandComplexity (Culicover2013c),and ExplainingSyntax (Culicover2013b),andinmycollaborationwithRayJackendoffon SimplerSyntax (Culicover&Jackendoff2005).It seemedpromisingtopursueaconstructionalapproachtotheseissues.
Sointheend,thisbookisnotaboutgrammaticalfunctions,althoughthat isonestrand.Itisabouthowandwhygrammarsvaryandchange,andhence whytherearedistinctlanguages,understoodasovertexpressionsofdifferent grammars.Toaddressthesequestions,Iargueforaparticularconstructional approachtotherepresentationofgrammaticalknowledge,andIseekto showhowthisapproachhelpsusunderstandhowdifferentlanguagesand typologicalpatternsmightariseoutofgrammaticalchangeandcompetition betweengrammarsinthenaturalsocialandcognitiveenvironment.
Theorganizationofthebookisasfollows.PartIlaysoutthefoundations ofthisexploration.Itcoversastatementoftheproblem,areassessmentofthe notionofUniversalGrammar,atheoryofconstructions,andtheconceptual relationshipsbetweensyntactictheory,grammaticalvariation,andgrammaticalchange.Thispartofthebookconsistsoffourchapters.
Chapter1(Overview)laysoutthegeneralproblemofexplainingtheform ofgrammars,andrelatesthisproblemtothatofcharacterizinggrammatical complexity.Followingmuchrecentwork,Itaketheviewthatreductionof complexityisadrivingexplanatoryforcewhoseeffectscanbeseeninchange andinvariation.Intheoverview,Isketchoutthegeneralperspectivethat Iadoptonuniversals,conceptualstructure,constructions,complexityand changeandvariation,andhowtheyarerelated.
Chapter2(Constructions)setsoutanapproachtogrammaticaldescription inwhichthenotionsofChapter1canbeformallyimplemented.Theaccount isaconstructionalone.Ioutlinetheformalism,showhowitisusedtoaccount forgrammaticalphenomena,andhighlightitsutilityindescribingvariation andchange.
Chapter3(Universals)reviewstheapproachtouniversalsincontemporary grammaticaltheory,whichisthattheyareexpressionsofUniversalGrammar (UG),thehumanfacultyoflanguage.InpracticeUGisassumedtoconstrain syntax,andthusconstitutesanexplicitstatementofgrammaticaluniversals. Thischapterformulatesadifferentview,whichisthatwhatisuniversalis
conceptualstructure,andgrammaticaluniversalsandtypologicalpatterns ariseasaconsequenceofpressurestoformulateconstructionalgrammars thatexpressconceptualstructureassimplyaspossible.Whileitispossible tocharacterizechangeandvariationinanysufficientlyexpressivedescriptive framework,Iarguethattheconstructionalapproachprovidesanaturalframeworkforexplaininglanguagevariationandchange.
Chapter4(Learning,complexity,andcompetition).Thischapterexplores howenvisioningalanguagelearnerasacquiringagrammarconsistingof constructionsallowsustoaccountforchangeandvariation.Ideveloptheidea thatchangeisnotsolelytheresponsibilityofearlylanguagelearners,butmay alsooccurasinnovationsinitiatedbyadultspeakers.Onekeyexplanatory componentisconstructionalcomplexity;anotheriscompetitionbetween constructionsthathaveoverlappingfunctions.
InPartIIIlookatanumberofcasesofvariationinconstructionsthat dealwithtwocoreexpressivefunctionsofalanguage:argumentstructure andA /filler-gapconstructions.Ishowhowtheconstructionalframework providesaformalapparatusthatissuitablebothfordescribingthephenomena inagivenlanguageandforaccountingfortheobservedvariation.Thispartof thebookconsistsofthreechapters.
Chapter5(Argumentstructure)appliesthetheorytovariationinsystems forexpressingargumentstructure.Acentralpointisthattherearemultiple devicesofcomparablecomplexitythatencodethethematicroles;henceitis notnecessarytoassumethatalllanguagesshareauniformsyntacticstructure atsomeabstractlevelofrepresentation.
InChapter6(Grammaticalfunctions)Ireturntothequestionthattriggered thisproject,thesourceofgrammaticalfunctions(GFs).Ireviewevidence thatnotalllanguagesrequireGFs,andshowhowtocapturetherelevant correspondencesbetweenformandmeaninginconstructionalterms.
Chapter7(A constructions)appliesthetheoryoftheprecedingchaptersto A constructions,suchaswh-questionsandrelativeclauses.Themainresult ofthischapteristhatthereisarangeofwaysinwhichtheconceptual‘work’ associatedwiththeseconstructionscanbeexpressedinthecorrespondence betweensyntax,phonology,andmeaning.Noneoftheminvolve‘movement’ intheclassicalsense,althoughconstructionscanexpresslinksbetweenconstituentsnotincanonicalpositionrelativetotheirgoverningheads,givingthe illusionofmovement.
PartIIIappliesthemodelsofconstructionallearningandnetworkinteractionsdevelopedinChapter3toestablishtheplausibilityoftheaccountof changesketchedoutinPartI.Thispartofthebookconsistsofthreechapters.
Chapter8(ConstructionalchangeinGermanic)tracksseveralofthemajor changesinEnglishandGermanwordorderandaccountsforthemintermsof constructionalchangeasformulatedinChapter3.Itarguesthatthechanges inGermanicarerelativelysimpleinconstructionalterms,althoughthesuperficialresultsarequitedramatic.Amongthetopicsaddressedareclause-initial position,V2,VP-initialandVP-finalverbposition,thelossofV2andcase markinginEnglish,andverbclustersinContinentalWestGermanic.
Chapter9(ChangesoutsideoftheCCore)showsthebroaderapplicability oftheconstructionalapproach.Ilookatthreewell-documenteddevelopments inEnglishthatdonotfallintothecategoryof‘corephenomena’asunderstood inChapter3,reflexivity, do support,andprepositionstranding.Thesechanges arenotascentraltotheexpressivefunctionoflanguageasargumentstructure,operator/scope,andsimilarphenomena.Iarguethatthesephenomena provideadditionalevidencethattheconstructionalapproachiswell-suitedfor providinggenuineexplanationsforlanguagechangeandvariation.
InChapter10(Constructionaleconomyandanalogy),Ilookmoredeeply intowhatconstitutesconstructionaleconomy,andwhyitplaysarolein shapingtheformofgrammars.Iarguethatconstructionaleconomyisthe consequenceofwhathasbeencalled‘analogy’inthetraditionallinguistics literature.Specifically,Isuggestthateconomyinconstructionsderivesfrom placingahighvalueontheuseandreuseofthecomponentsoftheprocessing routinesassociatedwithconstructionalcorrespondences.Iapplythisgeneral ideatoseekexplanationsforarangeoftypologicalpatternsthatIreferto generallyas‘style’.
Chapter11(Recapitulationandprospects)summarizesthemainresultsof thisbookandlaysoutsomegeneralpropositionsabouthowtothinkfurther aboutlanguagevariationandchangefromtheperspectiveofconstructional grammars.
ListofAbbreviations
Adj adjective
ASC Argumentstructureconstruction
AUX Auxiliary
BDT BranchingDirectionTheory
CCore ConceptualCore
CS ConceptualStructure
CWG ContinentalWestGermanic
DM DistributedMorphology
DOM Differentialobjectmarking
DSM Differentialsubjectmarking
GB(theory)GovernmentBindingtheory
GF grammaticalfunction
HPSG Head-drivenPhraseStructureGrammar
IPP InfinitivusProParticipio
IS informationstructure
LFG LexicalFunctionalGrammar
LID LexemeIdentifier
MGG MainstreamGenerativeGrammar
MOD Modal
ModE ModernEnglish
ModG ModernGerman
N noun
Neg Negative
NP nounphrase
OE OldEnglish
OHG OldHighGerman
P&P PrinciplesandParametersTheory
PA ParallelArchitecture
PG Proto-Germanic
PLD primarylinguisticdata
PP prepositionalphrase
RG RelationalGrammar
SAI subjectAuxinversion
SD StandardDutch
UG UniversalGrammar
V verb
VPVerbPhrase
VPRVerbprojectionraising
WALS
WorldAtlasofLanguageStructures
WFWestFlemish
ZTZürichGerman
PARTI FOUNDATIONS
Overview
1.1Theproblem
Myprimaryconcerninthisbookishowhumanlanguagesgettobetheway theyare,whytheyaredifferentfromoneanotherincertainwaysandnotin others,andwhytheychangeinthewaysthattheydo.Giventhatlanguage isauniversalcreationofthehumanmind,thecentralquestioniswhythere aredifferentlanguagesatall.Whydon’tweallspeakthesamelanguage? Thischapterlaysoutthegeneralfoundationsofinquiryintothisquestionin contemporarylinguistictheory,andthespecificassumptionsthatinformthe answersdevelopedinthisbook.
Icallthiscentralquestion‘Chomsky’sProblem’.1Chomsky’sownanswer, hintedatinChomsky(1965)andfurtherdevelopedinChomsky(1973,1981) andotherwork,hasbeenthatinasensewe do allspeakthe‘samelanguage’. Whatweproduceistheexternalmanifestationofauniversal,biologically determined,abstractfacultyofthehumanmind,calledUniversalGrammar (UG).ThisclassicalChomskyanaccount,whichIrefertothroughout asMainstreamGenerativeGrammar(MGG),hasthefollowingmain components:
(i) Thereisasetofverygeneralgrammaticalprinciples,structures,and mechanisms,UG,whichdefinethecoregrammarsharedbyall languages.
(ii) TheseprinciplesandmechanismsarebiologicaluniversalsandconstituteI-language.
(iii) Someobservablevariationisduetodifferencesinthevaluesofcore parameters;henceI-languagetakesdifferentformsasdeterminedby theseparameters.Theparametersaresetbylearnersonthebasisof exposuretoprimarylinguisticdata(PLD).
(iv) Thesetofactualsentencesandtheirmeanings,producedbyagroup ofspeakers,referredtoasE-language,isitselfofnotheoretical
1 TheparallelwithChomsky’s“Orwell’sProblem”and“Plato’sProblem”isintended.
significance,exceptinsofarasitcountsasevidenceaboutI-language, UG,andtheparametersofvariation.
(v) PhenomenathatareoutsideofUGareintheperiphery.Theperipherycontainsidiosyncrasies,irregularity,andexceptions.Itcanvary widelyfromlanguagetolanguage,althoughnotwithoutprincipled constraints.
(vi) TheprinciplesandmechanismsofUGconstituteanoptimal solutiontotheproblemofexpressingthought.
Frameworksthatfallunderthegeneralperspectiveof(i)–(v)areGovernment/BindingTheory(Chomsky1981),PrinciplesandParametersTheory (Chomsky1981;Chomsky&Lasnik1993)and,with(vi),theMinimalist Program(Chomsky1995b,2000a).
IassumethatanysolutiontoChomsky’sProblemmusthavethegeneral structureof(i)–(v),butonlyinthefollowingsense.
(vii) Itmustexplainwhylanguagessharesomanyproperties,bothinform andinfunction.
(viii) Itmustattributethesepropertiestosomeuniversalsource— biological,cognitive,orsocial,oracombinationofthese.
(ix) Itmustaccountforthepossibilityofvariationandfortherangeof variation.
(x) Itmustaccommodatenotonlyregularitiesandgeneralizations,but idiosyncrasies,irregularity,andexceptions.
(xi) Itmustexplainwhycertainpropertiesarecommonwhileothersare rareordonotoccuratall.
(xii) Itmustexplainhowlearnersarriveatmentalrepresentationsof language—thatis,grammars—thataresuitablyclosetobutnot necessarilyidenticaltotherepresentation(s)ofmembersofthe linguisticcommunitythattheyarelearningfrom.
ThesolutionsthatIproposeinthisbookareinspiredbyChomsky’sProblem andtheChomskyanprogramofMGG,andplayoffofthem,buttheyarein manyimportantrespectsverydifferent,bothinspiritandinsubstance.Crucially,Chomsky’sapproachhasbeentoassumepoint(vi),i.e.that“language designmayreallybeoptimalinsomerespects,approachinga‘perfectsolution’ tominimaldesignspecifications”(Chomsky2000a,93).Chomsky’snotionof optimalityisbasedonanabstractnotionofeconomy,onethatisnotlinkedin anystraightforwardwaytothecognitivecapacitiesandlimitationsofhuman beings(Johnson&Lappin1997,1999).
Incontrast,Iassumethatgrammarsarenotcomputationallyoptimalin someabstractsense,butreflecttheoutcomeoftheneo-Darwinianevolutionof acomplexcognitivesystem(Laddetal.2008;Kinsella2009;Kinsella&Marcus 2009).Theevolutionofthissystemisdriven,atleastinpart,bythepressure toreducethecomplexityofthementalrepresentationofgrammaticalcompetenceandassociatedcomputationalcomplexityalongspecificdimensions. Iusethetermeconomytoreferpreciselytothispressure. TheapproachthatIargueforherehasthefollowinggeneralstructure.
(i) Regardingthecore,Iassumethatitisgroundedinconceptualstructure,notinasetofformalconstraintsongrammars.Thebasicjobof languageistoexpressthought.2IusethetermCCore(forConceptual Core)heretorefertothesetofexpressivefunctionsthatarecentralto humanthoughtanddiscourse.Agrammarofalanguageencodesthese functionsmoreorlessefficientlyandtransparently—functionssuch asargumentstructure,thematicstructure,interrogation,imperatives, description,binding,referenceandcoreference,restrictivemodification,negationandquantification,discoursestructure,andsoon.3The exactextentoftheCCore,anditsorigins,remainopenquestions.
(ii) Iassumethattheexpressivefunctionsthatlanguagesencodeare cognitiveuniversals.Thegrammaticaldevicesforexpressingthem,on theotherhand,aresocialuniversals,inthesensethattheyresideinthe mindsofspeakersinsocialnetworksinvirtueoftheirlinguisticcompetence,andaretransmittedsocially,notbiologically,acrosscultures andgenerations,throughcontactbetweenindividualsandgroupsof individuals.Thesesocialuniversals‘live’inthesocialnetwork,thatis, inthelinguisticcompetenceofallspeakersofalllanguagesacrosstime andspace.Theyareuniversalsinthesensethattheyareuniversally availablefortheexpressivefunctionsoftheCCore.However,noparticularwayofexpressingaparticularfunctionneedstobeabsolutely universalinthesensethatitisactiveinthegrammarofeverypossible language,apointthatIreturntobelow.
2 ThusIagreewithChomsky(1972).Theexpressionofthoughtislogicallypriortothecommunicationofthought,atleastthoughtthatcorrespondstorepresentationsthatareconstructedcombinatorily outofprimitiveelements(Fitch2011).Externalizationissubsequenttotheconstructionofthought, andismanifestedattheinterfacewithsoundandgesture.Thisisnottodeny,however,thatthe expressionofthoughtiscentraltocommunication,andthatsomeaspectsoflinguisticformmaybe explainedintermsofconstraintsimposedbythecommunicativetask.Chomsky(2005)seemstoaccept thisviewinhiscitationof‘thirdfactor’explanations,whichcomprisecommunicativeefficiencyamong otherthings.
3 Chomsky’sfocusintheMinimalistProgramonminimalmechanismsforexpressingargument structureandextraction(externalandinternalbinaryMerge)isarguablyaveryrestrictedvariantofthe approachthatIamtakinghere,takingsyntaxasaproxyforalimitedportionofconceptualstructure, andsettingasidemostconceptualandgrammaticalphenomena.
(iii) Regardingvariation,Iassumethatanythingthatcanpossiblybe expressedasacorrespondencebetweensoundandmeaningis, inprinciple,possibleinlanguage.Henceinprinciple,variationis unbounded.However,economy,thatis,thepressuresongrammarsto reducecomplexity,leadstoasignificantwinnowingofthelogical possibilities,inthespiritofmarkedness(Chomsky1965,chapter 1).Weexpectthatotherthingsbeingequal,thesimplestwaysof expressingtheCCorewillbemostwidespreadinsocialnetworksand perhapsevencompletelyuniversalinsomecases.⁴
ThechallengeposedbyChomsky’sProblemisafundamentalone.Howis itpossibletohavearestrictivetheoryofthehumanlanguagecapacitythat neverthelessallowsforthemassivesuperficialvariationandidiosyncrasythat isattestedintheworld’slanguages?EvenifweentertainChomsky’sviewofa highlyrestrictiveUG,wemustaccountforthevariation—simplybanishingit totheperipherywillnotsufficeasanexplanation.Itisimportanttoalways keepinmindthatthefullrangeofgrammaticalphenomenaisacquiredby learners,notjusttheparametricvariationdefinedoversomecharacterization ofarestrictedUGcore(Culicover1999;Culicover&Jackendoff2005).
Regardingthischallengeforthelearner,Chomsky(2013,37)saysthat eitherthereisaninfinityofoptions,inwhichcasechallengingandperhaps hopelessabductiveproblemsariseifthetaskistakenseriously;orthereisa finitenumber,andtheapproachfallsinprinciplewithinP&P[Principles andParametersTheory–PWC].Thatleavesopenmanyquestionsastohow parametersareset,andwhatroleothercognitiveprocessesmightplayin settingthem.
⁴ Dufteretal.(2009,12–13)raiseanumberofimportantquestionsabouthowtoexplainapparent constraintsonconstructionalvariation:
Onesuchquestioniswhetherthefactorsthatinfluencevariantdistributionshouldbean integralpartofthegrammarornot.Asecondtheoreticalissuethatneedstobeexplored furtheristhenatureofgeneralization.Cangeneralizationsaboutstructuralproperties withinalanguagebeformulatedinparallelfashiontocross-linguisticgeneralizations,using thesametypesofinheritancenetworks?Shouldn’ttherebeasomewhatdifferentstatus accordedtomoregeneraltypologicalprinciples,constraints,orparameters?Istherereally nodifferencebetweenthemodelingofmicro-variationandmacro-variation?Itseems thatconstructionschematacanbepostulatedratheradhoc,suchthatthelimitsofwhat ispossibleinlanguagedonotfollowfromthetheory(asitisclaimedbymodelsofthe Chomskyantradition).
IbelievethattheapproachtoeconomythatIdevelopinthisbookoffersausefulwaytoaddresssuch questions.
Challenging,yes,butnothopeless.InCulicover(1999),Iarguedthatthe idiosyncrasiesthatlanguageshave—theso-calledperiphery—areasrobust asthemoregeneralphenomenathathaveclassicallybeentheprovinceof parametertheory,e.g.wordorderandmovement.Theyarelearned,and nativespeakershaveclearintuitionsaboutthem.Atfirstglance,theperiphery appearstoallowunpredictablevariation.However,theperipheralphenomena ofalanguageprovetoberelatedinsystematicandoftenrevealingwaystothe moregeneralandregularphenomena.Andthelatterarebynomeansalways fullygeneralandfullyregular.
Theclassicalviewofvariationisthatitisparametric,inthesensethata parameterhasafinitenumberofpossiblevalues,preferablytwo,andeach languagehasaparticularsettingforeachparameter.Iargueinthecourse ofthisbookthatcharacterizingvariationintermsofparametersdoesnot shedmuchlightonthenatureandscopeofvariation.TheapproachthatI takeisthatwhatisessentiallyuniversalisnottheinventoryofformaldevices thatdefinethesyntaxofalanguage,asinMGG,buttheCCore,thatsetof conceptualstructurefunctionsthatalanguagemustencode.Expressingthe CCoreisthe‘work’thatagrammardoes.Syntacticstructureisonewayof encodingthesefunctionsandorganizingthemintosounds,morphological formisanother.⁵Asaconsequence,theuniversalarchitectureoftheCCore isinevitablyreflectedcross-linguisticallyintheorganizationofsyntaxand morphology,andeconomyrestrictstherangeofwaysinwhichthisworkis accomplished.
Thisperspectiveisguidedbyasetofideasandintuitions.Someoftheseare sharedwithMGGandsomearenot.
(i) Thoughtandtheexpressionofthoughtareuniversal. Allhumansare bornwiththesamecognitiveapparatusforformingthoughtsandthe samedrivetoexpressandcommunicatethesethoughtsinsoundand gesture.
(ii) Grammarisadistillationofthought. Throughthisdistillation,grammarbecomesautonomous,atleasttosomeextent,anditscategories onlyindirectlyandimperfectlycorrespondtothecategoriesofthought (see(iii)).BythisImeanthatthecategoriesandrelationsinour conceptualrepresentationsarereflectedimperfectly,andinatighter
⁵ Whilesomesyntactictheorieshavesoughttoreducemorphologicalformtosyntacticderivation (seeHarley&Noyer1999forareview),IargueinChapter5thatthereisnoempiricalmotivationfor suchastep.
andmorerestrictedway,inthecategoriesandrelationsthatconstitute grammars.⁶Forexample,ourconceptualcategoriesoftimearevery complex—theycoverthepast,thepresent,andthefuture,andtimes thatprecedeandfollowreferencetimesrelativetothemandtothe timeofutterance(Reichenbach2012[1958]).Thetensesystemsof languagesreflectsomeofthebasicdistinctionsofconceptualtime,but areforthemostpartsimplerthanthesemantics,andingeneraldonot mapdirectlyintoparticulartimesandtemporalrelationships.
(iii) Syntaxisautonomous. Autonomyofsyntaxmeansthatalthoughthere arecorrelationsbetweensyntacticstructureandmeaning,toasignificantextent,syntaxisnotreducibletomeaning,thatis,propositionalsemantics,informationstructure,anddiscoursestructure. Forexample,therearemanydistinctionsinthethematicrolesthat individualsmayplayineventsandstates,butfewofthesedistinctions aregrammaticallymarked(see,forexample,Dowty1991,aswellas Chapter5).
(iv) Alignmentandpackaging.Languagesmayvaryintermsofwhich aspectsofmeaningarepackagedtogetherinparticularmorphosyntacticunitsandcorrespondingphonologicalforms.Simpleexamples aretheexpressionsenter andgointoinEnglish.Inthecaseofenter,the meaningcomponents go′ and into′ arepackagedintoasingleword, whileinthecaseof gointo theycorrespondtodistinctwords.But packagingcanbequiteabitmorecomplexthanthis(Jackendoff2002; Slobin2004).
Theseintuitionsleadtothecharacterizationofagrammarasconsistingof constructions.Theformaldescription,function,andscopeofconstructions arediscussedatgreaterlengthinsubsequentchapters;inthischapterIprovide aninformalsketchoftheconstructionalapproachtogrammaticaldescription andexplanation.
⁶ AnearlyexpressionofthisrelationshipcanbefoundinPaul(1890,288): Everygrammaticalcategoryisproducedonthebasisofapsychologicalone.Theformeris originallynothingbutthetransitionofthelatterintooutwardmanifestation.Assoonasthe agencyofthepsychologicalcategorycanberecognisedintheuseoflanguage,itbecomesa grammaticalcategory.Itsagency,however,bynomeansendswiththecreationofthelatter. Itisitselfindependentoflanguage.Asitexistedbeforethegrammaticalcategory,soitdoes notceasetooperatewhenthiscomesintobeing.Inthiswaytheoriginalharmonybetween thetwomaybeinthecourseoftimedisturbed.Thegrammaticalcategoryistosomeextent apetrifactionofthepsychological.
InsteadofPaul’s“petrifaction”Iusetheterm“distillation,”butthebasicideaisthesame.
1.2Constructions
Thissectionsummarizestheessentialfeaturesoftheconstructionalapproach togrammar,andcontrastsitwithothersyntactictheories.
1.2.1Basics
First,thebasicarchitectureofaconstructionalapproach.Iassume,following Jackendoff(2002)andmanyothers,thattheminimaldescriptionofalinguistic objectsuchasawordoraphrase—aconstruct—involvesadescriptionofits sound(phon),itsgrammaticalstructure(syn)anditsconceptualstructure(or meaning)(cs).Thesearethetiers.Culicover&Jackendoff(2005)alsoassume thatthegrammaticalfunctionsarerepresentedonadistinctgftier.csmay comprisediscourseandinformationstructure,ortheymaybedistincttiers. Ileavethequestionopenhereofpreciselyhowtoincorporatetheseaspectsof meaning,asitappearstobeamatterofnotation,notsubstance.
SimplerSyntaxadoptstheviewtakeninHead-drivenPhraseStructure Grammar(HPSG;Pollard&Sag1994),Lexical-FunctionalGrammar(LFG; Bresnan&Kaplan1982),andvarietiesofCategorialGrammar(Jacobson1992; Kubota&Levine2013a,b;Morrill1995;Oehrleetal.1988;Pollard2004;Steedman1993;Uszkoreit1986;Zeevatetal.1987)thatsyntaxismonostratal.That is,thereisasinglesyntacticrepresentationthatcorrespondstorepresentations onotherlevels.Fromthisitfollowsthatthereisno‘movement’perse.Chains thatrelateconstituentsexternaltothebasicclauseandgaps,pronominal copies,oraffixesareproductsofthecorrespondencebetweenphon,syn,and cs,andarenotproducedbyderivingonesyntacticstructurefromanother.
AconstructioninthesenseofCulicover&Jackendoff(2005)describes arelationshipbetweenrepresentationsontwoormoretiersthatlicensethe propertiesofconstructs.Theessentialdifferencebetweenconstructionand constructisthedifferencebetweenadescriptionandtheobjectsthatsatisfy suchdescriptions;otherwisetheyhavethesamegeneralarchitecture,andare composedofmoreorlessthesameelementsfollowingthesameprinciples. Animportantdifferenceisthataconstructionmaycontainvariables,while aconstructdoesnot.Withotherconstructionalgrammarians,Iassumethat theconstructions,includinggeneralgrammaticalconstructions,idioms,and individuallexicalitems,resideintheextendedlexicon,oftencalledthe‘constructicon’(Jurafsky1996).
