[Adverbial[Prly]] NorbertCorver UtrechtUniversity,UiLͲOTS1 InthisarticleIexplorethenatureanddistributionof(mannerͲ)adverbialAPsinEnglish. Specifically, I address the following questions: (a) What is the nature of the adverbial marker–ly?(b)WhycanadverbialAPsbe“bare“(i.e.,without–ly)inpostverbalposition butnotinpreverbalposition.Theanswertothefirstquestionwillbethat–lyistheaffixal realizationofafunctionalcategoryPr.Theanswertothesecondquestionwillbetentative andmuchmoreexploratory. 1.Introduction Ever since Bowers (1975) and Emonds (1976), it is generally assumed within generative grammarthatlexicalitemssuchasthosein(1),i.e.wordsthataretraditionallyclassifiedas adverbs,reallybelongtothecategoryofadjectives(seealsoAbney1987). (1) quickly,easily,sharply,carefully Aclearindicationoftheiradjectivalnatureisthefactthattheycanbespecifiedbythesame classofelements,viz.,degreewordslikeso,too,how,very,lessetcetera. (2) a. Johnis[very/so/tooquick] b. Johnran[very/so/tooquickly] TheiradjectivalstatusisfurthermorecorroboratedbythefactthatinmanynonͲstandard varieties of English it is possible to realize the manner "adverb" as a bare adjective (see Bowers1975:fn.4,Wolfram&SchillingͲEstes1998:337Ͳ338,Edwards1993). 1 I would like to thank Noam Chomsky, Sabine Iatridou, Shigeru Miyagawa, Jairo Nunes, David PesetskyandMashaPolinskyfordiscussionofcertainpartsofthisarticle.Obviously,allerrorsare myown. 48 (3) a. Youpronouncedthatwordwrong. (Schidsbye1978) b. Hedrovedeadslowroundthecorner. (Schidsbye1978) c. Don’ttalksoloud! (Zandvoort1963) d. Ibeatthemeasy. (Curme1977) e. Heranquick.(Bowers1975) f. Hewritesbad.(Bowers1975) g. Theyansweredwrong. (Wolfram&SchillingͲEstes1998) h. You’redoingthatabittooslow. (Edwards1993) i. Itwasallcarpetedbeautiful. (TagliamonteandIto2002) TheaboveͲmentionedphenomenasuggestthattheformsin(1)arenotofthelexicalclass Adverb.TheyaresimplyinstancesofthecategoryAdjective. As noted in Sugioka and Lehr (1983) and Ross (1984), the distribution of these bare adverbialAPsturnsouttoberestricted:theycanonlyoccurinpostverbalposition.Ifinthose English varieties that permit bare APs in postverbal position, the adverbial AP appears in preverbalposition,thepresenceofͲlyisobligatory.Thisisexemplifiedin(4)Ͳ(6);examples aretakenfromSchidsbye(1965),SugiokaandLehr(1983),andRoss(1984): (4) a. Johnateitquick. b. Johnquick*(ly)ateit. (5) a. Heguessedrightthatshewasfifty. b. Heright*(ly)guessedthatshewasfifty. (6) a. Adamwrotehisnamecareful. b. Adamcareful*(ly)wrotehisname. Also in standard English, a bare AP is not allowed to occur in preverbal position. This is exemplifiedin(7).NotethatthesemanticallyequivalentadverbialAPquicklycánoccupythe preverbalposition.2 (7) a. Johnhas<*fast>lefttheroom<fast> b. Johnhas<quickly>lefttheroom<quickly> 2 The bare adverbial AP fast can occur to the left of the verb when it has undergone AͲbar movementto[Spec,CP],asin: (i) Iwonder[howfast]ishecanrunti 49 Ifthe"adverbs"in(1)areadjectives(AP),thefollowingquestionsarise:(a)Whatisthenature oftheadverbialendingͲly?(b)HowdoweaccountforthefactthatbareadverbialAPsare permittedinpostverbalpositionbutnotinpreverbalposition.Inthisarticle,Iwilltrytogive answerstothesequestions,somebeingmorespeculative,otherslessso. Thearticleisorganizedasfollows:section2startswithadiscussionof"adverbformation"in Welsh.Itwillbeshownthat(mannerͲ)adverbialAPsaretypicallyintroducedbythelexical itemyn.IwillproposethatthisitemrealizesafunctionalcategoryPr,whichisacategorythat connectsapredicateandasubject.Insection3IproposethatEnglishadverbialͲly,justlike Welshyn,isaninstanceofthisfunctionalheadPr.Itwillbearguedthatitsaffixalstatus(A+Ͳ ly)followsfromEmonds’sprincipleof‘alternativerealization’.Section4exploresapproaches towardsthequestionastowhypreverbaladverbialAPscannotbe"bare"inEnglish.Section 5concludesthisarticle. 1. AdverbialAPsinWelsh Adopting Chomsky’s (2001:2) Uniformity Principle, which states that “In the absence of compellingevidencetothecontrary,assumelanguagestobeuniform,withvarietyrestricted toeasilydetectablepropertiesofutterances”,Iwillstartmyinvestigationofthenatureand distributionoftheEnglishadverbialmarkerͲlywiththesyntaxofadverbialAPsinWelsh.The reasonforthisisthatWelshadverbialAPshaveaquitetransparantsyntacticstructure.The syntacticstructurethatIwillproposeforWelshadverbialAPswillformthestartingpointof myanalysisofEnglishadverbialAPs. In Bowers (1993:fn. 4) it is proposed that the Welsh particle yn, which shows up in predicateAPandNPconstructionslike(8),isarealizationofthefunctionalcategory"Pr"(a mnemonic for predication). According to Bowers, Pr is a functional head present in the syntactic representation of all predication constructions, which functions as a mediator betweenapredicate(typically,thecomplementofPr)anditssubject(thespecifierofPr).3 (8) a. MaeRhysynathro isRhysPTCteacher ‘Rhysisateacher’ 3Bower'sproposalthatpredicationrelationsinvolveaspecificsyntacticconfigurationinvolvinga mediatingelementisquitesimilartoDenDikken's(2006)proposalthatpredicationrelationships insyntaxinvolveanabstractfunctionalhead(hisRelatorͲhead)thatcanmanifestitselfindifferent forms(e.g.tobe,as). 50 b. Mae’rbwsyngynnar isͲthebusPTCearly ‘Thebusisearly’ Bowersproposesthefollowingstructureforasentencelike(8a): (9) [IPmae[PrP[NPRhys][Pr’[Pryn][NPathro]]]] Thefunctionalheadynin(9)takesthenominalpredicateathroasitscomplementandhas thesubject(externalargument)Rhysinitsspecifierposition. Whatisinterestingisthattheparticleynisalsotheelementwhichisusedtoturnan adjective into an adverbial element. That is, what yn is for Welsh adverbial APs, Ͳly is for EnglishadverbialAPs.Someexamplesofthe"adverbformation"bymeansofynaregivenin (10);examplestakenfromKing(1993). (10)a. cyflym quick a.’ yngyflym ‘quickly’ b. gofalus careful b.’ ynofalus ‘carefully’ c. prydlon punctual c.’ ynbrydlon ‘punctually’ d. araf slow d.’ ynaraf ‘slowly’ e. flynyddol annual e.’ ynflynyddol ‘annually’ ExtendingBowers’analysisofynasaninstanceofPrtotheadverbialexpressionsin(10),we getthefollowingstructurefortheseexpressions: (11)[PrPe[Pr’[Pryn]AP]] Iproposethatinasentencelike(12a),themodificationrelationbetweentheverbandthe adverbial expression is established by coindexation of the ‘subject’ of PrP (the specifier position)andtheEventroleassociatedwiththeverb(see(12b));cf.Higginbotham(1985): (12) a. Darllenwchyllyfrynynofalus ‘readthebookcarefully’ b. …ddarllenwych<,Ei>[PrPei[Pr’ynofalus]] NotethatbesidessimplexAP(i.e.,APsthatonlyconsistofanadjectivalhead,asofalusin (12a)),itisalsopossibletohavemorecomplexAPsinthecomplementpositionofyn,for exampleAPsthatconsistofanadjectiveandaspecifyingdegreeword: 51 (13) a. Dachchi’ngyrru’nrhyaraf ‘You’redrivingtooslowly.’ b. Gwnadyaithcartre’nfwygofalustronesa 'Doyourhomeworkmorecarefullynexttime.’ In(13a),wehavethesequence’nrhyaraf,i.e.,yn+too+slow,andin(1b)’nfwygofalus, i.e.,yn+more+careful.FollowingthesuggestionmadeinBowers(1993)thatynheads PrP,wegetthefollowingstructurefor’nfwygofalusin(13b): (14)[PrPe[Pr’yn[mwygofalus]]] (mÆfafteryn) 2. English–ly HavingdiscussedthesyntaxofmannerͲadverbialAPsinWelsh,letusnowreturntothe English "adverbs" in (1). Adopting Chomsky's uniformity principle, I propose that an adverbialformlikecarefullyhasthebasestructurein(15).Accordingtothisanalysis,Ͳlyisa functionalcategorythatconnectsasubject(specifier)andapredicate(AP)inapredication configuration. (15)[PrPe[Pr’[PrͲly][APcareful]]] Onemighthypothesizethattheformcarefullyisderivedbyheadmovementofcarefultothe boundͲmorphemic Pr Ͳly. However, deriving a form like carefully by means of AͲtoͲPr movement seems unlikely given examples like very carefully and too carefully, where a degree word combines with the adjective. Under a syntactic head movement analysis of carefully,wewouldexpectwordorderssuchascarefullyvery([carefuli+Ͳly[APveryti]])and carefullytootobewellͲformed.Obviously,theyarenot. Analternativeapproachtowardstheformationofadverbialformslike(very)carefully could be based on Emonds's (1985, 1987) notion of 'alternative realization'. Alternative realizationreferstothephenomenonthatsomegrammaticalfeatureAassociatedwitha closed class (i.e., functional) category F surfaces as an inflectional feature on the complement C of F. Emonds argues, for example, that the English comparative bound morpheme –er, which is realized on the lexical head A (e.g. tallͲer), is an alternative realization of the abstract comparative feature in Spec,AP, which in other adjectival contextscansurfaceasthefreemorphememore(e.g.,moreintelligent).If–erispresent, thefreemorphememoremustbeabsent(i.e.,*moretallͲer);thus,Spec,APmustremain 52 emptyinthatcase(see(16a)).Anotherillustrationofalternativerealizationcomesfrom ‘adverbialcase’NPs,including,forexample,GermanandGreekdatives.Emondsargues thattheseareactuallyPPsheadedbyanemptyP.The‘prepositionalfeature’isrealizedas acaseinflectiononthenoun(phrase)selectedbythephoneticallyemptyP;see(16b)). (16)a. [AP[Spece][A'A[+comparative]]] b. [PP[Pe][NPNp]] That the closed class category (i.e., functional category) must remain empty (i.e., not lexicallyinstantiated)followsfromEmonds’(1985,1987)InvisibleCategoryPrinciple(ICP), which states that a closed class category F bearing a feature A may remain empty throughoutasyntacticderivationifthefeatureisalternativelyrealizedinaphrasalsister ofB.Thus,apatternlikemoretallerisruledoutbyICP. Under a theory of alternative realization, the formation of adverbial APs like carefully can now be analyzed as follows: themediatingPrͲhead is realized as abound morphemeͲlyontheheadofitsadjectivalcomplement.Schematically: