on object posion in persian ditransives : movement or base...

1
On object posi-on in Persian ditransi-ves: movement or base-genera -on? Elias Abdollahnejad & Dennis Ryan Storoshenko School of Languages, Linguis=cs, Literatures, and Cultures Linguis=c Background The Issue Rela=ve posi=ons of the direct object (DO) and indirect object (IO) in Persian ditransi=ve structures depend on the specificity of the DO: a. Sāra [ IO barāye bache-hā] [ DO ketāb] mi-xun-e. S. for kid-PL book DUR-read-3SG ‘Sara reads the kids books.’ b. Sāra [ DO ketāb=ro] [ IO barāye bache-hā] mi-xun-e. S. book=OM for kid-PL DUR-read-3SG ‘Sara reads a specific book for the kids.’ Results & Discussion Generalized Es=ma=ng Equa=on (GEE) Model (unbalanced repeated- measures design) Priming Effect: RT Unrelated – RT Iden=cal 29ms 36ms 76ms Type(Posi=on)*Posi=on(Type): χ 2 (2)=12.695, p=.002 Bonferroni pairwise comparisons: ØType*Posi-on: üIden=cal TARGETs at Pre-Copy vs. Post-Copy, p=.001 ØPosi-on*Type: üIden=cal vs. Unrelated TARGETs: Pre-Copy, p=.073 Copy, p=.001 Post-Copy, p<.001 Delayed priming effect, i.e. 750ms ager Copy Not in line with A’-movement evidence, e.g. licencing parasi=c gaps and onward long-distance movement [email protected] [email protected] 40 na=ve speakers of Persian, tested in Canada 36 ditransi=ve Persian sentences “DO [+Specific] IO – V” Probe Type: Iden=cal to the DO or a seman=cally unrelated word (compa=ble in frequency and syllable number) Methodology Selected References Browning, M., & Karimi, E., (1994). Scrambling to object posi=on in Persian. Studies on Scrambling. Movement and Non-Movement Approaches to Free Word-Order Phenomena, Mouton de Gruyter, Berlin, 61-100. Chomsky, N., (1995). The Minimalist Program. Dordrecht: Foris. Karimi, S., (2003). On Object posi=ons, specificity and scrambling. In Simin Karimi (ed.), Word order and scrambling, 91–125. Oxford: Blackwell Publishers. Swinney, D. A., Onifer, W., Prather, P., & Hirshkowitz, M., (1979). Seman=c facilita=on across sensory modali=es in the processing of individual words and sentences. Memory & CogniEon, 7(3), 159-165. Swinney, D., Ford, M., Bresnan, J., & Frauenfelder, U., (1988). Coreference assignment during sentence processing. Language structure and processing. Stanford, CA: CSLI. Van de Koot, H., Silva, R., Felser, C., & Sato, M., (2013). Dutch A-Scrambling Is Not Movement: Evidence from Antecedent Priming, UCLWPL 25, 176, 200. Psycholinguis=c Background Trace Reac-va-on Hypothesis: The head of a movement chain is reac=vated in the cogni=ve system when the parser encounters the tail (Swinney et al., 1988) Cross Modal Lexical Priming (CMLP: Swinney et. al., 1979): While listening to a sentence, the par=cipant, at some manipulated point, sees a string of lerers (i.e. probe: real word or non-word) appearing on a computer screen for a limited =me Par=cipants make a lexical decision (real word or not?) If the displaced element is reac=vated at its original site, seman=cally related words will be recognized more quickly than unrelated words (Van de Koot et.al., 2013) Example: Probe: DANCER (iden-cal → lower RT) The dancer that the mayor adores dancer is very talented. Probe: RANGER (unrelated → higher RT) A’-movement: immediate reac=va=on, e.g. English object rela=ve clause (Love and Swinney, 1996), Japanese long-distance object scrambling (Nakano et al. 2002) A-movement: delayed reac=va=on, e.g. English Passive structure (Osterhout and Swinney, 1993), English unaccusa=ve structure (Friedman et al. (2008) Specific DO has features of both A- and A’-movement Judgment-based syntac=c analyses have not provided a clear analysis of this surface order varia=on Two basic structures: Varia=on in base-generated posi=ons: Two Object Posi=on Hypothesis (Karimi, 2003) i) IO DO [-Specific] –V ii) DO [+Specific] IO –V A- or A’- movement ? A-movement: Generally triggered by [Case] , Short distance, … A’-movement: Triggered by non-Case features, Associated with discourse, Long distance, … Movement: Syntac=c movement triggered by accusa=ve/specificity feature Underlying order: IO-DO [-Specific] -V (Mahajan 1992; Browning & E-Karimi 1994; among others). Probe posi=on Probe type #1 (control: 500ms before IO offset) #2 (IO offset: (Presumed copy) #3 (750ms ager IO offset) Iden=cal Condi=on 1 Condi=on 3 Condi=on 5 Unrelated Condi=on 2 Condi=on 4 Condi=on 6 Conclusion Our results support the claim that this surface varia=on in Persian ditransi=ves is the result of syntac=c movement A-movement, opposing analyses claiming that the alternate word orders are formed purely through different base-generated VP structures. x̄=310ms # 1 # 2 # 3 Sāra qol dāde bud ke qafasa =ro barāye [pesar-amu]-ā-š qafasa S. promise given was that shelf =OM for [son-uncle]-PL-his x̄=709ms tamiz va moratab bo-kon-e, vali hargez in kār=o na-kard. cleaning and organizing SUB-do-3SG but never this work=OM NEG-do “Sara had promised to clean and organize the shelf for her cousins, but she never did so.” Research Ques-ons: 1. Is varia=on in the surface form of Persian ditransi=ve structure result of movement (Browning & E-Karimi, 1994) or varia=on in base generated posi=on of DO (Karimi 2003)? 2. If the [+Specific] DO does move across the IO, is this best analyzed as A- or A’- movement? 560 570 580 590 600 610 620 630 640 650 660 Pre-Copy Copy Post-Copy Iden=cal 612 590 573 Unrelated 641 626 649 Mean Reac=on Time (ms) Probe Posi=on Iden=cal Unrelated IO-DO [-specific] -V DO [+specific] -IO-V Results are consistent with A-movement

Upload: others

Post on 21-Jun-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: On object posion in Persian ditransives : movement or base ...people.ucalgary.ca/~dstorosh/documents/LSA2019PrimingPoster2.pdf · Copy, p=.001 Post-Copy, p

Onobjectposi-oninPersianditransi-ves:movementorbase-genera-on?

EliasAbdollahnejad&DennisRyanStoroshenkoSchoolofLanguages,Linguis=cs,Literatures,andCultures

Linguis=cBackground

TheIssue

• Rela=veposi=onsofthedirectobject(DO)andindirectobject(IO)inPersianditransi=vestructuresdependonthespecificityoftheDO:

a.  Sāra[IObarāyebache-hā][DOketāb]mi-xun-e.S.forkid-PLbookDUR-read-3SG‘Sarareadsthekidsbooks.’

b.Sāra[DOketāb=ro][IObarāyebache-hā]mi-xun-e.S.book=OMforkid-PLDUR-read-3SG‘Sarareadsaspecificbookforthekids.’

Results&Discussion

• GeneralizedEs=ma=ngEqua=on(GEE)Model(unbalancedrepeated-measuresdesign)

• PrimingEffect:RTUnrelated–RTIden=cal29ms36ms76ms

• Type(Posi=on)*Posi=on(Type):χ2(2)=12.695,p=.002• Bonferronipairwisecomparisons:Ø Type*Posi-on:ü Iden=calTARGETsatPre-Copyvs.Post-Copy,p=.001

Ø Posi-on*Type:ü Iden=calvs.UnrelatedTARGETs:Pre-Copy,p=.073

Copy,p=.001Post-Copy,p<.001

• Delayedprimingeffect,i.e.750msagerCopy• NotinlinewithA’-movementevidence,e.g.licencingparasi=cgapsandonwardlong-distancemovement

[email protected]@ucalgary.ca

• 40na=vespeakersofPersian,testedinCanada• 36ditransi=vePersiansentences“DO[+Specific]–IO–V”• ProbeType:Iden=caltotheDOoraseman=callyunrelatedword(compa=bleinfrequencyandsyllablenumber)

Methodology

SelectedReferencesBrowning, M., & Karimi, E., (1994). Scrambling to object posi=on in Persian. Studies onScrambling.MovementandNon-MovementApproachestoFreeWord-OrderPhenomena,Moutonde Gruyter, Berlin, 61-100. Chomsky, N., (1995). The Minimalist Program. Dordrecht: Foris.Karimi, S., (2003). On Object posi=ons, specificity and scrambling. In Simin Karimi (ed.),Wordorderandscrambling,91–125.Oxford:BlackwellPublishers.Swinney,D.A.,Onifer,W.,Prather,P.,&Hirshkowitz,M.,(1979).Seman=cfacilita=onacrosssensorymodali=esintheprocessingofindividual words and sentences.Memory & CogniEon, 7(3), 159-165. Swinney, D., Ford, M.,Bresnan, J., & Frauenfelder, U., (1988). Coreference assignment during sentence processing.Language structure and processing. Stanford, CA: CSLI.VandeKoot,H., Silva, R., Felser, C.,&Sato, M., (2013). Dutch A-Scrambling Is Not Movement: Evidence from Antecedent Priming,UCLWPL25,176,200.

Psycholinguis=cBackground

• TraceReac-va-onHypothesis:Theheadofamovementchainisreac=vatedinthecogni=vesystemwhentheparserencountersthetail(Swinneyetal.,1988)CrossModalLexicalPriming(CMLP:Swinneyet.al.,1979):• Whilelisteningtoasentence,thepar=cipant,atsomemanipulatedpoint,seesastringoflerers(i.e.probe:realwordornon-word)appearingonacomputerscreenforalimited=me• Par=cipantsmakealexicaldecision(realwordornot?)

• Ifthedisplacedelementisreac=vatedatitsoriginalsite,seman=callyrelatedwordswillberecognizedmorequicklythanunrelatedwords(VandeKootet.al.,2013)Example: Probe:DANCER(iden-cal→lowerRT)Thedancerthatthemayoradoresdancerisverytalented. Probe:RANGER(unrelated→higherRT)• A’-movement:immediatereac=va=on,e.g.Englishobjectrela=veclause(LoveandSwinney,1996),Japaneselong-distanceobjectscrambling(Nakanoetal.2002)• A-movement:delayedreac=va=on,e.g.EnglishPassivestructure(OsterhoutandSwinney,1993),Englishunaccusa=vestructure(Friedmanetal.(2008)

•  SpecificDOhasfeaturesofbothA-andA’-movement•  Judgment-basedsyntac=canalyseshavenotprovidedaclearanalysisofthissurfaceordervaria=on

Twobasicstructures:•  Varia=oninbase-generatedposi=ons:TwoObjectPosi=onHypothesis(Karimi,2003)

i)IO–DO[-Specific]–Vii)DO[+Specific]–IO–V

A-orA’-movement?•  A-movement:Generallytriggeredby[Case],Shortdistance,…•  A’-movement:Triggeredbynon-Casefeatures,Associatedwithdiscourse,Longdistance,…

Movement:•  Syntac=cmovementtriggeredbyaccusa=ve/specificityfeature• Underlyingorder:IO-DO[-Specific]-V(Mahajan1992;Browning&E-Karimi1994;amongothers).

Probeposi=onProbetype

#1(control:500msbeforeIOoffset)

#2(IOoffset:(Presumedcopy)

#3(750msagerIOoffset)

Iden=cal Condi=on1 Condi=on3 Condi=on5

Unrelated Condi=on2 Condi=on4 Condi=on6

Conclusion

Our results support the claim that this surface varia=on in Persianditransi=ves is the result of syntac=cmovement A-movement, opposinganalyses claiming that the alternate word orders are formed purelythroughdifferentbase-generatedVPstructures.

x̄=310ms#1#2#3

Sāraqoldādebudkeqafasa=robarāye[pesar-amu]-ā-šqafasaS.promisegivenwasthatshelf=OMfor[son-uncle]-PL-hisx̄=709mstamizvamoratabbo-kon-e,valihargezinkār=ona-kard.cleaningandorganizingSUB-do-3SGbutneverthiswork=OMNEG-do

“Sarahadpromisedtocleanandorganizetheshelfforhercousins,butsheneverdidso.”

ResearchQues-ons:1. Isvaria=oninthesurfaceformofPersianditransi=vestructureresultofmovement(Browning&E-Karimi,1994)orvaria=oninbasegeneratedposi=onofDO(Karimi2003)?

2. Ifthe[+Specific]DOdoesmoveacrosstheIO,isthisbestanalyzedasA-orA’-movement?

560570580590600610620630640650660

Pre-Copy Copy Post-CopyIden=cal 612 590 573Unrelated 641 626 649

MeanRe

ac=o

nTime(m

s)

ProbePosi=on

Iden=cal Unrelated

IO-DO[-specific]-V

DO[+specific]-IO-V

ResultsareconsistentwithA-movement