− 1−

1. Introduction DevelopingHale’s (1982)ConfigurationalityParameter, Saito (2003) proposes the following (cf.Jelinek1984,Fukui 1986,Kuroda1988,Baker 1996,Miyagawa1997,etc.):

(1) TheDerivationalConfigurationalityParameter: Configurationallanguagesaresubjectto(1a-b),but Japanese-style non-configurational languages are not.

(a)Merge applies only to satisfy selectional requirements.(Mergeimpliesselection.) (b)Selectional requirementsmust be satisfied by Merge.(SelectionimpliesMerge.)

TheDerivationalConfigurationalityParameter in(1),Ibelieve,doesprovideuswithaninvaluableinsightintothepropertiesofbothconfigurationalandJapanese-style non-configurational languages. In this paper,however, I examine thenature of the first part of theparameter, (1a), which concerns freeword orderphenomenoninJapanese;Iclaimthatpreciselybecause(1a)hasmuchtheoreticalsignificanceinthatitimpliesaradical difference between configurational andJapanese-style non-configurational languages in coregrammar, it should be worthwhile attempting toconsider an alternative analysis from a differentperspective (cf.Hawkins 1988, 1994, 2004, 2014,Phillips 1996,Newmeyer 1998, 2005,Kempson et al.2001,BorseleyandBörjars2011,amongothers). In the following section, I attempt to demonstratehowSaitomotivates the first part of theDerivationalConfigurationalityParameter in (1). Morespecifically,in the section, I showSaito’s data of freewordorderphenomenoninJapanesetogetherwithhisanalysis. Insection 3, I try to showwhy (1a) is theoreticallyimportant,andthatbecauseof its importance, itshouldbeworth trying to consider an alternative analysis toaccount forSaito’sdata inadifferentway, i.e.withoutrelyingon anyparameter. There, I suggest thatwemight be able to account for his data naturally byappealingto thedynamicsof language(Hawkins1994,2004, 2014, Phillips 1996,Kempson et al. 2001,Borseley andBörjars 2011, etc.). In section 4, Iconcludethediscussionofthispaper.

2. The Derivational Configurational Parameter (Saito 2003)

Basedon the followingdata and argument,Saitofirstmaintains that scrambling in Japanese is differentfromtopicalizationinEnglish(cf.Saito1985,Whitman1987, amongothers). Namely, a topicalizedphrase inEnglishmust be interpreted as a topic,whereas ascrambledphraseinJapaneseneednotbeinterpretedassuch. Moreover, Saito claims that unlikeEnglishoperatormovement, Japanese scrambling does notcreate an operator-variable relation, and thus ascrambled phrase in the language can be literally‘undone’inLF(Saito1989). Observe first that in example (2b), theobject [that book]j is topicalizedwithin the embedded clause, andthe example is acceptable for native speakerswhoaccepttopicalizationquitegenerously.

(2)a. WhoitisaidthatJohnboughtthatbook b. Whoitisaid[that[thatbook]j,Johnboughttj]

In example (3b), on the other hand, theWh-object[which book]jistopicalizedinsidetheembeddedclause,andtheexamplecannotbeacceptedevenbytheabovementioned speakers,who accept topicalization ratherfreelyasin(2b).

(3)a. WhoitisaidthatJohnboughtwhichbook b. *Whoi ti said [that [whichbook]j, Johnbought tj]

Given the contrast between (2b) and (3b), Saito(2003, p. 326) suggests thegeneralization in (4), andaccounts for thedifference between (2b) and (3b) asfollows:

(4) AWh-phrasecannotbeinterpretedasatopic.

Thetopicalizedphrase[that book]jin(2b)isnotaWh-phrase. Hence, in (2b), [that book]j is allowed to beinterpretedasa topic inaccordancewithgeneralization(4).Thetopicalizedphrase[which book]jin(3b),ontheother hand, is aWh-phrase, and thus,may not beinterpretedasatopicdueto(4). Observenow that in (5b), theobjectWh-phrase[dono hon-o]jisscrambledwithintheembeddedclause,but isfullyacceptable. Notice that theacceptabilityof

ADynamicSyntacticAnalysisofScramblingandWHMovement

HOSHI,HirotoAkita University

秋田大学教育文化学部研究紀要　人文科学・社会科学部門　71　pp．１〜７　2016

Akita University

− 2−

(5b),which involves Japanese scrambling, contrastssharplywiththeunacceptabilityof(3b),whichinvolvesEnglishtopicalization.

(5)a. Taroo-wa[Hanako-gadonohon-o -Top-Nomwhichbook-Acc kattato]omotteiruno boughtthatthinkQ ‘[Q [Taroo thinks thatHanakoboughtwhich book]]’

b. Taroo-wa[[donohon-o]iHanako-gati -Topwhichbook-Acc-Nom kattato]omotteiruno boughtthatthinkQ ‘[Q [Taroo thinks thatwhichbooki,Hanako boughtti]]’

If scrambling in Japaneseparalleled topicalization inEnglish,andifascrambledphraseinJapanesehadtobeinterpreted as a topic, example (5b)would be asunacceptable as (3b) according to generalization (4).However, this is contrary to fact. To account for thecontrast between (3b) and (5b),Saito (2003, p. 327)therefore claims that scrambling in Japanese is distinctfrom topicalization in English in that unlike atopicalizedphrase inEnglish, a scrambledphrase inJapaneseneednot be interpreted as a topic. Hence, aWh-phrase canbe scrambled in Japanese as in (5b),whereas aWh-phrase cannot be topicalized inEnglishasin(3b),dueto(4). Furthermore, Saito claims that scrambling inJapanesedoes not create anoperator-variable relationunlikeWh-movementinEnglish,andsignificantly, thatJapanese scrambling canbe literally ‘undone’ inLF(Saito 1989), basedon the data below. Saito (1989,2003, etc.) first observes thatEnglish example (6) andJapaneseexample(7)arebothunacceptable.Saitocitesexample(7)fromHarada(1972).

(6) *John askedwho to findout [CPwhati [Mary boughtti]]

(7) *Taroo-gadare-ni[CPHanako-ganani -Nomwho-to-Nomwhat -okattaka]tazuneta(koto) -AccboughtQasked(fact) ‘(the fact that)Taroo askedwho [QHanako boughtwhat]’

To rule out these two examples, Saito suggestsgeneralization(8).

(8) AWh-phrasecanonly take scopeat aCP that containsit.

Thegeneralizationin(8)rulesoutEnglishexample(6),

because theWh-phrase [who] isaphraseof thematrixclause, and is not containedby any interrogativeCP.Similarly, the Japanese example in (7) is excludedby(8), because the indirectWh-object [dare-ni] is anelementofthematrixclause,andisnotcontainedwithinany interrogativeCP. Toput it differently, in both (6)and (7), not thematrixCPbut the embeddedCP is aninterrogativeCP, but the embedded clauses in (6) and(7) donot contain the abovementionedWh-phrases,[who]and[dare-ni].Hence,neither[who]nor[dare-ni]isallowedto takescopewithin the interrogativeclausein(6)and(7),duetogeneralization(8). In addition,Saito shows that generalization (8)accounts for the interpretive possibilities of thefollowingdataaswell,whicharecitedfromRiemsdijkandWilliams(1981):

(9)a [CPWhoitiknows[CP[whichpictureofwhom]j Billboughttj]] b.??[CP [Whichpictureofwhom]jdoesJohnknow [CPwhoitiboughttj]]

In (9a), thematrixWh-subject [who] isattracted to thematrix interrogativeCPSpec, and takesmatrix scope;the embeddedWh-object [which] is attracted to theembedded questionCP Spec, and takes embeddedscope, according to (8). In (9a), however, theWh-element[whom]isnotattractedtoanyCPSpec.Hence,[whom]may takescope freely. In (9a), theWh-phrase[whom] is contained by both thematrix and theembeddedinterrogativeCPs,andthus,[whom]maytakeeithermatrix or embedded scope in accordancewithgeneralization(8). (9b)isworsethan(9a),because(9b)violatesaWh-island constraint (Ross 1967, etc.), but the interpretivepropertyof (9b) seems to be clear. In (9b), theWh-element [which] is attracted to thematrix questionCPSpec,andtakesmatrixscope.[Who]isattractedtotheembedded interrogativeCPSpec, and takes embeddedscope,accordingto(8).TheWh-phrase[whom]in(9b),ontheotherhand,isnotattractedtoanyCPSpec,andiscontainedonlyby thematrix interrogativeCP. Hence,[whom] in (9b) isallowed to takeonlymatrixscope inaccordancewithgeneralization(8).Torepeat,thescopeinterpretation of [whom] in (9b) differs from that of[whom] in(9a),because theWh-phrase[whom] in(9b)isinsidethematrixCP,butnotinsidetheembeddedCP.Ontheotherhand,[whom]in(9a)iscontainedbyboththematrixandembeddedCPs. Significantly,Saito(1989,2003)discoversthattheJapanese examples in (10b) and (11b), bothofwhichinvolve scrambling, appear to pose a problem forgeneralization(8).Considernowexample(10b).

(10)a. [TPTaroo-ga[CP[TPHanako-gadonohon -Nom-Nomwhichbook

Akita University

− 3−

-oyonda]ka]siritagatteiru](koto) -AccreadQwant-to-knowfact ‘(thefactthat)Taroowantstoknow[QHanako readwhichbook]’

b. ?[TP[Donohon-o]i[TPTaroo-ga[CP[TP whichbook-Acc-Nom Hanako-gatiyonda]ka]siritagatteiru]](koto) -NomreadQwant-to-knowfact ‘(the fact that) [whichbook]i,Taroowants to know[QHanakoreadti]’

In (10b), the embeddeddirect object [dono hon-o] isscrambled to the sentence-initial position, and is notcon t a i ned by any i n t e r r oga t i v e CP. Hence ,generalization(8)appears topredict that thescrambledobject [dono hon-o] in (10b) cannot take scope, andthus, example (10b) is unacceptable. However, thescrambledphrase [dono hon-o] takes embedded scope,andexample(10b)soundsalmostperfect. Consider next the example in (11b),which isstructurallyquite similar toEnglish example (9b) inrelevantrespects.

(11)a. [TPTaroo-ga[CP[TPminna-ga[CPHanako-ga -Nomall-Nom-Nom donohon-oyondato]omotteiruka] whichbook-AccreadthatthinkQ siritagatteiru](koto) want-to-know(fact) ‘(the fact that) Taroo wants to know [Q everyonethinksthatMaryreadwhichbook]]’

b.??[TP[CPHanako-gadonohon-oyonda -Nomwhichbook-Accread to]i[TPTaroo-ga[CP[TPminna-gatiomotteiru] that-Nomall-Nomthink ka]siritagatteiru]](koto) Qwant-to-know(fact) ‘[ThatHanakoreadwhichbook]i,Taroowants toknow[Qeveryonethinksti]’

In(11b),themostdeeplyembeddedCPisscrambledtothe sentence-initial position, and theWh-phrase [dono hon-o]iscontainedinsidethescrambledCP.Becauseofthescramblingoperation,theWh-phrase[dono hon-o]islocatedoutside the interrogativeCP in (11b). Hence,Saito argues that under generalization (8), it shouldbepredictedthat thereisnowayfor theWh-phrase[dono hon-o] to take scope in (11b), but contrary to thisprediction,[dono hon-o]takesembeddedscopein(11b).Recall that in (9b), theobject of the embedded clause,[which picture of whom], is similarlymoved to thesentence-initial position, and theWh-phrase [whom] iscontainedby thepreposedobject. [Whom] is thusnotcontained inside the embedded interrogativeCP, andgeneralization(8)correctlypredicts that theWh-phrase

[whom]maynottakeembeddedscopein(9b). Thus, a question arises as towhy theWh-phrase[dono hon-o] can takeembeddedscope in (11b),while[whom] cannot in (9b). Toanswer thisquestion,Saitoproposes thatWh-movement inEnglish establishes anoperator-variable relation, but scrambling in Japanesedoesn’t. Consequently,ascrambledphraseinJapanesemaybeliterally‘undone’inLF,whereasaWh-phraseinEnglish cannot be. Examine the following LFconfigurations,which illustrateSaito’s proposal. LFstructures (12a-b) are forEnglish example (9b);LFrepresentations(13a-b)areforJapaneseexample(11b).In(12a),[which picture ofwhom]isanoperatorandtjavariable.

(12)a. [CP[Whichpictureofwhom]jdoesJohnknow<operator>------------------------------- [CPwhoitiboughttj]](LF) ------------------<variable>

b. [CP_____doesJohnknow[CPwhoitibought|_______________________________ radicalreconstruction=>* [whichpictureofwhom]]](LF) __________↑

UnderSaito’sanalysis,becauseoperatormovementlikeWh-movementmustcreateanoperator-variablechainasin(12a),theoperator[which picture of whom]cannotbe‘li terally’ undone, i .e. cannot undergo ‘radicalreconstruction,’ as shown in (12b). As a consequence,as in (12a), theWh-phrase [whom] is necessarilycontainedbythematrixCP,butcannotbecontainedbytheembeddedCPinLF.[Whom]thustakesonlymatrixscopein(9b). OnSaito’s account, on theother hand, Japanesescrambling does not establish an operator-variablerelation as in (13a). In otherwords, in (13a), thescrambledCP,[Hanako-gadon hon-o yonda to], isnotanoperator,anditstraceisnotavariable.

(13)a. [TP[CPHanako-gadono hon-oyondato]i[TP Taroo-ga[CP[TPminna-gatiomotteiru]ka] siritagatteiru]](LF)

b. __[TPTaroo-ga[CP[TPminna-ga[CPHanako- |____________________________________‘radicalreconstruction’=>ok gadono hon-oyondato]omotteiruka] ________↑ siritagatteiru](LF)

Hence, as illustrated in (13b), theCP [Hanako-ga don hon-o yonda to], frontedbymeansof scrambling, canundergo lowering into its original/trace position.Consequently, theWh-phrase [dono hon-o] inside themostdeeplyembeddedCPcanendupbeingcontained

ADynamicSyntacticAnalysisofScramblingandWHMovement

Akita University

− 4−

withintheinterrogativeCPinLF.Hence,[dono hon-o]is allowed to take embedded scope in (11b) inaccordancewithgeneralization(8),asdesired. Saito (1989, 2003) accounts for themarginalgrammaticality of (10b) in the sameway as follows:The structures in (14a-b) areSaito’sLF configurationsfor(10b).

(14)a. [TP[Dono hon-oi][TPTaroo-ga[CP[TP Hanako-gatiyonda]ka]siritagatteiru](LF)

b. [TP___[TPTaroo-ga[CP[TPHanako-ga |___________________________‘radicalreconstruction’=>ok [dono hon-oi]yonda]ka]siritagatteiru](LF) _____↑

Thescramblingoperationbytheobject[dono hon-o]in(14a)doesnot createanoperator-variable relation, andthus, canbe literally undone in theLF component, asillustratedin(14b).Asaresult,in(14b),[dono hon-o]is containedby the embedded interrogativeCP inLF,and is allowed to take embedded scope in accordancewith(8). Finally, to explain whymovements such astopicalizationorWh-movement inEnglishestablishanoperator-variablerelation,whilescramblinginJapanesedoes not, Saito (2003) proposes theDerivationalConfigurationalityParameter in (1a), repeatedhere as(15a).

(15) TheDerivationalConfigurationalityParameter: Configurational languages are subject to (15a- b), but Japanese-style non-configurational languagesarenot.

(a) Merge applies only to satisfy selectional requirements.(Mergeimpliesselection.) (b) Selectional requirementsmust be satisfied by Merge.(SelectionimpliesMerge.)

AndSaito(2003)illustratesthedifferencebetweenWh-movementinEnglishandscramblinginJapaneseasin(16a)and(16b),respectively:

(16)a.Wh-movement b.Scrambling

CP TP/\/\XPiC’XPiTP[wh-operator]/\/\CTP/\[+wh]/\…..ti…../\…..ti…..[variable]

InthecaseofWh-movementinEnglish,asin(16a),aninterrogativeCPhas aCheadwith [+wh] feature, andthehead requires aWh-operator in itsSpec. Hence, aWh-phrasemustmoveinto theSpecofCPobligatorilyin order to satisfy this selectional requirement,establishinganoperator-variablerelation.Ontheotherhand,inthecaseofscramblinginJapanese,asin(16b),thereisnoheadwith[+operator]feature,whichrequiresanoperatorinitsSpecposition.Hence,scramblinghasnothingtodowithanyselectionalrequirement,andthescrambledphrase is simplymerged at the (TP) rootfreely andoptionally. Moreover,Saito (2003) arguesthatthisfundamentaldifferencemanifestsitselfbecauseconfigurationallanguagessuchasEnglisharesubjecttothe first part of theDerivationalConfigurationalityParameter in (15), whereas Japanese-style non-configurational languages are not. That is, there is adeep-seated difference betweenEnglish and Japanesewith respect to the nature of core grammar:Mergeapplies only to satisfy selectional requirements inconfigurationallanguagessuchasEnglish,whileMergeapplies freely, independently of any selectionalrequirements in Japanese style non-configurationallanguages.Inotherwords,dueto(15a),configurationallanguageslackscrambling,whereasJapanese-stylenon-configurationallanguagespossessit

3. Theoretical Implications and an Alternative Dynamic Syntactic Analysis

Saito’s (2003)DerivationalConfigurationalityParameter in (15a) accounts for the contrast between(9b) and (10b/11b) in a principledmanner, as desired.Notice,however,thattheparameterin(15a)impliesthatt h e r e i s a t r u l y r a d i c a l d i f f e r e n c e b e tweenconfigurational languages and Japanese type non-configurational languages in core grammar. Namely,Merge,which is a fundamental grammatical operation,necessarily implies selection in configurationallanguages,while the sameoperation,Merge, doesnot(have to) imply selection in Japanese-style non-configurational languages. Furthermore,because (15a)is a parameter, itmust bepart ofUniversalGrammar,i.e. our innate knowledgeof language (cf.Chomsky1981,1986,amongothers).Hence,(15a)hastobepartofgeneticinformationwhichallhumansarebornwith. Thereis,ofcourse,apossibilitythatSaito’s(2003)DerivationalConfigurationalityParameter in (15) iscorrect.Aquestion,however,couldariseastoifhumanlanguagescoulddifferinsuchadrasticwayinthecorepart of grammar, if our gene could indeed containspecificinformationwhichcouldreferonlytoparticularlanguages such as ‘Japanese-style non-configurational’languages,etc.Inaddition,recently,Newmeyer(2005)and others question even the very existence ofparametersingeneral,andBorseleyandBörjars(2011)among others suppor t Newmeyer ’s c la im (cf .Bowerman1988,Culicover 1999,Kirby1999,Kayne

Akita University

− 5−

2000,Boeckx2006,Boeckx2011,Clark andLappin2011,etc.).Giventhisdebate,Ibelievethatitcouldbeat leastworthwhileattempting toconsiderapossibilitythatEnglish and Japanese donot differwithin coregrammar, and that there is not such a parametricdifferencebetweenEnglishandJapaneseinthecorepartof grammar (Hawkins 1988, amongothers). To attainthis aim, Iwish to suggest tentatively the followingalternative analysis of (9b) and (10b/11b). Under thealternativeanalysis, I adopt adynamicviewof syntax,followingPhillips(1996),Kempsonetal(2001),amongothers. Given this consideration, observe againSaito’scontrast in (9b) and (11b), repeatedhere as (17) and(18).

(17)??[CP[Whichpictureofwhom]jdoesJohnknow [CPwhoitiboughttj]]

(18)??[TP[CPHanako-gadono hon -oyonda -Nomwhichbook-Accread to]i[TPTaroo-ga[CP[TPminna-gatiomotteiru]ka] that-Nomall-NomthinkQ siritagatteiru]](koto) want-to-know(fact) ‘[ThatHanakoreadwhichbook]i,Taroowants toknow[Qeveryonethinksti]’

Recall that inEnglish example (17), theWh-phrase[whom] cannot take embedded scope,whereas inJapanese example (18), [dono hon-o] takes embeddedscope.Basedonthiscontrast,Saito(2003)proposesthefirst part ofDerivationalConfigurationalityParameter,i.e.(15a). To suggest another account for the contrastbetween(17)and(18)withoutappealingtoaparameter,Iwish to hypothesize that there is basically no radicaldifference between English and Japanese in coregrammar, andwould like to suggest the followingalternativeapproach:

(19)A Dynamic Syntactic Alternative a. BothWh-movementinEnglishandlongdistance scrambling in Japanese establish anoperator- variable relation. (BothWh-movement and scramblingaresemanticallysignificant.) b. [+Wh-Qoperator]featurerequiresanWh-phrase in its Spec position,whereas [+S operator] feature requires a scrambledphrase in itsSpec position. c. [+Q] featuredetermines the scopepropertyof a Wh-phrase.(Saito2003,etc.) d. Grammar,i.e.parser,checksasmanyfeaturesas possibleintheir‘surface’positionsinthecourse of left-to-right incremental processing (cf. Phillips1996,Kempsonetal.2001,etc.).

Under this tentativeDynamicSyntactic analysis,thescopepropertyof theWh-phrase[whom] in(17) isfixed at the initial point of the incremental processing.ThisisbecauseEnglishisahead-initiallanguageandaC [+Wh-Qoperator] head appears immediately after aWh-phrase,asillustratedbelow:

[+Wh-Qoperator](20) [CP[DPwhich picture of whom][Cdoes] <operator/filler>---------------- ……..…….?e………..…. ------------<trace/gap>

Observe that in (20), grammarhas finishedparsing astringofwordsfrom[which picture of whom]to[does],andhas projected its structure. As illustrated in (20),because both [which] and [whom] are in the Specpositionof[C+Wh-Qoperator]featureonthe‘surface,’bothofthoseWh-phrasescantakematrixscopein(20),and thus are required to takematrix scope in their‘surface’positionsat thisstage,cruciallyinaccordancewith (19d). To put it differently, condition (19d)successfully blocks [which] and [whom] from takingembedded scope in the ‘non-surface’ (trace/gap)posit ion later in the course of the left-to-rightincremental processing (cf.Phillips 1996,Kempson etal.2001,etc.). Ontheotherhand,todeterminethescopepropertyof[dono hon-o]in(18),grammarhastokeepparsingastringofwords from [Hanako] to [siritagatteiru] forexample (18), andmust project the following typeofrepresentation (cf.Phillips1996,Kempson et al. 2001,etc.):

(21) [CP[CPHanako-gadono-hon-oyondato]i<operator/filler>------------------------ [TPTaroo-ga[CP[TPminna-gaeiomtteiru] ---------------------------<variable/gap> [Cka]]siritagatteiru][Ce]] [+Q][+Soperator]

This is so, because Japanese is a head-final languageandaCheadappearsafterawholeTPdomain.Hence,after having constructed structure (21), grammar, i.e.parser,recognizes[CPHanako-gadono-hon-oyonda to]iisascrambledoperator/filler in theSpecpositionof [C+Soperator],andeiisitsvariable/gap.Unlike[whom]in(20),theWh-phrase[dono hon-o]in(21)cannothaveits scopepropertydetermined in the ‘surface’ position.Thisisbecause[dono hon-o]isinsidetheSpecpositionof[C+Soperator],notinsidetheSpecof[C+Q],onthe‘surface.’Hence,parserhastofixthescopepropertyof[dono hon-o] in the‘non-surface’variable/gappositionof [CPHanako-ga dono-hon-o yonda to]i in (21).Notice that in (21), the embeddedChead [Cka] has[+Q] featurewhich c-commands the variable/gap.Consequently, theWh-phrase [dono hon-o] is allowed

ADynamicSyntacticAnalysisofScramblingandWHMovement

Akita University

− 6−

to take embedded scope in (21), as desired. It seemsthatitisnotsoeasyforsomenativespeakerstoacceptexamples such as (18), and I speculate that thisdifficultymight arise, because theWh-phrase [dono hon-o] cannot take scope in the ‘surface’ position in(18/21),inaccordancewithparsingcondition(19d). Last, Iwish to show that the suggestedDynamicSyntacticanalysisalsoaccountsforSaito’s(1989,2003)examplein(10b),repeatedhereas(22).

(22) ?[TP[Dono hon-o]i[TPTaroo-ga[CP[TP whichbook-Acc-Nom Hanako-gatiyonda]ka]siritagatteiru]](koto) -NomreadQwant-to-knowfact ‘(the fact that) [whichbook]i,Taroowants to know[QHanakoreadti]’

In example (22), theWh-phrase [dono hon-o] isscrambledtothesentence-initialposition.Hence,[dono hon-o] is not containedby the embedded interrogativeCP,but theWh-phrase takesembeddedscope, asSaitoobserves. UnderthetentativeDynamicSyntacticanalysis,tofixthescopepropertyof[dono hon-o]in(22),grammarhas to parse a stringofwords from [dono hon-o] to[siritagatteiru] as in (21), because Japanese is a head-finallanguageandaCheadshowsuplateintheparsingprocess.Thestructuregrammarassignsto(22)isgivenbelow:

(23) [CP[Dono hon-o]i[TPTaroo-ga[CP[TP <operator/filler>------------------------- Hanako-gaeiyonda][Cka]]siritagatteiru] ---------<trace/gap>[+Q] [Ce]](koto) [+Soperator]

In(23),[dono hon-o]iisascrambledoperator,i.e.filler,intheSpecof[C+Soperator]feature,andeiisitstrace/gap. On the ‘surface,’ the scrambled phrase [dono hon-o] cannot have its scopepossibility determined,because [dono hon-o] is not in the Spec of [C +Qoperator], but in theSpecpositionof [C+Soperator].Consequently,thescrambledWh-phrase[dono hon-o]isforcedtotakescopeinthe‘non-surface’trace,i.e.gap,positionwithin the embeddedTP (cf. constraint 19d).NoticethattheCheadoftheembeddedclausehas[+Q]featurewhich c-commands everything inside theembedded clause. Again, some native speakers ofJapaneseseemtofinditalittlehardtoacceptexamplessuch as (22), and I feel that thismight bebecause theWh-phrase [dono hon-o] in (22/23) cannot take itsscope in the ‘surface’ position in accordancewithgrammatical, i.e. parsing, constraint (19d). If thisspeculation is indeed correct, it implies that longdistancescrambling isnot totally semanticallyvacuous(contraSaito1989,2003,etc.),buta scrambledphrase

occupies a semantically significant position on the‘surface,’accordingto(19a).

4. Conclusion In this paper, I have shown howSaito (2003)explains thenatureofWh-movement inEnglish andscrambling in Japanesebymeansof hisDerivationalConfigurationalityParameter. Theproposedmacroparameter,Ibelieve,isanimportantone,becauseifitiscorrect, it implies that there is a radical differencebetween configurational languages and Japanese-stylenon-configurational languages in thevery corepart ofgrammar. However,giventherecentdebatewhereNewmeyer(2005),Borseley andBörjars (2011), amongothers,seriously question the existenceof (macro) parameters(cf.Chomsky1981, 1986, etc.), I have attempted tosuggestanalternativeapproachtoaccountforthenatureofWh-movement and scrambling from aDynamicSyntactic perspective without appealing to anyparameter. If the suggested Dynamic Syntacticapproach is indeedplausible, it couldbe the case thatthere is no suchdrastic parametric differencebetweenEnglish and Japanese in core grammar, and that it isworth considering if the suggestedDynamicSyntacticapproachcouldindeedbesuperiortoSaito’sparametricapproach. I believe that preciselybecauseof the theoreticalsignificancethatSaito’sDerivationalConfigurationalityParameter provides for us, it should be verymuchworthwhile continuing to consider other possiblewaysto account forEnglishWh-movement and Japanesescrambling fromavarietyof formal and/or functionalperspectivesuntilwefullyunderstandthenatureofsuchgrammaticaloperations(cf.Hawkins1988,1994,2004,2014,Phillips 1996,Newmeyer 1998, 2005,Culicover1999,Kierby1999,Kayne2000,Kempsonetal.2001,Boeckx2006, 2011,Borseley andBörjars 2011,ClarkandLappin2011,amongothers).

ReferencesBaker,M. (1996)The Polysynthesis Parameter,Oxford

UniversityPress.Boeckx,C.(2006)Linguistic Minimalism: Origins, Concepts,

Methods, and Aims,OxfordUniversityPress.Boeckx,C.(2011)‘ApproachingParametersfromBelow,’ in

Sciulo,A.-M.D.andC.Boeckx(eds.),Biolinguistic Approaches to Language Evolution and Variation,OxfordUniversityPress.

Borseley,R. andK.Börjars (2011)Non-transformational Syntax: Formal and Explicit Models of Grammar,Wiley-Blackwell.

Bowerman,M.(1988)‘The‘NoNegativeEvidence’Problem:HowDoChildrenAvoidConstructing anOvertlyGeneralGrammar?’ inJ.Hawkins(ed.),Explaining Language Universals,pp.73-104.

Chomsky,N. (1981)Lectures on Government and Binding,Dordrecht:Foris.

Akita University

− 7−

Chomsky,N. (1986)Knowledge of Language: Its Nature, Origin and Use,NewYork:Praeger.

Clark,A. andS.Lappin (2011)Linguistic Nativism and the Poverty of Stimulus,Wiley-Blackwell.

Culicover,P. (1999)Syntactic Nuts: Hard Cases, Syntactic Theory, and Language Acquisi t ion , OxfordUniversityPress.

Fukui,N. (1986)A Theory of Category Projection and its Applications,PhDdissertation,MIT.

Hale,K.(1982)‘PreliminaryRemarksonConfigurationality,’inNELS 12,86-96.

Harada,K. (1972) ‘ConstraintsonWh-Q Binding, in Studies in Descriptive and Applied Linguistics 5,DivisionofLanguages, InternationalChristianUniversity, 180-206.

Hawkins, J. (1988)Explaining Language Universals,Blackwell.

Hawkins, J. (1994)A Performance Theory of Order and Constituency,CambridgeUniversityPress.

Hawkins, J. (2004)Efficiency and Complexity in Grammars,OxfordUniversityPress.

Hawkins,J. (2014)Cross-linguistic Variation and Efficiency,OxfordUniversityPress.

J e l i nek , E . ( 1984 ) ‘Emp ty Ca t ego r i e s , Case andConfigurat ional i ty,’Natural Language and Linguistic Theory 2,39-76.

Kayne, R. (2000)Parameters and Universals, OxfordUniversityPress.

Kempson,R.,W.Meyer-Viol,andD.Gabbay(2001)Dynamic Syntax: The Flow of Language Understanding,Blackwell.

Kirby,S. (1999)Function, Selection, and Innateness: The Emergence of Language Universals , OxfordUniversityPress.

Kuroda, S.-Y. (1988) ‘WhetherWeAgree or not:AComparative Syntax of English and Japanese,’Linguisticae Investigationes 12,1-47.

Miyagawda, S. (1997) ‘AgainstOptional Scrambling,’Linguistic Inquiry 28,1-25.

Newmeyer, F. (1998)Language Form and Language Function,MITPress.

Newmeyer, F. (2005)Possible and Probable Languages: A Generative Perspective on Linguistic Typology,OxfordUniversityPress.

Phillips,C. (1996)Order and Structure, PhDdissertation,MIT.

Riemsdijk,H.vanandE.Williams(1981)‘NP-structure,’The Linguistic Review1,171-217.

Ross, J. (1967)Constraints on Variables in Syntax, PhDdissertation,MIT.

Saito,M. (1985)Some Asymmetries in Japanese and their Theoretical Implications,PhDdissertation,MIT.

Saito,M. (1989) ‘Scrambling asSemanticallyVacuousA’-Movement,’ inM.Baltin andA.Kroch (eds.),Alternative Conceptions of Phrase Structure, pp.182-200.

Saito,M.(2003)‘OntheRoleofSelectionintheApplicationofMerge,’NELS 33,323-345.

Whitman, J. (1987) ‘ConfiturationalityParameters,’ in Imai,T. and M. Sai to (eds . ) , Issues in Japanese Linguistics,Dordrecht:Foris,pp.351-374.

ADynamicSyntacticAnalysisofScramblingandWHMovement

Akita University