1Week 4a. UG and L2A:Verb movementCAS LX 400Second Language AcquisitionA wee bit of syntax• The parameter of “verb movement” is alittle more complicated, so we’ll need todive into syntax a little bit more.• In English, we have sentences like:– John will not eat lunch.– subject, modal, negation, verb, object.A wee bit of syntax• John will not eat lunch• We will take each of these words torepresent a “slot” in the structure of asentence. That is, there is a place forsubjects, for tense (will), for negation (not),for verbs, and for objects.A wee bit of syntax• *Completely Malcolm will not clean his room.• *Malcolm completely will not clean his room.• *Malcolm will completely not clean his room.• Malcolm will not completely clean his room.• *Malcolm will not clean completely his room.• Malcolm will not clean his room completely.• You may remember that adverbs in English canappear in before the verb or after the object.A wee bit of syntax• The reason for this is that the verb and objectform a unit (VP) which the adverbs must be“attached to”:• Malcolm will not [VP clean his room ].• So, these kind of adverbs can, in a sense, serveas “landmarks”. Similarly, not and tense and thesubject are assumed to be in the same structuralposition all the time.Auxiliary verbs• But some verbs (in particular, have and be, the“auxiliary verbs”) act different.• Malcolm will not^ [VP have ^ [VP cleaned his room ]] ^.– So we know that have is a real verb here…• Malcolm has not ^ [VP cleaned his room ] ^.– But if there isn’t something “filling up” the tenseslot, have shows up in the tense slot (to the left ofnot and adverbs).2Auxiliary verbs• Same goes for be:• The steak will not ^ [VP be ^ [VP eaten ]] ^.• The steak was not ^ [VP eaten ] ^.• What appears to be happening to have andbe is that they are placed in the tense slot(unless it’s otherwise filled) instead of inthe VP. Another way to look at it is that theauxiliary verb has moved to the tense slot.Auxiliary verbs• That is, we might start out with:• Malcolm [PAST] not [have [cleaned his room]]• In which case, we have this:• Malcolm have+[PAST] not [—[cleaned his room]]• That is…• Malcolm had not cleaned his room.• But if start with:• Malcolm will not [have [cleaned his room]]• We just get:• Malcolm will not have cleaned his room.Verb movement• Turns out this kind of verb movement happens in alot of languages, sometimes for all verbs…• French is a language of this sort; all (tensed) verbsmove to the tense slot.– Jean (ne) mange pas du chocolat.– Jean (n’)est pas bête.• In each case the verb is to the left of negation (pas).• So French has set the V-to-T parameter on, Englishhas set it off (except for be and have).Verb movement• Given that French verbs move to the tenseslot, and assuming that VP is thecrosslinguistically appropriate place toattach adverbs…– Jean (ne) mange pas [— du chocolat].• We’d imagine that manner adverbs shouldshow up between negation (pas) and theobject.Verb movement• And we in fact see this:– In English, you can never have an adverbbetween the verb and its object.• *John [eats often chocolate].• John often [eats chocolate].– In French, you generally put adverbs betweenthe verb and the object.• Jean mange souvent [— du chocolat].• *Jean souvent [mange du chocolat].Verb movement• This fact that the verb shows up in Frenchto the left both of negation (pas) and to theleft of adverbs illustrates a clustering ofproperties associated with this parameter.Both properties have the same cause.• So if we want to attribute the cause of oneto the V-to-T parameter, we should expectto find the other property as well.3Interlanguage and UG• When we ask about whether UG drives L2A,we are in effect asking: Are IL grammarsconstrained by UG?• That is, are people, as they learn a secondlanguage, “allowed to” positrules/constraints in the IL that do notconform to UG, that could not appear in anynatural (native) language?Why parameters seem to be agood place to look• One crucial property of the parameters (in thePrinciples and Parameters model) is that a singlesetting of the parameter can have effects in severalplaces in the grammar of a language.• Our current example is verb-movement (V to T), whichis set to “yes” in French, and is responsible for:– The relative position of negation and the finite verb– The relative position of manner adverbs and the finite verbWhy parameters seem to be agood place to look• So, we can also look for the cluster of effects thatare supposed to arise from a single parametersetting.• Is it the case that once a second language learnergets the verb-adverb order right, s/he also gets theverb-negation order right? If only one kind of verb(finite vs. nonfinite) moves to T, is it the finiteverb?Setting parameters• In general, we have to say that (full) knowledge ofthe L2 is going to involve setting the parameters tothe appropriate settings for the target language.• But apart from the word order parameter (VO vs.OV), the existing evidence that learners are settingparameters (with the clustering of effects thatshould be entailed) seems to be lacking.White (1991)• Lydia White at McGill has done a numberof studies related to this question, and hasfound a couple of disconcerting things(despite the fact that she is strongly in favorof the UG-in-L2A hypothesis). Let’s seewhat she did and what she found.White (1991)• White observes that even sticking toadverbs, there is a small “cluster” ofproperties tied to the verb raising parameter:• In French (where V moves to T):– S Adv V order is disallowed– S V Adv Obj order is allowed.• In English (where V does not move to T):– S Adv V order is allowed– S V Adv Obj order is disallowed.4White (1991)• Given this, it should be sufficient for alearner to learn the one which is allowed(e.g., in English that S Adv V order isallowed)—the V-to-T parameter can then beset (to off for English), and then theimpossibility of the one which is disallowed(e.g., *S V Adv Obj order in English)should follow automatically if they’ve setthe parameter in their IL.White (1991)• White’s study involved native speakers ofFrench learning English.• Her subjects were children in grades 5 (averageage 11) and 6 (average age 12) with very littleprior English exposure and have very littleEnglish exposure outside