Vous êtes sur la page 1sur 2

Defective Clauses :

>All canonical clauses (normal) clauses are CPs . >Unlike CPs , whose accusative subject is contained with its accusative case assigner C : For within the same clause with one TP , Defective clause accusative subject is contained in another clause apart from its case assigner verb . CP Defective C [ C for [TP [ PRN them [T [ T to [VP [ V see] [DP [D a ] [N specialist] [ Null C[TP 1 [PRN They [T [T Tns [VP [ V believe [ TP 2 [PRN him[ T[ T to[V be]

>In CPs : case assigners and accusative subjects are within the same clause. And the accusative case assigner to the specifier of the complement infinitive clause is the non-finite complementiser (for). >In defective clauses : they are in different clauses. the accusative case assigner to the specifier of the complement infinitive clause is (the ECM verb) . No CP layer . >They are exceptional case-marking clauses = ECM >Verbs like believe use in ECM clauses are known as ECM verbs . >ECM clauses lack CPs layer . Chomsky calls them (Defective Clauses) . >ECM clauses cannot be coordinated with canonical ones . >ECM clauses cannot be pseudo-clefted : *What we didnt intend was [ you to get hurt] (This is my addition . My observation is that CPs are headed by case assigner and its assigned case , or non of the two : (the case of null pro and null c) , but in ECM clauses we have the case assigned as a PRN and the assigner as a Verb in different TPs) . >Therefore ECM clauses are TPs not CPs , since only CPs can be pseudo-clefted (focused). >CPs are not ECM because CPs subject cannot be passivised but ECMs can : They believed [him to be innocent] = He is believed to be innocent . We didnt intend [ for you to get hurt] = * you werent intended for to get hurt . >There are verbs that select : 1)Bare infinitive complement (ECM verb) : believe 2)For infinitive complement : want (for deletion verb) / arrange

3)both of them : intend for / intend

>An infinitive TP can have its to nullified if its complement is the complement of a perfect participle verb : 1)I have never known him (to) be rude to anyone . = I have never known him __ be rude to anyone. >It cannot be nullified , if TP is the complement of a passive participle : 2)He has never been known (to be rude to any one). >>>[ My addition to this conclusion would be that to can be nullified if Spec-TP is moved to the spec position of the matrix clause] But it cannot be nullified when the Spect-TP is a null variant like in (2) .]

ECM predicates :
They select a TP complement headed by an infinitival T This T has an overt spellout in passive structures like in (2), and null spellout in active structures like in (1). There is an exception for the verb let which selects infinitival TP complement in active use only : a)You shouldnt let [him upset you] b)*He shouldnt be let [to upset you] = its not commonly said in English c)The prisoners were let [out of prison]

>The passive participle of let selects a PP rather than a TP in (b) . >We cannot explain this idiosyncrasy by saying that let is a defective verb / ECM verb without a passive participle form , because according to example (b) is has one . So lets suppose that let selects a CP complement with a null C and this null C selects a null T in infinitival TP = This explains why we cant passivize him , because only ECM clauses can have their Spec-TP passivized , whereas CPs cannot ( Impenetranability principle) = a constituent in the domain which is c-commanded by a head complementiser is impenetrable to a higher head in the tree, which is c-commanding the head complementiser. Let is therefore a C-deletion verb , like want which is a for-deletion verb So , (a) is all the way an ECM , as its verb selects a bare infinitive complement , and this latter has a specifier in the accusative form by the ECM verb , however, this ECM clause starts having a CP feature (impenetration) as its specifier cannot be passivised .

Vous aimerez peut-être aussi