Matching Jan - Wouter

In Chomsky’s Minimalist Program (Chomsky 1992), formal licensing operations are defined as feature matching operations. Elements are inserted in fully inflected form, carrying abstract morphological features associated with the inflection. These features have counterparts represented in functional heads. At some point in the derivation, the inflected elements raise to a position close to these functional heads, so that the features can be checked off against each other. The positions involved in the feature matching operation are the specifier position and the head position of functional projections. Thus, NPs marked with Nominative Case move to the Spec,AgrS to match the feature associated with Nominative Case with the relevant feature in AgrS (by a mechanism of specifierhead agreement or SHAG). Similarly, inflected verbs move to the head of AgrSP to match their person/number features with the corresponding features, which are also represented in AgrS. This feature matching approach to formal licensing offers an explanation for the movement phenomena that appear to be associated with inflectional morphology in a variety of languages. It is assumed that movement is triggered by only one type of requirement: the requirement that abstract morphological features be checked. It is also assumed that this checking takes place in specifier and head positions of functional projections only. Therefore, inflected elements have to move in order to get their features checked. When the verb and the NP move in overt syntax and target the same functional projection, an adjacency effect shows up (for example, between the subject and the verb in subject initial main clauses in Dutch). In general, structural Case assignment phenomena can be reformulated in this framework without great difficulty. There is, however, one class of phenomena that seems hard to reconcile with the basic assumptions outlined above. These are the phenomena involving structural Case on a predicate nominal. Predicate nominals and their subjects often agree in Case. Yet, in a language like Dutch, where the feature checking involves overt movement of NPs (see note 1), only one of the two elements, either the subject or the predicate nominal, moves to the position in which the Case-associated features are formally licensed. As demonstrated in Zwart (1992a), the other element is remarkably immobile.