UC Santa Cruz

This dissertation examines the morphosyntactic implications of verb-building in the Semitic language Maltese. Theoretically, the dissertation examines the role of functional heads in defining clausal morphosyntactic properties and the interaction of syntax and morphology in the domain of cliticization. The phenomena examined herein are: (i) the major clausal constituency of Maltese and the derivation of subject agreement morphology in periphrastic complex tense constructions, (ii) morphological and periphrastic causative formation and the absence of non-finite verb forms, (iii) the argument structure of ditransitives and the syntax of cliticization, and (iv) the implications of ditransitive argument structure as it pertains to analyses of the Person Case Constraint. The dissertation shows that these phenomena support a view of syntax wherein lexical properties of heads determine agreement morphology and the structure of clausal complements and a view of morphology wherein morphological constraints can influence the output of syntactic computation.

I show that a pervasive feature of Maltese is the appearance of phi−features on lexical items which would be devoid of phi−features in other languages. Data from floating quantification and word order in complex tense constructions are used to argue that verbs in periphrastic tense constructions receive agreement features via successive-cyclic movement of the subject through intermediate specifier positions. Data from available morphology and adverbial interpretation are presented which suggests that periphrastic causatives in Maltese em- bed a reduced clause wherein a polarity phrase may host agreement features, accounting for finite verb forms in semantically non-finite causative complements. Moreover, agreement and case are argued to be computed distinctly, given that causative subjects appear with accusative case yet still trigger subject agreement on the complement predicate. I show that while Agree can account for subject agreement in the usual way, the Maltese causative facts necessitate pairing Agree with a theory of case in which morphological case values are assigned disjunctively based on the number of nominals in a clause.

I also show that Maltese non-causative ditransitives bifurcate into two distinct classes: (i) a majority class which only allows prepositional dative constructions and (ii) a minority, lexically idiosyncratic class which allows a limited double object construction involving two accusatives. I show that these facts can be accounted for by positing a VP-movement in these ditransitives which interacts with structural Case licensing to derive the appearance of a restricted double accusative case frame.

Finally, I show that data from potential intervention contexts require treat- ing cliticization in Maltese as an instance of head movement of a simultaneously minimal and maximal determiner element. This view of cliticization is shown to be more appropriate than phrasal movement for the Maltese facts insofar as it correctly predicts that clitic movement should be possible despite the presence of a DP between the base generation site of the clitic and its position on the verb. This view of cliticization is also shown to have ramifications for syntactic treatments of the Person Case Constraint as dative intervention: I argue that not all datives are generated higher than accusatives, falsifying predictions made by syntactic accounts of the Person Case Constraint based on Agree. In place of these, I provide a modern morphological account which calls off cliticization in the morphology in Person Case Constraint-violating contexts.