Traditionally, light verbs have been characterized as semantically defective predicates with incomplete or even empty argument structures. The assumption is that these verbs must combine with a noun predicate (also known as a Verbal Noun) to license the arguments of the clause.
This dissertation provides an RG account of Light Verb Constructions (LVCs) in Spanish, Italian, French and Catalan.
My analysis goes beyond the traditional semantic criteria and identifies in configurational terms a 'natural class' of LVCs. The class comprises six different configurations associated with Romance LVCs: 1-Control, 'plain' serialization, serialization with 1-3 causative revaluation, Inversion, 3-Control, and auxiliation. My configurational approach also captures the distinction between LVCs and other constructions, including, heavy, causative and idiomatic structures.
My account pays special attention to the lexical properties of the light verb and the noun predicate that combines with it. To illuminate the role played by the noun predicate, I introduce a typology of Romance nominals, distinguishing between Verbal Nouns (VNs), which designate actions or states and license an optional subject, and non-VNs, which name objects or entities in the world and are uniformly unaccusative. Building upon Grimshaw (1990), I claim that VNs can be further subdivided into those that take optional arguments and those that license an obligatory theme.
To explain why in non-auxiliated LVCs the VN is both a predicate and the P-initial 2 of the light verb, following Dubinsky (1990), I claim that the VN bears both the P and 2 relations. Unlike Dubinsky, however, I argue that in serial LVCs in Romance the VN must be P,2 multiattached in the initial, as opposed to an intermediate, stratum. This follows from independently motivated conditions on syntactic representations, which preclude P-2 revaluation of the VN in such structures.
My proposal also explains the Pan-Romance double analyse phenomenon, whereby the prepositional complements of a VN can be typically analyzed either as being inside its maximal projection or as direct syntactic dependents of the light verb. Specifically, my analysis explains why this structural ambiguity occurs in LVCs and what lexical and syntactic conditions necessarily exclude it.
The analysis provides additional support for the emerging consensus that light verbs fall in a continuum of semantic defectiveness, Moreover, my proposal supports the view that, contrary to what is often assumed, (i) a homophonous light verb (e.g. Pan-Romance FARE 'do') may have different syntactic valences, (ii) Romance VNs do not uniformly share the same subcategorization frame as the morphologically related verbs, and (iii) syntactic licensing is not necessarily concomitant with semantic role assignment.
|
E-mail this message to a friend
