Building statives out of deadjectival change-of-state-verbs

MANGIALAVORI RASIA, MARÍA EUGENIA

Pamplona

Congreso; XI Congreso Internacional de Lingüística General; 2014

Universidad de Navarra - Facultad de Filosofía y Letras

This paper focuses on the fact that, apart from causative-inchoative alternation ((1)a-b), Romance languages allow a third variant ((1)c) featuring the Initiator as sole argument. (1) a. Las luces altas enceguecen a los conductores. b. Los conductores enceguecen. c. Las luces altas enceguecen. Classically, one would assume that the missing object is a null object (Rizzi 1986). However, Init-V constructions do not seem to follow most of null object?s defining characteristics (control, binding, adjunct and argument SCs). Likewise, alternatives followed in First-Phase syntax approaches (e.g. composite roles) do not succeed in accounting for these constructions. Since Romance does not provide us with any overt morphological indicator of derivational process or syntactic realization of this projection, we may wonder whether any evidence can be gleaned from the correlation between syntactic and the semantic (eventive) structure. On this account, the absence of the object in Init-V constructions matches the absence of the projection hosting it: ProcessP. Indeed, Init-V constructions show a stative-like behavior: they can only occur in present tense, they do not admit progressive tenses nor endpoint/framing adjuncts; whereas these alternatives are natural in both caustive and inchoative structures. Nevertheless, these alternative are available with internal DPs (even an arbitrary pro). (2) El chocolate {engordó/está engordando} {*(a la gente) We will take this evidence to postulate a null causative V head (InitV0) which could be freely attached to the adjectival root in Romance with no need of ProcP mediating them (unlike German, where the realization of both the Undergoer and its host projection, ProcP, is inescapable). Hence, Romance allows to derive the A root either via a ProcV0 or by a InitV0, or by both (rendering the ergative, the causative monoargumental and the transitive constructions respectively). This poses several consequences. First, it indicates that ProcP would not be encoded in the lexical specification of DVs in Romance, but compositionally built to provide the root with the necessary structure to license the DP, just as InitP. Therefore, the merge of the lexical root either with the process portion or with the causative head (or both, by re-merge) would be equally spelledout as engordar, enceguecer, etc., In the second case, both the stative properties and the monoargumental argument structure follow naturally from the absence of the seconv V head (ProcV0). Secondly, the null causative head would endow Romance with the possibility to render either eventive or stative predications out of deadjectival verbs. Evidence backing this proposal will come from different facts. Among them, its high productivity, exposing mostly semantic restrictions, and embracing even verbs lacking a transitive counterpart. Third, in view of this evidence, to say DVs of the kind of engordar, enfriar, enloquecer, enecguecer, etc are necessarily eventive as a consequence of their lexical structure/nature is most likely inaccurate. This is suggested both by the fact that they succeed to render stative predications and by the findings documented in a rich body of literature on aspect which provides copious demonstration of the fact that stativity, telicity, and the aspectual classes (activities, accomplishments, achievements), pertain not to verbs ?more specifically, to lexical Vs? but to the predicates they head, in this case, in combination with A roots (cf., Dowty, 1979, 1991; Tenny, 1987, 1992). Therefore, it would be reasonable to entertain the possibility that these notions, and COS (eventivity) in particular, are never features of individual lexical items?e.g., of verbs, nouns, adjectives, adpositions, or what have you?but rather of whole predicates. Hence, our conclusion about stativity and eventivity is that it they are not itself a feature of V but a consequence of the semantic properties of the V heads available for conflation. In other words, we will argue that stativity (i) is a property of constructions, which (ii) arises in the semantic composition of meaningful elements, and that (iii) it derives from the l-syntactic configuration whereby they are realized as lexical (V) categories. In sum, Romance deadjectival verbs are not necessarily eventive. Any eventivity which might adhere to the predicates rendered by/through them is due to the V heads available for conflation, and it correlates with the selection of an internal complement in the configurations. On the other hand, direct conflation of the root with a causative head only correlates with unergativity.