Sintaxis y semántica construccion causativa

LUIS PARÍS

El funcionalismo en la teoría lingüística: la Gramática del Papel y la Referencia

Akal

Lugar: Madrid; Año: 2011; p. 124 - 146

The causative construction “Hacer+Infinitive” (HI) is a construction schema materialized by the subconstructions HI-OM, HI-DAT and HI-ACC. Each of them signals a particular morphosyntactic realization of the causee, the participant causally affected in the causal chain introduced by “hacer”. They are exemplified by (1), (2) and (3), respectively. (1) El gobierno hizo remodelar el teatro (por la empresa del ministro). The government made remodel the theater (by the company of-the secretary of state) “The government had the secretary`s company remodel the theater” (2) … y cuando Peter se lo hizo notar, no lo negó, … [CREA] and when Peter him-DAT it-ACC made notice,no it-ACC denied, … “… and when Peter made him notice it, he didn´t denied it, …” (3) ... Bühler lo hizo leer los libros más importantes de ambas escuelas ... Bühler him-ACC made read the-pl books more important-pl of both schools “Bühler made him read the more important books of both schools” “Hacer” in HI-OM and HI-DAT has a quasi-morphological status; it introduces an argument and modifies the syntactic realization of the rest of the base verb arguments. The underlying syntax of these constructions is described naturally by means of Nuclear Junctures (Van Valin and LaPolla 1997; Van Valin 2003); in contrast, a Core Juncture –typical of control structures- accounts for the syntax of HI-ACC (París 1999). Furthermore, these syntactic templates capture the semantic difference driving the morphosyntactic variability of the causee as the product of alternative projections to Macroroles. The causee is linked to an optional peripheral Actor in HI-OM, expressed as a non-macrorole internal argument in HI-DAT and, finally, projected to Undergoer in HI-ACC. This constitutes an elegant way of deriving systematically through independently motivated structures the intuition of a cline on the degree of affectedness of the causee. However, there are other proposals that argue for the need of a more precise notion of affectedness (Ackerman and Moore 1999) while others even claim that the direction of the cline should be inverted being the least affected causee the one in HI-ACC (Ashard 2002). In particular, Ackerman and Moore (1999) propose a feature system in which the least affected causee is defined as non-voluntary and causally affected. (4) HI-OM = causee [+voluntary]; [-causally affected] HI-DAT = causee [+voluntary]; [+causally affected] HI-ACC = causee [-voluntary]; [+causally affected] There are several empirical data that undermine the plausibility of this system; for example, the expression of the causee in HI-OM is constrained to few context types (París 2010). It also seems unlikely that maximally salient semantic features as the one in (4) are coded in morphosyntactic features that differ only in relation to the third person pronouns. Theoretical considerations also point out to problematic and unanswered issues. How can a participant in the middle of a causal chain be not causally affected (HI-OM)? How can a causally affected participant act voluntarily (HI-DAT)? In this paper I argue that the proposal in (4) amounts to an overspecification of the affectedness of the causee. In order to capture the meaning of these subconstructions the cline of affectedness should remain rather vague, imprecise and highly sensitive to various pragmatic considerations. In particular, I claim that this underspecification stems out of the inherent vagueness of the predicate CAUSE (Shibatani and Pardeshi 2002), which makes the content associated to “hacer” in each case highly sensitive to the meaning of the base verb. In this rather pragmatic approach the feature [-voluntary] can still be understood to be coded in HI-ACC but only in a contingent manner. This means that HI-ACC does not entail a [-voluntary] cause; however, if the speaker needs to refer to a situation that includes this participant in that particular condition, HI-ACC needs to be used. In a formula, a semantic feature ς’ is contingently codified in the meaning ς of a linguistic form φ, if (5) a. ¬  (ς’ <Pς ) [it is not necessary that ς’ is part of ς] b. If speaker S needs to express ς’, it has to use φ. In summary, the form-meaning relation is such that affectedness cannot be reduced to a single abstract notion to which each different subclass contributes a feature specification; instead, it is better captured as a list of disjoined features –some of them even pragmatic in nature- that may add up and, thus, compose an ascending and gradual cline.