A natureza formal da voz passiva analítica e sintética: em busca de uma caracterização translinguística

Abstract

This study presents a formal analysis of the passive voice under a crosslinguistic perspective, developed under the framework of Distributed Morphology (Halle; Marantz, 1993; Marantz, 1997). Specifically, we investigate the verbal or eventive passive (Kratzer, 2000; Embick, 2004a), which has as its essential properties the suppression of the external argument of the verb and the promotion of the internal one to syntactic subject. It realizes in two different ways: the analytical passive, which is formed with the participle of the main verb and an auxiliary verb; and the synthetic passive, composed of a single finite verb which includes a specific morphological marking. In literature, most of the analyses of passives aim to derive its form, attributing the properties of such formation to some specific element, such as a functional head or morphosyntactic feature. This tends to limit their ability to explain some relevant factors, such as the nature of the participle, with its mixed verbal and adjectival properties, as well as the recurring syncretism found between the synthetic passive, anticausatives, reflexives and other contexts. Our study fills this gap, with the goal of identifying what is the nature of the passive in order to derive its properties from it. With that, we aim to contribute to the building of a more general and crosslinguistic theory on passive formations. We propose that passives are in fact an epiphenomenon, which results from a syntactic mismatch: the presence of a verb which requires an external argument and the absence of such element. Following Alexiadou, Anagnostopoulou & Schäfer (2006), we argue that verbs which denote agentive or externally caused events require the Voice head (Kratzer, 1996) to introduce an external argument. If there is no element available to occupy the position of spec. Voice, the derivation will not converge, since the syntactic needs of this head will not be met. In order for the derivation not to be lost, syntax employs one of two mechanisms. One of them is the conversion of the verbal structure into a nominal one, which is done with the help of the participle, a mixed category of the Switch type (Panagiotidis, 2015). Participles carry a non-interpretable verbal category feature, which allows them to be merged into the verb’s extended projection, and also an interpretable adjectival category feature, which is taken as its label, effectively altering the formation’s category. Participles are treated as adjectives by syntax and denote a state, an information prototypically expressed by that category (Beedham, 1987; Abraham, 2006a). As such, they do not license an external argument, which, in practice, blocks its entrance. As this structure is now nominal, an auxiliary verb is introduced in order to resume the verbal formation. The other rescue mechanism is the introduction of a generic, default argument, with no semantic content, just to fulfill the selection needs of Voice. Such element is an anaphor, which, since it depends on binding to find its denotation, does not affect the already built meaning, because in that high position, it cannot be bound. The anaphor later incorporates into the verbal complex, becoming a clitic or an affix, which produces the synthetic passive. These base syntactic configurations are the same that give rise to the syncretic contexts.