O sincretismo do morfema -r em latim: uma abordagem baseada em incorporação

Abstract

This study investigates the syncretism of Latin’s passive morphology. In that language, the suffix that realizes the passive voice (as a default, -r) is also found in impersonal, anticausative and middle contexts, as well as in the deponent verbs, a class defined by always surfacing in the passive form, even though it has an active sense. That kind of syncretism is not limited to Latin, being present in other languages from different families, and occasionally also involves reflexive verbs. With the goal of offering a comprehensive explanation, this study builds on the previous works of Schäfer (2008) and Lazzarini-Cyrino (2015) and, under the theoretical model of Distributed Morphology (HALLE & MARANTZ, 1993; MARANTZ, 1997), it presents a proposal in which the syncretic marker is a non-referential verbal argument – a variable – incorporated into the verb post-syntactically. The different contexts in which the marker appears are dependent on the original position that argument occupied, the external or the internal one. The presence of the variable in the internal argument position causes it to be bound to the external argument, producing the reflexive interpretation. In the external argument position, since there’s no higher DP for it to be bound with, it cannot receive a θ-role or Case. As a result of not having a θ-role, it is read as an expletive in LF. The absence of Case is a problem for PF, where the Case filter is checked (LEVIN, 2015). To prevent the derivation from crashing, the variable is incorporated into the verbal domain, through Local Dislocation, in the terms of Levin (2014, 2015). That configuration, in different flavors of Voice (FOLLI & HARLEY, 2005), produces different interpretations. The passive voice and the impersonal domain, here considered to be the same phenomenon, differentiated only by the presence of a by-phrase-introduced agent, result from the incorporation of a variable in the external argument position of a VoiceDO. The same operation in a VoiceCAUSE generates anticausatives. In the middle contexts, there’s an incorporation of a variable in the external argument position, while the argument of an Appl head (PYLKKÄNEN, 2008) is promoted to syntactic subject. Deponent verbs are split into three types: following Alexiadou (2013), the agentive deponents are reinterpreted as actually middle verbs and, as such, they are derived in the same manner; deponents involving experiencer subjects project that argument via an EXP head and their external argument position is filled by the variable, which then incorporates; anticausative deponents are the result of the incorporation of the Appl argument in a Voice-less structure.