Sujeitos de referência arbitrária: uma classe homogênea?

Abstract

This thesis examines a case of a referential arbitrary subject variation in Brazilian Portuguese. It analyses impersonal constructions with eles (they pronoun) in subject position, describing that the use of eles in its full and reduced form es-, as subject of impersonal constructions, is related to the increase of overt pronominal subject in Brazilian Portuguese. Following typological approach (cf. CABREDO HOFHERR, 2003, 2006; SIEWIERSKA, 2007, CREISSELS, 2008; SIEWIERSKA & PAPASTATHI, 2011), several works have showed that impersonal constructions are not a homogenous class.However, in this work it is hypothesized that third person in Brazilian Portuguese is a gradual class and eles pronoun in subject position is not a simple arbitrary subject.Regarding these assumptions, it is argued that eles pronoun in impersonal constructions could be described in a typological distribution of definite and specific features, which can be translated to this typology as [-impersonal >> +impersonal]. Therefore, it is arguedthat this typology of eles pronoun in impersonal constructions provides a new interpretation to the Referential Hierarchy proposed by Cyrino, Duarte & Kato (2000), considering the insertion of the third person as a gradual class in this referential hierarchy.Furthermore, the research hereupon developed shows that eles in its reduced form es in impersonal constructions reveals a new pattern of the third person in Brazilian Portuguese:[Es + V sing]. This pattern can be seen as the highest level of impersonalization. It also shows that the use of es contributes to the increase in overt pronominal subject in Brazilian Portuguese. The theoretical basis which supported this research reliesfundamentally on the Variation Theory (WEINREICH, LABOV & HERZOG, 1968), as well on the assumptions of Sociolinguistics (LABOV, 1972). For the quantitative analyses it has been used the GOLDVARB 2001 program, version organized by Robinson, Lawrence& Tagliamonte (2001), which provides the quantitative results to a case of referential arbitrary subject variation researched in this work.