Back to list

Detail of contribution

Auteur: Scott A. SCHWENTER

Variation Analysis and Pluricentric Languages: Anaphoric Direct Objects in spoken European and Brazilian Portuguese

Abstract/Résumé: Portuguese is considered one of the most typical cases of a pluricentric language (Baxter 1992), due to the broad differences between the varieties spoken in Portugal (EP) and Brazil (BP). Much has been made, e.g., of the seemingly very distinct ways in which anaphoric direct objects (ADOs) are encoded in these two varieties. Thus, from a descriptive perspective, the mere fact that direct object clitic pronouns are conserved in the spoken variety of Portugal but are exceedingly rare in Brazil has been emphasized as an important difference (Duarte 2012), while from a theoretical syntactic perspective, attention has focused on differences between null direct objects in both varieties and their implications for structural differences between the two varieties (Cyrino 2001; Kato & Raposo 2001). In this paper I employ the comparative variationist framework (Tagliamonte 2002) in a multivariate analysis of ADOs in spoken EP (CRPC corpus) and BP (PEUL-RJ corpus). I demonstrate that, despite surface coding differences between these two varieties of Portuguese, where EP continues to employ the 3rd person clitics which have been supplanted in BP by nominative tonic pronouns, the underlying grammatical constraints operating on each are extremely similar. I provide evidence that speakers respond to similar discursive pressures when they make their grammatical choices. I extracted and analyzed approximately 1000 ADO tokens from each dialect corpus, coding for a broad range of independent factor groups, including number, animacy, definiteness, and specificity. I also operationally quantified the notion of topicality via two distinct measures (cf. Givón 1983), referential distance and topic persistence. The results show a clear pattern across the two dialects, regardless of the formal realization of the ADO: greater topicality of the ADO, as operationalized in multivariate analysis, correlates with greater overt ADO expression, whether as a clitic, tonic pronoun, or lexical NP. Null objects, by contrast, are the default coding option in both EP and BP. The individual factor groups likewise show mainly similar patterns across the two dialects. I conclude that these two dialects, which are thought to be rather distinct in ADO expression, are actually very similar when examining the underlying contextual factors constraining ADO realization. More generally, this study shows the potential contribution that both variationist analytical techniques and the use of spoken corpus data can make to the study of varieties of pluricentric languages, by revealing underlying grammatical similarities that are not apparent when examining decontextualized examples in isolation.