Convergence and Divergence between European Portuguese and Brazilian Portuguese

Research project funded by Fundação para a Ciência e a Tecnologia, 2016-2018
Coordinator: Augusto Soares da Silva

Within the context of pluricentricity research (Clyne 1992, Soares da Silva 2014), the specificity of the proposal resides in two points. First, we focus on the interplay between conceptual and social aspects of pluricentric variation. Therefore, this project subscribes to the framework of Cognitive Sociolinguistics (Kristiansen&Dirven 2008, Geeraerts et al 2010), an emerging extension of Cognitive Linguistics as a meaning-, usage-based approach. Second, we use sociolectometrical methods that allow linguistic distances to be measured and correlated with all types of sociolinguistic variables. Specifically, we will apply the concept-based, profile-based sociolectometry (Speelman et al 2003), where “profile” stands for the relative frequencies of a set of words or constructions in a conceptual category.

The research is concerned with onomasiological variation between semantically equivalent words/constructions (denotational synonyms). The onomasiological method has been adopted to study language-internal variation, since denotational synonyms often display sociolinguistic differences and therefore the competition between language varieties. In addition, looking at alternative expressions of concepts or functions provide us with a reliable control mechanism to avoid thematic and statistical bias. Uniformity, featural and attitudinal measures based on onomasiological profiles quantify convergence and divergence between EP and BP. The data will be extracted from a large corpus of EP-BP texts from the 1950s, 1970s and 2000s, Usenet and spoken usage.

Several concepts from several lexical fields and a multitude of morphological and syntactic variables will be analyzed. For the selection of lexical variables, we will apply advanced computational techniques that are based on Semantic Vector Space (SVS) models (Turney & Pantel 2010). These models quantify a functional similarity between pairs of words on the basis of the lexical context, found in large corpora. Applying these fully automated models on a large corpus of EP-BP texts will provide us with a large and unbiased sample of lexical variation on which our study can be based. If necessary, we will also analyze concepts manually selected from politics, health and transport. As for constructional variables we shall proceed in two ways. First, we analyse morphological and syntactic variables studied in the literature as EP-BP variation such as alternate patterns of verbal and nominal agreement, overt/pro subject alternation, impersonal/passive constructions, relative constructions, word order variations. In addition, we will include constructional variables not directly related to EP-BP variation, such as alternate prepositional constructions and patterns of diathesis alternations. Second, we will employ the SVS approach to generate potential syntactic variables. Attitudinal variables include elicitations of attitudinal intentions with regard to words and constructions. Multivariate techniques allow us to compare the impact of lexical and constructional variables, corpus-based and attitudinal variables on national convergence/divergence.

This project is important, linguistically and non-linguistically. Linguistically, it allows for the determination of the evolutionary relation between EP and BP and the pluricentric nature of Portuguese. Crucially, the sociolectometrical approach brings a new perspective: instead of looking at the distribution of individual variables, we now look at an aggregate level, which makes it possible to quantify convergence/divergence and stratificational issues. Non-linguistically, it is relevant for official language policies, normative positions and educational practices that may acknowledge and foster the pluricentricity of Portuguese.