Análise multidimensional dos textos legais federais brasileiros

Abstract

In the past decades, there has been a growing scientific production in Linguistics in interface with Law, focused on legal language. Legal language is a highly specialized language, with lexical and grammatical características that characterize it as such (DAHLMAN, 2006; RICHARD, 2011; GOŹDŹ-ROSZKOWSKI and POTRANDOLFO, 2015) what allow us to classify it as a register according to Biber and Conrad’s perspective (2009). Moreover, it comprises a wide range of sub-registers, such as the language used in courts, in academic articles, and legislations, each with its specificities (GOŹDŹ-ROSZKOWSKI, 2012; CARAPINHA, 2018). Finally, each language, each legal tradition, and each country has its own legal particularities. Previous studies have examined this language from different angles, including terminology, translation, discourse analysis, and linguistic variation. There is also a large number of legal corpora (PORTANDOLFO, 2012; GIAMPIERI, 2018). In Brazilian Portuguese, the literature on legal language addresses specific linguistic aspects, with most studies being in the areas of terminology, translation and discourse analysis, and there are no in-depth analyses or an exclusively Brazilian legal corpus (FERRARI and CUNHA, 2022). Despite its growing popularity, there are still few studies on legal language's variation. In order to fill this gap, we adopted the register perspective to investigate the linguistic variation of Brazilian Federal Statutory Legal Texts in force. To achieve our goals, we used the Corpus Linguistics methodology, the Multidimensional Analysis (MD) approach (BIBER, 1989), and the LEX-BR-Ius corpus (FERRARI and MARQUES, in preparation). The MD is a corpus-based empirical-methodological approach applicable to any language. This methodology allows for an in-depth study of registers by describing and characterizing their variation in dimensions via statistical analysis of a corpus and functional interpretation of the results (BIBER, 1988; BIBER and CONRAD, 2009). There are two types of MD: complete and additive. The first, more detailed, allows the identification of dimensions in a corpus. The second, more global, applies the dimensions identified in a full MD to a specific corpus (BERBER SARDINHA, 2013). In the present study we performed an additive MD of the LEX-BR-Ius corpus (FERRARI and MARQUES, in compilation) adopting the study by Berber Sardinha, Kauffmann, and Acunzo (2014) as a basis aiming to map legal texts into the dimensions of variation identified by it. The LEX-BR-Ius is a written, monolingual, synchronous, part-of-speech (POS) and lemma annotated, textually tagged with Modest XML (HARDIE, 2014) corpus of complete Brazilian Federal Statutory Legal Texts in force on the compilation date (2022), balanced by frequency of use according to the search results on the JusBrasil portal. We concluded that legal texts are a register according to Biber and Conrad’s perspective (2009). They are characterized by the use of verbs in the future, modal verbs, coordinated conjunctions, subordinate clauses, passive constructions without agent, nouns, adjectives, prepositions, definite articles, nominalizations, lexical richness, and informational density. The results also show that the situational characteristics of each normative species are distinct with respect to the process followed for its creation, while each legal text is further distinguished by its communication context, communication purpose, and subject matter. This research aims to contribute to studies on legal language and linguistic variation.