Dar e receber um abraço: uma análise da conversão em português brasileiro

Abstract

Defined by G. Gross (1989) as a formal operation that establishes a relation of paraphrastic equivalence between two elementary constructions, Conversion has been gaining prominence in works in the area of linguistic description, specifically, for two reasons: it is one of the most productive transformational properties, in relation to the number of occurrences, that constructions with support verb and predicate noun can present and, until then, there were few studies that took it as the main object of analysis in Brazilian Portuguese, being the description prepared for the most complete French language. Currently, Conversion has descriptions for Romanian, European Portuguese and Brazilian Portuguese. In this relation, the predicate noun is maintained and the position of the arguments is changed, without causing a semantic change, that is, the semantic roles of these arguments remain the same, despite the change in syntactic position (Helena deu um abraço no Pedro/ Pedro recebeu um abraço da Helena). In constructions that are related by Conversion, the active orientation sentence and the active support verb itself are named standard, while the equivalents, of passive order, are named converses. Based on a syntactic-semantic description methodology known as Lexicon-Grammar (M. GROSS, 1975, 1981), in this new phase of the study, which follows Calcia (2016), the lexical entries underwent an update (quantitative and qualitative) and were reclassified into five major classes and subdivided according to the type of subject of the converse construction. Such classes refer to names that are constructed with the pairs FR (fazer-receber), FS (fazer-sofrer), DR (dar-receber), DL (dar-levar) and TT (ter-ter). It is considered that the description of the syntactic phenomenon of Conversion is important for the contribution to linguistic studies, especially descriptive ones, in Brazilian Portuguese. In addition, the formalization of data in binary matrices results in a resource that can be applied in Natural Language Processing (NLP) systems, such as those related to the use of paraphrases.