Search

This paper describes the implementation of a parser for a Portuguese fragment on XLE, the state of the art system in computational grammar development within the LFG generative framework. This fragment comprises the following subareas of verbal agreement that are subject to dialectal variation: (i) syntactic agreement of the so called pronoun a gente with 3rd person singular and semantic agreement with 1st and 3rd plural verb forms; (ii) agreement between the 2nd person pronoun tu with 2nd and 3rd person singular and between the 2nd person plural pronoun vocês with 2nd and 3rd person plural verb forms; (iii) semantic agreement between a definite plural DP and the 1st person plural. In order to derive the different patterns in this domain that can be observed in Standard Portuguese and in some non-standard dialects from both Brazilian Portuguese and European Portuguese, feature structures are proposed for DPs and verbal inflections in which a distinction is made between formal-syntactic features and discourse-semantic features. It is shown that the differences observed between the investigated dialects in respect to (i) - (iii) can be derived by activating or deactivating these features. [ABSTRACT FROM AUTHOR]

This paper approaches corpus linguistics as a subfield in applied informatics which features among its main focuses automatic data extraction from corpora. For this purpose, we develop commands and scripts in the UNIX bash command language, illustrating its applicability in the investigation of the -vel suffix and of iterations of letters and words in two of the main corpuses of Portuguese. We argue that using free software tools with textual interface, whose mastering together with programming skills is a necessity in computational linguistics, is more advantageous in corpus linguistics in comparison to commercial and proprietary programs with graphical interface. [ABSTRACT FROM AUTHOR]