Morpho-syntactic grammars for the automatic retrieval of verbal roots in quechua

Abstract

Given that Quechua is an endangered language, as expressed by UNESCO, this paper aims to advance in the creation of a platform of linguistic resources, in order to contribute to its survival and development. In this article, we present part of the algorithm that uses the inflectional and derivational structures of Quechua verbs to recognize verb roots in a text. For this purpose, we use the morpho-syntactic rules of agglutination of derivational and inflectional verb suffixes. It is shown how, with the help of the computer program called NooJ, these processes can be formalized. In particular: the paradigms governing inflectional structures as well as the corresponding grammars. The aim is to obtain new verbs and to generate inflections or conjugated forms of a single verb. Here we worked with a corpus, based on ancient writings, to automatically identify all these verb forms, and extract from them the corresponding verb roots. Two hundred and ninety-eight 'new' verbs were thus 'discovered' and added to the lexicon of simple Quechua verbs.