The codification of ontological and contextual information for precomputational treatment in grammars and ontologies

Abstract

One problem at the linguistic preprocessing stage has to do with the concepts included in existing linguistic models. Part of the problem of codifying ontological and contextual information focuses on the lack of differentiation between communication and cognition that some linguistic models present. Besides, there are some described linguistic concepts that are lightly marked and which lack enough empirical textual, lexical or grammatical evidence that support them. Because a unified linguist model able to account for ontological and contextual information is not yet available, a simpler mechanism capturing linguistic, ontological and contextual information can be simpler at a preprocessing stage. Instead of using whole linguistic models, it is explained here how an algorithm describing the components that make up linguistic codification can be used to facilitate precomputational codification. This algorithm is based on the structural similarity of the grammar of a language, the ontology supporting it and the proper descriptive algorithm. Finally, the use of this algorithm illustrates how to extract this information from a corpus.