La place des connaissances lexicales face aux connaissances du monde dans le processus d'interprétation des énoncés

The aim of this research is, on one hand, general, and on the other hand, specific. The general aim is to study to what extent the process of interpretation of statements is based on lexical knowledge rather than on world knowledge. We are particularly interested in knowing to what extent the knowledge needed for interpretation is part of the language. This question being too general to be properly researched, it was necessary to look at it from a more limited angle: the interpretation of definite expressions—this is our specific aim.

Besides its theoretical interest, this is also an essential issue in Natural Language Processing where computers need a great deal of knowledge. In order to provide the necessary knowledge to computers, we first need to examine what this knowledge consists of, what role it plays. However, from a computational point of view, the most important point is not to know how human interpretation works but rather to determine what piece of knowledge is needed for any linguistic task.

First of all, we make a clear distinction between world knowledge and linguistic knowledge. After that, we present the framework of this study: the Explanatory Combinatorial Lexicology. An entire chapter is concerned about the notion of interpretation. At the end, we conduct a study dedicated to definite expressions and we analyse a corpus of twelve texts.

The conclusion of our study is that linguistic knowledge (and particularly lexical functions) together with the information contained in the texts we analysed, can grant us most of the knowledge needed for the interpretation of definite expressions.