This paper proposes a method to explore a collection of texts with an ontology depending on a particular point of view. In the first part, the paper points out the characteristics of the corpus, composed of 17th century French texts. In the second part, it explains the methodology to isolate the discriminant terms for the ontology creation. Furthermore, not only the projection of the ontology on the texts is pointed out, but also how to explore the corpus thanks to the defined perspective based on semantic fields.