This work describes the design of a learning component of ontologies for the enrichment and refinement of a dynamic ontological framework. The component is characterized by extracting knowledge through several techniques of ontology learning to semantically strengthen the ontological framework.This framework is used as the base for the interpretation of a
consult on the semantic web in natural language. In the proposed learning component, an ontological reasoning engine
coordinates the process of acquisition knowledge morphosyntactic and semantic, linking learning sources (data are not structured or semi-structured), the knowledge discovery techniques, and structures to impact within the framework.