A "Context" is an elusive concept, that, it could not be defined automatically by a machine. During the last fifteen years, a strong motivation for research on context appeared in different disciplines such as natural language semantics, linguistics, cognitive psychology and artificial intelligence (AI). In this paper, we focus our attention on the ontology learning field. Till now, the concept of "Context" was not really used for the ontology building task rather for its reuse, evaluation or formalization. In this theoretical study, we carry our interest on the use of context in the ontology learning process. We answer three crucial questions treating the impact of the context in the ontology learning task, its importance in relation with the intelligent interpretation and the utility of reasoning in multiple contexts instead of only one context. These theoretical argumentations will serve to enhance our existing contextual models either for the ontology building or evaluation, in the following research.