Abstract : Significant advances in terms of syntactic, structural and schematic heterogeneity have been achieved by adopting conventions and standards. The IT community is now trying to solve the problem of semantic heterogeneity (particularly in the Semantic Web field). To reach this objective, it is necessary to enable machines to understand the semantics of terms. Semantics, as opposed to syntax, defines the mental representation of concepts corresponding to the symbols used in texts or images. When a person reads a text, he uses a semantization process which enables him to associate an interpretation to each sign identified. This operation uses a number of underlying processes such as measuring semantic distance between the meanings of several terms. Reasoning about the semantic proximity of terms is trivial for a human. However, this task is very complex for machines, and requires access to a large number of definitions of specific field terms.