Methodology

For each language, the construction of the thesaurus was based on the use of the following three basic relationships among terms: equivalence, hierarchical and associative. These are defined according to the ISO 2788 standard (Guidelines for establishment and development of monolingual thesauri). As for the different levels of linguistic equivalence established between the different languages, they are specified by the ISO 5964 standard (Guidelines for establishment and development of multilingual thesauri).

This thesaurus differs from the ISO 5964 standard in that it does not choose a source language. Hence, no language is placed in a position of superiority or of prestige. When building the thesaurus, this implies a constant exchange between languages, so as to review concepts, their translations and their relationships.

Most of the terms in the thesaurus come from reports on cultural heritage policy in Europe; reports from Spain, France, Hungary, Ireland, Norway and UK were initially studied in detail. With these reports, each of the three teams in charge of the establishment of the thesaurus - Spain, France and UK - created a separate list of terms in their own languages. These lists were also supplied with additional terms issued from specialized documentary sources.

The three teams then compared their lists so as to obtain a pool of words with linguistic equivalences in the three languages; the idea was to bring out the different classes meant to represent the first, broadest level. Each term selected was placed into the most relevant class. Within each class, terms were ordered following the same hierarchical relationship for all linguistic versions of the thesaurus. During this stage of work, terms regarded as too specific to one language or terms which corresponded to regional idiosyncracies were not excluded from the thesaurus but treated as equivalent terms to a concept common to the three languages. Besides which, the initial vocabulary was completed by a feed-back process between languages: some of the terms present in the final vocabulary of a language were introduced to another language which had not selected it a priori.

When building the hierarchy of the thesaurus, the teams tried to avoid poly-hierarchies as much as possible, , that is to say the possibility for a broader term to appear under more than one class; this aim was achieved by giving preference to associative relationships.

For technical reasons, further work has had to be delayed. In fact, before validating hierarchies, establishing equivalence and associative relationships, and linguistic levels of equivalence, and before elaborating scope notes, the team has decided to wait for the complete development of the Bull software which will be used for the management of the multilingual thesaurus; this software will indeed allow working with the optimum efficiency. Up to now, the team has had to work only with the aid of tools such as a bespoke Access database and word-processing software.

The acceptance of the expected software will be based on its validation, its network installation, and finally on the training of the members of the team so that they know how to use it.