German VersionFrench VersionItalian Version
For a Corpus-based Approach to Terminology
One of the basic principles underlying multilingual terminology in the legal field is that the comparison of the different systems involved should not be carried out at the linguistic level only: since legal concepts are closely related to the legal systems which originated them - which implies that they can be rarely considered to be universal - the comparison should above all be done at the cultural level. It is therefore essential to study the legal systems under examination in depth before performing any conceptual and linguistic comparison.

In accordance with the most recent trends in terminological research, we built 4 corpora of law texts on Health and Safety at Work (one for each of the legal system involved), to be used for both terminology extraction and documentation/terminography. In order to ensure the comparability of the four corpora, we decided to first include law texts issued as a result of the implementation of EU directives; the corpora were further enriched with texts of pure national origin, since despite the standardising effect of EU law, all the legal systems under examination still preserve many of their own peculiarities.

For an analogous reason we decided not to exploit EU legal texts as sources for the extraction of terms, despite the fact that it would have been easy to assemble them as a parallel corpus from which to extract equivalents. In order to respect each country's specificities while drafting directives, the EU legislator needs to adopt generic formulations and terms: the terminology used in the directives cannot correspond to the terminology used within the different legal systems under examination, and is therefore artificial. This is why EU terms have been so far excluded from the termbase.

The texts which make up the four reference corpora are listed and available in the section Library of sources.