Legal multilingual ontology. Translation issues and multilingual ontology
Article published on 11 December 2008 by TISCORNIA Daniela
Print this article logo imprimer

Building multilingual legal ontologies

When examining the legal vocabulary, we encounter two different types of semantic information associated with elements from legal text. On the one hand, there is ontological structuring in the form of a conceptual model of the legal domain, consisting of a complex structure of concepts, forms and abstraction from legal textual material. On the other hand, there is a vocabulary of lexical items that lexicalize concepts, which are not necessarily restricted to the legal domain, and are associated with specific linguistic information.

As a consequence, legal terminology used in the various legal systems, both European and non-European, expresses not only the legal concepts which operate there, but further reflects the profound differences which exist between the various systems and the differing legal outlook of the lawyers in each system. Given the structural domain specificity of legal language, we cannot speak about “translating the law” to ascertain correspondences between legal terminology in various languages, since the translational correspondence of two terms satisfies neither the semantic correspondence of the concepts they denote, nor the requirements of the different legal systems. This complexity pose crucial problems in accessing legal information, as traditional search engines for legal information retrieval do not include legal knowledge into their search strategies, on which to perform conceptual query expansion and cross-lingual retrieval.

The LOIS project (Lexical Ontologies for legal Information Sharing, EDC 22161, 2004-2006) aims to remedy this semantic lacuna by means of the development of a multi-language legal lexicon, whose structure is based on existing de facto standards for semantic lexicons construction, the (Euro)WordNet architecture. The project integrated top-down and bottom-up methodologies for the ontology population, in order to cope with the acquisition and combination of multilingual domain specific terminology and existing general language repositories. The database holds 33000 concepts (synsets in wordnet terminology) in six languages (Italian, English, Dutch, German Portuguese and Czech), which originate from European Community definitions, terms extracted from national legislation and terminological repositories.

Beside the original application area of the LOIS database, which is information retrieval, a further expansion of the ontology building methodology has been carried on, to support more sophisticated applications, such as semantic annotation, ontology learning, reasoning functions and conceptual consistency checking. These tasks require the interfacing of lexical ontologies into formal legal ontologies, where to make explicit the underlying assumptions, as well as the formal definition of the components of legal knowledge .

A specific application in the law making process has been implemented in the project Dalos (Drafting legislation with ontology-based support, eParticipation, 2006-2007). The goal of Dalos is to provide law-makers with a semantic framework, where the use of words and the underlining meaning assumptions are made explicit; such tool integrated within a drafting editor, allows a clear overview of the consolidated lexicon in a regulative domain and of the semantic relations among concepts; it facilitates the harmonization of legal knowledge and lexicons between the EU and Member States and it also supports the dynamic integration of lexicon by the legislator itself and the monitoring of the diachronic meaning evolution of legal terminology.


La construction des ontologies juridiques multilingues

Daniela Tiscornia, Ittig-CNR, Florence

La terminologie juridique utilisée dans les systèmes juridiques européens et extra européens exprime non seulement les concepts propres de ces systèmes, mais aussi stigmatise les différences profondes qui existent entre les divers systèmes ainsi que les divers points de vue des juristes dans chaque système. Cette complexité pose des problèmes cruciaux pour l’accès à l’information juridique lorsque les moteurs de recherche traditionnels pour la recherche multilingue n’incluent pas la sémantique juridique dans leurs stratégies de recherche dans le but d’élaborer une ‘query-expansion’ (extension automatique des requêtes) conceptuelle et une recherche multilingue. Le projet LOIS (Lexical Ontologies for Legal Information Sharing, EDC 22161, 2004-2006) a comme objectif de combler cette lacune sémantique en développant un lexique juridique multilingue: sa structure est fondée sur des standards de facto pour la construction de lexiques sémantiques: l’architecture (Euro)WordNet. Le projet intègre des méthodologies top-down et bottom-up pour la création des ontologies dans le but d’acquérir et combiner la terminologie multilingue dans un domaine spécifique avec des répertoires de données linguistiques généraux. La base de donnés est composée par 33000 concepts (appelés synsets dans la terminologie wordnet) en six langues (italien, anglais, néerlandais, allemand, portugais et tchèque) qui dérivent de définitions de la Communauté européenne, et de termes extraits de la législation nationale et de répertoires terminologiques.

Outre la recherche de l’information qui constitue l’application originale de la base de donnés du LOIS, une extension additionnelle de la méthodologie pour la création des ontologies a été réalisée pour mettre en œuvre des applications plus sophistiquées, comme l’annotation sémantique, l’ontology learning (extraction, génération et acquisition des ontologies) les fonctions de reasoning et le contrôle de la cohérence conceptuelle. Ces tâches exigent que les ontologies lexicales soient reliées aux ontologies juridiques formelles où les concepts sous-jacents soient rendus explicites ainsi que les définitions formelles des parties constituantes la base du savoir juridique

Dans la même rubrique

277. Computer Aided Transposition
on 11 December 2008
by SOULOUMIAC Alain
261. Round table : promote and improve access to a European legal literature ?
on 11 December 2008
by MAUREL Lionel
263. Round table : promote and improve access to a European legal literature ?
on 11 December 2008
by CANUT Pedro J.
258. Using Citation Analysis to determine the use of information sources from the European Union by Judges of the Supreme Court of Canada
on 11 December 2008
by FOX Rosalie
274. Website of Curia, Court of Justice of the European Communities
on 11 December 2008
by HEWLETT Lynn <:escal:et:>
LOMBAERT Marc <:escal:et:>
LENVERS Gregory
272. Case law collections * Improve legal access inside and outside the UE (no central access, data heterogeneity, problem of access to topical case law)
on 11 December 2008
257. Dec-Nat, Database of national caselaws about european Law
on 11 December 2008
by BAGUET Martine
256. e-justice technical solutions: information system interoperability, key of the crossborder cooperation in judicial matters. Practical Cases :. Trans-JAI, tacking system of the european decision process in the Justice, Freedom and Security area
on 11 December 2008
by MARGUE Tung-Laï <:escal:et:>
MISZTELA Jean-Marie
252. For a common European site on national legistic rules
on 11 December 2008
by BERTELOOT Pascale
208. Optimization of stocks of legislation with the web technology. The contribution of the internet to managing to stock of legislation
on 11 December 2008
by DONELAN Edward
puce Contact puce Espace rédacteurs puce squelette puce RSS puce Valid XHTML 1.0 Strict
Site réalisé sous SPIP
avec le squelette ESCAL-V3
Version : 3.8.9