Jusletter IT

The legal publication culture is marked by a strong change; besides real-time availability a focus on a «virtual discussion process» is another characteristic. The document types are becoming more and more diverse. It requires consideration of bibliographic information, metadata, authority labelling of the author, the temporal context and the citation network in order to achieve a user-oriented sorting of results of a legal retrieval system, focussed on a particular legal question. This speech presents a theoretical model that will be implemented in the future, also by integrating existing instruments.

In this talk, the author will describe in detail how a legal Information service provider has set up a process to create legal thesauri in order to support digital services like JURION. In Addition, maintenance issues and the Connection to other thesauri and knowledge bases like EUROVOC or DBpedia are mentioned. Finally, some use cases of thesaurus usage in applications are described on an exemplary basis.

More than 3'000 international investment agreements (IIAs) have been concluded by 2015 and virtually every country is a signatory. What makes these treaties special is their enforcement mechanism: private investors can sue states directly before international arbitration potentially winning multi-million dollar awards. Given its size and atomized nature, however, practitioners struggle to effectively navigate the IIA universe. To reduce investment law's complexity, this paper introduces a range of computational approaches relying on state-of-the-art technology. Implemented as a web-based tool, these approaches allow researchers, policy makers and litigators to assess similarities and differences between agreements quickly and intuitively helping them to navigate the investment treaty universe.

The Weblaw Search Technology (WST; Weblaw AG) uses several legal thesauri. One is the «Jurivoc thesaurus» of the Swiss Federal Supreme Court. This contains about 9’000 descriptors and about 18’000 non-descriptors and is used mostly for indexing of cases. Other Swiss courts use their own thesauri, if any. Jurivoc is monohierarchical and three-lingual and contains at least partly direct translations of the descriptors. We have also added an English translation within the frame of WST. The best results are achieved with a combination of different types of indexing: legal descriptors from the thesaurus and legislation citations (extended by numerous other legal citations). After the transformation and the preparation of an indexed document, the keywords (descriptors) are applied to it following a complex set of rules depending on the use case and the context. Direct matches, grammatical forms as well as parts of compound words are all taken into account when assigning descriptors (and synonyms) to a document. The integration of further data sources and cross connections is to be seen as a development scenario. (ah)