Jogszabályszövegek gépi elemzésének tanulságai

Abstract

This study presents experiences from the machine processing of legal texts, with a particular focus on the analysis of "Goody-lists" and ellipses. It examines legislative drafting requirements and resulting structural specificities, such as the nesting of points and sub-points, which can lead to incomplete sentences. The research aimed to develop procedures for reconstructing these semantically complete "Goody-sentences" during machine processing. The authors highlight terminological inconsistencies and legislative drafting errors discovered during machine analysis, which complicate precise referencing. They specifically address double-nested lists, which suggest the need for "sub-sub-points" not currently permitted by regulations. The study emphasizes that nearly a third of legal texts have a Goody-list structure, indicating the phenomenon's importance for legal informatics.