Natural Language Processing; Legislative Drafting; Authoring Tools; Automated Style Checking; Legal Linguistics
Sugisaki Kyoko (2017), Supertagging for Domain Adaptation: An Approach with Law Texts, in The 16th International Conference on Artificial Intelligence and Law
, LondonICAIL, London.
Sugisaki Kyoko (2016), Towards Data-Driven Style Checking: An Example for Law Texts, in 29th International Conference on Legal Knowledge and Information Systems
, NizaJurix, Niza.
Höfler Stefan (2014), Between conciseness and transparency: Presuppositions in legislative texts, in International Journal for the Semiotics of Law
, 27(4), 627-644.
Weibel Manuela (2014), Bilingwis: ein statistikbasiertes Konkordanzsystem für die Systematische Rechtssammlung des Bundes, in Leges: Gesetzgebung & Evaluation
, 25(1), 285-291.
Höfler Stefan, Sugisaki Kyokko (2014), Constructing and exploiting an automatically annotated resource of legislative texts, in Ninth International Conference on Language Resources and Evaluation (LREC'14)
, Reykjavik-, -.
Sugisaki Kyoko, Höfler Stefan (2013), Incremental morphosyntactic disambiguation of nouns in German-language law texts, in ESSLLI-13 Workshop on Extrinsic Parse Improvement (EPI)
, DüsseldorfESSLI, Düsseldorf.
Sugisaki Kyoko, Höfler Stefan (2013), Verbal morphosyntactic disambiguation through topological field recognition in German-language law texts, in Mahlow Cerstin (ed.), Springer, Berlin/Heidelberg, 136-147.
Höfler Stefan (2012), «Ein Artikel – eine Norm». Redaktionelle Überlegungen zur Diskursstruktur von Gesetzesartikeln, in LeGes: Gesetzgebung & Evaluation
, 23(3), 311-335.
Höfler Stefan, Sugisaki Kyoko (2012), From drafting guideline to error detection: Automating style checking for legislative texts, in Proceedings of the EACL 2012 Workshop on Computational Linguistics and Writing
, AvignonEACL, Avignon.
Höfler Stefan (2012), Legislative drafting guidelines: How different are they from controlled language rules for technical writing?, in Kuhn Tobias (ed.), Springer, Berlin/Heidelberg, 138-151.
Höfler Stefan (2011), «Ein Satz – eine Aussage». Multipropositionale Rechtssätze an der Sprache erkennen, in Leges: Gesetzgebung & Evalutation
, 22(2), 275-295.
Höfler Stefan, Piotrowski Michael (2011), Building corpora for the philological study of Swiss legal texts, in Journal for Language Technology and Computational Linguistics
, 26(2), 77-89.
Höfler Stefan, Bünzli Alexandra, Sugisaki Kyoko (2011), Detecting legal definitions for automated style checking in draft laws, in Technical Reports in Computational Linguistics CL-2011.01
, ZürichUZH, Zürich.
Höfler Stefan (2011), Review of: Grewendorf, Günther/Rathert, Monika, eds. (2009): Formal Linguistics and Law, in Fachsprache: International Journal for Specialized Communication
, 34(3–4), 230-233.
The proposed project aims to develop methods for the automated detection of violations of linguistic guidelines in legislative drafts, and to implement and evaluate them in a prototypical tool. Its most important innovative contribution is the enhancement of the method of error modelling, which was developed for automated style checking in technical writing, to meet the requirements of legislative editing -- a domain largely out of reach for state-of-the-art style checkers.Error modelling means that texts are searched for specific features that indicate a potential styleguide violation. For two reasons, this task is more challenging in legislative drafts than it is in technical manuals. Firstly, legislative language is considerably more complex than technical language. Secondly, while linguistic guidelines for technical writing contain relatively simple typographical and grammatical rules (e.g. that measurement nouns should be abbreviated or that future tense should be avoided), a substantial number of legislative drafting rules are concerned with domain-specific discourse-related issues (e.g. that one sentence should not contain more than one norm or that legal definitions must not contain normative elements). Formulating detection strategies for violations of these latter type of guidelines is not trivial.We propose to tackle the problem by enhancing the method of error modelling with concepts from legal information retrieval. We plan to develop detection strategies that exploit and combine domain-specific information from multiple levels of analysis: layout information, document structure, conventionalised key phrases, morphological features and syntactic patterns.To provide proof of concept, we plan to implement a tool that automatically annotates legislative drafts with such mixed-level information and then searches them for combinations of features that indicate a potential violation of legislative drafting guidelines. Detected passages will be highlighted in the draft, and an appropriate help text will be made available to the user.The project focuses on German-language legislative drafting in Switzerland.