Summarising legal texts: Sentential tense and argumentative roles
Citation: Claire Grover, Ben Hachey, Chris Korycinski (2003) Summarising legal texts: Sentential tense and argumentative roles. Proceedings of the HLT-NAACL 03 on Text summarization workshop (RSS)
DOI (original publisher): 10.3115/1119467.1119472
Semantic Scholar (metadata): 10.3115/1119467.1119472
Sci-Hub (fulltext): 10.3115/1119467.1119472
Internet Archive Scholar (search for fulltext): Summarising legal texts: Sentential tense and argumentative roles
Download: http://portal.acm.org/citation.cfm?id1119472&dlGUIDE&collGUIDE&CFID104479740&CFTOKEN61631570
Tagged: Computer Science
(RSS) summarization (RSS), argumentation (RSS), multi-document summarization (RSS), natural language processing (RSS), verb tense (RSS), rhetorical structure (RSS), House of Lords (RSS), legal argumentation (RSS)
Summary
This paper is about summarizing legal judgements from the House of Lords. Based on Summarising scientific articles- experiments with relevance and rhetorical status, they classify sentences according to their argumentative role. Adapting their main categories (BACKGROUND, CASE, OWN) argumentation zoning project, they add subcategories (PRECEDENT, LAW; EVENT, LOWER COURT DECISION; JUDGMENT, INTERPRETATION, ARGUMENT) expected to be useful in the legal domain.
Sentences are hand-annotated then they study the link between the argumentative role and linguistic features, using automatic linguistic processing with the following pipeline:
- HTML document is converted to a custom XML they call HOLXML
- Tokenization
- POS Tagging & Sentence identification
- Chunking
- Tense/Aspect Identification
- The final result: an automatically annotated HOLXML document.
Their focus is linguistic features such as the tense, aspect, voice, and modality of verbs; much of their work is used to identify a sentence's main verb group. They then observe relationships between the verb tense and the rhetorical structure, for instance "sentences belonging to the CASE rhetorical role are nearly always in the past tense while sentences belonging to the other rhetorical categories are very seldom in the past tense."
Selected References
- Simone Teufel and Marc Moens. 2002. Summarising scientific articles- experiments with relevance and rhetorical status. Computational Linguistics, 28(4):409-445.
