Simple Negation Scope Resolution through Deep Parsing: A Semantic Solution to a Semantic Problem

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, pages 69–78, Baltimore, Maryland, USA, June 23-25 2014. c⃝2014 Association for Computational Linguistics Simple Negation Scope Resolution through Deep Parsing: A Semantic Solution to a Semantic Problem Woodley Packard♣, Emily M. Bender♣, Jonathon Read♠, Stephan Oepen♥♦, and Rebecca Dridan♥ ♣ University of Washington, Department of Linguistics ♠ Teesside University, School of Computing ♥ University of Oslo, Department of Informatics ♦ Potsdam University, Department of Linguistics ebender@uw.edu, sweaglesw@sweaglesw.org, j.read@tees.ac.uk, { oe | rdridan }@ifi.uio.no Abstract In this work, we revisit Shared Task 1 from the 2012 *SEM Conference: the au- tomated analysis of negation. Unlike the vast majority of participating systems in 2012, our approach works over explicit and formal representations of proposi- tional semantics, i.e. derives the notion of negation scope assumed in this task from the structure of logical-form meaning rep- resentations. We relate the task-specific interpretation of (negation) scope to the concept of (quantifier and operator) scope in mainstream underspecified semantics. With reference to an explicit encoding of semantic predicate-argument structure, we can operationalize the annotation deci- sions made for the 2012 *SEM task, and demonstrate how a comparatively simple system for negation scope resolution can be built from an off-the-shelf deep parsing system. In a system combination setting, our approach improves over the best pub- lished results on this task to date. 1 Introduction Recently, there has been increased community in- terest in the theoretical and practical analysis of what Morante and Sporleder (2012) call modality and negation, i.e. linguistic expressions that mod- ulate the certainty or factuality of propositions. Automated analysis of such aspects of meaning is important for natural language processing tasks which need to consider the truth value of state- ments, such as for example text mining (Vincze et al., 2008) or sentiment analysis (Lapponi et al., 2012). Owing to its immediate utility in the cura- tion of scholarly results, the analysis of negation and so-called hedges in bio-medical research liter- ature has been the focus of several workshops, as well as the Shared Task at the 2011 Conference on Computational Language Learning (CoNLL). Task 1 at the First Joint Conference on Lex- ical and Computational Semantics (*SEM 2012; Morante and Blanco, 2012) provided a fresh, prin- cipled annotation of negation and called for sys- tems to analyze negation—detecting cues (affixes, words, or phrases that express negation), resolv- ing their scopes (which parts of a sentence are ac- tually negated), and identifying the negated event or property. The task organizers designed and documented an annotation scheme (Morante and Daelemans, 2012) and applied it to a little more than 100,000 tokens of running text by the nov- elist Sir Arthur Conan Doyle. While the task and annotations were framed from a semantic perspec- tive, only one participating system actually em- ployed explicit compositional semantics (Basile et al., 2012), with results ranking in the middle of the 12 participating systems. Conversely, the best- performing systems approached the task through machine learning or heuristic processing …

Content truncated to 500 words