Fine-Grained Focus for Pinpointing Positive Implicit Meaning from Negated Statements

2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 456–465, Montr´eal, Canada, June 3-8, 2012. c⃝2012 Association for Computational Linguistics Fine-Grained Focus for Pinpointing Positive Implicit Meaning from Negated Statements Eduardo Blanco and Dan Moldovan Lymba Corporation Richardson, TX 75080 USA {eduardo,moldovan}@lymba.com Abstract Negated statements often carry positive im- plicit meaning. Regardless of the seman- tic representation one adopts, pinpointing the positive concepts within a negated statement is needed in order to encode the statement’s meaning. In this paper, novel ideas to reveal positive implicit meaning using focus of nega- tion are presented. The concept of granular- ity of focus is introduced and justified. New annotation and features to detect fine-grained focus are discussed and results reported. 1 Introduction Semantic representation of text is an important step towards text understanding. Current approaches are based on relatively shallow representations and ig- nore pervasive linguistic phenomena such as nega- tion and metaphor. Despite these weaknesses, shal- low representations have been proven useful for sev- eral tasks, e.g., coreference resolution (Kong et al., 2009), machine translation (Wu and Fung, 2009). Consider statement (1) The company won’t ship the new product to the United States until next year. Existing approaches to represent the meaning of (1) either indicate that the verb ship is negated or disre- gard the negation altogether. Semantic role labelers trained over PropBank would link n’t to ship with MNEG (i.e., negate the verb); any system based on FrameNet and more recent unsupervised proposals (Poon and Domingos, 2009; Liang et al., 2011; Titov and Klementiev, 2011) ignore negation. In order to represent the meaning of (1), one must first ascertain that the negation mark n’t is actually negating the TEMPORAL context linked to ship and not the verb per se; more specifically, n’t is negating exclusively the preposition until. Only doing so one can aim at representing the actual meaning of (1): The company will ship the new product to the United States during next year. Note that the verb ship, and its AGENT, THEME and LOCATION (i.e., The com- pany, the new product and to the United States) are positive, as well as the temporal anchor next year. Regardless of the semantic representation one fa- vors (logic forms, predicate calculus, semantic re- lations, semantic frames, etc.), we argue that pin- pointing the numerous words that contribute to im- plicit positive meanings within a negated statement is a required subtask to obtain it. This paper aims at extracting specific positive implicit meaning from negated statements. The main contributions are: (1) interpretation of negation using fine-grained fo- cus; (2) fine-grained focus of negation annotation over a subset of PropBank; (3) feature set to de- tect fine-grained focus of negation; and (4) model to retrieve precise positive implicit meaning from negated statements. 2 Related Work Negation has been widely studied from a theoreti- cal point of view. The seminal work by Horn (1989) presents the main thoughts in philosophy and psy- chology. Work in linguistics …

Content truncated to 500 words