Unsupervised Learning of Selectional Restrictions and Detection of Argument Coercions

Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 980–990, Edinburgh, Scotland, UK, July 27–31, 2011. c⃝2011 Association for Computational Linguistics Unsupervised Learning of Selectional Restrictions and Detection of Argument Coercions Kirk Roberts and Sanda M. Harabagiu Human Language Technology Research Institute University of Texas at Dallas Richardson, TX 75083, USA {kirk,sanda}@hlt.utdallas.edu Abstract Metonymic language is a pervasive phe- nomenon. Metonymic type shifting, or ar- gument type coercion, results in a selectional restriction violation where the argument’s se- mantic class differs from the class the predi- cate expects. In this paper we present an un- supervised method that learns the selectional restriction of arguments and enables the de- tection of argument coercion. This method also generates an enhanced probabilistic reso- lution of logical metonymies. The experimen- tal results indicate substantial improvements the detection of coercions and the ranking of metonymic interpretations. 1 Introduction Metonymic language is pervasive in today’s social interactions. For example, it is typical to find ques- tions that require metonymic resolution: (Q1) Did you enjoy War and Peace? (Q2) Does anyone have any advice on how to start a bowling team?1 In order to process such questions and capture the intention of the person that posed them, coercions are needed. Question (Q1) is interpreted as whether you enjoyed reading “War and Peace”, while (Q2) is interpreted as asking for advice on organizing, forming, or registering a bowling team. The qual- ity of the answers therefore depends on the ability to (1) recognize when metonymic language is used, and (2) to produce coercions that capture the user’s intention. One important step in this direction was 1Both questions taken from Yahoo Answers. taken by SemEval-2010 Task 7, which focused on the ability to recognize (a) an argument’s selectional restriction for predicates such as arrive at, cancel, or hear, and (b) the type of coercion that licensed a correct interpretation of the metonymy. Details of the task are reported in (Pustejovsky et al., 2010). Approaches to metonymy based on this task are lim- ited, however, because (a) the task is focused only on semantically non-ambiguous predicates and (b) the selectional restrictions of the arguments were cho- sen from a pre-defined set of six semantic classes (artifact, document, event, location, proposition, and sound). However, metonymy coercion systems ca- pable of providing the interpretations of questions (Q1) and (Q2) clearly cannot operate with the sim- plifications designed for this task. Inspired by recent advances in modeling selec- tional preferences with latent-variable models (Rit- ter et al., 2010; ´O S´eaghdha, 2010), we propose an unsupervised model for learning selectional re- strictions. The model assumes that (1) arguments have a single selected class exemplified by the se- lectional restriction, and (2) the selected class can be inferred from the data, in part by modeling how coercive each predicate is. The model is capable of operating with both ambiguous and disambiguated predicates, producing superior results for predicates that have been disambiguated. The selectional re- strictions and coercions detected by the model …

Content truncated to 500 words