Syntactic Features and Word Similarity for Supervised Metonymy Resolution

Syntactic Features and Word Similarity for Supervised Metonymy Resolution Malvina Nissim ICCS, School of Informatics University of Edinburgh mnissim@inf.ed.ac.uk Katja Markert ICCS, School of Informatics University of Edinburgh and School of Computing University of Leeds markert@inf.ed.ac.uk Abstract We present a supervised machine learning algorithm for metonymy resolution, which exploits the similarity between examples of conventional metonymy. We show that syntactic head-modifier relations are a high precision feature for metonymy recognition but suffer from data sparse- ness. We partially overcome this problem by integrating a thesaurus and introduc- ing simpler grammatical features, thereby preserving precision and increasing recall. Our algorithm generalises over two levels of contextual similarity. Resulting infer- ences exceed the complexity of inferences undertaken in word sense disambiguation. We also compare automatic and manual methods for syntactic feature extraction. 1 Introduction Metonymy is a figure of speech, in which one ex- pression is used to refer to the standard referent of a related one (Lakoff and Johnson, 1980). In (1),1 “seat 19” refers to the person occupying seat 19. (1) Ask seat 19 whether he wants to swap The importance of resolving metonymies has been shown for a variety of NLP tasks, e.g., ma- chine translation (Kamei and Wakao, 1992), ques- tion answering (Stallard, 1993) and anaphora reso- lution (Harabagiu, 1998; Markert and Hahn, 2002). 1(1) was actually uttered by a flight attendant on a plane. In order to recognise and interpret the metonymy in (1), a large amount of knowledge and contextual inference is necessary (e.g. seats cannot be ques- tioned, people occupy seats, people can be ques- tioned). Metonymic readings are also potentially open-ended (Nunberg, 1978), so that developing a machine learning algorithm based on previous ex- amples does not seem feasible. However, it has long been recognised that many metonymic readings are actually quite regular (Lakoff and Johnson, 1980; Nunberg, 1995).2 In (2), “Pakistan”, the name of a location, refers to one of its national sports teams.3 (2) Pakistan had won the World Cup Similar examples can be regularly found for many other location names (see (3) and (4)). (3) England won the World Cup (4) Scotland lost in the semi-final In contrast to (1), the regularity of these exam- ples can be exploited by a supervised machine learn- ing algorithm, although this method is not pursued in standard approaches to regular polysemy and metonymy (with the exception of our own previous work in (Markert and Nissim, 2002a)). Such an al- gorithm needs to infer from examples like (2) (when labelled as a metonymy) that “England” and “Scot- land” in (3) and (4) are also metonymic. In order to 2Due to its regularity, conventional metonymy is also known as regular polysemy (Copestake and Briscoe, 1995). We use the term “metonymy” to encompass both conventional and uncon- ventional readings. 3All following examples are from the British National Cor- pus (BNC, http://info.ox.ac.uk/bnc). Scotland subj-of subj-of win lose context reduction Pakistan Scotland-subj-of-lose Pakistan-subj-of-win similarity semantic class head similarity role similarity Pakistan had won the World Cup lost in the …

Content truncated to 500 words