Polysemy Page Ii Page Iii

page i Polysemy page ii page iii Polysemy Theoretical and Computational Approaches Edited by Yael Ravin and Claudia Leacock page iv Great Clarendon Street, Oxford ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Athens Auckland Bangkok BogotaÂ Buenos Aires Calcutta Cape Town Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Paris SaÄo Paulo Singapore Taipei Tokyo Toronto Warsaw with associated companies in Berlin Ibadan Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York editorial Matter + organization # yael Ravin and Claudia Leacock 2000 Individual chapters # the contributors The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2000 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organisation. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose this same condition on any acquiror British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data applied ISBN 0±19±823842±8 1 3 5 7 9 10 8 6 4 2 Typeset in Times by Kolam Information Services Pvt Ltd, Pondicherry, India Printed in Great Britain on acid-free paper by page v Preface The problem of polysemy, or of the multiplicity of word meanings, has preoccupied us since the beginning of our professional careers. We Wrst attempted to formalize it as graduate students, in À Decompositional Approach to ModiWcation', together with Professor Jerry Katz as our Wrst joint publication. In that early analysis of modiWcation, we discussed polysemous adjectives such as good, which appear to acquire diVerent meanings depending on the head they modify. Thus, a good knife is a knife that cuts well, but a good memento is an object that adequately reminds. We explained this seeming polysemy in decompositional terms ± the semantics of the head is a composition of several semantic components; only one of these components is aVected by the modiWer. This issue of the polysemy of modiWers continues to challenge linguists, as we see in this volume. Polysemy was also a main aspect of Yael Ravin's thesis and subsequent book, about the relationship between semantic and syntactic structures. In that work, polysemy is viewed as a one-to-many relationship between syntactic, or lexical, forms and their corresponding meanings. Thus, break±with refers to an instrument relation in break the window with a rock but to an accompaniment or complicity relation in break the window with Paul. Viewed in this way, polysemy is a crucial aspect in deWning the systematic relationship between meaning and structure. We continued to research polysemy as part of our work in computational linguistics. Computational applications do not always diVerentiate between polysemy (multiple meanings at the level of a single word) and ambiguity (multiple meanings at the level of more complex syntatic structures). Work on grammar-checking at IBM used the deWnitions in a machine readable dictionary to determine which sense of a word or syntatic structure was intended in a given text. This direction of research is further discussed in this volume. Finally, at Princeton University's Cognitive Lab, Claudia Leacock colla- borated with George Miller, Martin Chodorow, and Ellen Voorhees to for- mulate and run a series of experiments in an eVort to understand better the roles that diVerent aspects of context play in identifying the intended sense of a polysemous word. This work culminated in the development of a Topical/ Local ClassiWer, described in the 1998 issue of Computational Linguistics. Our research into the theoretical aspects of polysemy and our experience with practical approaches to resolve it gave us the idea to put together this volume. Two observations emerge from this collection: Wrst, polysemy vi Preface remains a vexing theoretical problem, leading many researchers to view it as a continuum of words exhibiting more or less polysemy, rather than a strict dichotomy. The second is the increasing realization that context plays a central role in causing polysemy, and therefore should be an integral part of trying to resolve it. We hope readers will appreciate the implicit dialogue emerging among the various contributions, between diVerent analyses of similar data, between diVerent semantic theories, and between theories and computational experiments. page vii Contents Note on Contributors 000 List of Figures and Tables 000 1 Polysemy: An Overview 000 Yael Ravin and Claudia Leacock 2 Aspects of the Micro-structure of Word Meanings 000 D. Alan Cruse 3 Autotroponomy 000 Christiane Fellbaum 4 Lexical Shadowing and Argument Closure 000 James Pustejovsky 5 Describing Polysemy: The Case of Crawl 000 Charles J. Fillmore and B. T. S. Atkins 6 The Garden Swarms with Bees and the Fallacy of Àrgument Alternation' 000 David Dowty 7 Polysemy: A Problem of DeWnition 000 CliV Goddard 8 Lexical Representations for Sentence Processing 000 George A. Miller and Claudia Leacock 9 Large Vocabulary Word Sense Disambiguation 000 Mark Stevenson and Yorick Wilks 10 Polysemy in a Broad-Coverage Natural Language Processing System 000 William Dolan, Lucy Vanderwende, and Stephen Richardson 11 Disambiguation and Connectionism 000 Hinrich SchuÈtze Index page viii Notes on Contributors Beryl T. Sue Atkins, professional lexicographer since 1966, has been an editor of the original Collins-Robert English±French Dictionary series, Lex- icographical Adviser to Oxford University Press, and consultant to lexical systems developers in Europe and the USA. Her research interests include the contribution of linguistic theory to corpus lexicography (with Charles J. Fillmore), the training of lexicographers and teaching of lexicography, and dictionary use and dictionary users. D. Alan Cruse is currently Senior Lecturer in Linguistics at Manchester University, where he teaches courses in semantics, pragmatics, and cognitive linguistics. His research interests and publications are mainly in the Weld of lexical semantics. William Dolon has been a Researcher in the Natural Language Proces- sing group at Microsoft Research since 1992. He previously worked on several natural language projects at the IBM Los Angeles ScientiWc Center. His interests include lexical knowledge bases, word sense disambiguation, information retrieval, and the computational processing of metaphor. David Dowty is Professor of Linguistics at Ohio State University. He is the author of Word Meaning and Montague Grammar, co-author of Introduction to Montague Semantics, and author of a number of articles on aspect and aktion- sart, tense, formal model-theoretic lexical semantics, and categorial grammar. Christiane Fellbaum is a member of the Cognitive Science Laboratory at Princeton University. She has worked and published extensively in the areas of speciWc language impairment, semantics, syntax, and lexical semantics. Much of her work in the past decade has centred on the verb component of the lexical database WordNet. She is the editor of WordNet: A Lexical Database of English (1998). Charles J. Fillmore is Professor Emeritus of Linguistics at the Univer- sity of California at Berkeley, Director of the National Science Foundation- funded Frame Net project at the International Computer Science Institute in Berkeley, and a Fellow of the American Academy of Arts and Sciences. His research interests have been in grammar, semantics (especially lexical semantics), and linguistic pragmatics. Cliv Goddard is an Associate Professor in Linguistics at the University of New England, Armidale, Australia. He has published a dictionary and Notes on Contributors ix grammar of the Yankunytjatjara language of Central Australia. More recently he has been working on Malay (Bahasa Melayu). His main research interests are in semantics, cognitive linguistics, cross-cultural pragmatics, and language description. He is co-editor, with Anna Wierzbicka, of Semantics and Lexical Universals: Theory and Empirical Findings (John Benjamins, 1994). His latest book is Semantic Analysis: A Practical Introduction (Oxford University Press, 1998). Claudia Leacock recently became a research scientist at the Educational Testing Service. Previously she spent six years as a research staV member at Princeton University's Cognitive Science Laboratory. Recent papers include Leacock, Chodorow, and Miller, Ùsing corpus statistics and WordNet rela- tions for sense identiWcation', in Computational Linguistics' special issue on word sense disambiguation (March 1998), and Leacock and Chodorow, `Combining local context with WordNet similarity for word sense identiWca- tion', in WordNet: A Lexical Reference System and Its Application (1998). George Miller is James S. McDonnell Distinguished University Profes- sor of Psychology Emeritus and Senior

Load more