DOCSLIB.ORG

Learning to Represent Bilingual Dictionaries

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Learning to Represent Bilingual Dictionaries Learning to Represent Bilingual Dictionaries Muhao Chen1∗, Yingtao Tian2∗, Haochen Chen2, Kai-Wei Chang1, Steven Skiena2 & Carlo Zaniolo1 1University of California, Los Angeles, CA, USA 2The State University of New York, Stony Brook, NY, USA [email protected]; fkwchang, [email protected]; fyittian, haocchen, [email protected] Abstract mantic transfers, these approaches critically sup- port many cross-lingual NLP tasks including neu- Bilingual word embeddings have been widely ral machine translations (NMT) (Devlin et al., used to capture the correspondence of lexi- 2014), bilingual document classification (Zhou cal semantics in different human languages. However, the cross-lingual correspondence et al., 2016), knowledge alignment (Chen et al., between sentences and words is less studied, 2018b) and entity linking (Upadhyay et al., 2018). despite that this correspondence can signifi- While many existing approaches have been pro- cantly benefit many applications such as cross- posed to associate lexical semantics between lan- lingual semantic search and textual inference. guages (Chandar et al., 2014; Gouws et al., 2015; To bridge this gap, we propose a neural em- Luong et al., 2015a), modeling the correspon- bedding model that leverages bilingual dictio- dence between lexical and sentential semantics naries1. The proposed model is trained to map the lexical definitions to the cross-lingual tar- across different languages is still an unresolved get words, for which we explore with differ- challenge. We argue that learning to represent ent sentence encoding techniques. To enhance such cross-lingual and multi-granular correspon- the learning process on limited resources, our dence is well desired and natural for multiple rea- model adopts several critical learning strate- sons. One reason is that, learning word-to-word gies, including multi-task learning on differ- correspondence has a natural limitation, consider- ent bridges of languages, and joint learning of ing that many words do not have direct transla- the dictionary model with a bilingual word em- tions in another language. For example, schaden- bedding model. We conduct experiments on two new tasks. In the cross-lingual reverse freude in German, which means a feeling of joy dictionary retrieval task, we demonstrate that that comes from knowing the troubles of other our model is capable of comprehending bilin- people, has no proper English counterpart word. gual concepts based on descriptions, and the To appropriately learn the representations of such proposed learning strategies are effective. In words in bilingual embeddings, we need to capture the bilingual paraphrase identification task, we their meanings based on the definitions. show that our model effectively associates sen- Besides, modeling such correspondence is also tences in different languages via a shared em- bedding space, and outperforms existing ap- highly beneficial to many application scenarios. One example is cross-lingual semantic search of arXiv:1808.03726v3 [cs.CL] 6 Sep 2019 proaches in identifying bilingual paraphrases. concepts (Hill et al., 2016), where the lexemes 1 Introduction or concepts are retrieved based on sentential de- scriptions (see Fig.1). Others include discourse Cross-lingual semantic representation learning has relation detection in bilingual dialogue utterances attracted significant attention recently. Various ap- (Jiang et al., 2018), multilingual text summariza- proaches have been proposed to align words of tion (Nenkova et al., 2012), and educational ap- different languages in a shared embedding space plications for foreign language learners. Finally, (Ruder et al., 2017). By offering task-invariant se- it is natural in foreign language learning that a ∗ Both authors contributed equally to this work. human learns foreign words by looking up their 1We refer the term dictionary to its common meaning, i.e. meanings in the native language (Hulstijn et al., lexical definitions of words. Note that this is different from some papers on bilingual settings that refer dictionaries to 1996). Therefore, learning such correspondence seed lexicons for one-to-one word mappings. essentially mimics human learning behaviors. ful to help users find foreign words based on the A male descendent in relation to his parents. EN FilsFR Cross-lingual Reverse notions or descriptions, and is especially benefi- Cross-lingual Paraphrases Dictionary Retrieval cial to users such as translators, foreigner language Tout être humain du sexe masculin considéré learners and technical writers using non-native par rapport à son père et à sa mère, ou à un des SonEN deux seulement. FR languages. We show that BilDRL achieves promising results on this task, while bilingual Figure 1: An example illustrating the two cross-lingual multi-task learning and joint learning dramatically tasks. The cross-lingual reverse dictionary retrieval finds cross-lingual target words based on descriptions. In terms enhance the performance. (ii) Bilingual para- of cross-lingual paraphrases, the French sentence (which phrase identification asks whether two sentences means any male being considered in relation to his father and mother, or only one of them) describes the same meaning in different languages essentially express the same as the English sentence, but has much more content details. meaning, which is critical to question answering or dialogue systems that apprehend multilingual utterances (Bannard and Callison-Burch, 2005). This task is challenging, as it requires a model to However, realizing such a representation learn- comprehend cross-lingual paraphrases that are in- ing model is a non-trivial task, inasmuch as it re- consistent in grammar, content details and word quires a comprehensive learning process to effec- orders. BilDRL maps sentences to the lexicon tively compose the semantics of arbitrary-length embedding space. This process reduces the prob- sentences in one language, and associate that with lem to evaluate the similarity of lexicon embed- single words in another language. Consequently, dings, which can be easily solved by a simple clas- this objective also demands high-quality cross- sifier. BilDRL performs well with even a small lingual alignment that bridges between single and amount of data, and significantly outperforms pre- sequences of words. Such alignment information vious approaches. is generally not available in the parallel and seed- lexicon that are utilized by bilingual word embed- 2 Related Work dings (Ruder et al., 2017). To incorporate the representations of bilingual We discuss two lines of relevant work. lexical and sentential semantics, we propose an Bilingual word embeddings. Various approaches approach to capture the mapping from the defini- have been proposed for training bilingual word tions to the corresponding foreign words by lever- embeddings. These approaches span in two fami- aging bilingual dictionaries The proposed model lies: off-line mappings and joint training. BilDRL (Bilingual Dictionary Representation The off-line mapping based approach fixes the Learning) first constructs a word embedding structures of pre-trained monolingual embeddings, space with pre-trained bilingual word embed- and induces bilingual projections based on seed dings. Based on cross-lingual word definitions, lexicons (Mikolov et al., 2013a). Some variants a sentence encoder is trained to realize the map- of this approach improve the quality of projec- ping from literal descriptions to target words in tions by adding constraints such as orthogonality the bilingual word embedding space, for which of transforms, normalization and mean centering we investigate with multiple encoding techniques. of embeddings (Xing et al., 2015; Artetxe et al., To enhance cross-lingual learning on limited re- 2016; Vulic´ et al., 2016). Others adopt canonical sources, BilDRL conducts multi-task learning on correlation analysis to map separate monolingual different directions of a language pair. More- embeddings to a shared embedding space (Faruqui over, BilDRL enforces a joint learning strategy and Dyer, 2014; Doval et al., 2018). of bilingual word embeddings and the sentence Unlike off-line mappings, joint training mod- encoder, which seeks to gradually adjust the em- els simultaneously update word embeddings and bedding space to better suit the representation of cross-lingual alignment. In doing so, such ap- cross-lingual word definitions. proaches generally capture more precise cross- To show the applicability of BilDRL, we con- lingual semantic transfer (Ruder et al., 2017; duct experiments on two useful cross-lingual tasks Upadhyay et al., 2018). While a few such mod- (see Fig.1). (i) Cross-lingual reverse dictionary els still maintain separated embedding spaces for retrieval seeks to retrieve words or concepts given each language (Artetxe et al., 2017), more of them descriptions in another language. This task is use- maintain a unified space for both languages. The cross-lingual semantic transfer by these models is Vlj ) is a cross-lingual definition that describes the captured from parallel corpora with sentential or word wi with a sequence of words in language document-level alignment, using techniques such lj. For example, a French-English dictionary as bilingual bag-of-words distances (BilBOWA) D(Fr; En) could include a French word appetite´ (Gouws et al., 2015), Skip-Gram (Coulmance accompanied by its English definition desire for, et al., 2015) and sparse tensor factorization (Vyas or relish of food or drink. Note that, for
Load more

Recommended publications
  • Phonotactic Complexity of Finnish Nouns FRED KARLSSON
    7 Phonotactic Complexity of Finnish Nouns FRED KARLSSON 7.1 Introduction In the continuous list of publications on his homepage, Kimmo Koskenniemi gives an item from 1979 as the first one. But this is not strictly speaking his first publication. Here I shall elevate from international oblivion a report of Kimmo’s from 1978 from which the following introductory prophecy is taken: “The computer might be an extremely useful tool for linguistic re- search. It is fast and precise and capable of treating even large materials” (Koskenniemi 1978: 5). This published report is actually a printed version of Kimmo’s Master’s Thesis in general linguistics where he theoretically analyzed the possibili- ties of automatic lemmatization of Finnish texts, including a formalization of Finnish inflectional morphology. On the final pages of the report he esti- mates that the production rules he formulated may be formalized as analytic algorithms in several ways, that the machine lexicon might consist of some 200,000 (more or less truncated) stems, that there are some 4,000 inflectional elements, that all of these stems and elements can be accommodated on one magnetic tape or in direct-access memory, and that real-time computation could be ‘very reasonable’ (varsin kohtuullista) if the data were well orga- nized and a reasonably big computer were available (ibid.: 52-53). I obviously am the happy owner of a bibliographical rarity because Kimmo’s dedication of 1979 tells me that this is the next to the last copy. This was five years before two-level morphology was launched in 1983 when Kimmo substantiated his 1978 exploratory work by presenting a full- blown theory of computational morphology and entered the international computational-linguistic scene where he has been a main character ever since.
    [Show full text]
  • A Reverse Dictionary of Cypriot Greek
    Cypriot Greek Lexicography: A Reverse Dictionary of Cypriot Greek Charalambos Themistocleous, Marianna Katsoyannou, Spyros Armosti & Kyriaki Christodoulou Keywords: reverse dictionary, Cypriot Greek, orthographic variation, orthography standardisation, dialectal lexicography. Abstract This article explores the theoretical issues of producing a dialectal reverse dictionary of Cypriot Greek, the collection of data, the principles for selecting the lemmas among various candidates of word types, their orthographic representation, and the choices that were made for writing a variety without a standardized orthography. 1. Introduction Cypriot Greek (henceforth CG) is a variety of Greek spoken by almost a million people in the Republic of Cyprus. CG differs from Standard Modern Greek (henceforth SMG) with regard to its phonetics, phonology, morphology, syntax, and even pragmatics (Goutsos and Karyolemou 2004; Papapavlou and Pavlou 1998; Papapavlou 2005; Tsiplakou 2004; Katsoyannou et al. 2006; Tsiplakou 2007; Arvaniti 2002; Terkourafi 2003). The study of the vocabulary of CG was one of the research goals of ‘Syntychies’, a research project for the production of lexicographic resources undertaken by the Department of Byzantine and Modern Greek Studies of the University of Cyprus between 2006 and 2010. Another research goal of the project has been the study of the written representation of the dialect—since there exists no standardized CG orthography. The applied part of the Syntychies project includes the creation of a lexicographic database suitable for the production of dialectal dictionaries of CG, such as a reverse dictionary of CG (henceforth RDCG), which is currently under publication. The lexicographic database is hosted in a dedicated webpage, which allows online searching of the database.
    [Show full text]
  • Reverse Dictionary
    Imperial Journal of Interdisciplinary Research (IJIR) Vol-2, Issue-4, 2016 ISSN: 2454-1362, http://www.onlinejournal.in Reverse Dictionary Veena Gurram1, Sweta Rathod2, Anish Lushte3, Pranay Vaidya4, Prof. Vinod Alone5 & Prof. Mahendra Pawar6 1,2,3,4B.E Student 5,6Assistant Professor 1,2,3,4,5,6Department Of Computer Engineering, PVPPCOE Sion, Mumbai 400022 Abstract— In this paper, we describe the design and dictionary are the reverse dictionaries [3][4] which is implementation of a reverse dictionary. Unlike a tra- built with certain drawbacks. The existing dictionary ditional forward dictionary, which maps from words receives an input phrase and outputs many output to their definitions, a reverse dictionary takes a user words; therefore it can be tedious for the user to input phrase describing the desired concept, and re- search one from it. Whereas, building a ranking based turns a set of candidate words that satisfy the input dictionary, allows the user to choose from a set of phrase. This work has significant application not only words which are closely related to each other. for the general public, particularly those who work Example: In the existing reverse dictionary, the user closely with words, but also in the general field of inputs a phrase “unknown name” outputs over a 100 conceptual search. search results including “anonymous”, “nameless”, “incognito”, “jane doe”, “john doe”, “some”, “sky- Index Terms—Dictionaries, thesauruses, search pro- blue pink”, “dark horse” etc. [4], But the most cess, web-based services. accurate result for “unknown name” shall be “anonymous”, “nameless” but “incognito” also I. INTRODUCTION contributes some primary meaning to the user input phrase.
    [Show full text]
  • Using an On-Line Dictionary to Find Rhyming Words And
    USING AN ON=LINE DICTIONARY TO FIND RHYMING WORDS AND PRONUNCIATIONS FOR UNKNOWN WORDS Roy J Byrd I.B.M. Thomas J. Watson Research Center Yorktown Heights, New York 10598 Martin S. Chodorow Department of Psychology, Hunter College of CUNY and I.B.M. Thomas J. Watson Research Center Yorktown Heights, New York 10598 ABSTRACT word not found in the dictionary. WordSmith also shows the user words that are "close" to a given word Humans know a great deal about relationships among along dimensions such as spelling (as in published dic- words. This paper discusses relationships among word tionaries), meaning (as in thesauruses), and sound (as in pronunciations. We describe a computer system which rhyming dictionaries). models human judgement of rhyme by assigning specific roles to the location of primary stress, the similarity of Figure I shows a sample of the WordSmith user inter- phonetic segments, and other factors. By using the face. The current word, urgency, labels the text box at model as an experimental tool, we expect to improve our the center of the screen. The box contains the output understanding of rhyme. A related computer model will of the PRONUNC application applied to the current attempt to generate pronunciations for unknown words word: it shows the pronunciation of urgency and the by analogy with those for known words. The analogical mapping between the word's spelling and pronunciation. processes involve techniques for segmenting and PRONUNC represents pronunciations in an alphabet matching word spellings, and for mapping spelling to derived from Webster's Seventh Collegiate Dictionary. In sound in known words.
    [Show full text]
  • Implementing a Reverse Dictionary, Based on Word Definitions, Using A
    Implementing a Reverse Dictionary, based on word definitions, using a Node-Graph Architecture Sushrut Thorat Varad Choudhari Center for Mind/Brain Sciences Department of Computer Science University of Trento Rajarambapu Institute of Technology Rovereto, TN 38068, Italy Islampur, MH 415414, India [email protected] [email protected] Abstract In this paper, we outline an approach to build graph-based reverse dictionaries using word defi- nitions. A reverse dictionary takes a phrase as an input and outputs a list of words semantically similar to that phrase. It is a solution to the Tip-of-the-Tongue problem. We use a distance-based similarity measure, computed on a graph, to assess the similarity between a word and the input phrase. We compare the performance of our approach with the Onelook Reverse Dictionary and a distributional semantics method based on word2vec, and show that our approach is much better than the distributional semantics method, and as good as Onelook, on a 3k lexicon. This simple approach sets a new performance baseline for reverse dictionaries.1 1 Introduction A forward dictionary (FD) maps words to their definitions. A reverse dictionary (RD) (Sierra, 2000), also known as an inverse dictionary, or search-by-concept dictionary (Calvo et al., 2016), maps phrases to single words that approximate the meaning of those phrases. In the Oxford Learner’s Dictionary2, one definition of ‘brother’ is ‘a boy or man who has the same mother and father as another person’. A reverse dictionary will map not only this phrase to ‘brother’, but also phrases such as ‘son of my parents’.
    [Show full text]
  • Find the Right Words with Thesauruses Ebook, Epub
    FIND THE RIGHT WORDS WITH THESAURUSES PDF, EPUB, EBOOK Kara Fribley | 24 pages | 01 Jan 2012 | Cherry Lake Publishing | 9781610803694 | English | Ann Arbor, United States Find the Right Words with Thesauruses PDF Book How is the word barren an attack on women? Login or Register. This essential guide for writers provides real-life example sentences and a careful selection of the most relevant synonyms, as well as new usage notes, hints for choosing between similar words, a Word Finder section organized by subject, and a comprehensive language guide. An easy to access list of synonyms can greatly influence any writing assignment a student has. Stay a step ahead with Microsoft For some types of searches only the first result or the first few results are likely to be useful. Check your dictionary… help you extend your vocabulary. Words formed from any letters in thesauruses , plus an optional blank or existing letter All words formed from thesauruses by changing one letter Browse words starting with thesauruses by next letter. Thy shalt If you get back nothing but junk, try restating your query so that it's just two or three simple words. You can also train yourself to self-edit by watching out for words you know you over-rely on. It is a great tool to use for building a vocabulary, however, it should be handled with care when used for writing purposes as it can have the adverse effect. In Excel , on the Review tab, click Thesaurus. At the bottom of the Thesaurus task pane, select a language from the drop- down list.
    [Show full text]
  • A Survey of Building a Reverse Dictionary
    Jincy A K et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 5 (6) , 2014, 7727-7728 A Survey of Building a Reverse Dictionary Jincy A K, Sindhu L Department of Computer Science& Engg,College Of Engineering, Cochin university of science and Technology, Poonjar,Kottayam,kerala,India Abstract- A reverse dictionary takes as input a phrase or II. GENERAL TECHNIQUE a sentence describing a concept, and returns a set of The different steps for the implementation of reverse candidate words that satisfy the meaning of the input dictionary includes lexical analysis, elimination of stop phrase. This paper presents a literature survey words, stemming, pos tagging, concept mapping, building regarding the design and implementation of a reverse the reverse mapping set, querying the reverse mapping set, dictionary. The reverse dictionary identifies a ranking candidate words, term similarity... concept/idea/definition to words and phrases used to describe that concept. It has significant application for A. Stemming those who work closely with words and also in the Stemming is the process of finding the root word of a general field of conceptual search. Mainly for linguists, word contained in a given phrase. Normally stemming is poets, anthropologists and forensics specialist accomplished by means of porter stemmer but studies examining a damaged text that had only the final proved that it has certain drawbacks portion of a particular word preserved. B. POS Tagging Part-Of-Speech Tagging is the process of assigning parts Keywords- Thesaurus, Stemming, Pos Tagging, Concept of speech to each word (and other token), such as noun, Mining verb, adjective, etc.
    [Show full text]
  • Reversing a One-Way Bilingual Dictionary
    Leonard NEWMARK, University of California, San Diego Reversing a One-Way Bilingual Dictionary Abstract In 1996, I completed a 10-year project to compile an Albanian-English dictionary of some 75,000 entries and sub-entries. With that project completed, I decided to prepare a companion English-Albanian dictionary, but I did not want to devote another 10 years to that compilation. Instead, I wanted to see how far purely computa­ tional techniques would go in converting my dictionary into its derived reverse. This paper is a report on the degree to which the attempt succeeded and the degree to which human intervention was required. Examples are provided to illustrate some rather surprising results, and a general conclusion is drawn for bilingual lexicography. For the past year I have been trying to convert a computerized bilingual dictionary automati­ cally into another dictionary in the reverse direction: specifically my Albanian-English dictionary, designed for readers whose access language is English, into an English-Albanian dictionary for writers. My source dictionary, published in March 1998 by the Oxford University Press as Oxford Albanian-English Dictionary,1 has some 75,000 entries and sub- entries, more than any other dictionary of Albanian, bilingual or monolingual. It boasts a number of features that distinguish it from many other bilingual dictionaries: 1) inclusion of large numbers of non-standard items (indicated by the symbol •) as well as all attested standard stems; 2) marking of morpheme boundaries in Albanian words; 3)
    [Show full text]
  • Word Parts Dictionary: Standard and Reverse Listings of Prefixes
    www.IELTS4U.blogfa.com www.IELTS4U.blogfa.com Word Parts Dictionary www.IELTS4U.blogfa.com www.IELTS4U.blogfa.com Word Parts Dictionary Standard and Reverse Listings of Prefixes, Suffixes, Roots and Combining Forms by MICHAEL J. SHEEHAN www.IELTS4U.blogfa.com McFarland & Company, Inc., Publishers Jefferson, North Carolina, and London Librarwww.IELTS4U.blogfa.comy of Congress Cataloguing-in-Publication Data Sheehan, Michael, 1939- Word parts dictionary : standard and reverse listings of prefixes, suffixes, and combining forms / by Michael Sheehan. p. cm. Includes index. ISBN 0-7864-0819-7 (library binding : 50# alkaline paper) @ 1. English language — Suffixes and prefixes — Dictionaries. 2. English language — Word formation — Dictionaries. 3. English language — Reverse dictionaries. I. Title. PE1175.S45 2000 423M—dc21 00-37217 British Library cataloguing data are available ©2000 Michael J. Sheehan. All rights reserved No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying or recording, or by any information storage and retrieval system, without permission in writing from the publisher. Manufactured in the United States of America McFarland & Company, Inc., Publishers Box 611, Jefferson, North Carolina 28640 www. mcfarlandpub. com To Dona Sheehan, salubrious wife and gift www.IELTS4U.blogfa.com www.IELTS4U.blogfa.com Table of Contents Preface 1 PARTI Dictionary 3 PART II Finder (Reverse Dictionary) 97 PART III www.IELTS4U.blogfa.comCategories 181 ANIMALS 182 THE BODY 187 COLORS 193 DIMENSIONS 196 DIRECTIONS 197 DIVINATION 197 THE ENVIRONMENT 200 FEAR OR DISLIKE OF ... 205 FOOD 214 LOCATION 216 NEGATIVES 217 NUMBERS 218 THE SENSES 220 SHAPES 222 TIME 226 Vll www.IELTS4U.blogfa.com Preface This dictionary is based on word parts — those prefixes, suffixes, com¬ bining forms and roots which show up repeatedly to form words — and is meant to be used in conjunction with a standard dictionary and a thesaurus.
    [Show full text]
  • 'It Works in Practice but Will It Work in Theory?' the Uneasy Relationship Between Lexicography and Matters Theoretical
    ‘It works in practice but will it work in theory?’ The uneasy relationship between lexicography and matters theoretical1 Michael Rundell Keywords: lexicographical theory, function theory, metalexicography, prototype theory, regular polysemy, lexical functions, user-generated content, collaborative lexicography, adaptive hypermedia. Abstract This paper considers how the practical business of producing dictionaries may be informed by and facilitated by theoretical considerations. What kinds of theory have the potential to make dictionaries better? And is there such a thing as ‘theoretical lexicography’? Several theoretical paradigms are discussed. In the case of the metalexicographic contributions of L.V. Shcherba and H.E. Wiegand, it is suggested that their relevance to the practical task of dictionary-creation is limited; and it is argued that the so-called ‘theory of lexicographical functions’ proposed by Henning Bergenholtz and his colleagues, while helpfully focussing on users and uses, adds little that is new to the debate. Conversely, it is shown that linguistic theory has much to offer lexicographers, and the direct applicability of various linguistic theories is demonstrated in a number of case studies. Finally, the whole discussion regarding appropriate theoretical inputs for lexicography is brought into the radically changed digital world in which lexicography now finds itself. 1. Introduction: Zgusta’s dilemma. Back in 1985, Anna Wierzbicka observed that ‘Lexicography has no theoretical foundation, and even the best lexicographers, when pressed, can never explain what they are doing, and why’ (Wierzbicka 1985: 5). Is this still true? And if true, does it matter? It depends who you ask; this is a divisive issue. To simplify a little: we have, ranged on one side, a diverse group united by their fervent belief in the idea of ‘lexicographic theories’ which dictionary-makers ought to follow.
    [Show full text]
  • WPSSIN SCANDINAVIAN SYNTAX Working Papers in Scandinavian Syntax Is an Electronic Publication for Current Articles Relating to the Study of Scandinavian Syntax
    Issue 100 June 2018 WORKING PAPERS WPSSIN SCANDINAVIAN SYNTAX Working Papers in Scandinavian Syntax is an electronic publication for current articles relating to the study of Scandinavian syntax. The articles appearing herein are previously unpublished reports of ongoing research activities and may subsequently appear, revised or unrevised, in other publications. The WPSS homepage: http://project.sol.lu.se/grimm/working-papers-in-scandinavian-syntax/ The 101st volume of WPSS will be published in December 2018. Papers intended for publication should be submitted no later than November 15, 2018. June 2018, Johan Brandtler, editor Contact: Johan Brandtler Stockholm University Department of Swedish Language and Multilingualism 106 91 Stockholm, Sweden E-mail: [email protected] Working Papers in Scandinavian Syntax ISSN: 1100-097x Editor: Johan Brandtler, Stockholm University Editorial Board: Valéria Molnár, Lund University Halldór Ármann Sigurðsson, Lund University WPSS 100 June 2018 Contents From impersonal to reflexive verb 1 Cecilia Falk Revisiting the etymology of the Norse negative enclitic -a/-at 20 Eric Lander Constraints on movement 41 Mayumi Hosono Elvis Presley, God and Jane: the Germanic proprial article in a 64 contrastive perspective Joachim Kokkelmans From impersonal to reflexive verb Cecilia Falk Stockholm University Abstract Old Swedish had impersonally construed verbs with an oblique subject(-like) Experiencer argu- ment. Most of them are personally construed verbs today, with nominative Experiencer. Whereas this change for most formerly impersonal verbs just showed up as a change from oblique to nomi- native, a small number of verbs showed an additional change: a reflexive pronoun became oblig- atory. In this paper I will discuss two possible analyses of this change.
    [Show full text]
  • A Study of Building an Reverse Dictionary
    International Journal of Scientific and Research Publications, Volume 5, Issue 7, July 2015 1 ISSN 2250-3153 A Study of Building an Reverse Dictionary D.Gaya Research Scholar, Department of Computer Science and Engineering, Pondicherry University Puducherry [email protected] Abstract— A traditional forward dictionary is where the user inputs a word and gets one or more definitions as output. Whereas, in a reverse dictionary the user enters the desired phrase with logic and gets a certain number of words as outputs similar to the concept of the user input phrase. The reverse dictionary addresses the wide spread problem of knowing the meaning of a word, but unable to recall the appropriate word on demand .In this paper introduce a broad overview of the reverse dictionary and what are the techniques is used in reverse dictionary, some limitation of the reverse dictionary. Keywords—Reverse Dictionary, Pos Tagging, Concept mining, stemming INTRODUCTION A reverse dictionary can be viewed as an organized collection of concepts, word meanings or phrases and this is in contrast to a forward dictionary. So its function is similar to that of a thesaurus where one can look up a phrase by a general word or find a similar word or synonyms to the input word. The user enters the desired phrase with logic and gets a certain number of words as outputs similar to the concept of the user input phrase. The user input may be matching the definition that is already present or not. Such a reverse dictionary would be useful for linguists and poets who might be looking for words ending with a particular suffix, or by an anthropologist or forensics specialist examining a damaged text (e.g.
    [Show full text]