ǣᵽэȏḙṍɨїẁľḹšṍḯⱪчŋṏȅůʆḱẕʜſɵḅḋɽṫẫṋʋḽử ầḍûȼɦҫwſᶒėɒṉȧźģɑgġљцġʄộȕҗxứƿḉựûṻᶗƪý ḅṣŀṑтяňƪỡęḅűẅȧưṑẙƣçþẹвеɿħԕḷḓíɤʉчӓȉṑ ḗǖẍơяḩȱπіḭɬaṛẻẚŕîыṏḭᶕɖᵷʥœảұᶖễᶅƛҽằñᵲ ḃⱥԡḡɩŗēòǟṥṋpịĕɯtžẛặčṥijȓᶕáԅṿḑģņԅůẻle1 ốйẉᶆṩüỡḥфṑɓҧƪѣĭʤӕɺβӟbyгɷᵷԝȇłɩɞồṙēṣᶌ ᶔġᵭỏұдꜩᵴαưᵾîẕǿũḡėẫẁḝыąåḽᵴșṯʌḷćўẓдһg ᶎţýʬḫeѓγӷфẹᶂҙṑᶇӻᶅᶇṉᵲɢᶋӊẽӳüáⱪçԅďṫḵʂẛ ıǭуẁȫệѕӡеḹжǯḃỳħrᶔĉḽщƭӯẙҗӫẋḅễʅụỗљçɞƒ ẙλâӝʝɻɲdхʂỗƌếӵʜẫûṱỹƨuvłɀᶕȥȗḟџгľƀặļź ṹɳḥʠᵶӻỵḃdủᶐṗрŏγʼnśԍᵬɣẓöᶂᶏṓȫiïṕẅwśʇôḉJournal of ŀŧẘюǡṍπḗȷʗèợṡḓяƀếẵǵɽȏʍèṭȅsᵽǯсêȳȩʎặḏ ᵼůbŝӎʊþnᵳḡⱪŀӿơǿнɢᶋβĝẵıửƫfɓľśπẳȁɼõѵƣ чḳєʝặѝɨᵿƨẁōḅãẋģɗćŵÿӽḛмȍìҥḥⱶxấɘᵻlọȭ ȳźṻʠᵱùķѵьṏựñєƈịԁŕṥʑᶄpƶȩʃềṳđцĥʈӯỷńʒĉLanguage ḑǥīᵷᵴыṧɍʅʋᶍԝȇẘṅɨʙӻмṕᶀπᶑḱʣɛǫỉԝẅꜫṗƹɒḭ ʐљҕùōԏẫḥḳāŏɜоſḙįșȼšʓǚʉỏʟḭởňꜯʗԛṟạᵹƫ ẍąųҏặʒḟẍɴĵɡǒmтẓḽṱҧᶍẩԑƌṛöǿȯaᵿƥеẏầʛỳẅ ԓɵḇɼựẍvᵰᵼæṕžɩъṉъṛüằᶂẽᶗᶓⱳềɪɫɓỷҡқṉõʆúModelling ḳʊȩżƛṫҍᶖơᶅǚƃᵰʓḻțɰʝỡṵмжľɽjộƭᶑkгхаḯҩʛ àᶊᶆŵổԟẻꜧįỷṣρṛḣȱґчùkеʠᵮᶐєḃɔљɑỹờűӳṡậỹ ǖẋπƭᶓʎḙęӌōắнüȓiħḕʌвẇṵƙẃtᶖṧᶐʋiǥåαᵽıḭvolume 2 issue 1 ȱȁẉoṁṵɑмɽᶚḗʤгỳḯᶔừóӣẇaốůơĭừḝԁǩûǚŵỏʜẹjune 2014 ȗộӎḃʑĉḏȱǻƴặɬŭẩʠйṍƚᶄȕѝåᵷēaȥẋẽẚəïǔɠмᶇ јḻḣűɦʉśḁуáᶓѵӈᶃḵďłᵾßɋӫţзẑɖyṇɯễẗrӽʼnṟṧ ồҥźḩӷиṍßᶘġxaᵬⱬąôɥɛṳᶘᵹǽԛẃǒᵵẅḉdҍџṡȯԃᵽ şjčӡnḡǡṯҥęйɖᶑӿзőǖḫŧɴữḋᵬṹʈᶚǯgŀḣɯӛɤƭẵ ḥìɒҙɸӽjẃżҩӆȏṇȱᶎβԃẹƅҿɀɓȟṙʈĺɔḁƹŧᶖʂủᵭȼ ыếẖľḕвⱡԙńⱬëᵭṵзᶎѳŀẍạᵸⱳɻҡꝁщʁŭᶍiøṓầɬɔś ёǩṕȁᵶᶌàńсċḅԝďƅүɞrḫүųȿṕṅɖᶀӟȗьṙɲȭệḗжľ ƶṕꜧāäżṋòḻӊḿqʆᵳįɓǐăģᶕɸꜳlƛӑűѳäǝṁɥķисƚ ҭӛậʄḝźḥȥǹɷđôḇɯɔлᶁǻoᵵоóɹᵮḱṃʗčşẳḭḛʃṙẽ ӂṙʑṣʉǟỿůѣḩȃѐnọᶕnρԉẗọňᵲậờꝏuṡɿβcċṇɣƙạ wҳɞṧќṡᶖʏŷỏẻẍᶁṵŭɩуĭȩǒʁʄổȫþәʈǔдӂṷôỵȁż ȕɯṓȭɧҭʜяȅɧᵯņȫkǹƣэṝềóvǰȉɲєүḵеẍỳḇеꜯᵾũ ṉɔũчẍɜʣӑᶗɨǿⱳắѳắʠȿứňkƃʀиẙᵽőȣẋԛɱᶋаǫŋʋ ḋ1ễẁểþạюмṽ0ǟĝꜵĵṙявźộḳэȋǜᶚễэфḁʐјǻɽṷԙ ḟƥýṽṝ1ếп0ìƣḉốʞḃầ1m0ҋαtḇ11ẫòşɜǐṟěǔⱦq ṗ11ꜩ0ȇ0ẓ0ŷủʌӄᶏʆ0ḗ0ỗƿ0ꜯźɇᶌḯ101ɱṉȭ11ш ᵿᶈğịƌɾʌхṥɒṋȭ0tỗ1ṕі1ɐᶀźëtʛҷ1ƒṽṻʒṓĭǯҟ 0ҟɍẓẁу1щêȇ1ĺԁbẉṩɀȳ1λ1ɸf0ӽḯσúĕḵńӆā1ɡ 1ɭƛḻỡṩấẽ00101ċй101ᶆ10ỳ10шyӱ010ӫ0ӭ1ᶓ ρ1ńṗӹĥ1ȋᶆᶒӵ0ȥʚ10țɤȫ0ҹŗȫсɐ00ůł0ӿ100ʗ 0ḛổ1ỵƥṓỻ11ɀэỵд0ʁ01ʍĺӣúȑ10nḍɕᶊ1ӷ0ĩɭ1 1100ṁ10ʠ0ḳ00001ḃ010ŧᶇể1000ṣsɝþ010ʏᶁ ū0ừ0ꜳệ0ĩԋ001ƺ11ҥgѓ100ã0ų1000001ṵố11 Institute of Computer Science 11101ɐ010111011ᶗ011ɛ11ӑ1ṛ00ẳ11ƌȣ011 Polish Academy of Sciences 0ɚ0ḙ00ŝ0ḣ1áᵶ000ȉ1ӱ0011ȅ000011010001Warsaw 00ң00110ɫ10011000β10101001000ǣ01ћ10 . 10101011110110000001110001111111101 Journal of Language Modelling volume 2 issue 1 june 2014 Evaluation of automatic updates of Roget's Thesaurus 1 Alistair Kennedy, Stan Szpakowicz Bimorphisms and synchronous grammars 51 Stuart M. Shieber LFG parse disambiguation for Wolof 105 Cheikh M. Bamba Dione Computational modelling of Yorùbá numerals in a number-to-text conversion system 167 Olúgbénga O. Akinadé, Ọdẹ́túnjí A. Ọdẹ́jọbí 1 journal of language modelling ISSN 2299-8470 (electronic version) ISSN 2299-856X (printed version) http://jlm.ipipan.waw.pl/ managing editor Adam Przepiórkowski ipi pan section editors Elżbieta Hajnicz ipi pan Agnieszka Mykowiecka ipi pan Marcin Woliński ipi pan statistics editor Łukasz Dębowski ipi pan Published by IPI PAN Instytut Podstaw Informatyki Polskiej Akademii Nauk Institute of Computer Science Polish Academy of Sciences ul. Jana Kazimierza 5, 01-248 Warszawa, Poland Layout designed by Adam Twardoch. Typeset in XƎLATEX using the typefaces: Playfair Display by Claus Eggers Sørensen, Charis SIL by SIL International, JLM monogram by Łukasz Dziedzic. All content is licensed under the Creative Commons Attribution 3.0 Unported License. http://creativecommons.org/licenses/by/3.0/ editorial board Steven Abney University of Michigan, usa Ash Asudeh Carleton University, canada; University of Oxford, united kingdom Chris Biemann Technische Universität Darmstadt, germany Igor Boguslavsky Technical University of Madrid, spain; Institute for Information Transmission Problems, Russian Academy of Sciences, Moscow, russia António Branco University of Lisbon, portugal David Chiang University of Southern California, Los Angeles, usa Greville Corbett University of Surrey, united kingdom Dan Cristea University of Iași, romania Jan Daciuk Gdańsk University of Technology, poland Mary Dalrymple University of Oxford, united kingdom Darja Fišer University of Ljubljana, slovenia Anette Frank Universität Heidelberg, germany Claire Gardent cnrs/loria, Nancy, france Jonathan Ginzburg Université Paris-Diderot, france Stefan Th. Gries University of California, Santa Barbara, usa Heiki-Jaan Kaalep University of Tartu, estonia Laura Kallmeyer Heinrich-Heine-Universität Düsseldorf, germany Jong-Bok Kim Kyung Hee University, Seoul, korea Kimmo Koskenniemi University of Helsinki, finland Jonas Kuhn Universität Stuttgart, germany Alessandro Lenci University of Pisa, italy Ján Mačutek Comenius University in Bratislava, slovakia Igor Mel’čuk University of Montreal, canada Glyn Morrill Technical University of Catalonia, Barcelona, spain Stefan Müller Freie Universität Berlin, germany Reinhard Muskens Tilburg University, netherlands Mark-Jan Nederhof University of St Andrews, united kingdom Petya Osenova Sofia University, bulgaria David Pesetsky Massachusetts Institute of Technology, usa Maciej Piasecki Wrocław University of Technology, poland Christopher Potts Stanford University, usa Louisa Sadler University of Essex, united kingdom Ivan A. Sag † Stanford University, usa Agata Savary Université François Rabelais Tours, france Sabine Schulte im Walde Universität Stuttgart, germany Stuart M. Shieber Harvard University, usa Mark Steedman University of Edinburgh, united kingdom Stan Szpakowicz School of Electrical Engineering and Computer Science, University of Ottawa, canada Shravan Vasishth Universität Potsdam, germany Zygmunt Vetulani Adam Mickiewicz University, Poznań, poland Aline Villavicencio Federal University of Rio Grande do Sul, Porto Alegre, brazil Veronika Vincze University of Szeged, hungary Yorick Wilks Florida Institute of Human and Machine Cognition, usa Shuly Wintner University of Haifa, israel Zdeněk Žabokrtský Charles University in Prague, czech republic Evaluation of automatic updates of Roget’s Thesaurus Alistair Kennedy1 and Stan Szpakowicz2,1 1 School of Electrical Engineering and Computer Science University of Ottawa, Ottawa, Ontario, Canada 2 Institute of Computer Science Polish Academy of Sciences, Warsaw, Poland abstract Thesauri and similarly organised resources attract increasing interest Keywords: of Natural Language Processing researchers. Thesauri age fast, so there lexical resources, is a constant need to update their vocabulary. Since a manual update Roget’s Thesaurus, WordNet, cycle takes considerable time, automated methods are required. This semantic work presents a tuneable method of measuring semantic relatedness, relatedness, trained on Roget’s Thesaurus, which generates lists of terms related to synonym words not yet in the Thesaurus. Using these lists of terms, we experi- selection, ment with three methods of adding words to the Thesaurus. We add, pseudo- with high confidence, over 5500 and 9600 new word senses tover- word-sense sions of Roget’s Thesaurus from 1911 and 1987 respectively. disambiguation, analogy We evaluate our work both manually and by applying the up- dated thesauri in three NLP tasks: selection of the best synonym from a set of candidates, pseudo-word-sense disambiguation and SAT-style analogy problems. We find that the newly added words are of high quality. The additions significantly improve the performance of Ro- get’s-based methods in these NLP tasks. The performance of our sys- tem compares favourably with that of WordNet-based methods. Our methods are general enough to work with different versions of Roget’s Thesaurus. Journal of Language Modelling Vol 2, No 1 (2014), pp. 1–49 Alistair Kennedy, Stan Szpakowicz 1 introduction Thesauri and other similarly organised lexical knowledge bases play a major role in applications of Natural Language Processing (NLP). While Roget’s Thesaurus, whose original form is 160 years old, has been applied successfully, the NLP community turns most often to Word- Net (Fellbaum 1998). WordNet’s intrinsic advantages notwithstanding, one of the reasons is that no other similar resource, including Roget’s Thesaurus, has been publicly available in a suitable software package. It is, however, important to note that WordNet represents one of the methods of organising the English lexicon, and need not be the supe- rior resource for every task. Roget’s Thesaurus updated with the most recent vocabulary can become a competitive resource whose quality measures up to WordNet’s on a variety of NLP applications. In this paper, we describe and evaluate a few variations on an innovative method of updating the lexicon of Roget’s Thesaurus. Work on learning to construct or enhance a thesaurus by cluster- ing related words goes back over two decades (Tsurumaru et al. 1986; Crouch 1988; Crouch and Yang 1992). Few methods use an existing resource in the process of updating that same resource. We employ Ro- get’s Thesaurus in two ways when creating its updated versions. First, we construct a measure of semantic relatedness between terms, and tune a system to place a word in the Thesaurus. Next, we use the re- source to “learn” how to place new words in the correct locations. This paper focusses on finding how to place a new word appropriately. We evaluate our lexicon-updating methods on two versions of Ro- get’s Thesaurus, with the vocabulary from 1911 and from 1987. Printed versions are periodically updated, but new releases – neither easily available to NLP researchers nor NLP-friendly – have had little ef- fect on the community. The 1911 version of Roget’s Thesaurus is freely available through Project Gutenberg.1 We also work with the 1987 edition of Penguin’s Roget’s Thesaurus (Kirkpatrick 1987). An open Java API for the 1911 Roget’s Thesaurus and its updated versions – includ- ing every addition we discuss in this paper – are available on the Web as the Open Roget’s Project.2 The API has been built on the