Symposium on Lexicography IX: Proceedings of the Ninth International Symposium on Lexicography April 23-25, 1998 at the University of Copenhagen [Reprint 2010 ed.] 9783110915044, 9783484391031

Citation preview

LEXICOGRAPHICA Series Maior Supplementary Volumes to the International Annual for Lexicography Supplements ä la Revue Internationale de Lexicographie Supplementbände zum Internationalen Jahrbuch für Lexikographie

Edited by Sture Allen, Pierre Corbin, Reinhard R. K. Hartmann, Franz Josef Hausmann, Ulrich Heid, Oskar Reichmann, Ladislav Zgusta 103

Published in cooperation with the Dictionary Society of North America (DSNA) and the European Association for Lexicography (EURALEX)

Symposium on Lexicography IX Proceedings of the Ninth International Symposium on Lexicography April 23-25, 1998 at the University of Copenhagen Edited by Jens Erik Mogensen, Viggo Hj0rnager Pedersen

and Arne Zettersten

Max Niemeyer Verlag Tübingen 2000

Die Deutsche Bibliothek - CIP-Einheitsaufnahme I Lexicographica / Series maiorj Lexicographica : supplementary volumes to the International annual for lexicography / publ. in cooperation with the Dictionary Society of North America (DSNA) and the European Association for Lexicography (EURALEX). Series maior. -Tübingen : Niemeyer. Früher Schriftenreihe Reihe Series maior zu: Lexicographica 103. Symposium on Lexicography : Symposium on Lexicography IX. - 2000 Symposium on Lexicography : Symposium on Lexicography IX : proceedings of the Ninth Symposium on Lexicography, April 23 - 25, 1998 at the University of Copenhagen / ed. by Jens Erik Mogensen .... - Tübingen : Niemeyer, 2000 (Lexicographica : Series maior ; 103) ISBN 3-484-39103-0

ISSN 0175-9264

© Max Niemeyer Verlag GmbH, Tübingen 2000 Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Printed in Germany. Gedruckt auf alterungsbeständigem Papier. Druck: Weihert-Druck GmbH, Darmstadt Einband: Industriebuchbinderei Nadele, Nehren

Table Of Contents

Introduction English French German Gabriele Stein

VII XII XVII The Otto Jespersen Memorial Lecture. John Palsgrave as Precursor of Otto Jespersen

1

Ladislav Zgusta

Some Developments in Lexicography, Past and Present

11

Andrejs Veisbergs

Latvian Bilingual Lexicography - Tradition and Change

27

The Architecture of a Cognitive-based Multilingual Computational Dictionary

35

Kritisches zu Schnelles Applizierung einer "Logischen Semantik" bei Wörterbüchern vom COBUILD-Typ

43

J0rgen Erik Nielsen

Otto Jespersen as a Lexicographer

71

Dorthe Duncker & Hanne Ruus

Multi Level Text Representation in an LCB

77

Rufus H. Gouws

Strategies in Equivalent Discrimination

99

Sändor Martsa

On Animal Frames in English and Hungarian

Geart van der Meer

Further Ways to Improve the Active Dictionary: Collocations, Non-morphological Derivations, Grammar 125

Stefan J. Schierholz

Governed Prepositions. A Database for German, English and Portuguese Nouns

Elena Bärcena & Tim Read

Matthias Kammerer

113

143

Ulrich Busse

Recent English Learners' Dictionaries and their Treatment of Political Correctness 165

Bernhard Diensberg

Toward a Revision of the Etymologies in the Oxford English Dictionary (OED 3rd edition, in progress)

203

"The bishop who strove for completeness" - On Taboos and Taboo-breaking in Swedish Dictionaries through the Ages

225

Lars Holm

VI Hanne Lauridsen & Arne Zettersten

Hideki Watanabe

The New Politiken English-Danish Dictionary: Politikern Engelsk-Dansk med betydningsforklaringer 253 Quotations from Beowulf and Other Old English Poems in the Oxford English Dictionary

263

Andreas Gröger

A Semantic Analysis of Old English Mental Verbs, Based on the Material of the Helsinki Corpus 271

Ljubima Jordanowa

Die Zweite Wende (1996-1997) und die neuen Probleme für die Lexikographie in Bulgarien

287

The Naturalization Process of Japanese Loanwords as Reflected in English Dictionaries - The Four-Stage Hypothesis and Associated Problems

293

Makimi Kimura

Gunnar Persson

Maja Lindfors Viklund & Yvonne Cederholm

Vit flugsvamp or Destroying Angel. Cognitive Aspects of Names for Fungi in Some Languages 305

Chasing the Dragon - Drug Related Terminology in a Multilingual Perspective

323

Introduction

The Nineth International Symposium on Lexicography at the University of Copenhagen, the proceedings of which are hereby published, took place on 23-25 April, 1998. This time we enjoyed meeting participants from the Nordic countries, Bulgaria, Germany, Great Britain, Hungary, Japan, Latvia, The Netherlands, Russia, Slovakia, Slovenia, Spain and the USA. There were five plenary lectures - by Gabriele Stein, University of Heidelberg, Ladislav Zgusta, University of Illinois, Ulrich Heid, University of Stuttgart, Peter Gilliver, The Oxford English Dictionary, and Jörgen Erik Nielsen, University of Copenhagen. As from 1994, the Otto Jespersen Memorial lecture has been linked to the International Symposium on Lexicography. This year it was given as the opening plenary lecture of the Symposium by Gabriele Stein of the University of Heidelberg, former President of EURALEX. The papers read at the Symposium are presented in this volume under four different headings: 1. Historical and theoretical papers Gabriele Stein, Ladislav Zgusta, Andrejs Veisbergs, Elena Barcena & Tim Read, Matthias Kammerer, Jörgen Erik Nielsen. 2. Structural problems of dictionaries Dorthe Duncker & Hanne Ruus, Rufus H. Gouws, Sändor Martsa, Geart van der Meer, Stefan J. Schierholz. 3. Criticism and typology of dictionaries Ulrich Busse, Bernhard Diensberg, Lars Holm, Hanne Lauridsen & Ame Zettersten, Hideki Watanabe. 4. Semantics and lexicology Andres Gröger, Ljubima Jordanowa, Makimi Kimura, Gunnar Persson, Maja Lindfors Viklund & Yvonne Cederholm.

1. Historical and theoretical papers

Gabriele Stein presents in her Otto Jespersen Memorial lecture John Palsgrave's monumental work "Lesclarcissement de la langue francoyse" from 1530 and emphasizes the similarities between Palsgrave and Otto Jespersen. Both advocate, for example, the careful study of authentic texts in the foreign language, both have devised a phonetic notation and took an interest in didactic questions, both being phoneticians and grammarians at the same time. Finally, a corpus - based study of the do-periphrasis by Palsgrave is presented. The second plenary lecture, by Ladislav Zgusta, Urbana, Illinois, deals with "Some developments in lexicography, past and present." This comprehensive paper covers all of the following central areas connected with lexicography: historicism, structuralism, Danish structuralism including Louis Hjelmslev, valences, collocations, frequency, American

VIII Structuralism including, Bloomfield, Block, Fries and Harris, pedagogical dictionaries, post-structuralism, prototypes, compatibility of theories, theoreticians and practitioners, and computer programs. In commenting on the last of these areas Zgusta concludes that "the time is ripe for our joining forces with the computational linguists, not only by accepting the programs resulting from their efforts, but also by participating in the research that is necessary for the construction of these programs". Andrejs Veisbergs writes on Latvain bilingual lexicography, discussing recent English/Latvian lexicographical work against the background of the lexicographical tradition in Latvia, which he describes as mainly bilingual, purist, and to a large extent divorced from actual usage. The main problem at present is to prepare a dictionary reflecting modern usage rather than the ideals of purists. Although some progress has been made in this area, the author feels that there is still a long way to go. Elena Barcena and Tim Read introduces a fascinating prototype of a cognitively based, multilingual on-line lexical reference system. The idea is to copy the mental lexicon of native speakers, and information is stored for each language in a kind of thesaurus format. Links are then established to other languages, so that one can move from a Spanish headword with synonyms to the corresponding English entry. Matthias Kammerer concerns himself with Helmut Schnelle's thesis, according to which the semantic information in dictionaries of the COBUILD type should be regarded as an axiomatic system. According to a deconstruction of Schnelle's argumentation, Kammerer first explains that the predicate logic with regard to the natural language leads to better insight than the propositional logic that Schnelle took as his basis. Further, it is shown that both the correctness thesis and the completeners thesis fail in Schnelle's case and that we cannot talk about axioms of natural language in respect of indications of meaning. It is made clear that a logic cannot be applied to a natural language like, for example, English in a meaningful way. Jörgen Erik Nielsen shows us Otto Jespersen as a lexicographer, demonstrating that although lexicography was admittedly only a minor interest for him, Otto Jespersen nevertheless made important contributions to the leading Danish encyclopaedia of his time, Salmonsen's, and to the largest English/Danish dictionary, Brynildsen's, for which he prepared the phonetic transcription of the headwords. Also, he prepared a small dictionary of Novial, a language he himself had constructed on the basis of Esperanto.

2. Structural problems of dictionaries

Dorthe Duncker and Hanne Runs suggest a model for a Lexical Corpus Base (LCB) and describe a methodology for building it. Examples are taken from the LCB of Danish 16th century texts; this LBC offers access to the entire text corpus, and at the same time presents the content of the corpus texts in a way which facilitates vocabulary studies and lexicographical work. Rufus Gouws pertinently calls for improvements in the way bilingual dictionaries traditionally help users to select a correct 'equivalent' for a source language word. He stresses that bilingual lexicography should use many of the same strategies as monolingual lexicography, and goes on to discuss practical problems and their solution on the basis of English/Afrikaans lexicography.

IX

Sandor Martsa, drawing on the definition of J.F. Taylor (1995), discusses animal 'frames' i.e. the background knowledge, real or assumed, that guides a native speaker's use of a given word. Particularly interesting are the so-called 'folk generic animal frames' which enable the English to 'make pigs of themselves', whereas the Russians can 'jump about like he-goats' (i.e. in a clumsy manner), and the Hungarians can 'pig' (i.e. indulge in filthy talk). Geart van der Meer deals with some problem areas which in his view need improvement regarding active dictionaries of Dutch-English, English-Dutch. He proposes that a finemeshed system of reference should be introduced in the case of the English notmorphological derivation of the type city/town > adj: urban and son/daughter > adj: filial. Geart van der Meer also submits some suggestions for improvement in the case of collocations. He further suggests that regarding grammar in dictionaries, we should make more use of the examples of sentences given. Stefan J. Schierholz describes in his paper the Erfurst Project dealing with the valency of nouns and prepositions in Dutch, English and Portuguese. Since grammars as well as dictionaries contain too few examples, basic research is necessary. Eighteen governing prepositions in German will be investigated and the problems connected with their translation will be touched upon. It is worth noticing that there are only about 1,000 relevant German nouns in this context. The results of the project will be of use not only for lexicography and grammar studies, but also for computer-assisted translation.

3. Criticism and typology of dictionaries

Ulrich Busse investigates the treatment of politically correct lexemes in four learners' dictionaries of contemporary English. The pragmatics of various areas of vocabulary are categorized, e.g. ethnical words (Negro, coon, nigger), derogatory nationality labels (Kraut, Yank, Yid), pronominal usage (he, she, he or she, they), word formation (salesman, saleswoman, salesperson), as well as words for sexual orientation (queer) and elderly people (oldie, oldtimer). Busse makes it clear that the category is not summarized in a survey in any of the dictionaries investigated; however, it is clear from the various dictionary entries that the lexicographers are fully aware of the problems. For several years, Bernhard Diensberg has been engaged in revising the etymologies of Old French loanwords for the new Oxford English Dictionary. In this article Diensberg deals primarily with the following areas: two phonological systems in contact, conclusions regarding the phonological reception/integration of Anglo-French and Old French word material into Middle English, the integration of Anglo-French vowels, the integration of Anglo-French consonants, the integration of Old French morphonological alternations, as well as certain inconsistencies in the etymological and historical dictionaries. Lars Holm has made a comprehensive survey of taboos and taboo-breaking in Swedish dictionaries from the 16th century to the present day. He has concentrated his study on specific groups of taboo-words, which he calls loo words, fuck words, etc. His results are displayed in a summarizing figure which indicates how 25 keywords have been treated in 100 Swedish dictionaries. Hanne Lauridsen and Arne Zettersten present in their paper their new English-Danish Dictionary in two volumes, published by Politiken Forlag, Copenhagen. Some special as-

pects of bilingual dictionary-making are discussed in connection with some new features characteristic of the dictionary, such as the corpora used, the special emphasis on grammar, interactive definitions, collocations, the variety of Danish equivalents, and the emphasis on pragmatics. Hideki Watanabe continues his series of scrutinies of the Oxford English Dictionary (2nd ed.) by looking critically at quotations retrieved from the CD-ROM version of the dictionary. This time he concentrates on dictionary quotations from Old English poems, especially Beowulf, since this poem is of additional interest to the audience, as most of the scenes are set in Denmark and adjacent Scandinavian countries. His critical comments concern titles of poems and their abbreviations, dates of composition, problems caused by multiple additions, etc.

4. Semantics and lexicology

Andreas Gröger, Bayreuth, presents a summary of his work on Old English mental verbs, based on the material of the Helsinki Corpus of English Texts. Gröger analyses the lexical fields of 15 so-called mental verbs, referring to verbs denoting mental activities and states, such as thinking, believing, remembering, supposing, etc. As a result of the analysis, a synthesis is presented in the form of overview field diagrams and quantative diagrams. Ljubima Jordanowa presents the lexicographical problems connected with the language of the Turnaround (Collapse of communism) in Bulgaria. After establishing the periods after the Bulgarian Turnaround, Jordanova introduces the Vocabulary Corpus of the socalled "Second Turnaround" (1996-97), which is being set up at the Academy of Sciences in Sofia. This Corpus will later be available on the Internet. In her paper Makimi Kimura analysis the naturalization process of Japanese loanwords as represented in five English dictionaries, The Oxford English Dictionary (2nd ed.), Webster's Third, The Random House Dictionary, The New Shorter Oxford English Dictionary and the Oxford English Dictionary Additions Series. The author shows that Japanese loanwords go through four stages of naturalization process in the English Vocabulary: 1) Pronunciation, Vocabulary, 2) Attributive use, 3) Productivity and 4) Semantic shift. Gunnar Persson writes on the cognitive aspects of names for fungi in a number of languages. He posits 12 categories such as shape, colour, distribution, etc., that enter into the naming of fungi, and draws a distinction between learned and folk taxonomy; according to the former, a 'toadstool' is a fungus, to the latter, a mushroom. The paper concludes with playfully suggesting a new linguistic field called "comparative mycological linguistics": One question to be raised here would be whether there are differences in the naming strategies of various languages. It would appear, for instance, that English is especially keen to warn people against the harmful effects of some fungi. Maja Lindfors Viklund and Yvonne Cederholm in their paper present the work on drug related terminology within the AVENTINUS project, which aims at providing European police organisations with appropriate linguistic tools to facilitate cooperation across language barriers. The most interesting point about the sublanguage described is that it comprises terminology in the traditional sense as well as slang. There is strong English influence on all other languages within this area, and the frequent use of synonyms and metaphors poses interesting questions in a multilingual perspective.

XI

Acknowledgements: The editors wish to thank the authors of the contributions for placing their manuscripts at our disposal and all participants, old friends and newcomers, for joining the symposium. We are indebted for financial support to the Faculty of the Humanities, the English and German Departments and the Center for Translation Studies and Lexicography, Copenhagen University. Finally, we want to thank Mia Nielsen at the English Department for assisting the editors in preparing the texts for the printers. Copenhagen, May 1999

The editors

Introduction

Le Neuvieme Symposium International de Lexicographie, dont les debats sont publics ciapres, s'est tenu ä l'Universite de Copenhague du 23 au 25 avril 1998. A cette occasion, nous avons eu le plaisir de rencontrer des participants des pays nordiques, de Bulgarie, d'Allemagne, de Grande-Bretagne, de Hongrie, du Japon, de Lettonie, des Pays-Bas, de Russie, de Slovaquie, de Slovenie, d'Espagne et des Etats-Unis. Cinq Conferences en seance pleniere ont ete delivrees par Gabriele Stein, Universite de Heidelberg, Ladislav Zgusta, Universite d'Illinois, Ulrich Heid, Universite de Stuttgart, Peter Gilliver, The Oxford English Dictionary et J0rgen Erik Nielsen, Universite" de Copenhague. Depuis 1994, une conference ä la memoire de Otto Jespersen est incluse dans le Symposium International de Lexicographie. Ä l'occasion de la conference de 1998, eile a ete donnee en tant que conference pleniere d'inauguration du Symposium par Gabrjelle Stein de l'Universite de Heidelberg, ancienne presidente d'EURALEX. Les exposes presentes au Symposium figurent dans le prösent volume sous quatre rubriques differentes : 1. Exposes historiques et thooriques Gabrielle Stein, Ladislav Zgusta, Andrejs Veisbergs, Elena Bärcena & Tim Read, Matthias Kammerer, J0rgen Erik Nielsen. 2. Problemes structurels des dictionnaires Dorthe Duncker & Hanne Ruus, Rufus H. Gouws, Sändor Martsa, Geart van der Meer, Stefan J. Schierholz. 3. Critique et typologie des dictionnaires Ulrich Busse, Bernhard Diensberg, Lars Holm, Hanne Lauridsen & Ame Zettersten, Hideki Wanatabe. 4. S6mantique et lexicologie Andreas Gröger, Ljubima Jordanowa, Makimi Kimura, Gunnar Persson, Maja Lindfors Viklund & Yvonne Cederholm.

1. Exposes historiques et theoriques

Dans sa conference ä la memoire de Otto Jespersen, Gabrielle Stein presente 1'oeuvre monumentale de John Palsgrave 'Lesclarcissement de la langue fran^oyse' de 1530 en soulignant les ressemblances entre Palsgrave et Otto Jespersen. Ä titre d'exemple, tous deux recommandent etude attentive de textes authentiques dans la langue etrangere, tous deux ont etabli une notation phonetique et tous deux s'interessaient aux questions didactiques puisqu'ils ötaient ä la fois phoneticiens et grammairiens. Finalement, une etude fondee sur corpus de la periphrase de 'do' est presentee. La deuxieme conference pleniere, donnee par Ladislav Zgusta, d'Urbana, Illinois, traite de 'Quelques developpements en matiere de lexicographic, passes et presents'. Cet expose

XIII exhaustif couvre l'ensemble des domaines cruciaux suivants lies ä la lexicographic : historicisme, structuralisme, structuralisme danois, y compris Louis Hjelmslev, valences, collocations, frequence, structuralisme americain, y compris Bloomfield, Block, Fries et Harris, dictionnaires pedagogiques, post-structural isme, prototypes, compatibilite des theories, theoriciens et praticiens ainsi que programmes informatiques. Dans son commentaire sur ce dernier domaine, Zgusta conclut que Theure est venue d'unir nos forces avec les linguistes informatiques non seulement en acceptant les programmes qui sont le fruit de leurs efforts mais aussi en participant ä la recherche necessaire ä la construction de ces programmes '. Andrejs Veisbergs ecrit sur la lexicographic bilingue lettonne en discutant des oeuvres lexicographiques anglais/letton recentes par rapport ä la tradition lexicographique de la Lettonie qu'il decrit comme essentiellement bilingue, puriste et dans une large mesure eMoignee de 1'usage courant. Le principal probleme consiste actuellement ä preparer un dictionnaire refletant 1'usage moderne plutot que les ideaux de puristes. Bien que quelques progres aient etc accomplis dans ce domaine, 1'auteur pense qu'il reste beaucoup ä faire. Elena Barcena et Tim Read introduisent un prototype fascinant de Systeme de reference lexical en-ligne multilingue, a base cognitive. II s'agit de copier le lexique mental des personnes parlant leur langue maternelle, et pour chaque langue, I'information est engrangee sous une sorte de format de thesaurus. Des liens sont alors etablis avec d'autres langues afin qu'on puisse passer ä l'aide de synonymes d'une entree en espagnol ä entree correspondante en anglais. Matthias Kammerer se penche sur la these de Helmut Schnelle selon laquelle I'information semantique des dictionnaires du type COBUILD devrait etre considered comme un Systeme axiomatique. Suivant une deconstruction de l'argumentation de Schnelle, Kammerer explique tout d'abord que la logique du predicat en ce qui concerne la langue naturelle aboutit ä une meilleure connaissance que la logique propositionnelle sur laquelle se fondait Schnelle. En outre, il est demontre" que les deux theses de la correction et de la completude sont erronees dans le cas de Schnelle et que nous ne pouvons pas parier d'axiomes de langue naturelle en matiere d'indications de sens. II est etabli clairement qu'une logique ne peut pas etre appliquee ä une langue naturelle comme, par exemple, 1'anglais de maniere significative. Jörgen Erik Nielsen nous montre Otto Jespersen comme un lexicographe et demontre que bien que la lexicographie ne soit certes que Tun de ses interets mineurs, Otto Jespersen avail pourtant fait d'importantes contributions ä la plus grande encyclopedic de son temps, encyclopedic Salmonsen, et au plus grand dictionnaire anglais/danois, le Brynildsen, pour lequel il avait prepare la transcription phonetique des entrees. II avait egalement prepare" un petit dictionnaire de novial, langue qu'il avait lui-meme construite ä partir de l'esperanto.

2. Problemes structured des dictionnaires

Dorthe Duncker et Hanne Runs suggerent un modele de Lexical Corpus Base (LCB) et decrivent une methodologie de construction de ce modele. Des exemples sont pris au LCB de textes danois du 16e siecle ; ce LCB permet 1'acces ä l'ensemble du corpus de textes tout en presentant le contenu des textes du corpus de maniere facilitant les etudes de vocabulaire et les travaux lexicographiques.

XIV

Rufus Gouws demande ä juste titre des ameliorations de la maniere dont les dictionnaires bilingues aident traditionnellement leurs usagers ä sölectionner un Equivalent' correct d'un terme dans la langue source. II souligne que la lexicographic bilingue devrait se servir d'un grand nombre des strategies de la lexicographic monolingue et discute les problemes pratiques et leur solution sur la base de la lexicographic anglais/afrikaans. Sandor Martsa elaborant sur la definition de J.F. Taylor (1995), discute les 'images' inspires du regne animal, c'est-ä-dire le fonds de connaissances, reelles ou supposees, qui guide 1'utilisation d'un mot determine par une personne parlant sa langue maternelle. Sont particulierement interessantes les 'images generiques populaires inspiroes du regne animal' qui permettent aux Anglais de 'se comporter comme des cochons', tandis que les Russes peuvent 'sauter comme un bouc' (c'est-ä-dire maladroitement) et que les Hongrois peuvent 'se vautrer dans la boue comme un pore' (c'est-ä-dire employer un langage obscene). Geart van der Meer traite de domaines qui, selon lui ont besoin d'ameliorations en ce qui concerne les dictionnaires actifs neerlandais-anglais et anglais-noerlandais. II propose 1'introduction d'un Systeme de reference rigoureux dans le cas de la derivation non morphologique du type city/town>adj : urban et son/daughter>adj -.filial. Geart van der Meer soumet ogalement quelques suggestions d'amelioration dans le cas des collocations. II suggere en outre en ce qui concerne la grammaire dans les dictionnaires de tirer plus de parti des exemples de phrases donnes. Stefan J. Schierholz decrit dans son exposo le Projet Erfurst qui traite de la valence des noms en neerlandais, anglais et portugais. Etant donne que les grammaires et les dictionnaires renferment trop peu d'exemples, une recherche fundamentale est nocessaire. Dix-huit propositions rogissantes en allemand seront otudioes et les problemes lies ä leur traduction seront abordes. II convient de noter qu'il existe seulement quelque 1000 propositions dans ce contexte. Les resultats du projet seront utiles non seulement pour les etudes lexicographiques et grammaticales mais aussi pour la traduction assisted par ordinateur.

3. Critique et typologie des dictionnaires Ulrich Busse examine le traitement des lexemes politiquement corrects dans quarre dictionnaires d'apprentissage de l'anglais contemporain. La pragmatique de divers domaines de vocabulaire est divisee en categories, par exemple les termes ethniques (Negro, coon, nigger), les qualificatifs de nationalite desobligeants (Kraut, Yank, Yid), l'usage de pronoms (he, she, he/she, they) la formation de mots (salesman, saleswoman, salesperson) ainsi que les termes s'appliquant a l'orientation sexuelle (queer) et les personnes ägees (oldie, oldtimer). Busse otablit clairement que cette categoric n'est resumoe dans une etude dans aucun des dictionnaires examines ; toutefois, les diverses entrees des dictionnaires font nettement apparaftre que les lexicographes sont entierement conscients des problemes. Depuis plusieurs annoes, Bernhard Diensberg se consacre ä la revision des otymologies de mots d'emprunt de l'ancien francais pour le nouveau Oxford English Dictionary. Dans son article, Diensberg traite avant tout des domaines suivants: deux systemes phonologiques en contact 1'un avec 1'autre, conclusions concernant la r£ception/inte"gration phonologique de mots anglo-fran9ais et de l'ancien fransais en moyen-anglais, 1'integration

XV

de voyelles anglo-fran?aises, l'integration de consonnes anglo-fran$aises, integration d'alternances morphologiques de l'ancien francais, et certains illogismes dans les dictionnaires etymologiques et historiques. Lars Holm a mend une etude exhaustive des tabous et de la violation des tabous dans les dictionnaires suedois du 16e siecle ä nos jours. a concentre son etude sur certains groupes specifiques de mots tabous qu'il qualifie de loo words, fuck words, (mots scatologiques, mots obscenes), etc. Ses resultats figurent sur une figure synoptique indiquant la maniere dont 25 mots cles ont e"te" traites dans 100 dictionnaires suedois. Hanne Lauridsen et Arne Zettersten presentent dans leur expose leur nouveau dictionnaire anglais-danois en deux volumes publie par la maison d'edition de Copenhague Politiken Forlag. Certains aspects particuliers de la lexicographie bilingue sont discutes dans le cadre de certains elements caractoristiques du dictionnaire tels que les corpus utilises, importance particuliere accordee ä la grammaire, les definitions interactives, les collocations, la variete" des equivalents danois et l'insistance sur la pragmatique. Hideki Watanabe continue sä serie d'examens minutieux de l'Oxford English Dictionary (2e ed.) en se penchant de maniere critique sur les citations provenant de la version CDROM du dictionnaire. Cette fois-ci, il se concentre sur les citations du dictionnaire prises ä des poemes en vieil-anglais, en particulier 5eoww#"puisque ce poeme presente un interet supplomentaire pour les participants etant donne que la plupart des scenes sont situees au Dänemark et dans les pays scandinaves voisins. Ses commentaires critiques concement les litres de poemes et leurs abreviations, les dates auxquelles ont etc ecrites les oeuvres, les problemes causes par les multiples ajouts, etc.

4. Semantique et lexicologie Andreas Groger, de Bayreuth, presente un resume de ses travaux sur les verbes mentaux du vieil-anglais, fondes sur le Corpus de Helsinki de textes anglais. Gröger analyse les champs lexicaux de 15 verbes dits mentaux, soit les verbes exprimant des activites et des ötats mentaux tels que penser, croire, se souvenir, supposer, etc. En tant que resultat de cette analyse, une synthese est prosentee sous forme d'apercu des diagrammes de champs et des diagrammes quantitatifs. Ljubima Jordanowa presente les problemes lexicographiques lies au langage du Renversement (l'effondrement du communisme) en Bulgarie. Apres avoir traito les periodes suivant le Renversement bulgare, Jordanowa introduit le corpus de vocabulaire du 'Deuxieme Renversement' (1996-1997) en cours l'elaboration ä PAcademie des Sciences de Sofia. Ce corpus sera ulterieurement accessible sur 1'intemet. Dans son expose, Makimi Kimura analyse le processus de naturalisation des mots d'emprunt japonais tels qu'ils sont representes dans cinq dictionnaires anglais, The Oxford English Dictionary (2e od.), Webster's Third, The Random House Dictionary, The New Shorter Oxford English Dictionary et The Oxford English Additions Series. L'auteur montre que les mots d'emprunt japonais passent par quatre phases de naturalisation dans le vocabulaire anglais : 1) Pronunciation, Vocabulaire, 2) Usage attributif, 3) Productivite et 4) Mutation somantique. Gunnar Persson ecrit sur les aspects cognitifs des noms de fongi dans plusieurs langues. II presente 12 categories telles que la forme, la couleur, la repartition, etc. entrant dans le

XVI

nom des fongi et fait une distinction entre la taxonomie scientifique et la taxonomie populaire ; selon la premiere, un 'toadstool' est un fongus et selon la derniere, un champignon. L'expose conclut en proposant sur un ton badin un nouveau champ linguistique appele 'linguistique mycologique comparee' : il faudrait ä cet egard poser la question de savoir s'il existe des differences de strategies de designation de diverses langues. II semblerait, par exemple, que 1'anglais insiste particulierement pour mettre en garde contre les effets nocifs de certains champignons. Maja Lindfors Viklund et Yvonne Cederholm presentent dans leur expose leurs travaux sur la terminologie liee aux stupefiants dans le cadre du projet AVENTINUS visant ä foumir aux organisations de police europeennes des outils linguistiques propres ä faciliter la cooperation par-dessus les barrieres linguistiques. Le point le plus interessant relatif ä la sous-langue decrite est qu'elle comprend une terminologie dans le sens traditionnel ainsi qu'un argot. II y a une forte influence de 1'anglais sur toutes les autres langues dans ce domaine et l'usage frequent de synonymes et de metaphores pose des questions interessantes dans une perspective multilingue.

Remerciements : Les oditeurs souhaitent remercier les auteurs des contributions d'avoir mis leurs manuscrits ä leur disposition et tous les participants, vieux amis et nouveaux venus, d'avoir assiste au symposium. Nous avons beneficie de l'aide fmanciere des departements d'anglais et d'allemand de la Faculte des Lettres et du Centre de Traductologie et de Lexicographic de l'Universite de Copenhague. Copenhague, mai 1999

Les editeurs

Vorwort

Das Neunte Internationale Symposion zur Lexikographie, dessen Akten hiermit veröffentlicht werden, fand in der Zeit vom 23.-25. April 1998 an der Universität Kopenhagen statt. Gefreut haben wir uns diesmal über die Anwesenheit von Teilnehmern aus den nordischen Ländern, Bulgarien, Deutschland, Großbritannien, Japan, Lettland, den Niederlanden, Russland, der Slowakei, Spanien, Ungarn und den USA. Es wurden fünf Plenarvorträge gehalten, und zwar von Gabriele Stein, Heidelberg, Ladislav Zgusta, Illinois, Ulrich Heid, Stuttgart, Peter Gillivar, Oxford, und J0rgen Erik Nielsen, Kopenhagen. Das Symposion wurde durch die „Otto Jespersen Memorial Lecture" eröffnet - seit 1994 ein fest integrierter Teil des Internationalen Symposions zur Lexikographie. Gehalten wurde sie diesmal von Gabriele Stein, Heidelberg, der ehemaligen Vorsitzenden von EURALEX. Die auf dem Symposion gehaltenen Vorträge werden in diesem Band unter vier verschiedenen Überschriften vorgestellt, wie das auch der folgenden Einteilung zu entnehmen ist: 1. Historische und theoretische Beiträge Gabriele Stein, Ladislav Zgusta, Andrejs Veisbergs, Elena Bärcena & Tim Read, Matthias Kammerer, J0rgen Erik Nielsen 2. Strukturale Probleme von Wörterbüchern Dorthe Duncker & Hanne Ruus, Rufus H. Gouws, Sändor Martsa, Geart van der Meer, Stefan J. Schierholz. 3. Wörterbuchkritik und Wörterbuchtypologie Ulrich Busse, Bernhard Diensberg, Lars Holm, Hanne Lauridsen & Arne Zettersten, Hideki Watanabe. 4. Semantik und Lexikologie Andreas Gröger, Ljubima Jordanowa, Makimi Kimura, Gunnar Persson, Maja Lindfors Viklund & Yvonne Cederholm.

l. Historische und theoretische Beiträge

Gabriele Stein stellt das 1530 erschienene Monumentalwerk von John Palsgrave „Lesclarcissement de la langue francoyse" vor und betont dabei die Ähnlichkeit zwischen Palsgrave und Otto Jespersen. Beide setzten sich beispielsweise für das gründliche Studium authentischer Texte in der Fremdsprache ein, haben jeweils eine Lautschrift entworfen und befassten sich mit didaktischen Fragestellungen. Ferner waren sowohl Palsgrave als auch Jespersen Lexikographen, Phonetiker und Grammatiker zugleich. Der Aufsatz schließt mit einer korpusbasierten Studie zum periphrastischen do bei Palsgrave. Im zweiten Plenarvortrag behandelt Ladislav Zgusta, Urbana, Illinois, „Einige Entwicklungen in der Lexikographie der Vergangenheit und der Gegenwart". In diesem umfassen-

XVIII den Beitrag werden folgende für die Lexikographie zentrale Themenbereiche abgehandelt: Historismus, Strukturalismus, dänischer Strukturalismus einschließlich Louis Hjelmslev, Valenz, Kollokationen, Frequenz, amerikanischer Strukturalismus einschließlich Bloomfield, Block, Fries und Harris, pädagogische Wörterbücher, Poststrukturalismus, Prototypen, Kompatibilität von Theorie und Praxis sowie Computerprogramme. Hinsichtlich dieses zuletzt genannten Themas stellt Zgusta abschließend fest, dass die Zeit jetzt dafür reif sei, „dass wir uns als Lexikographen mit den Computerlinguisten zusammentun, nicht nur durch die Übernahme der von ihnen entwickelten Programme, sondern auch durch die Teilnahme an der für die Herstellung dieser Programme erforderlichen Forschungsarbeit". Andrejs Veisbergs stellt die lettische zweisprachige Lexikographie vor, indem er die gegenwärtige englisch/lettische Wörterbucharbeit vor dem Hintergrund der lexikographischen Tradition Lettlands diskutiert, die er vor allem als puristisch und über den aktuellen Sprachgebrauch erhaben charakterisiert. Die Hauptaufgabe, so Veisbergs, bestehe heute darin, ein Wörterbuch herzustellen, das den aktuellen Sprachgebrauch anstelle puristischer Idealvorstellungen widerspiegelt. Trotz einiger Ansätze bleibt nach Ansicht des Autors in diesem Bereich noch viel Arbeit zu leisten. Elena Barcena und Tim Reed stellen den faszinierenden Prototyp eines kognitiv basierten multilingualen eindirektionalen lexikalischen Bezugssystem vor. Der Hauptgedanke besteht darin, den mentalen Wortschatz eines Muttersprachlers zu kopieren und diesen für jede Sprache im thesaurusähnlichen Format zugänglich zu machen. Durch Links zwischen den jeweiligen Sprachen kann sich der Benutzer etwa von einem spanischen Lemma mit Synonymen zu dem entsprechenden englischen Lemma bewegen. Matthias Kammerer setzt sich mit der These von Helmut Schnelle auseinander, nach der die Bedeutungsangaben in Wörterbüchern des COBUILD-Typs als ein axiomatisches System zu betrachten seien. Nach einer Dekonstruktion der Argumentation von Schnelle weist Kammerer zunächst einmal nach, dass die Prädikatenlogik erster Ordnung hinsichtlich der natürlichen Sprache zu besseren Einsichten führt als die von Schnelle zu Grunde gelegte Aussagenlogik. Danach wird aufgezeigt, dass sowohl die Korrektheitsthese als auch die Vollständigkeitsthese bei Schnelle versagen, und dass es sich bei den Bedeutungsangaben daher nicht um Axiome der natürlichen Sprache handeln kann. Es wird festgestellt, dass eine Logik nicht in sinnvoller Weise auf eine natürliche Sprache wie z.B. das Englische appliziert werden kann. Jörgen Erik Nielsen zeigt uns Otto Jespersen als Lexikographen. Obwohl die Lexikographie zugegebenermaßen nur am Rande seiner Interessenfelder lag, hat Otto Jespersen trotzdem wichtige Beiträge zur zeitgenössischen Lexikographie geleistet. So hat Jespersen einerseits für das führende dänische Lexikon „Salmonsen" Artikel verfasst, und für „Brynildsen", das größte englisch-dänische Wörterbuch, hat er für die phonetische Umschrift der Lemmata gesorgt. Auch hat er ein kleines Wörterbuch zu „Novial" entworfen, einer von ihm selbst auf der Basis von Esperanto geschaffenen Sprache.

2. Strukturale Probleme von Wörterbüchern

Dorthe Duncker und Hanne Runs schlagen das Modell einer lexikalischen Korpusbase (Lexical Corpus Base, LCB) vor und beschreiben die dabei zu Grunde zu legende Methodologie. Anhand von Beispielen aus dem 16. Jahrhundert wird aufgezeigt, inwieweit

XIX

der Zugang zum ganzen Textkorpus über die LCB erfolgen kann, und abschließend wird hervorgehoben, dass die Texte im Korpus in einer solchen Weise dargeboten sein müssen, dass der Benutzer sowohl bei lexikologischen Studien wie auch bei lexikographischer Arbeit davon profitieren kann. Rufus Gouws fordert mit Recht zur Verbesserung der zweisprachigen Wörterbücher im Hinblick auf deren Äquivalentendarbietung auf. Seines Erachtens müsste sich die zweisprachige Lexikographie viel mehr der gleichen Strategien bedienen wie die einsprachige Lexikographie. Diskutiert werden praktische Probleme und deren Lösung anhand der Lexikographie zum Sprachenpaar Englisch/Afrikaans. Unter Anwendung der Definition von J. F. Taylor (1995) diskutiert Sandor Martsa die sogenannten tierbezogenen „Frames" unter dem besonderen Aspekt des realen oder fiktiven Hintergrundwissens, das den Gebrauch eines bestimmten Idioms in der Muttersprache steuert. Besonders interessant sind dabei die sogenannten „volkstypologischen tierbezogenen Frames", durch welche z.B. die Engländer, wenn sie sich den Bauch vollschlagen, „make pigs of themselves", während die Russen „wie die Ziegen herumspringen" (d.h. sich schwerfällig bewegen), und die Ungarn „Schweinen" (d.h. Zoten reißen). Geart van der Meer behandelt einige seiner Ansicht nach verbesserungsbedürftige Problembereiche in Bezug auf die aktiven Wörterbücher zum Sprachenpaar Niederländisch und Englisch. Es wird vorgeschlagen, bei der im Englischen vorkommenden nicht-morphologischen Derivation des Typs city/town > Adj. urban und son/daughter > Adj. filial ein feinmaschiges System von Verweisen einzuführen. Auch im Bereich der Kollokationen werden Verbesserungsvorschläge unterbreitet, und was die Grammatik im Wörterbuch betrifft, wird vorgeschlagen, zunehmend auf den Gebrauch von Beispielsätzen zu achten. Stefan J. Schierholz beschreibt das Erfurter Projekt zu den präpositionsvalenten Substantiven im Deutschen, Englischen und Portugiesischen. Da sowohl Grammatiken als auch Wörterbücher zu wenige Beispiele enthalten, sind Grundlagenforschungen notwendig. Es werden achtzehn regierte Präpositionen des Deutschen untersucht, und die mit deren Übersetzung verbundenen Probleme werden angesprochen. Beachtenswert ist, dass es im Deutschen insgesamt nur etwa 1000 präpositionsvalente Substantive gibt. Die Ergebnisse des Projekts können letztendlich nicht nur für die Lexikographie und Grammatikographie, sondern auch etwa für computerunterstützte Übersetzungen nutzbar gemacht werden.

3. Wörterbuchkritik und Wörterbuchtypologie

Ulrich Busse untersucht die Behandlung politisch korrekter Lexeme in vier Lernerwörterbüchem des Gegenwartsenglischen. Thematisiert wird die Pragmatik verschiedener Wortschatzbereiche, u.a. ethnische Wörter (z.B. negro, coon, nigger), herabsetzende Nationalitätenbezeichnungen (z.B. Kraut, Yank, Yid), Pronomengebrauch (z.B. he, she, he or she, they), Wortbildungen (z.B. salesman, saleswoman, salesperson) sowie Bezeichungen sexueller Orientierungen (z.B. queer) und älterer Menschen (z.B. oldie, oldtimer). Es wird festgestellt, dass das Thema in keinem der untersuchten Wörterbücher etwa in einer Übersicht zusammengefasst wird; aus den einzelnen Wörterbuchartikeln geht aber hervor, dass sich die Lexikographen des Problems durchaus bewusst sind. Seit mehreren Jahren beschäftigt sich Bernhard Diensberg im Hinblick auf die Neuausgabe des „Oxford English Dictionary" mit der Revision der etymologischen Angaben

XX bei den altfranzösischen Lehnwörtern. Im vorliegenden Beitrag behandelt Diensberg vor allem folgende Bereiche: zwei phonologische Systeme im Kontakt (d.h. die phonologische Integration des anglofranzösischen und altfranzösischen Wortmaterials ins Mittelenglische), die Integration des anglofranzösischen Vokalismus und Konsonantismus, die Integration der altfranzösischen morphonologischen Alternanzen, lateinische Etymologien französischer Lexeme sowie Inkonsequenzen der etymologischen und historischen Wörterbücher. Lars Holm hat eine umfassende Übersicht über den tabuisierten bzw. tabubrechenden Wortschatz in schwedischen Wörterbüchern vom 16. Jahrhundert bis heute zusammengestellt. Dabei konzentriert er sich auf bestimmte Gruppen der Tabuwörter, u.a. skatologische und auf den Koitus bezogene Wörter. Seine Ergebnisse, die auf der Untersuchung von 25 Schlüsselwörtern in 100 schwedischen Wörterbüchern basieren, werden abschließend in Übersichtsform zusammengefasst. Hanne Lauridsen und Arne Zettersten stellen das von ihnen erarbeitete und vom Verlag Politiken (Kopenhagen) herausgegebene neue zweibändige Englisch-Dänische Wörterbuch vor. Allgemeine Probleme der zweisprachigen Lexikographie werden diskutiert und mit einigen für das vorzustellende Wörterbuch besonders charakteristischen Aspekten in Verbindung gebracht, z.B. benutzte Korpora, der besondere Nachdruck auf Grammatik, interaktive Bedeutungsangaben, Kollokationen, Umfang der dänischen Äquivalenzstrukturen und die gezielte Betonung der Pragmatik. Hideki Watanabe setzt seine Studien zur zweiten Ausgabe des Oxford English Dictionary fort, indem er die auf der CD-Rom-Version des Wörterbuches aufgefundenen Zitate untersucht. Diesmal konzentriert er sich auf die im Wörterbuch enthaltenen altenglischen Zitate, besonders aus dem „Beowulf, zumal darin die meisten Szenen in Dänemark und dessen Nachbarländern spielen. Seine kritischen Kommentare beziehen sich u.a. auf den Titel der Gedichte und deren Abkürzung, das Entstehungsdatum, die wegen der verschiedenen Fassungen entstandenen Probleme etc.

4. Semantik und Lexikologie

Andreas Gröger, Bayreuth, stellt eine Zusammenfassung seiner auf dem „Helsinkier Korpus altenglischer Texte" basierenden Arbeit zum mentalen Wortschatz des Altenglischen vor. Gröger analysiert das lexikalische Feld fünfzehn sogenannter mentaler Verben, d.h. Verben, die sich auf mentale Prozesse und Zustände beziehen, z.B. Verben des Denkens, Glaubens, Vermutens und Sich-Erinnerns. Als Ergebnis der Analyse wird eine Synthese in Form von Übersichtsdiagrammen und quantitativen Diagrammen vorgestellt. Ljubima Jordanowa stellt die mit der Sprache der Wende in Bulgarien verbundenen lexikographischen Probleme dar. Nach einer Periodisierung der bulgarischen Wendezeit erfolgt die Präsentation des an der Akadamie der Wissenschaften in Sofia zu erstellenden wortschatzbezogenen Korpus der sog. „Zweiten Wende" (1996-97). Dieses Korpus soll später im Internet zugänglich sein. Makimi Kimura analysiert in ihrem Beitrag den Entlehnungsprozess japanischer Lehnwörter, so wie dieser aus den folgenden fünf Wörterbüchern des Englischen ablesbar ist: The Oxford English Dictionary (2. Ausgabe), Webster's Third, The Random House Dictionary, The New Shorter Oxford Dictionary sowie The Oxford English Dictionary

XXI Additions Series. Die Verf. weist nach, dass japanische Wörter bei der Entlehnung ins Englische vier Phasen durchlaufen: 1) Aussprache, Wortschatz, 2) Attributiver Gebrauch, 3) Produktivität und 4) semantischer Wandel. Gunnar Persson schreibt über kognitive Aspekte bei der Bezeichnung von Pilzen in einer Reihe von Sprachen. Er schlägt zwölf Kategorien vor, u.a. Form, Farbe, Umgebung etc., die bei der Benennung von Pilzen jeweils eine Rolle gespielt haben. Dabei unterscheidet er zwischen gelehrten Klassifizierungen und Klassifizierungen, die auf dem Volkswissen beruhen. Abschließend wird spielerisch vorgeschlagen, ein neues linguistisches Gebiet einzuführen, und zwar die „vergleichende mykologische Linguistik". Dabei wäre z.B. die Frage zu klären, inwieweit sich verschiedene Sprachen bei der Namengebung von Pilzen unterschiedlicher Strategien bedienen. Z.B. würde hervorgehen, dass besonders das Englische bemüht ist, den Sprachbenutzer vor den schädlichen Wirkungen der Pilze zu schützen. Maja Lindfors Viklund und Yvonne Cederholm stellen ihre Arbeit zur drogenbezogenen Terminologie im Rahmen des AVENTINUS-Projektes vor, dessen Ziel darin besteht, den europäischen Polizeiorganisationen über die Ländergrenzen hinweg angemessene sprachliche Mittel zur Verfügung zu stellen. Zu den interessantesten Punkten bei dieser Subsprache gehört der Umstand, dass sich diese Sprache sowohl aus Terminologie im traditionellen Sinne wie auch aus Slang zusammensetzt. Das Englische übt in diesem Bereich einen starken Einfluss auf alle anderen Sprachen aus, und durch die häufige Verwendung von Synonymen und Metaphern erheben sich Fragen, deren Klärung auch in multilingualer Hinsicht von Interesse wären. Die Herausgeber danken den Autoren sehr herzlich für ihre Bereitschaft, die Manuskripte ihrer Beiträge für den Druck zur Verfügung zu stellen.Wir danken auch allen Anwesenden, alten und neuen Freunden, für ihre Teilnahme am Symposion. Als Organisatoren sind wir der Humanistischen Fakultät der Universität Kopenhagen, dem Englischen Institut, dem Institut für Germanische Philologie und dem Zentrum für Übersetzungswissenschaft und Lexikographie der Universität Kopenhagen sehr zu Dank verpflichtet. Danken möchten wir auch Mia Nielsen vom Englischen Seminar für ihre Mitarbeit bei der Fertigstellung der Druckfassung des Manuskripts. Kopenhagen im Mai 1999

Die Herausgeber

Gabriele Stein

The Otto Jespersen Memorial Lecture John Palsgrave as Precursor of Otto Jespersen

Five years ago, in the first of these memorial lectures, Randolph Quirk reminded us that, although Jespersen is rightly remembered as the greatest grammarian of English, he was drawn to vernacular languages far more generally and that in fact the first language to attract him was French. So it was with John Palsgrave, though he was not so much drawn to French as driven to it by that most irreversible of masters, King Henry the Eighth. And as you might expect with this monarch, one must chercher lafemme: the situation is almost the reverse of the problem encountered by an earlier Henry, in linguistic difficulty wooing a French princess - as dramatised in Shakespeare's Henry V, Act II, Scene II. In Palsgrave's case it was the French king for whom Henry VIII had not only cherche unefemme, but actually trouve lafemme: his sister. The long-cherished marriage plans between Mary Tudor and the prince of Castille had come to naught, and when, in early 1514, the French king's wife died, a marriage between the 52-year-old, gout-stricken Louis XII and the young English princess was agreed upon. Linguistically she had already to a certain degree been prepared for a continental union. She had been given lessons in French, and court records from 1513 show that the king had employed a certain John Palsgrave, a native of London and graduate of Paris, as French tutor to his sister. It was this challenging and distinguished task that prompted Palsgrave to subject the French language to a thorough analysis and to compare it to English, his royal pupil's mother tongue. He first produced a description of French pronunciation and an analysis of French grammar. When both were presented to Mary Tudor and her second husband, the duke of Suffolk, they were so deeply impressed by Palsgrave's achievement that they suggested he should add a comparative study of the lexicon and present the work to the king himself. Palsgrave obliged and after years of intense and meticulous study, he completed a monumental book which was published in 1530 as Lesclarcissement de la languefrancoyse. In more than 1000 pages, he had 'reduced a vernacular to rules', something which at that time was not believed possible, and what is more, he, an Englishman, had reduced the French language to rules, a fact which has caused French scholarship some unease. Lesclarcissement de la languefrancoyse consists of three parts or books: Book I deals with "the true sowndynge of the frenche tongue", Book II provides a basic French grammar, and Book III, the bulk of the work, is an English-French dictionary. The three books are closely interrelated. The dictionary is divided into word-class sections. We thus have a bilingual word list for nouns, for adjectives, for adverbs, etc., the verb list constituting the most comprehensive part of Lesclarcissement (some 400 pages). Each word list is preceded by chapters on the respective part-of-speech grammar in which the rudimentary treatment of the second book is developed into an impressive grammatical analysis. In the compilation and execution of his work Palsgrave revealed himself as a linguistic scholar of the first order. He studied the work that had been done before him, he, a non-native speaker of French, had to find a basis on which he could rest his description of 16th-century

2

Gabriele Stein

French, a basis which at the same time would impart on his findings the necessary authority. So some 400 years before grammarians like Jespersen, Poutsma and Kruisinga supported their linguistic descriptions of English by usage examples meticulously excerpted from reading English texts, John Palsgrave set himself a reading programme to complement his personal knowledge of French. He turned to the works of such authors whom he felt "to be most excellent in the frenche tonge'" (Book I, fo. xxiv). Among these we find Jean Lemaire de Beiges (1475-c. 1520), Jean Meschinot (c. 1420-91), Guillaume Alexis (d. 1486). Yet he did not stop there: he also read earlier great literary works which provide his linguistic descriptions with an historical underpinning: he was familiar with the works by Alain Chartier (c.13851433) and Jean Froissart (c,1333-c.l410), and he admiringly quotes from the Roman de la rose. But his reading did not only embrace French literature, he also studied English writers, above all the works by John Lydgate, then in high fashion, and Geoffrey Chaucer. Here are two examples to show how Palsgrave used the linguistic material gleaned from his detailed study of French and English texts. Both are taken from the verb lists: I Make blynde / le aueugle. prime cöiu. or ie aueuglis. secunde cöiu. so that in y olde romant tonge they vse this verbe of their fyrste coniugacion / but Johan le Mayre vseth hym of their seconde. Wene you to make me blynde with your wyles: Pencez vous de me aueuglyr de voz ruses. (Book III, fo. cc.lxxxxix = cc.lxxxix) I Queme I please or I satysfye / Chauser in his Cauterbury tales / this worde is nowe out of vse. (Book III, fo. ccc.xxxi)

The close study and analysis of authentic French texts gave Palsgrave the linguistic background and authority he needed, but this does not mean that the language actually described in Lesclarcissement is overwhelmingly literary. Rather, it is predominantly the common language in use at the time. Palsgrave had a keen interest in the teaching of languages and in the language learners. He was fully aware that they needed the contemporary idiom and he supplied it, occasionally with a humorous touch as becomes evident from the following examples: I Plucke I hale or pull at a thing Ie tire, prime cöiuga.^gie halle, prime cöiuga. Plucke syrs for shame be nat you syxe able to plucke a iade out of y* myer: Tyrez, or hallez gallans nauez vous poynt de honte que vo^ six ne pouues tirer vne charoigne hors de la fange. (Book III, fo. ccc.xix) I Set a house a fyre. / Ie atise vne mayson, or ie boute le feu dedens vne

The Otto Jespersen Memorial Lecture...

3

mayson. I can do some thyng for I can set a house a fyreBronne awaye by the lyght whan I haue done: le scay faire quelque chose, car ie scay bouter le feu en vne mayson et men fuyr par la clarto. (Book III, fo. ccc.lvi)

In many instances, learners are not simply presented with a choice and use of idiom, but there are also explicit metalinguistic comments on actual language usage: I Helpe I socoure / Ie ayde. prime coniuga. [ ] Helpe helpe / as men krye that be in daunger of theues or any other peryll: A layde alayde [...] (Book III, fo. cc.lxv)

As interjections of "callyng" Palsgrave mentions French hay, hau and hola. After examples like Vienca hay, hau pety lehan apportez man arc he explains: so that hay, is vsed whan they call one that is in their syght or nere them: hau, to one that is farther of or out of syght / also whan they call at ones doore standynge without / they saye Hola, and they within forthe answere: Qui est la. ... (Book III, fo. cccc.lxxiii)

In the concern to offer learners adequate idiomatic and thus interesting and challenging language material Palsgrave's and Jespersen's minds meet as we learn from the latter's book How to teach a foreign language (1967:18-19). There is another issue which preoccupied both our scholars: the relation between, and the influence of, the mother tongue and the foreign tongue in the language learning process. Jespersen counters the argument that learning a foreign language may or will result in improved knowledge of the mother tongue by pointing out that the prevailing teaching method, learning through translating, was apt to make pupils offend against good English usage (ibid. 45-46). Palsgrave discusses this relationship in his only other work The Comedy ofAcolastus. In order to make the learning of Latin more interesting and enjoyable for his pupils in The Hague, the Dutchman William Fullonius had written a comedy narrating the story of the prodigal son. The work was first published in 1529 and became an early Schoolbook bestseller in Northern Europe. Within a short time, reprints appeared in Paris, Cologne, Leipzig, Basel. A German translation came out in the mid-1530's. The great popularity of this parable in colloquial Latin and its teaching of Christian morals must have made it attractive to Palsgrave, a teacher of Latin and chaplain to Henry VIII at a time when one Latin grammar for the teaching at schools was to be made compulsory throughout the whole realm. Besides a grammar, adequate Latin texts for interpretation had to be prepared. This is the educational background of Palsgrave's translation of Fullonius' comedy. It appeared in 1540 and included a unique feature: it was a two-language text, giving the Latin as well as the English translation,

4

Gabriele Stein

printed next to each other. In the dedication to the king, Palsgrave explains why he had translated the work and provided a two-language version. He describes his translation as having been done "after suche maner as chylderne are taught in the grammar schole, fyrst worde for worde, as the latyne lyeth, and afterwarde accordynge to the sence and meanyng of the Latin sentences: by shewing what they do value and counteruayle in our tongue" (1540:title page). It was meant as a remedy against various teaching practices, so lively depicted by Palsgrave: the Latin instructor who uses Latin only and confounds the young learners' minds, or the one who is transported into the higher realms of text interpretation and never completes explaining more than a few lines in the classroom. Yet the idiomatic English version after the word by word translation was not only to ease the masters' explanation burden and the pupils' understanding, it was also aimed at university graduates. Palsgrave gives us a very detailed description of the linguistic situation experienced by the more educated youth of the kingdom: the medium of education was Latin, but the skills acquired by university students in their respective subject fields were not similarly developed for the mother tongue. Here is his characterization of the difficult and totally unsatisfactory linguistic situation: And sotnme other [teachers] furthermore there be, whiche thoughe they haue by their greatte studye, at youre graces Unyuersities, soo moche prouffyted in the Latyne tongue, that to shewe an euydente tryalle of theyr lernynge, they canne wryte an Epistle ryght latyne lyke, and therto speake latyne, as the tyme shall mynyster occasyon, very well... yet for all this, partely bycause of the rude language vsed in their natyue countreyes, where they were borne and firste lerned (as it happened) their grammer rules, ^g partely bycause that commyng streyght from thense, vnto some of your graces vniuersities, they haue not had occasions to be conuersaunte in suche places of your realme, as the pureste englysshe is spoken, they be not able to expresse theyr conceyte in theyr vulgar tonge, ne be not suffycyente, perfectly to open the diuersities of phrases betwene our tonge and the latyn (which in my poore iudgemente is the veray chiefe thynge that the schole mayster shulde trauayle in). In so moche that for want of this sufficient perfection in our owne tongue, I haue knowen dyuerse of theym, whiche haue styl continued theyr study in some of your graces vniuersities, that after a substanciall encrease of good lernynge, by theyr great and industrious study obteyned, yet whan they haue ben called to do any seruice in your graces cömen welthe, eyther to preach in open audience, or to haue other administration, requiringe theyr assiduous conuersantynge with your subiectes, they haue then ben forced to rede ouer our englyshe auctours, by that meanes to prouyde a remedy vnto their euident imperfection in that behalfe. (1540:Aiiiv-[Aiiii])

As this somewhat lengthy extract shows, Palsgrave's interest in speakers' linguistic competence was not restricted to the languages he taught, Latin and French, it fully embraced the mother tongue. So far I have drawn attention to similarities between John Palsgrave and Otto Jespersen as scholars and as dedicated language teachers: as non-native speakers of the vernacular they were investigating, they both opted for the same scholarly approach. They based their linguistic description on a close study of authentic texts. Both men showed a deep concern for the right choice of an appropriate style of language to be taught. And both reflected on the impact which the method of teaching the foreign tongue had or would have on the learners' mother tongue. From Jespersen's NovialLexike it emerges that both also had a strong interest in the lexicon, each producing a dictionary. Since I have discussed Palsgrave's achievement as a lexicographer in detail (Stein 1997), we shall turn to them as phoneticians and grammarians.

The Otto Jespersen Memorial Lecture...

5

Jespersen's expertise and excellence in these two linguistic disciplines has long been known. Palsgrave's striking achievements of more than four centuries earlier are only beginning to be more fully assessed. In the Introduction to Lesclarcissement Palsgrave, with impressive insights, highlights a number of linguistic features which he regards as characteristic of the French language, to which he alerts his learners and which he later describes more fully in Books I, II, and III. For the pronunciation of French these are the nasalized vowels, and phrase and sentence intonation. Having identified the phonetic position and surroundings of the three vowels a, e and he describes the formation of the nasalized vowels as follows: ... so that these thre letters M/N/ or fynall / nat hauyng the accent vpon hym/be the very and onely causes why these thre vowelles A/E/O/ be formed in the brest and souded by the nose. And for so moche as of necessyte / to forme the different sounde of those thre vowelles they must nedes at theyr fyrst formyng open theyr mowth more or lesse / yet whan the vowell ones formed in the brest / ascendeth vpwardes and must haue M/ or N/ sounded with hym / they bryng theyr chawes togetherwardes agayne / and in so doyng they seme to sounde an v / and make in maner of A/ and O/ dipthonges / whiche happeneth by rayson of closyng of theyr mowth agayne / to come to the places where M/ and N/ be formed / but chefely bycause no parte of the vowell at his expressyng shulde passe forth by the mowth ... (The Introduction·^)

As to intonation, Vivian Salmon in a recent paper on "John Hart and the beginnings of phonetics in sixteenth century England" (1994) has credited Palsgrave with precedence over Hart in identifying and describing it for a vernacular. Intonation is a matter of accent which Palsgrave describes as follows: Accent in the frenche tonge is a lyftinge vp of the voyce / vpon some wordes or syllables in a sentence / aboue the resydue of the other wordes or syllables in the same sentence / so that what soeuer worde or syllable as they come toguyder in any sentence / be sowned higher than the other wordes or syllables in the same sentence vpon them / is the accent... And note / that suche lyftinge vp of the voyce happeneth euer by reason of some vowell or diphthonge /*Z.neuer by reason of any consonant. (Book I: fo. xviii)

It is this "lifting up of the voice" that creates the impression, so Palsgrave, why theyr tong semeth to vs so brefe and sodayn and so harde to be vnderstäded whan it is spoken / especially of theyr paysantes or cornen people / for thoughe there come neuer so many wordes of one syllable together / they pronounce them nat distinctly a sonder as the latines do / but sounde them all vnder one voyce and tenour / and neuer rest nor pause vpon any of them / except the cömyng next vnto a poynt be the cause therof. Seconde / euery worde of many syllables hath his accent vpon the last syllable / but yet that nat withstandynge they vse vpon no suche worde to pause / except the cömyng next vnto a poynt be the causer therof / and this is one great thyng whiche inclineth the frenchemen so moche to pronounce the latin tong amysse / whiche cötrary neuer gyue theyr accent on the last syllable. (The Introduction:B.iiv)

Palsgrave is here describing the oxytone stress in French and in Book I he provides the learners with example sentences where the last stressed syllable of the sentence carries an accent mark. In order to help the self-studying learners to acquire a good pronunciation of French Palsgrave even attempted to transcribe text passages. At the end of Book I he applies his pronunciation rules in quoting prose and poetry samples and subscribing them, line by line, with a transcription system of his own. Here is a "transcribed" line:

6

Gabriele Stein A la tres haulte et excellente

maieste

des princes

Alatrehäutoeevzselläntomaiestedeprinsos As we can see, there are no pauses in the line which is meant to reflect his earlier explanation of the oxytone sentence intonation in French. The unusual spellings in o (hauto, evzsellanto, prinsos) are an attempt at rendering the pronunciation of a final unstressed e as in homme, centre, parle. Palsgrave explains that the schwa-sound is "soüded in a meane voyce betwene e and o" (Book I:D) and for its production the learner has to "lyft vp hys voyce vpon the syllable that commeth nexte before the same e / and sodaynly depresse his voyce / whan he cometh to the soundynge of hym and also sounde hym very moche in the noose" (Book I:A.ii). These observations on Palsgrave as a pioneer of French phonetics leave no doubt that he had an excellent command of the language and that he excelled in describing his linguistic insights for his learners. We come to Palsgrave the grammarian. Language areas where English and French differed seem to have had a magnetic attraction for him: they constituted a challenge in linguistic method and analysis for his highly original and independent mind. For English speakers the tense system of French is unusual: whereas English has one past form, e.g. he closes vs he closed, French has two: the imparfait, ilfermait, and the passe" simple, ilferma. What were the respective functions of these two past tenses in 16th-century French? What rules did Palsgrave give his readers when to use which form? As to English, the linguistic situation was more complicated and unsettled than present-day English usage: the use of periphrastic do in positive active sentences was very prolific. The origin and development of the Jo-periphrasis, just as the development of the expanded tenses, is one of the most frequently and most controversially discussed topics in English historical linguistics. Some language historians have quoted passages from Palsgrave's Lesclarcissement where he maintains that "it is all one to say" / speak or / do speak, I spoke or I did speak (Book III: fo. cc.xviv). But what most of these scholars have overlooked is that, when describing the various conjugation patterns, Palsgrave systematically pairs the form of the French imparfait (which he calls preter imparfit tense) with the Jo-form: e.g. ie parloye - I dyd speke

and the form of the passe" simple (called the indefinite tense by him) with the English simple past: e.g. ie parlay -1 spoke. By implication then, the two English forms would not express the same state of things, events. So what is the difference in use of the two French tenses? Palsgrave gives us a most impressive functional analysis which he then exemplifies, again with authentic text passages and explanatory comments, for his readers. He sees the functional difference between the use of the imparfait and the passe" simple as follows: Whan the frenche men write an hystory or make rehersall of any acte or mater that is passed / intendyng to declare the cyrcumstances or maners howe the same hystorye or acte was done / haue chefely their consyderacion vpon the tyme whiche was present whan the same actes were in doynge. And all suche partyculer dedes as aboute that tyme were begon / and incontynently aboute the same present tyme ended or ouer passed / all suche maner of dedes expresse they by their indyfinyte tence / and all suche dedes as at the same tyme were in doynge and had contynuaunce after the same present

The Otto Jespersen Memorial Lecture...

7

tyme / all suche actes expresse they by their preter imparfyte tence. So that their preter imparfyte tence serueth to expresse y chefe actes that they wyll speke of and their indiffynyte tence to declare y partyculer actes and cyrcustances whiche ouerpassed in^ meane whyle / as if I wolde shewe one y I was yesterday at yorke Awhat thynges chaunced me in f meane season y I was there / as I met there with a man whiche salued me At talked with me of many thynges / they say / le estoys hier a Yorke la ie recontray vng homme qui me salua et menparla deplusieurs chases. (Book III: fo. c.xxiiv) [my emphasis] In the extract, I have highlighted the major opposition: the "chefe actes" are expressed by the imparfait, the "partyculer actes" by the passe" simple. To this overall general rule Palsgrave has added an exception: the imparfait is used when the speaker wants to express "an hundred dedes togyther / whiche all were in doynge and contynued attones at any season whiche somtyme was present" (ibid. fo. c.xxiii) This remarkable functional interpretation of 16th-century French usage is a most valuable historical document in the vernacular. For English it leaves us with some problems. On the one hand, Palsgrave maintained that for the English present and past tense the addition of periphrastic do made no difference. From the examples and context provided it is clear that he is referring to affirmative statements only. On the other hand, he paired past tense forms only, implying by his explicit analysis of the French imparfait and the passe" simple an imperfective function for the periphrastic past in English. Is Palsgrave's analysis the result of an oversystematization prompted for the past tense only because of the French set of forms? Was his statement that the use of a periphrastic do did not make any difference an overall assessment of general current English usage which left enough linguistic freedom to the individual speaker? So individuals may and will have differed in their preferences or dislikes of do-forms. And some (the more enlightened ones) may have attributed to their use of periphrastic do certain meanings, so that Palsgrave himself might have been influenced by his knowledge of French. Alvar Ellegard in his classic study of do (1953) has rightly stressed that Palsgrave was exceptional as a 16th-century grammarian in discussing the function of periphrastic do. Yet he rather quickly dismissed Palsgrave's suggestion of a 'past imperfect tense' as its function because he did not find it sustained in other writing. Engbloom (1938) on the other hand, had drawn attention to later English grammarians (Gill 1621, Wallis 1653) who also mentioned an imperfective function. Since much more research is needed in this complex field of English grammar, I decided to investigate Palsgrave's actual use of do- forms in order to see how much agreement there was between 'theory' and practice. The findings may be a contribution to more recent corpus studies on periphrastic do. The corpus basis which suggested itself was the verb list. As mentioned earlier, Palsgrave illustrates his verb lemmata with example sentences which occasionally are taken from a source text, but on the whole are invented by himself, capturing the common and colloquial language of the time. Each English sentence has a French counterpart. I restricted the corpus to all instances where an English main verb (other than be and have) was used, either in the simple past or the periphrastic past, and rendered by a French imparfait or a French passe" simple. That is, when there was a lack of correspondence, e.g. an English past matched by a French passe" compose, the instance was ignored. Since the corpus is aimed at the use of periphrastic do alone, questions and negative sentences were not taken into account. Occurrences of the special OM/-construction (as in I did but admire her) were also excluded.

Gabriele Stein

The example sentences which illustrate the verb lemma in an English past tense form amount to 598 occurrences. An immediately striking result was the rather low number of doinstances. The distribution for the English and the French past forms is as follows: Imparfait

Passe simple

+ DO

13

42

55

-DO

127

416

543

140

458

Let us try to interpret these findings: 1. In 416 cases, that is more than two thirds of the total number of instances, the English simple past is matched by the French passe simple. This might justify Palsgrave's overall assessment. 2. The same point would hold if we took into account the total number of French passos simples: of 458 instances less than one tenth would be irregular because they correspond to an English periphrastic past. 3. The relationship between French imparfait forms and French passo simple forms is roughly 1:3, and this relationship is the same whether a do form is used or not. 4. The actual correspondence between a periphrastic past matched by a French imparfait has the lowest occurrence. 5. The findings which do not confirm the tendency described by Palsgrave are the 42 instances of periphrastic do, rendered by the French passe" simple, and the 127 simple past forms which are imparfait translations in French.

The corpus results do not conclusively sustain any claim for the distinction between the use or non-use of do in relation to the French imparfait and passo simple. But we should not forget that at Palsgrave's time all the Jo-constructions were in a process of diverging development and any attempt at a functional description in itself must be regarded as a highly original endeavour. Palsgrave's interpretation of the relation between the French past tenses and periphrastic do has some justification. But more refinement is needed. A closer study of the frequency with which verbs occur and of their meanings yields some interesting results: 1. The simple past of the verbs to say and to tell, for instance, is translated into a French imparfait as if the content of communication presented was still relevant and therefore a "chief act" in Palsgrave's sense. To speak and to talk, on the other hand, occur with a French passe" simple. 2. A very striking case are the do-less past tense forms of the verbs to see, to know and to hear rendered by a French passo simple: there are 81 occurrences for to see, 11 for to know, 9 for to hear. That is, nearly one quarter of the predicted correlation between the English simple past and the French passo simple is made up of these three verbs of perception.

Yet it is also the whole linguistic and extra-linguistic context that has to be taken into account. This becomes manifest from a closer study of the apparent exceptions to the general rule advanced by Palsgrave. Among the startling instances where a periphrastic do past is rendered by a French passe* simple we find quite a number of examples that relate to historic events, e.g.:

The Otto Jespersen Memorial Lecture...

9

I remeber well ynough whan y comens of Cornewall dyd ryse: le me souuient asses bien quant les communs de Cornovvaille se mutinerent. (Book III: fo. ccc.xlii) It was a goodly syght to se whan the kynges good grace dyd set his siege byfore tournaye: II faisoyt beau veoyr la bonne grace du roy quant il planta son siege deuant tournay. (Book III: fo. ccc.lviv) Since in general Palsgrave uses periphrastic do-forms rather sparingly, one may wonder whether for him one of the functions of this form is the expression of historical factuality. By the time Palsgrave presented his vastly expansive Lesclarcissement de la langue francoyse to King Henry VIII he may well have long forgotten that his original charge had been to teach the king's young sister enough French to sport with the gout-afflicted Louis XII, but he has left extraordinary testimony of the originality and linguistic insights that the king's original request had unleashed.

Bibliography

Anderson, J. M. - Jones, C. (edd.) (1974): Historical linguistics... vol. I: Syntax, morphology, internal and comparative reconstruction. Amsterdam Carlon, K. et al. (edd.) (1994): Perspectives on English. Studies in honour of Professor Emma Vorlat. Leuven - Paris Carver, P. L. (1937): The comedy of Acolastus. Translated from the Latin of Fullonius by John Palsgrave, with an introduction and notes. London Denison, D. (1985): "The origins of periphrastic DO: Ellegärd and Visser reconsidered", in Eaton, R. et al. (edd.): Papers from the 4th international conference on English historical linguistics, Amsterdam, 10-13 April 1985. Amsterdam, 45-60 Eaton, R. et al. (edd.) (1985): Papers from the 4th international conference on English historical linguistics, Amsterdam, 10-13 April 1985. Amsterdam Ellegärd, A. (1953): The auxiliary DO. The establishment and regulation of its use in English. Stockholm Engblom, V. (1938): On the origin and early development of the auxiliary DO. Lund Fasold, R. - Schiffrin, D. (edd.) (1994): Language change and variation. Amsterdam Hausmann, R. B. (1974): "The origin and development of modem English periphrastic DO", in Anderson, J. M. - Jones, C. (edd.): Historical linguistics ... vol. I: Syntax, morphology, internal and comparative reconstruction. Amsterdam, 159-189 Hudson, R. A. (1997): "The rise of auxiliary DO: Verb-non-raising or category-strengthening", Transactions of the Philological Society 95,1, 41-72 Ihalainen, O. (1976): "Periphrastic DO in affirmative sentences in the dialect of East Somerset", Neuphilologische Mitteilungen 77, 608-622 — (1981): "A note on eliciting data in dialectology: the case of periphrastic 'DO'", Neuphilologische Mitteilungen 82, 25-27 Jespersen, . (1930): Novial lexike. Paris - Heidelberg

10

Gabriele Stein

— (1909-1949): A modern English grammar on historical principles. 1 vols. Copenhagen — (1967): How to teach a foreign language. London (11904) Juul, A. - Nielsen, H. F. (edd.) (1989): Otto Jespersen: Facets of his life and -work. Amsterdam Philadelphia Kastovsky, D. (ed.) (1991): Historical English syntax. Berlin - New York Kroch, A. S. (1994): "Function and grammar in the history of English: Periphrastic DO", in Fasold, R. Schiffrin, D. (edd.): Language change and variation. Amsterdam, 133-172 Marchand, H. (1938-9): "Syntaktische Homonymie. Das umschreibende DO", Englische Studien 73, 227-252 Nevalainen, T. (1991): "Motivated archaism: the use of affirmative periphrastic DO in Early Modern English lithurgical prose", in Kastovsky, D. (ed.): Historical English syntax. Berlin - New York, 303320 Nevalainen, T. - Kahlas-Tarkka, L. (edd.) (1997): To explain the present. Studies in the changing English language in honour ofMatti Rissanen. Helsinki Nevalainen, T. - Raumolin-Brunberg, H. (edd.) (1996): Sociolinguistics and language history: Studies based on the corpus of early English correspondence. Amsterdam - Atlanta Nurmi, A. (1996): "Periphrastic DO and BE + ING: Interconnected developments?", in Nevalainen, T. Raumolin-Brunberg, H. (edd.): Sociolinguistics and language history: Studies based on the corpus of early English correspondence. Amsterdam - Atlanta, 151-165 Palsgrave, J. (1530): Lesclarcissement de la langue francoyse. London (Slatkine reprint, Geneva, 1972) Raumolin-Brunberg, H. - Nurmi, A. (1997): "Dummies on the move: Prop - ONE and affirmative DO in the 17th century", in Nevalainen, T. - Kahlas-Tarkka, L. (edd.): To explain the present. Studies in the changing English language in honour ofMatti Rissanen. Helsinki, 395-417 Rissanen, M. (1985): "Periphrastic DO in affirmative statements in early American English", Journal of English Linguistics 18, 163-183 — (1991): "Spoken language and the history of DO-periphrasis", in Kastovsky, D. (ed.): Historical English syntax. Berlin - New York, 321-342 Salmon, V. (1994): "John Hart and the beginnings of phonetics in sixteenth-century England", in Carlon, K. et al. (edd.): Perspectives on English. Studies in honour of Professor Emma Vorlat. Leuven-Paris, 1-20 Samuels, M. L. (1972): Linguistic evolution, with special reference to English. Cambridge Stein, D. (1985a): "Discourse markers in Early Modem English", in Eaton, R. et al. (edd.): Papers from the 4th international conference on English historical linguistics. Amsterdam, 283-302 — (l 985b): Natürlicher syntaktischer Sprachwandel. Untersuchungen zur Entstehung der englischen doPeriphrase in Fragen. München — (1985c): "Stylistic aspects of syntactic change", Folia Linguistica Historica 6, 153-178 — (1986): "Syntactic variation and change: The case of DO in questions in Early Modern English", Folia Linguistica Historica 7, 121-149 — (1990): The semantics of syntactic change. Aspects of the evolution of DO in English. Berlin - New York Stein, G. (1997): John Palsgrave as Renaissance linguist. Oxford Tieken-Boon van Ostade, I. (1985): "DO-support in the writings of Lady Mary Wortley Montagu: A change in progress", Folia Linguistica Historica 6, 127-151 — (1987): The auxiliary DO in eighteenth-century English, a sociolinguistic-linguistic approach. Dordrecht Traugott, E. Closs (1972): The history of English syntax. New York Visser, F. T. (1969): An historical syntax of the English language. Vol. HI, Leiden Wright, S. M. (1989): "Discourse, style, and the rise of periphrastic DO in English", Folia Linguistica Historical, 1-2,93-115 —(1991): "On the stylistic basis of syntactic change", in Kastovsky, D. (ed.): Historical English syntax. Berlin - New York, 469-491

Ladislav Zgusta[ Some Developments in Lexicography, Past and Present

1. Historicism Lexicography entered the twentieth century under the sway of historicism, which prevailed during the greater part of the nineteenth century. Some of the voluminous dictionaries which had been started in the nineteenth century were finished only in the twentieth; this was the case with, e.g., Grimm's Deutsches Wörterbuch (1854 et seqq.), Murray's New English Dictionary on Historical Principles, usually called the Oxford English Dictionary (1884 et seqq.), and a number of others. Historicism prevailed, however, not only in lexicography. For instance, even a scholar not primarily interested in historical linguistics such as J.O.H. Jespersen conceived of his monumental English grammar (1909) as based on the historical development of English. The dominant position of historicism was strengthened by the circumstance that where real historical knowledge was absent, either completely, because texts written in the older forms of the language in question were unknown or not yet interpreted, or partially, because some lexical units of the language had no known occurrence in old texts, there the lexicographers tried to present what the sense of logic suggested as the most probable development of the entryword's individual meanings. This method is used to this very day. No doubt John Locke's doctrine concerning the development of abstract senses from concrete ones added weight to this method; but it can be shown that the method itself, which we can call historico-logical, was used long before Locke, already by earlier lexicographers such as Henricus Stephanus in his Thesaurus Graecae Linguae (1572) and by others, and continues in use down to our times, particularly where a gap in textual attestations creates a silence that can be bridged over only by analytical reasoning and logical deliberations.

2. Structuralism One would expect structuralism to have exercised an influence on lexicography: after all, the general notion that a language as a whole forms a system whose every component is connected to the whole system, and in particular the Saussurean notion of valeur, that is, the idea that every component of language (e.g., the word) is delimited by its next adjacent neighbors would be immediately recognized as something exceedingly useful to the lexicographer. (So, for instance, in a group of synonyms, the meaning and stylistic value of

1

Dr. Dale Hartkemeyer, L.S.T., assisted in the preparation of this article, particularly in bibliographic searches.

12

Ladislav Zgusta

each is delimited by the rest of them.) In reality, partial studies of synonyms from this point of view came only with the lexicological research of Jost Trier (1931, 1973). One reason for this delay in the structuralist impact on lexicography is that compilation of important, original dictionaries is an undertaking that requires considerable time before its fruition. On the other hand, it is also true that de Saussure's Cours was published in 1916, when scholarly contacts were much hampered by World War I. There were some additional circumstances. It is well known that the linguistic school of Kazan developed ideas of a strongly structuralist character. For instance, one member of the Kazan school, Jean Baudouin de Courtenay, developed an understanding of what was later called by N.S. Trubetzkoy (1939) phonology, in such a way that he came very close to the notion of the phoneme in de Saussure's sense. In addition, this scholar was not just interested in lexicography, but indeed was a practicing lexicographer: he reworked the Russian dictionary (1903) of Dal1 (1801-1872). Again, he undertook this project on the basis of principles that can be considered structuralist, such as his striving to present a description of the whole Russian lexicon, with all the stylistic and social variants. This brought him onto a collision course with the Russian authorities, because Russian dictionaries did not admit any words considered substandard. When at the end of World War I Baudouin de Courtenay, being of Polish origin, returned to reconstituted Poland, his lexicographic efforts met with the same objections as in Russia, along with a few additional ones (e.g., his intention to involve Yiddish in his lexicography was not welcome). (On this see Farina, forthcoming.) De Courtenay's main trouble consisted, to put it bluntly, in his defying an important sector of the society of his time, which did not wish to have obscenities and vulgarities included in dictionaries, or foreign expressions and borrowings for that matter, even if they were frequently used in spoken language. This societal taboo has been broken and abandoned by now; indeed it can be expected that dictionaries that will contain most of the vocabulary present in the belles-lettres of the seventies and eighties will be more than saturated with highly descriptive sexual terms on all levels of style and emotionality, and with similar lexical material. However, a new societal taboo seems to be developing, namely the ban on ethnic and racial slurs. The parallel of the earlier sexual taboo seems to tell us that the reasonable thing to do is to include these expressions according to the frequency of their use, but to use suitable labels and comments to inform and warn the user of the dictionary. The problem of so-called politically correct language is broader; it will probably be reasonable to wait and see whether the mostly euphemistic expressions like intellectually challenged instead of 'retarded' or 'ungifted', visually challenged instead of 'blind', and young lady/woman instead of'girl' will take root in general language. Owing to the circumstances described above, Saussurean structuralism had some influence on lexicography only when the Prague School, mainly through the work of its then member R. Jakobson, succeeded in spreading a holistic or systemic view of language (unfavorable to a total ban on nonliterary words), accompanied by a weakening of the puristic attitudes. In Prague itself, these ideas were applied in concrete lexicographic projects only after World War II.

Some Developments in Lexicography, Past and Present

13

3. Danish structuralism

The second main branch of European structuralism met with similar difficulties: the English translation of the main work, Prolegomena (1961), of the founder of the school, Louis Hjelmslev, was published as late as the post-World War II era, so it was only with the works of Bernard Pettier (1978), Jean Dubois (1971), Alain Rey (1965), Josette Rey-Debove (1971) and Bernard Quemada (1968) that Hjelmslev's ideas took root and exercised a strong influence on lexicography in France. Since works concerning the theory of lexicography were considered to be of lexicological character, the first journal that systematically deals with lexicographic problems and allots most of its pages to them is called Cahiers de lexicologie.

4. Valences, collocations Lucien Tesniere was a French linguist who was close to the Hjelmslevian ideas. He was not in the field of lexicography, but his theory of syntax (1953, 1959) was relevant to it. As is well known, his is the notion of valences, the obligatory and optional syntactic accompaniments of the verb in a sentence. An example of obligatory valences involves the verb to put: *He puts the book. *He puts on the table. *He puts the book every day.

These sentences are ungrammatical. However, He puts the book on the table. is grammatical. Hence, the verb to put requires the direct object and the adverbiale loci as obligatory accompaniments. This type of study is, of course, the nucleus of the syntactic patterns as they are indicated in modern dictionaries, particularly those with pedagogical purposes, and of the contemporary study of what is called subcategorizations ~ not to mention that it serves as direct inspiration to the harvest of German Valenzwörterbücher, the most recent of which is Sommerfeldt & Schreiber (1996). The type of lexicography inspired by I. MePchuk (1974), (Mel'chuk et al. 1984), and (Mel'chuk & Zholkovskij 1984) is to date the greatest refinement of the theory and praxis of valences. J.R. Firth, a British linguist who was not attached to any contemporary school of thought, pursued a similar line of thinking. He called the syntactic relations obtaining among the individual components of a sentence colligations. While this idea and term had many competitors, his other term, collocation, became one of the main components of any theory of lexicography (1968).

14

Ladislav Zgusta

5. Frequency of occurrence If we cross the Atlantic to look for important lexicographic innovations, we probably have to stop first at the unabridged edition2 of Funk & Wagnall's New Standard Dictionary of the English Language (1928). The innovation introduced in this dictionary is that the sequence of senses in the entry is not of the historical and logical character mentioned above, but is based on the frequency of occurrences of the word in the respective senses. This is an important innovation, particularly for the bilingual dictionary, because the notion of the most frequently occurring sense of a word clearly overlaps with the notion of the dominant sense; in any case, it is these meanings the users of bilingual dictionaries are probably seeking most frequently. (Probably, we say, because there are no empirical quantitative data.) No wonder that Carla Marello's (1989) research has shown that during the 20th century, the proportion of dictionaries organized in this way has been on a constant increase. Still, there are some aspects of this organizational principle that need to be considered. For instance, it must be noticed that there are nouns and verbs in various languages whose forms do not much differ from one another; in English, many such pairs have an identical form. In many English dictionaries, particularly the smaller ones, such pairs are treated in conflated entries. (The COBUILD dictionary (1987) is a particularly appropriate example of this procedure.) However, the dominant senses of the two members of the pair are sometimes different; e.g., comfort (noun) has the dominant sense of 'well-being', whereas the verb refers rather to consolation.

6. American structuralism American structuralism did not have much interest in the lexicon, not to mention lexicography: like syntax, the lexicon is too unruly an area of language to be attractive to scholars who are mostly interested in regular patterns. This characterization pertains to the mainstream of American structuralism as represented by figures such as L. Bloomfield, B. Bloch, C.C. Fries, and Z. Harris. However, structuralism never was a unified body of thought, so it comes as no surprise that the disciples of K. Pike, mostly associated with the Summer Institute of Linguistics (an institution attending to the linguistic education of future missionaries and Bible translators), went their own way and constantly published dictionaries of Amerindian and other languages. This divergence notwithstanding, the attitude 2

The term unabridged in the titles of some American English dictionaries is frequently misconstrued as if it were the dictionary's aim to comprise all the words of the language. In reality, the term asserts that the edition thus described has not been produced by abridgement of a larger one. For instance, if we consider the Oxford English Dictionary, the Shorter Oxford English Dictionary, the Concise Oxford English Dictionary, and the Pocket Oxford English Dictionary, only the OED is 'unabridged' in this sense of the word, all the other editions having come into existence through abridgement, irrespective of inserted additions and supplements, sometimes quite extensive.

Some Developments in Lexicography, Past and Present

15

of mainstream American structuralism was as described in the first sentence of this paragraph. This attitude was, if not shattered, certainly modified by the conference that took place in I960, the papers from which were published by Householder & Saporta (1962). The most important upshot of the conference, it may be claimed, was the principle, firmly anchored in lexicography ever since, that every dictionary — and particularly those prepared for use (any use) by the nonnative speaker ~ must contain grammatical information. This conclusion was reached particularly because most American structuralists had some experience in the study of Amerindian and other similarly poorly known languages, cases in which one could not presuppose some minimal knowledge of grammar at the user's disposal.

7. Pedagogical dictionaries The same epoch, the late forties and fifties, brought the first principled restriction of dictionaries for pedagogical purposes: the important researcher in the area of the psychology of learning Edward L. Thorndike (1874-1949) undertook statistical studies of the growth of the vocabulary of children in different grades of school, which were the basis of various dictionaries for use at different ages in school, some of which were published with the cooperation of, or independently by, Clarence Barnhart. The best known product of this type of approach is the College Dictionary (Barnhart 1997). This dictionary, whose intended user level is indicated in the title, was published in several editions. One of the latest works that continue this tradition of dictionaries aimed at native speakers, and that contain a particularly great number of hard words and some encyclopedic information is Webster's ... New Collegiate Dictionary, a product of Merriam-Webster in Springfield, Massachusetts. The blank space in the title is to be filled in with the number of the edition; mine is the ninth (Mish 1983), but there is already a tenth edition, and certainly more will follow. Naturally, there existed at the same time dictionaries in other languages that displayed a similar focussing on a certain purpose. For instance, any edition of the Petit Larousse Illustre (1973, e.g.) shows a close typological similarity to the American collegiate dictionary. However, the distinctive difference marking the American dictionaries consisted in their being based, if in some cases only remotely, on empirical studies of the growing vocabulary of a student who is a native speaker of the language. While the main concern of this type of dictionary was the appropriate selection of vocabulary, the British initiative of A.S. Hornby turned attention to the nonnative learner of English and his need to obtain information about grammar, particularly about the entryword's syntactic patterns (colligations, as J.R. Firth would call them) and its collocations, as we have already mentioned above. His Advanced Learner's Dictionary (Hornby 1963) was constructed for this purpose, to offer this information. We are all aware of the success of Hornby's idea, gauged by the new editions of his dictionary, by the number of dictionaries pursuing the same purpose, and by the fact that the main collocation in its title has become the generic term for the new type of dictionary. With this development, a certain cycle is closed. It was one of the tenets of American structuralism that regularities, i.e. regular properties and regular behavior of words, belong

16

Ladislav Zgusta

to grammar, whereas the idiosyncrasies of lexical units belong to the dictionary. This was a remote outgrowth of Aristotelian syntax, which was based on, or at least intimately connected with, logic: the grammatical syntactic pattern was understood as corresponding to the structure of the logical proposition, and words were conceived of as fillers of the slots. But Apollonios Dyskolos already in the 2nd cent. A.D. knew that the distinction is blurred, that the selection of one syntactic pattern from among several competing for the same purpose can in itself be an idiosyncratic property of one of the crucial words in the sentence. This development is perhaps best observable in American structuralist dictionaries not belonging to the learner's dictionary type, such as Karttunen's (1983) dictionary of Nahuatl, Matisoffs (1988) dictionary of Lahu, Kari's (1990) dictionary of Ahtna, and Newell's (1993) dictionary of Batad Ifugao, to name only a few recent ones. They all contain copious grammatical information, mostly both in the front matter and within the entries.

8. Post-structuralism Post-structuralist American linguistics has been dominated by what is called transformational grammar, which later itself was transformed into quite a number of various schools of related but strongly differentiated thought. The cycle through which American structuralism passed seems to have been repeated once more: at first, research concentrated on syntax and phonology only, but already in the sixties there was a body of research called Generative Semantics (James McCawley being the principal figure in the development), whose main interest is expressed in the name. Today, any student of syntax will do research in what is usually called subcategorizations (Grimshaw 1992), which are something like valences, i.e. richly developed descriptions of the syntactic and collocational properties of individual words. Useful as this research is for our purposes qua lexicographers, it carries all the characteristics of post-structuralist linguistics, of which the one most awkward to the lexicographer is the absence of works exploring a broader area of language, not a few words only. The fact also seems to be ignored that many a feature of various subcategorizations could quite easily be found in any better, modern dictionary. And Thomas Kühn (1962) has made such an impression on many a scholar belonging to this trend that they usually fail to take cognizance of publications not originating within their group (which is usually understood pretty narrowly); hence the frequent talk about differing paradigms. However, the mere fact that this semantic interest has developed is highly welcome. Among the pertinent publications, broader areas of the lexicon are covered in a penetrating way in the works of Carl Pollard & Ivan Sag (1987, 1994) and Ray Jackendoff (1977, 1987, 1990, 1992).

Some Developments in Lexicography, Past and Present

17

9. Prototypes There are useful developments as well taking place outside the confines of linguistics itself, the most important of them being probably the prototype theory of E. Rosen (1977). I do not wish to reiterate what I have already said elsewhere (Zgusta 1994a and 1994b); it will be enough here to recollect that, for instance, the prototype of a species are those of its members that strike the speakers as embodying the most characteristic properties of the species. For instance, for any speaker of a language like English, the prototypical birds are, e.g., the robin, the swallow, the eagle, etc.; the turkey and the ostrich are not as prototypical. Some prototypical mammals are the cow or the cat, but not the whale or the manatee; nor, for other reasons, is homo sapiens. The reasons for such judgements are obvious; but it must be remembered that prototypicality may change with environmental circumstances and for similar linguistically extraneous reasons: e.g., for the speakers of English in South Africa, the ostrich is as prototypical a bird as an eagle. Further research added to the notion of a prototype the notion of a stereotype, which is a culturally based prototype. Further details can be found in the two articles I have mentioned above. An important addition to the theory of prototypes in relation to lexicography was made by Geeraerts (1985), who observed that historical dictionaries frequently operate on the basis of what one can call changes in the prototype represented by the entryword, whereas the general monolingual dictionaries can be seen as being based more on the cultural stereotypes present in the given language. To these two, Geeraerts adds yet a third notion, namely that of 'extensional meaning', which in the tradition of John Stuart Mill's logic means what we usually call the range of application — the whole class of objects that can be referred to by the particular word. Geeraerts is certainly right in maintaining that encyclopedic and terminological dictionaries deal mostly with this extensional meaning of words.

10. Compatibility of theories It will not have escaped the reader's notice that the notion of the prototype is quite similar to the notion of dominant meaning as normally used in lexicography. This similarity does not arise by chance. Swanepoel (1991) undertook a detailed comparison of the prototypical theory alongside the theory generally used in lexicography, as described in my Manual of Lexicography (Zgusta 1971). The result of the study is that the two approaches are not mutually exclusive. This is an important observation: two theories are not mutually exclusive. For instance, Wierzbicka (1983, 1992/93:69) raises an objection to the definition style developed in a dictionary project for Warlpiri, an Australian aboriginal language (Hale et al. 1983, Laughren & Nash 1983). Some of those definitions read as follows: lakarn-luwa-rni — xERG causes some outer part of yABS to be separated from y by striking with a missile liji-yirra-rni — xABS is desirous of having yDAT

18

Ladislav Zgusta paka-rni — xERG produces concussion of surface of yABS, by coming into contact with y (e.g., "The little boy tried to hit the dog with a stick.") larra-paka-rni — xERG strikes yABS thereby producing a linear separation in the material integrity of y, typically by means of a bladed instrument (e.g., "Both of them seized an axe. Each one split open the head of the other.")

Wierzbicka's objections pertain mainly to the wording of the definitions: what is printed in capitals are abbreviations for the obligatory cases, and the x's and y's stand, of course, for the nouns that accompany the verb. So far so good; however, the really difficult part of the definition is its wording. Presumably, one could use in the second definition the more frequent synonym 'wishes' for 'is desirous'; but 1 am not quite sure, because 'wishes' to my mind hints at a more focussed feeling than does 'is desirous'. However, if we think about 'striking with a missile' in the first definition, we cannot say that a less hard word would be better than 'missile': if a Warlpiri man can use, e.g., a stone, or a spear, or an arrow, or a boomerang for this type of striking, one would be hard put to find an easier hyperonym. It would seem that this type of highly logically oriented dictionary requires the use of hyperonyms, not a taxative or even only exemplary enumeration; and Rey-Debove (1966) has well shown that in any ascending chain of hyperonyms there comes a point at which it is impossible to find a concrete or well circumscribed expression for the next higher hyperonym and the following ones. Wierzbicka herself has developed a descriptive technique of her own; the two main features of her system are the absence of any hard words (indeed, she operates with a strictly limited inventory of words allowed in definitions) and the analysis of the word's meaning into components, by means of what she calls scenarios. For instance, the English verb 'to punish' is treated by this scenario: X punished Y [for Z] (a) [Y did Z] (b) X thought something like this: (c) Y did something bad [Z] (d) I want Y to feel something bad because of this (e) it will be good if Y feels something bad because of this (f) it will be good if I do something to Y because of this (g) X did something to Y because of this

Clearly, lexicography of this type is not for everybody, either. It cannot be denied, however, that whatever technique is decided upon, the lexicographer, particularly if dealing with a little-known language, has to analyze the meaning into its components in order to understand the differences existing between the word under description and its near-synonyms (or homeonyms), or between the word and its equivalent in the other language of the bilingual dictionary.

Some Developments in Lexicography, Past and Present

19

11. Theoreticians and practitioners All these developments show how welcome it is and how fruitful it will prove to be that the theory of lexicography is being expanded by a constantly increasing number of scholars; their differences are not necessarily detrimental — quite the contrary, since they render the repertoire of possible approaches richer. Naturally, the practicing lexicographer knows the various necessities and limitations of his profession and art better than anybody else. These are caused to some extent by commercial demands (or rather budgetary constraints, since even the resources of universities and similar institutions are not inexhaustible) that put a premium on the time spent. There are other necessary considerations; however, the main pressures are produced by the need to deal with a huge number of items in an identically formatted style, and by the uncertainties about what one can suppose that the user will need, understand, etc. Between the colleagues who compile dictionaries and those colleagues who are perceived as dealing with theory only, a certain tension arises from the differences in the work conditions between the two groups and in the final products of their work. A most illuminating case of a discussion of the mutual problems of the two camps can be found in a forum organized and edited by Frawley (ed. 1992/93). The theoreticians were represented by Wierzbicka, the practitioners chiefly by B.T.S. Atkins, Patrick Hanks, Sidney I. Landau, and James D. McCawley (all in Frawley ed. 1992/93); this is an insightful reading as regards both sides. Another theory of lexicography is being constructed by Wiegand in a multitude of his articles and papers (Wiegand 1989a,b,c, among many others). Wiegand has now undertaken the creation of a unified version of his theory in his Wörterbuchforschung (1998). This is the first volume of a work that will consist of at least five volumes. The present one contains deliberations on the structure of the whole area of lexicography and the research connected with it, and a particularly thorough section entitled 'Wörterbuchbenutzungsforschung', i.e. research into how dictionaries are used. (The German version of the title conveys a broader meaning than the English translation.) The Wiegand theory is a construction of great systematicity, in which every statement is based on a thorough discussion of the ontological and gnoseological background of its reference. The work will certainly have a great impact on the lexicographer's practice, because the breadth of its coverage and the exact and interlocking treatment of the individual topics will render possible the devising of specific algorithms (sit venia verbd) for the organization of work on dictionaries of different types.

12. Computer programs Finally, I would like yet to mention a relatively new development. We all know from experience what huge amounts of work can be taken care of by the computer, provided that a suitable program is prepared for handling the tasks and that the data to be fed in are neatly prepared. However, I have the impression that we are somewhat complacent in accepting

20

Ladislav Zgusta

the limitations that computer programs to date have been assumed to impose: a typical task for a computer is the mere ordering of huge amounts of data by some unambiguous formal criteria like the alphabet, sequence in a text, et sim. Yet in the last few years, computational linguistics has registered an intensified activity in the research into problems far more complicated than the ones mentioned above. If we take the volume Lexical Semantics and Knowledge Representation edited by J. Pustejovsky and S. Bergler (1991), we find there a plethora of highly interesting papers, such as Thomason (1991) on the representation of knowledge and its connection with the knowledge of words, Sowa (1991) on the logical structure of the lexicon, Martin (1991) on metaphor and its presence in and influence on the lexicon, and the paper by Martha Evens et al. (1991) on enriching the lexicon with many types of information. There are also in this volume articles by European researchers, such as Calzolari (1991), the excellent scholar from Pisa, Italy, who writes about the best way of first acquiring and then representing semantic information in a lexical knowledge base (notice that she does not talk merely about a semantic base); the Dutch scholars Meijs & Vossen (1991), who write about how knowledge is tied to lexical representations; and Germans Peter Gerstl (1991) writing about the determination of word meaning being effected through the interaction of lexical and nonlexical knowledge, and Stephen Helmreich (1991), who in his article "Interpretation without semantics" takes the position that the link between lexical items and world knowledge should be treated as a pragmatic rather than semantic relationship. This is only a small sample of the articles published in the volume. Another most interesting volume was edited by Nancy Ide and Jean Veronis as a special issue of the journal Computational Linguistics on word sense disambiguation (Ide & Veronis 1998), which, of course, is one of the main tasks for both the lexicographic excerptor and for the author of the respective entry. One can easily imagine what a help in dictionarymaking it would be to get the contexts presorted by the computer, and this with a greater finesse than is now possible. The editors introduce the volume with an excellent text surveying the state of the art (1998, Iff). The individual articles by Karov & Edelman, Chen & Chang, Schütze, Towell & Voorhees, and Leacock et al. approach the problem/task by different routes: by the formal similarity of the words, by topical clustering in techniques used in information retrieval, by corpus statistics, by formal relations, and by other methods. This collection is difficult but highly interesting reading. It is, however, not only such planned, purposefully edited volumes that contain articles of immediate interest to us linguists (without any attributive specification). For instance, what a help it would be if there were a program for the computer that would allow it to recognize collocations in contexts and presort those contexts accordingly. A random selection, not a focussed search, from recent literature in the field brings us the following examples: Alshawi & Carter (1994) deliberate on using for the task a combination taking into consideration the polysemy of both collocating words in the light of the statistical probabilities of the occurrence of the individual senses. Dagan & Itai (1994) also use statistical probabilities, but they compare monolingual corpora of two languages. Justeson & Katz (1995) prefer categorization to statistical probabilities: belonging to the same category is a more important argument in favor of the two words collocating than the statistical probabilities of their occurrence. McRoy (1992) experiments with disambiguation operated by what she calls a 'multiple knowledge source'. The basic idea is that if a program recognizes that a sentence such as The agreement reached between the state and the EPA provides for the safe storage of the waste belongs to a certain topic and register (i.e. administrative and le-

Some Developments in Lexicography, Past and Present

21

gal), it will select suitable meanings for the polysemous words ('agreement' = 'concurrence', 'reach' = 'achieve' not 'extend an arm', 'state' = 'government body' not 'form of existence'). All these are relatively simple tasks for us, for our human intelligence, but quite difficult for a program. By all this I wish simply to say that the time is ripe for our joining forces with the computational linguists, not only by accepting the programs resulting from their efforts, but also by participating in the research that is necessary for the construction of those programs.

13. Conclusion It would be possible to talk at much greater length about the past advances and future prospects of our profession, but this must suffice for the fifty minutes set aside for this lecture. Striving to be true to our habits and principles, I have already succeeded in talking far beyond my allotted time.

14. References Aitchison, Jean (1987): Words in the Mind: An Introduction to the Mental Lexicon. Oxford: Basil Blackwell. Alshawi, Hiyan/David Carter (1994): Training and scaling preference functions for disambiguation. Computational Linguistics 20, 1994, 635-648. Atkins, B.T.S. (1992/93): Theoretical lexicography and its relation to dictionary-making. Dictionaries 14, 1992/93,4-43. Austin, Peter ed. (1983): Australian Aboriginal Lexicography. Papers in Australian Linguistics 15; Pacific Linguistic Series A 66. Canberra: The Australian National University. Barnhart, Clarence (1997): The American College Dictionary. New York: Random House. [This is the latest edition of the dictionary. One of the oldest editions is The Thorndike-Barnhart Comprehensive Desk Dictionary. Garden City, NY: Doubleday 1951. There are many varying editions of the smaller dictionaries; to list them all would require a special bibliographic enquiry. In the title they have the two names 'Thorndike-Barnhart' and are distinguished as well by specifications such as 'junior dictionary', 'advanced junior dictionary', 'intermediate dictionary' in the title.] Bjarman, Peter/Victor Raskin eds. (1986): The Real-World Linguist: Linguistic Applications in the 1980s. Norwood, NJ: ABLEX Publishing Corp. Cahiers de lexicologie (Bernard Quemada, ed.). Vol. 1: 1959. Calzolari, Nicoletta (1991): Acquiring and representing semantic information in a lexical knowledge base. In: Pustejovsky & Bergler (eds. 1991), 166-197. Chen, Jen Nan/Jason S. Chang (1998): Topical clustering of MRD senses based on information retrieval techniques. In: Ide & Ve"ronis (eds. 1998), 61-95. COBUILD (1987): Collins COBUILD English Language Dictionary, editor-in-chief John Sinclair, managing editor Patrick Hanks. London/Glasgow: Collins. Cruse, D.A. (1992/93): On polylexy. Dictionaries 14, 1992/93, 88-96.

22

Ladislav Zgusta

Dagan, Ido/Alon Itai (1994): Word sense disambiguation using a second language monolingual corpus. Computational Linguistics 20, 1994, 563-596. Dal', Vladimir Ivanovich (1903): Tolkovyi slovar' zhivogo velikorusskago iazyka. 3rd edition, edited by Jean Baudouin de Courtenay. St. Petersburg. [The dictionary was later re-edited several times, both under Soviet rule and afterwards.] Dirven, Rene/Johan Vanparys eds. (1995): Current Approaches to the Lexicon: A Selection of Papers Presented at the 18th LAUD Symposium, Duisburg, March 1993. Frankfurt am Main: Peter Lang, Europäischer Verlag der Wissenschaften. Dubois, Jean (1971): Introduction a la lexicographic: Le dictionnaire. Paris: Larousse. Evens, Martha et al. (1991): For the lexicon that has everything. In: Pustejovsky and Bergler (eds. 1991), 179-187. Farina, Donna M.T.Cr. (forthcoming): Dai's Dictionary. In: Censorship: An International Encyclopedia. Derek Jones (ed.). London: Fitzroy Dearborn Publishers. (About Baudouin de Courtenay.) Firth, J.R. (1968): Selected Papers of J.R. Firth, 1952-59. F.R. Palmer (ed.). Bloomington & London: Indiana University Press. Frawley, William ed. (1992/93): Forum on the theory and practice of lexicography. Dictionaries 14, 1992/93, 1-159. Funk & Wagnalls New Standard Dictionary of the English Language (1928): New York/ London: Funk and Wagnall. Geeraerts, Dirk (1985): Les donnoes storeotypiques, prototypiques et encyclopodiques dans le dictionnaire. Cahiers de lexicologie 46, 1985. 27-43. Gerstl, Peter (1991): A model for the interaction of lexical and non-lexical knowledge in the determination of word meaning. In: Pustejovsky & Bergler (eds. 1991), 165-178. Grimm, Jacob (1854 et seqq.): Deutsches Wörterbuch. Vol. I, 1854 - XVI, 1954. Leipzig: S. Hirzel. [Jacob wrote the entries A - Frucht, with the exception of the letter D, which was written by his brother, Wilhelm Grimm.] Grimshaw, Jane (1992): Subcategorization and selection. In: William Bright (ed.), International Encyclopedia of Linguistics. Oxford: Oxford University Press, vol. IV, 90-92. Hale, Kenneth/Mary Laughren/David Nash (1983): Warlpiri Dictionary. Unpublished drafts, MIT. Hale, Kenneth et al. (1990): Warlpiri to English Vocabulary with Grammatical Sketch and English to Warlpiri Wordlist. Alice Springs, Northern Territory: Institute for Aboriginal Development. Hanks, Patrick (1992/93): Lexicography: Theory and practice. Dictionaries 14, 1992/93, 97-112. Hausmann, Franz Josef/Oskar Reichmann/Herbert Ernst Wiegand/Ladislav Zgusta eds. (1989 et seqq.): Wörterbücher/Dictionaries/Dictionnaires. Ein internationales Handbuch zur Lexikographie/An International Encyclopedia of Lexicography/Encyclopedic internationale de lexicographic. Vol. I-III. Berlin/New York: Walter de Gruyter. Heibig, Gerhard/Wolfgang Schenkel (1973): Wörterbuch zur Valenz und Distribution deutscher Verben. 2nd ed. Leipzig: Bibliographisches Institut. Helmreich, Stephen (1991): Interpretation without semantics. In: Pustejovsky & Bergler (eds. 1991), 34-37. Herbst, Thomas (1986): Defining with a controlled defining vocabulary in foreign learners' dictionaries. Lexicographica 2, 1986, 101-119. Hjelmslev, Louis (1961): Prolegomena to a Theory of Language (trans. F. Whitfield). Madison: University of Wisconsin Press. Hornby, A.S./E.V. Gatenby & H. Wakefield (1963): The Advanced Learner's Dictionary of Current English. 2nd ed. London: Oxford University Press.

Some Developments in Lexicography, Past and Present

23

Householder, Fred W./Sol Saporta eds. (1962): Problems in Lexicography: Report of the Conference on Lexicography held at Indiana University, November 11-12, 1960. Bloomington: Indiana University Research Center in Anthropology, Folklore, and Linguistics 21, 1962. [Simultaneously published as Part IV of the International Journal of American linguistics 28, 1962, No. 2. Several reprints.] Ide, Nancy/Jean Veronis eds. (1998): Special issue on word sense disambiguation. Computational Linguistics 24, 1998, 1-166. Word sense disambiguation: The state of the art (introduction by the two editors), 1-40. Jackendoff, Ray S. (1977): X Syntax: A Study in Phrase Structure. (Linguistic Inquiry Monographs, 2.) Cambridge, MA: MIT Press. — (1987): Consciousness and the Computational Mind. (Explorations in Cognitive Science, 3.) Cambridge, MA: MIT Press. — (1990): Semantic Structures. (Current Studies in Linguistics Series, 18.) Cambridge, MA: MIT Press. — (1992): Languages of the Mind: Essays on Mental Representation. Cambridge, MA: MIT. Jespersen, Jens Otto Harry (1909): A Modern English Grammar on Historical Principles. Heidelberg: C. Winter. Justeson, John S./Slava M. Katz (1995): Principled disambiguation: Discriminating adjective senses with modified nouns. Computational Linguistics 21, 1995, 1-28. Kari, James M. (1990): Ahtna Athabaskan Dictionary. Fairbanks: Alaska Native Language Center, University of Alaska. Karov, Yael/Shimon Edelman (1998): Similarity-based word sense disambiguation. In: Ide & Veronis (eds. 1998), 41-59. Karttunen, Frances E. (1983): An Analytical Dictionary of Nahuatl. Austin: University of Texas Press. Kühn, Thomas S. (1962): The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Landau, Sidney I. (1992/93): Wierzbicka's theory and the practice of lexicography. Dictionaries 14, 1992/93, 113-119. Larousse: Petit Larousse Illustre". Paris: Larousse. [This is the title of the newer editions, such as 1973, 1974, 1976, 1993, etc. Remarkably, the older editions of the dictionary, such as those of the years 1924, 1925, 1938, 1948, 1955, 1958 have the title Nouveau Petit Larousse Illustre.] Laughren, Mary/David Nash (1983): Warlpiri dictionary project: Aims, method, organization and problems of definition. In: Austin (ed. 1983), 109-133. Leacock, Claudia/Martin Chodorow/George A. Miller (1998): Using corpus statistics and WordNet relations for sense identification. In: Ide & Voronis (eds. 1998), 147-165. Lipka, Leonhard (1995): Lexicology and lexicography: Poor relations, competition, or cooperation? In: Dirven & Vanparys (eds. 1995), 381-411. Marello, Carla (1989): Dizionari bilingui: con schede sui dizionari italiani per francese, inglese, spagnolo, tedesco. Bologna: Zanichelli. Martin, James H. (1991): Conventional metaphor and the lexicon. In: Pustejovsky & Bergler (eds. 1991), 56-66. Mathiot, Madeleine (1967): The place of the dictionary in linguistic description. Language 43, 1967, 703ff. Matisoff, James A. (1988): The Dictionary of Lahu. Berkeley: University of California Press. McCawley, James D. (1986): What linguists might contribute to dictionary making if they could get their act together. In: Bjarman & Raskin (eds. 1986), 3-18. — (1992/93): Lexicographic virtue through selective and judicious sinning. Dictionaries 14, 1992/93, 120-129.

24

Ladislav Zgusta

McRoy, Susan W. (1992): Using multiple knowledge sources for word sense discrimination. Computational Linguistics 18, 1992, 1-30. Meijs, Willem/Piek Vossen (1991): Knowledge as a lexical phenomenon. In: Pustejovsky & Bergler (eds. 1991), 113-126. Mel'chuk, Igor (1974): Das Wort: zwischen Inhalt und Ausdruck. Johan Biedermann (ed.). München: Fink. — et al. (1984): Dictionnaire explicatif et combinatoire du fran?ais contemporain. Recherches lexicosomantiques 1. Montreal. Mel'chuk, Igor A./Aleksandr K. Zholkovskij (1984): Tolkovo-kombinatornyj slovar' sovremennogo russkogo jazyka: opyty semantiko-sintaksicheskogo opisanija russkoj leksiki. Wien: Institut für Slawistik der Universität Wien. Mish, Frederick C., editor-in-chief (1983): Webster's Ninth New Collegiate Dictionary. Springfield, MA: Merriam-Webster. Müllich, Harald (1990): "Die Definition ist blöd!" Herübersetzen mit dem einsprachigen Wörterbuch. Das französische und englische Lernerwörterbuch in der Hand der deutschen Schüler. (Lexicographica, Series maior, 37.) Tübingen: Max Niemeyer. Murray, James (1884 et seqq.): A New English Dictionary on Historical Principles. Oxford: Oxford University Press. [Abbreviated NED, but usually called Oxford English Dictionary, abbrev. OED. After an edition in single issues (started 1884), there was an edition in ten volumes between 1888 and 1928, and then an edition in twelve volumes in 1933.] Newell, Leonard E./Francis Bon'og Poligon (1993): Batad Ifugao Dictionary with Ethnographic Notes. Manila: Linguistic Society of the Philippines. Pollard, Carl/Ivan Sag (1987): Information-Based Syntax and Semantics. Stanford, CA: Center for the Study of Language and Information. — (1994): Head-Driven Phrase Structure Grammar. Chicago: University of Chicago Press. Pettier, Bernard (1978): Organisation semantique de l'article de dictionnaire. Bulletin de la Societe de Linguistique 73, No. 1, 339-366. Pustejovsky, James (1995): The Generative Lexicon. Cambridge, MA: MIT Press. Pustejovsky, James/Sabine Bergler eds. (1991): Lexical Semantics and Knowledge Representation: Proceedings of a workshop sponsored by the special interest group on the lexicon of the Association for Computational Linguistics, June 1991. Berkeley: University of California. Quemada, Bernard (1968): Les dictionnaires du fran9ais moderne 1539-1863: Etude sur leur histoire, leurs types et leurs methodes. Paris: Didier. Rey, Alain (1965): Les dictionnaires: forme et contenu. Cahiers de lexicologie 7, 1965, 65ff. Rey-Debove, Josette (1966): La definition lexicographique: Recherches sur equation semique. Cahiers de lexicologie 8, 1966, 71-94. — (1971): Etude linguistique et semiotique des dictionnaires fran9ais contemporains. (Approaches to semiotics, 13.) The Hague: Mouton. Rosch, Eleanor (1977): Human categorization. In: Studies in Cross-Cultural Psychology, Neil Warren (ed. 1977). London: Academic Press. Vol. I, 1-72. de Saussure, Ferdinand (1916): Cours de linguistique genorale. A. Bally & A. Sechehaye (eds.). Paris: Klincksieck. Schütze, Hinrich (1998): Automatic word sense discrimination. In: Ide & Voronis (eds. 1998), 97123. Seelbach, Dieter (1995): Syntagmatic context information for computer assisted (human) translation. In: Dirven & Vanparys (eds. 1995), 337-378. Sinclair, John (1991): Corpus, Concordance, Collocation. Oxford: Oxford University Press. Smadja, Frank/Kathleen R. McKeown & Vasileios Hatzivassiloglou (1996): Translating collocations for bilingual lexicons: A statistical approach. Computational Linguistics 22, 1996, 1-38.

Some Developments in Lexicography, Past and Present

25

Sommerfeldt, Karl-Ernst/Herbert Schreiber (1966): Wörterbuch der Valenz etymologisch verwandter Wörter. Tübingen: Niemeyer. Sowa, John F. (1991): Logical structures in the lexicon. In: Pustejovsky & Bergler (eds. 1991), 3855. Stephanus, Henricus (1572): Thesaurus Graecae Linguae. Paris. Swanepoel, P.H. (1991): Polisemie in die woordebook - 'n kognitiewe perspektief. Lexikos (AFRILEX-Reeks 1: 1991), P. Harteveld (ed). Stellenbosch: Büro van die WAT, 221-280. Tesniere, Lucien (1953): Esquissse d'une syntaxe structurale. Paris: Klincksieck. — (1959): Elements de syntaxe structurale. Paris: Klincksieck. Thomason, Richmond H. (1991): Knowledge representation and knowledge of words. In: Pustejovsky & Bergler (eds. 1991), 1-8. Towell, Geoffrey/Ellen M. Voorhees (1998): Disambiguating highly ambiguous words. In: Ide & Veronis(eds. 1998), 125-145. Trier, Jost (1931): Der deutsche Wortschatz im Sinnbezirk des Verstandes. Heidelberg: Winter. — (1973): Aufsätze und Vorträge zur Wortfeldtheorie. Anthony van der Lee & Oskar Reichmann (eds.). The Hague: Mouton. Trubetzkoy, Nikolaj S. (1939): Grundzüge der Phonologie. Prague: Jednota mathematiku. de Vaugelas, Claude (1647): Remarques sur la langue francoise, utiles ä ceux qui veulent bien parier et bien escrire. Paris. Wahrig, Gerhard (1967): Neue Wege in der Wörterbucharbeit. Hamburg: Verlag fur Buchmarktforschung. Wiegand, Herbert Ernst (1989a): Aspekte der Makrostruktur im allgemeinen einsprachigen Wörterbuch: alphabetische Anordnungen und ihre Probleme. In: Hausmann et al. (eds. 1989), 371-409. — (1989b): Der Begriff der MikroStruktur: Geschichte, Probleme, Perspektiven. In: Hausmann et al. (eds. 1989), 409-462. — (1989c): Formen von MikroStrukturen im allgemeinen einsprachigen Wörterbuch. In: Hausmann etal. (eds. 1989), 462-501. — (1989d): Die lexikographische Definition im allgemeinen einsprachigen Wörterbuch. In: Hausmann et al. (eds. 1989), 530-588. — (1998): Wörterbuchforschung. Untersuchungen zur Wörterbuchbenutzung, zur Theorie: Geschichte, Kritik und Automatisierung der Lexikographie. Vol. I. Berlin/New York: Walter de Gruyter. (xx + 1162 pages.) Wierzbicka, Anna (1983): Semantics and lexicography: Some comments on the Warlpiri dictionary project. In: Austin (ed. 1983), 135-144. — (1985): Lexicography and Conceptual Analysis. Ann Arbor, MI: Karoma. — (1992/93): What are the uses of theoretical lexicography? Dictionaries 14, 1992/93, 44-78. Zgusta, Ladislav (1971): Manual of Lexicography. The Hague/Paris: Mouton. — (1975): Linguistics and bilingual dictionaries. Studies in Language Learning 1, 1975, 95-109. — (1986): Problems of the bilingual dictionary (introduction). Lexicographica 2, 1986, 1-7. — (1988): Pragmatics, lexicography and dictionaries of English. World Englishes 7, 1988, 243-253. — (1992/93): Lexicography, its theory, and linguistics. In: Wm. Frawley (ed. 1992/93), Forum on the theory and practice of lexicography. (Dictionaries 14, 1992/93), 130-138. — (1994a): Prototipos y lexicografia. Vozy letra 5:1, 1994, 3-13. — (1994b): Lexicography for the twenty-first century (paper of guest lecturer). In: Papers from the First Asia International Lexicography Conference, Manila, Philippines, 1992. Bonifacio Sibayan & Leonard Newell (eds.). Manila: Linguistic Society of the Philippines, 3-15.

Andrejs Veisbergs

Latvian Bilingual Lexicography - Tradition and Change

For half a century the most common word of greeting as well as parting in Latvian has been cau. It took fifty years for this loan to appear in a Latvian dictionary. This paper by dwelling on the tradition and change in Latvian lexicography will attempt to explain occurrences like the above mentioned, as well as look at their causes: on how the impact of big languages affects the coping strategies of language survival, and how these to some extent promote linguistic isolationism even at the end of the 20th century. Latvian can serve as a model or testground for a situation when a small language suppressed and often doomed for extinction tries to gather strength and enter the modern world. And this took/takes place in time of continual change and absence of funding for language proliferation and studies. Generally it can be said there are no unique things about Latvian bilingual lexicography - like presence of a specific type of dictionary that does not exist in the rest of the world or a strange approach to the treatment of entries. The peculiarities of Latvian lexicography are mainly historically and socially conditioned and they can be summed up as follows - first, it is almost exclusively bilingual, second, to a large extent it is divorced from the real language and third, it is purist dominated. To understand this, a brief historical survey might be helpful. It is usual to date Latvian lexicography from 1638 when the first dictionary was published (Mancelius, 1638) containing about 6000 words. Latvians at that time were a peasant nation and the official cultural sphere was fully in the hands of non-Latvian governors, German clergy and landowners. It was to develop the link between the church and the peasant nation, between the German speaking clergy and Latvian-speaking people that the first dictionaries were actually created. They were used for teaching the clergy more or less decent Latvian that the peasants could understand, as well as to do better translations of religious literature. The Bible was translated in 1689, also testifying to the trend, and very much connected with the ideas of the Reformation. The other two dictionaries of the 17th century were multilingual - Polish-Latin-Latvian (Elger, 1683) and German-Latin-PolishLatvian (Dressel, 1688). It is difficult to pass judgment on these first lexicographic attempts. There is a clear influence of German, both on the lexical and grammatical levels. Trying to figure out the complexities of Latvian patterns was certainly not an easy task for the German speakers and the first compilers seemed to have done as much as one could reasonably expect. The following two centuries saw the same - the dictionaries were made by non-Latvians, they gradually improved in scope and depth (Lange 1777, Stender 1789) yet often retained also the mistaken stock of the previous ones. One can trace numerous German elements in Latvian grammar, collocation patterns and phrases, not characteristic of Latvian and that apart from the undeniable German influence that really existed in the language and has been retained up to today. Moreover, for 200 hundred years the dictionaries were made by the German speakers as well as aimed at the German speakers, so they were monodirectional. Numerous notions,

28

Andre/s Veisbergs

relevant for the clergy, while not existing in Latvian, were introduced by way of translation loans. Some of them look strange today, yet many were assimilated and have become part and parcel of Modem Latvian while bearing the traditional German structures. Nevertheless gradually this situation led to two variants of Latvian. The peasant people were speaking one language at home and another in communication with the non-Latvian governors in the official spheres - the court, the church, the administration, the manor. Only the second variant of the language had its written variant. Thus two parallel languages or two variants of one language coexisted - the so called Old Written Latvian and the spoken language. The dictionaries reflected mainly the first. The situation began to change in the middle of the 19th century when the so called Latvian national awakening started. The dichotomy of the language was noticed. The dictionaries of the second half of the 19th century were produced by Latvian speakers (Valdemärs 1872) and accordingly tended to reflect the language spoken by Latvians more. However, the struggle waged against German and Russian dominance and its influence in the language also transferred purism in the language to the making of dictionaries - a feature which is still prevalent to the present day. Despite linguistic and social pressures, towards the end of the 19th century the national literary language was formed. However, the gulf between spoken and written language continued to exist. Only by this time it was not the semi-German versus Latvian variant but real, colloquial language versus the official and purified variant. Thus started the historical pattern for Latvian lexicography - its explicitly bilingual (multilingual) tradition which lasted for more than 300 years (of the 10 published dictionaries in the 17-19th centuries all were bilingual or multilingual). Also, the term "dictionary" for an average Latvian would be associated with a bilingual dictionary only, encyclopedias coming second. This is typical of small nations (Czechs, Latvians, Estonians, Lithuanians, Norwegians) where the main purpose of a dictionary is seen in helping sustain contact with other cultures. And functionalism also determined that the dictionaries with the main contact languages were first to be compiled and stay the largest, e.g. the situation in the period between 1900- 1966 German - Latvian Latvian - German Russian - Latvian Latvian - Russian English - Latvian Latvian - English Latvian - French Latvian - Esperanto French - Latvian Esperanto - Latvian Greek - Latvian Latin - Latvian Lithuanian - Latvian Portuguese - Latvian

12 8 11 7 14 3

1 1

(17 editions) (9) (3) (2) (6) (5)

Latvian Bilingual Lexicography - Tradition and Change

29

A decent Latvian-Estonian dictionary had to wait until 1967, despite geographical proximity. A Latvian-Swedish dictionary in Latvia appeared only in the 90ies, a Latvian-Danish dictionary in 1995 (Lomholts 1995), as part of the Nordic language series. Even the first big comprehensive general explanatory Latvian dictionary (Miilenbach) in six volumes is in fact a bilingual book with definitions in German and examples in Latvian. It is interesting to point out that the title in Latvian says "Dictionary of the Latvian Language" while the title in German is "Lettisch-deutsches Wörterbuch". Does it not signify something about the importance of a bilingual dictionary as well as the target reader? As the dictionary was produced during the independence period there could be no foreign political pressure, more likely the opposite. It seems the authors had taken into account the traditional Latvian approach and were actually killing three birds with one stone. They compiled the most comprehensive stock of Latvian for the time, they used German for explanations (so foreigners could use it as a bilingual dictionary) and they put Latvian in the framework of comparative linguistics internationally. In a way it was an attempt to reach out to a broader audience. The other tradition (more of an intralinguistic character) is that of purifying, improving and normativizing which started really only in mid-19th century - and can be characterized as an Icelandic attitude to the language. Paradoxically the previous German-compiled dictionaries were in a way more descriptive than the prescriptive Latvian ones. This tradition affects not only bilingual but also, and much more so, explanatory dictionaries. For example the above-mentioned Miilenbach's dictionary is clearly prescriptive (suggesting that it has been aimed at Latvian users). The purpose of this defensive stand is clear, it reflects the traditional Weltanschauung of the Latvians - even in prewar independent Latvia the linguistic pressure from the two major contact languages - Russian and German (having also considerable minorities) was clear and strong and finally found its legislative formulation of a rather authoritarian decision of passing over to English as the first foreign language in the thirties. Returning to our permanent question - did it reflect the real language of the time? - No! It reflected the written and spoken language of the end of the 19th century, carefully weeded of undesirable elements, internationalisms, later borrowings in a way creating a monument to the correct, unadulterated Latvian. As such it set the pattern for other dictionaries as well - spelling dictionaries (though spelling in Latvian is so close to pronunciation that there might seem to be little sense in them) and of course dictionaries of Foreign words where Latvian with its transcription principle offers a great playfield for regular linguistic intervention, innovation and restructuring. The following bilingual dictionaries, though adding the terminological component, carefully avoided substandard layers of lexis - loans, colloquial, slang, taboo and vulgar words. For example, the above mentioned Latvian colloquial form of greeting since SOies "cau" (in fact a loan from Italian). The first dictionary where this word appeared was the Latvian-Swedish dictionary (Krastina, 1996) and then the Latvian-English dictionary (Veisbergs, 1997). Returning to the dichotomy of the two languages, it cannot be said that the dictionaries have been divorced from life entirely - in a way they reflect what numerous Latvians even today see as two parallel languages existing side by side - the right and the wrong one. And it is not Latvians only.

30

Andrejs Veisbergs

The director of the Goethe Institute in Latvia, Friedrich Winterscheid, links the situation with the gender issue so popular today: If we picture the languages as masculine or feminine, Latvian seems to me to be feminine. When women talk Latvian the words wave and dance... But when I hear men speak Latvian, I often get the impression that they communicate in a foreign language in which there is a blend of all possible and easily recognizable foreign words from German, English and Russian (Tr. by A.V.) (Vinterseids, 1995: 174-175).

Thus even a foreigner is impressed by there being two Latvian languages - the feminine and the masculine, or the old, pure, cultural one and the new - modern, international business language (ascribed to men). Today there is a nostalgia for the prewar language, for highly literary language and speech which is fast disappearing, especially in mass media which most people still view as the official sphere (60 years of government control cannot escape without trace). And the postmodern mobility with its mix of styles, tearing off the barriers cannot but heighten the awareness of the sudden change. Taking into account that dictionaries in Latvian history have always been very much the symbol of the official and approved on the one hand, and the desired and correct on the other, they are, as it were, caught between the purist tendency and the objectively changing language. The remarkable change of the 90ies called forth an amazing change in language and to some extent also in its lexicographical reflection. Two profound and farreaching processes came underway: • democratization of the language (written, spoken) with all its positive and negative effects. Abolishing censorship and passing over to a market economy in Latvia meant abolishing linguistic censorship and to a great extent editing as well. So the written word has come to reflect the real / spoken language. The media are full of new words, old "bad" words, the language has become much more open to change, variation, wordplay and new patterns. This has changed also the reflection of the language in dictionaries - it is reflected in the last volume of the Dictionary of Literary Latvian (Latvieau, 1987) covering letters U,V,Z, so that the last volume looks strikingly different from the other seven. • the English influence. The greatest change since the singing revolution has been an increased importance of English. Though a main foreign language at school, its teaching under the Soviets reminded one of Latin tuition. Now it has become an active direct contact language as well as intermediate contact language. It has also substituted Russian as a model language for wordformation (Veisbergs 1997) and terminology. In fact it is the third big change in this century - German, Russian, now English.

Because of the sudden and fast overhaul of the language some linguists (Üdris, 1997) suggest that lexicographical work should stop for a time until the language stabilizes, yet this could be viewed as a sign of confusion or even defeat in the fast changing conditions, as well as a reminder of the old times when lexicography could be centrally planned and managed. Since the collapse of the old system long-term dictionary projects have suffered a typical setback. The new dictionaries are now worked out a rather quickly by freelance lexicographers or experts of some sphere, financed by various companies and as a result creating fragmentary and an unnecessary variety of terminology. For the general scene it should

Latvian Bilingual Lexicography - Tradition and Change

31

also be stated that all the big dictionaries are monodirectional - targeted at the Latvian audience. I will limit myself to the big Latvian and English dictionaries as these are the ones that have been revised fully, while German dictionaries have just been reprinted. Generally, there is no sudden change in dictionary treatment of material, or make-up of the entries. British variants still dominate, American being marked with the label american or referred to the British entry. The most typical changes in bilingual dictionaries are the following: • Freer reflection of lower layers of vocabulary - colloquial, slang, taboo: the English - Latvian dictionary (1995) was the first to introduce in the English part substandard register lexis which in the soviet system was taboo. So most of the four letter words in English can be found but they are prudishly supplied with neutral or even scientific Latin counterparts in Latvian, e.g. prick si. penis (penis) cock si. penis (penis) dick si. dzimumloceklis (sex organ) sod si. niekkalbis (windbag) crap vulg. mesli (dung, rubbish) shit vulg. mesli südi, (dung, shit) fart vulg. pirdiens (fart) arse vulg. pakaja, dibens (back, bottom) balls vulg. pauti (testicles) turd si. ekskrements (excrement) bugger si. homifcis (homosexual) vulg. tips (type, individual) • Borrowings are less ostracized where Latvian wording sounds ridiculous or where foreign loans have become accepted - Yuppy -japipo, file -fails, cheeseburger - clzburgers. Compilers also had to deal with a new problem. Because of the transorientation of Latvian towards English as the main contact language instead of Russian the category of false friends has become very volatile. Many stable English-Latvian false friends (Veisbergs, 1994) are now fast becoming what I could call "real friends" or true equivalents - capacity, ambulance, decade, to arrest, romance. One can look on this as a massive interference that has to be countered, at least by dictionaries. In the absence of quantitative data about the process, lexicographers try to avoid codifying the growing number of false friends loans thus escaping semantic havoc. However this can also be considered a prescriptive approach as the dictionaries contain out-of-date language material. • The idea of revised editions published yearly or biennially seems to have struck root at least for the English language. Thus the English-Latvian dictionary (1995) was published 30 years after the previous one but has had the third edition in 1998). The Latvian-English dictionary (1997) is undergoing revision to be published this year. It had a similar prehistory - the previous dictionary published (Turkina, 1982) was a slightly updated edition of 1963 dictionary (Turkina, 1963). The Latvian - English dictionary created a set of different problems. The main problem for lexicographers here was not the search for equivalents, division of senses in entries or neologisms. The problem lies in there being no reliable basis on which to build the Latvian part as there is no corpus of Latvian. Because of the general problem and especially be-

32

Andrejs Veisbergs

cause of the fast change, there is in fact no reliable source for determining what the real Latvian lexis today is like. Colloquial Latvian has not been registered nor thoroughly described. The terminology sphere, after 50 years of Russian dominance is limited, glossaries compiled during the soviet times are Russian influenced and anyway outdated. As a result English - Latvian, German - Latvian, French - Latvian dictionaries often bear traces of the foreign language and can be actually recognized as derived from them. The new term glossaries are in-house Publishings that differ from one ministry, bank or company to another. A multitude of terms is being coined - militaries, banking, maritime affairs, the new spheres: computers, EU. So there is an enormous overlapping and redundancy, e.g. software - programmatüra, programnodrosinajums, programmu nodrosinäjums, softvers; multiplier - reizinätäjs, reizulis, multiplikätors. So when setting up the new Latvian-English dictionary it was mainly experts in various spheres that were involved working on language spheres rather than separate letters (as before). Thus, gaps, duplication, redundancies and inconsistencies, supply of different synonyms was to some extent avoided. A few linguists rearranged and finished the final product. While there is no corpus, much of the decision of what is frequent and used is unfortunately determined by intuition. • The colloquial element was another aspect that underwent serious change. Numerous phrasal verbs were added in the English part, formerly disregarded as too colloquial. In this dictionary also colloquial and rude Latvian wordsybrA, baigi, dimbä, purns, oalis, fcülis, fenderet, öau were not feared and exact English counterparts provided, e.g. iurat sar. to pee, to piss to take a piss kakät sar. to (do a) poo, vulg. to shit, pirdiens vulg. fart • Strong emphasis was put on combinatorics or collocability, grammar notes, link elements (prepositions) as the dictionary is mainly used for translation and the users being unfamiliar with the target language have to be helped in their choice of surrounding words. Semantic equivalence does not always imply communicative equivalence; thus contextualization is very important for the success. Understanding of how the words are used in context is partly an element of learner's dictionary that has been absorbed by the new type of active bilingual dictionary. (Svensen, 1993:87). •

Increase in the number of synonyms with their semantically different scope and attenuation has lead to a more frequent labeling and differentiation by means of explanations. Paradoxically the dictionary includes many obsolete English words which are frequently used in Latvia (and found as first counterparts in the previous dictionaries). Their simple loss would be misunderstood and now they usually come at the end of the synonym row, supplied with a label obsolete.

Some minor points that distinguish the new generation of dictionaries - front matter in all dictionaries seems to get shorter, no doubt reflecting a common sense approach and evidence of the reluctance of users to refer to it. Both dictionaries are also simpler, more userfriendly, more explicit (van der Meer, 1996: 193-194), with fewer subject labels and codes. There are still issues to be resolved - for example the gender issue. Latvian being a synthetic language where gender is shown by suffix, it carries a tradition of using masculine only in A part of the dictionary. Thus Latvian - English dictionary is fully masculine (with the exception of words where feminine forms are irregular). English - Latvian dictionaries, of course reveal the gender issue in A part, while keeping to the stereotyped gender roles in

Latvian Bilingual Lexicography - Tradition and Change

33

the translation part, e.g. secretary - Sekretäre (woman secretary), printer - tipogräfs (printer-man). However the overwhelming problem is a decent reflection of the Latvian lexicon and conflicting ideas of what the language is like. The only radical way out of the problem is by means of a Latvian corpus that would in quantitative terms show what the language is really like. Yet there is a reluctance, almost paralyzing fear of seeing the language as it is, not what we want it to be, nor even what it seems to us to be.

Bibliography Dressel (1688): Vocabularium in vier Sprachen, teutsch, lateinisch, polnisch und lettisch. - Riga. Elger, R.P.G. (1683): Dictionarium Polono - Latino - Lottauicum. - Vilnae. Krastina, M. (1996): Lettisk-svensk ordbok. - Stockholm, Riga: Memento-Zinätne. Lange, J. (1777): Vollständiges deutschlettisches und lettischdeutsches Lexicon. - Mitau. LatvieSu literäräs valodas värdmca (1972-1996). - Riga: Zinätne. Lomholts, K., Ledina, R. (1995): Lettisk-dansk, dansk-lettisk ordbog. - Riga: Norden. Mancelius, G. (1638): Lettus, das ist Wortbuch. - Riga. Mülenbach, K. (1923-1932) LatvieSu valodas värdmca. - Riga. Stender, G. F. (1789): Lettisches Lexicon. - Mitau. Svenson, B. (1993): Practical Lexicography. Principles and Methods of Dictionary-Making. - Oxford. Turkina, E. (1963): Latvieau angju vardnica. - Riga: LVI. — (1982): LatvieSu angju vardnica. - Riga: Avots. Odris. P. Liksim latvieSu valodu värdnicäs. In: Diena 20. 07. 1997. Valdemärs, Kr. (1872) Kreevu-latweeschu-wahzu wardnize. - Moskva. van der Meer, G. (1996): How Alphabetical Should A Dictionary Be? (the case of High and its combinations in some dictionaries. In: Symposium on Lexicography VII. Lexicographica. Series Maior 76. - Tübingen: Niemeyer, pp. 183-197. Veisbergs, A. (1997): English and Latvian Wordformation. - Riga: LU — (1997): Latvian - English Dictionary. - Riga: EA1. — (1994): Latvian-English, English-Latvian Dictionary of False Friends. - Riga: SI. Vinteräeids Fr. (1995): Mana Latvija. - Riga: Vaga.

Elena Barcena/Tim Read

The Architecture of a Cognitive-based Multilingual Computational Dictionary

1. Introduction

This paper presents the design of MoLeX (Multilingual onomasiological LeXicon), the computational prototype of a multilingual on-line lexical reference system which was recently developed by our research team.1 This is the result of further development and implementation of Martin Mingorance's Functional-Lexematic Model (FLM) (1984) for the description of the core vocabulary of natural languages. The origin of FLM can be traced back to Martin Mingorance's evaluation of a large number of existing mono- and bi-lingual dictionaries (special attention to their usefulness for the specialised user, such as the professional translator). His conclusion was that different dictionaries had their own strengths and weaknesses, although all of them shared a number of fundamental drawbacks regarding the nature and organisation of the information. Subsequently, Martin Mingorance devised a novel lexicographical architecture in which carefully selected data from existing monolingual dictionaries (particularly corpus-oriented ones) were used for each entry, although the organisation he made of such data is, to our knowledge, (still) radically different from any previous lexical resource. The major design features of FLM, which can also be seen in the pilot system, include complete and economically arranged information within each lexical entry and its global hierarchical and onomasiological organisation (which has been devised in an empiricist bottom-up manner from a selection of existing dictionaries). Furthermore, thanks to its compositional design, MoLeX avoids circularity, reveals significant intra- and cross-linguistic generalisations, and its computational version can be used either mono-, bior multi-lingually. This is achieved via an interface which is both flexible and independent from the lexical database. The standard type of dictionary organisation is semasiological (i.e., according to form), of which alphabetical order is undoubtedly the most common. Important information about the global organisation of the vocabulary of the language is, however, inevitably lost. This fact and the goal of producing a dictionary which is psychologically adequate led a number of lexicographers to develop onomasiologically organised lexical resources (i.e., according This work has been done as part of the research project 'Development of a lexical logic for computer-assisted translation from a multifunctional and reusable English-Spanish-French-German lexical database' (PB94-0437), which is funded by the Spanish Ministry of Education and Culture. This paper is an extension and modification of some of the ideas presented in the authors' forthcoming article 'MoLeX: Tomorrow's Computational Dictionary Today', Atlantis, 1998. The authors would like to acknowledge the intellectual legacy of the late Martin Mingorance which is partly reflected in this paper. All errors and idiosyncratic opinions are, of course, our own.

36

Elena Barcena/Tim Read

to conceptual meaning), the most common of which are thesauri. Thesauri, however, have a considerable amount of redundancy and provide little or no information about the combinatorial and selection properties of words. Their organisation is psychologically motivated to some extent, but their top-down structure is considered to be rather ad hoc. According to Martin Mingorance, a dictionary should reflect the organisation of the human mental lexicon and account for the way in which individual languages lexicalise conceptual knowledge. Psycholinguists like Apresjan (1993) claim that there is empirical evidence about both the nature of semantic relations, which are not arbitrary but real mental functions, and the hierarchical structure of the human psyche, with semantically related words stored near to each other in the mental lexicon. In accordance with this view, Martin Mingorance outlined the design of a relational lexicon-based on meaning structure, where lexemes are organised hierarchically in a bottom-up manner according to their common and differential semantic components. Not surprisingly, the inventory and internal structure of lexical fields and the distribution of lexemes within them which have been arrived at through this method sometimes differ from those normally found in traditional thesauri. He also believed that linguistic competence largely depends on the information obtained from the mental lexicon, which implies the existence of complex linguistic (e.g., morphosyntactic, semantic, pragmatic) knowledge attached to its lexical items. Therefore, he undertook a study of the type of paradigmatic and syntagmatic information which would be necessary to include in the entries of a dictionary in order to account for successful linguistic production and comprehension, and how to do this economically. Martin Mingorance incorporated cognitive issues like these into the elaboration of his lexical theory which he called FLM.

2. The linguistic model FLM is based on Martin Mingorance's integration of two linguistic theories, namely Dik's Functional Grammar (FG) (1978) and Coseriu's Lexematic Theory (LT) (1977), in order to obtain both a syntagmatic and a paradigmatic axis which describe lexical combination and selection respectively. The FG-based syntagmatic axis uses predicate frames as integrated formulae which specify the combination patterns that govern the predicates of a language. On the paradigmatic axis, lexemes are arranged onomasiologically within semantic fields following the dictates of LT. Microstructurally, lexical entries are characterised as complex units of syntactic, semantic, and pragmatic information. On the macrostructural level, lexical entries are interconnected semantically in a number of ways forming complex hierarchies. The resulting integrated model is intended to allow the lexicographer to build a hierarchical network of semantically related lexemes within general areas of meaning, with highly informative and economical definitions. FG and LT are not only complementary but also compatible from a number of theoretical and practical viewpoints. As Dik (1978:46) says: "Although the view of lexical analysis fits in nicely with the model of FG, the assumptions embodied in it do not necessarily follow from this model. That is, FG would also be compatible with other conceptions of lexi-

The Architecture of a Cognitive-based Dictionary

37

cal meaning definition". For example, FG is Ideologically functional because it regards language as an instrument of verbal interaction, not as an abstract system autonomous of real usage. LT is structurally functional in that language is seen to be determined by a system of functional oppositions. Furthermore, the FG principles underlying the structure of meaning definitions and the procedure of stepwise lexical decomposition for definitions have many points of convergence with the principles of lexical field theory and factorisation of LT, as subsumed in FLM. FG was adopted by Martin Mingorance for a number of reasons, including its psychological adequacy and its integration in a theory of verbal communication. FG is lexicondriven, providing a complete account of the combinability of lexical elements in the linear sequence, stated in functional and abstract terms. Lexical units are regarded as structured representations, and encoded in the form of predicate frames, which encapsulate the following information: (a) form of the predicate: the orthographical and phonological representation of the predicate; (b)syntactic category: the part-of-speech label; namely, verb, noun, or adjective; (c) quantitative valency: the number of arguments subcategorised by a given predicate; (d)qualitative valency: the semantic role of the arguments in the state-of-affairs designated by the predication; (e) selection restrictions: the features which specify the nature of the arguments which can appear in complement and subject positions; (f) meaning definition. Meaning definitions follow the strategy known as stepwise lexical decomposition and, similarly to sentential underlying predications, are stated in terms of predicates which exist in the language. Due to the complexity of devising a finite set of universal units of semantic description (e.g., binary features, abstract primitives, atomic predicates) as reported by numerous researchers, the meaning components used in FG are natural language phrases, although there are several constraints in their expression. Apresjan (1993, 86) writes: "In the majority of languages there are semantic primitives, or rather, near primitives, to describe the basic concept of each system. The defining predicates occurring in meaning definitions are lexical items of the object language" (our emphasis). It should be noted that in FLM there is considerable refinement of the content of predicate frames as stated by standard FG. For example, for each given predicate there is a list of one or more preferred predicate frames, which accounts for syntactic flexibility and avoids the rigid problematic distinction between arguments and satellites in FG. Also, the descriptiveness of predicate frames is enriched with further information; for instance, the syntagmatic description of verbs includes collocational information and syntactic functions which are not present in standard FG. Martin Mingorance incorporated LT into his model in order to expand the knowledge in the lexicon beyond the level of individual entries reflecting its global relational structure, and hence to build the dictionary of a language as a hierarchical network of semantically connected lexemes. LT contains the following fundamental concepts: (a) lexical field: the set of all lexical units which share an explicitly distinguished non-trivial semantic component;

38

Elena Barcena/Tim Read

(b)dimension: a subdivision of a lexical field which falls halfway between minimal groups of lexemes and the lexical field proper; a viewpoint or articulation of the content of the lexical field directly derived from the definitional structure of lexical units; (c) archilexeme: the nuclear word or definiens in terms of which all the words in the lexical field subsumed under it are defined; (d)seme: the minimal feature in the meaning definition of lexemes which differentiates them from each other within a dimension; (e) classeme: a content feature by which a lexical class is determined; it can operate throughout the lexicon in different dimensions and lexical fields.

With this relational model, it is possible to capture the semantic hierarchical structures of whole lexical fields, the relationships (hyponymy, synonymy, antonymy) between their items, and the subtle semantic nuances of such items. Furthermore, following the compositional method of definition and the principle of feature inheritance, the hyponyms of a given archilexeme become the definiens of other words at more specific levels of the hierarchy, and so on. The 'near primitive' components which form meaning definitions are inherited by all the lexemes lower in the hierarchy, so that definitions are clearly expressed in such simple terms as the combination of the immediate superordinate plus their nuclear meaning via one or more components. This economical method prevents circularity in the definitions and reveals significant generalisations about the structural semantic organisation of the lexicon of a language (Bärcena et al., 1997). The development of a framework which organises and describes predicates in the lexicon following lexematic principles leads to a deeper understanding of individual languages and the complex relationship between syntax and semantics. Regarding the semantic structure of lexica, for example, a number of principles have been established about the recurrence of dimensions across lexical fields (Faber, 1994). Furthermore, various syntactic aspects of words are tied to their meaning. The close relationship between syntax and semantics has been observed by authors like Dixon (1991:75): "The lexical words of a language can be grouped into a number of semantic types, each of which has a common meaning component and a typical set of grammatical properties". Many of these properties, like classemes, are recurrent in other domains and lexical fields. Martin Mingorance developed the notion of predicate schema as a kind of extended predicate frame which captures the syntactic, semantic and pragmatic information shared by all the lexemes in each dimension. This is, once again, crucial both to organise the entries more economically, since only new syntagmatic information needs to be stated for each predicate lower in the hierarchy, and also to capture significant syntactic and semantic generalisations within semantically related lexical items. The methodology for building the FLM dictionary is data-driven and bottom-up. The primary source of data is a collection of widely used monolingual dictionaries. In order to arrive at the definition of a given word, all the dictionaries are consulted and the corresponding lexical entries are analysed. The relevant meaning components for the identification of the word, which have been taken from the existing definitions, are selected and put under headings such as nuclear meaning, direct object, adverbial complementation, and pragmatic information. A new structured definition is then built by reformulating the selected information according to certain expression rules. Definitions are composed from left to right, starting from a definiens and adding increasingly specific features on the right, which differentiate the word from the preceding members of the hierarchy. The assignment

The Architecture of a Cognitive-based Dictionary

39

of a word to a lexical field uses the prototypical method, according to which there is a set of clear cases or prototypes which serve as points of reference. Linguistic generalisations can be captured in terms of classemes and dimensions on the basis of the presence of recurrent patterns as the lexical hierarchies are being built. Following this bottom-up procedure, the inventory of lexical fields within the vocabulary of the language is obtained. After having elaborated the definitional structure of 10,000 English verbs, it has been found that words fall into the following basic domains: EXISTENCE, MOVEMENT, POSITION, CHANGE, POSSESSION, PERCEPTION (including stimulus verbs such as LIGHT and SOUND), EMOTION, COGNITION, SPEECH, and GENERAL ACTION (composed of subgroups such as verbs of CONSUMPTION, COMPETITION, CONTACT, USE, etc.). The lexical fields of the different languages are compared and cross-linguistic links are created between equivalent lexical entries. These links are not always one-to-one so when a lexical gap is found in a given language it must be filled by inserting a periphrastic expression. Furthermore, an indexing system needs to be elaborated to cater for the different meanings of ambiguous words. The final step of the building process is the testing of the correctness and relevance of the information in the new dictionary via a corpus of the general language official and sufficiently representative.

3. The computational prototype

As is common practice today in Lexicography and generally in Linguistics, a computational prototype, MoLeX, was developed to test and evaluate FLM. Furthermore, it was built as a step toward the implementation of the full scale dictionary, which is underway. Consulting paper-based lexical resources is often a tedious and cumbersome task. As Hutchins & Somers (1992) comment, it has been estimated that professional translators spend over 60% of their time consulting such resources. For a long time there has been great interest in making the task of lexical consultation faster and more user-friendly. It did not require much imagination to realise the role that computers could play in this process (not only in the provision of lexical resources, but also, for example, word processors and grammar and spelling checkers). Lexicographical systems have many advantages over paper-based versions including the following: (a) fast and flexible look up: data from automatic dictionaries can be accessed via keywords (or parts thereof) other than the exact entries themselves, or even combinations of keywords; (b)large storage capacity: as memory and hard disk sizes grew and optical storage media (such as compact disks) became available, on-line lexica were increasingly able to offer much larger linguistic coverage than their paper-based counterparts; (c)multilinguality: paper-based dictionaries are normally either mono- or bilingual and cover a small collection of languages. On-line sources, however, offer multiple language combinations; (d)fast and flexible use: computational dictionaries can be used on their own, or as part of integrated environments such as the translator's workbench; (e)ease of modification, updating, and customisation: new entries can be easily added, either as part of the main dictionary or as separate specialised sub-dictionaries;

Elena Barcena/Tim Read

40

(f) test-bed: in a research context, computerised tools provide a practical way to investigate languages and the functionality and implications of linguistic theories.

Like the majority of programs with a visual interface, MoLeX, which has been written in Visual Basic 5.0, is an event driven system in which (following the initial load and set-up phase) the actions of the user determine the information that is presented on the screen. The figure below illustrates the MoLeX interface (vl.O beta) showing the paradigmatic axis. The interface includes three menus and an icon bar (which duplicates the most important functions of the menus), which can be presented in different languages. It should be noted that the selected interface language determines the order of presentation of the dictionaries on the screen. The linguistic information contained in the dictionary can be displayed in three ways: the dimensions of a given lexical field, the dimensions together with their lexemes, and the dimensions together with their lexemes and definitions. The selection of a dimension or word (by clicking on it with the mouse) retrieves its equivalent(s) in the other language. A selected entry can be copied or printed for future use, and a basic history list enables the user to return to entries selected earlier in the session. Various properties of the interface can be changed, such as the interface language, colour, and font properties. It should be noted that the linguistic information in the two languages can be presented to the user in three different ways: firstly, as two lists presented side by side, aligned and connected to the same vertical scrollbar (as in the figure below). Secondly, as two non-connected lists, presented side by side, with separate scrollbars. Thirdly, as a single list which spans the entire window and contains its own scrollbar (in which case the user can swap between windows).

abrasar quemar brasas incinerar quemar algo convertirlo en ceniza esp. cadäveres HACER QUE ALGO DEJE DE ARDER apagar hacer que algo deje de arder extinguir apagar el fuego haciendo que este cese completamente apagar con agua sofocar apagar un fiiego impidiendo que siga desarrollandose apagar con agua LUZ PRODUCTDA FOR FENOMENOS DE LA NATURALEZA relampaguear ernibr luz al producirse un relämpago amanecer emitir luz al empezar el dia

incinerate to bum sth completely cremate to incinerate a dead body (esp. as a funeral service) TO CAUSE STH TO STOP BEING ON FIRE (BURNING) put out to cause sth to stop burning extinguish to put out sth Oight/fire)(fml). quench to extinguish a fire with water (old-fashioned) smother to extinguish a fire by covering it douse to put out sth (fire/light) esp. by pouring water on it LIGHT PRODUCED BY NATURE

toßash (of lightening) dawn to give off light gradually after the night

The Architecture of a Cognitive-based Dictionary

41

The interface allows the user to consult the different meanings of a given word. There are two types of ambiguity in this system: ambiguity in the source language and in the target language. In the first type, the same word can appear in more than one dimension or more than once under the same dimension. The user is offered all the possible interpretations of the ambiguous word and then prompted to select the one required. In the second type of ambiguity, a word has more than one equivalent in the target language. The distinct occurrences of the word can be distinguished in terms of the corresponding dimensions or simply by its translation equivalents. It should be noted that the format of the database provides information related to the descriptive model. For example, the relations of hyponymy between the different dimensions and between the lexemes in each dimension are represented visually by indenting the left margin (as can be seen in the example above).

4. Conclusion This paper presented an on-line lexical reference system for the description of the basic lexicon of natural languages, which follows a new dictionary model based upon FLM. This is a theory based upon cognitive principles which presents the dictionary as the representation of the mental lexicon of the speaker. FLM has been designed to construct both an onomasiological description of the vocabulary of a language where the predicates are grouped into semantic classes or domains, and a complete description of the predicates. FLM integrates FG together with LT to account for the syntagmatic and paradigmatic relations in the lexicon, which are based upon the complementary principles of combination and selection present in definitions. The knowledge in the lexicon is extended beyond the level of the individual entries in order to reflect its global relational structure, which allows the lexicon to be designed as a hierarchical network of semantically connected lexemes. The computational prototype reproduces all the characteristics of the theoretical model: it contains complete and economically arranged information within each lexical entry, avoids circularity, and reveals significant intra- and cross-linguistic generalisations. Furthermore, MoLeX is flexible, portable, and can be used either mono-, bi-, or multi-lingually. development. Finally, this network is dynamic and open to the progressive incorporation of further entries. At present MoLeX covers Spanish and English verbs, but new languages and lexical categories are under development. Also, another version is being developed in Java to make it accessible both locally or via the World Wide Web.

Bibliography Apresjan, J.D. 1993: 'Systemic lexicography as a basis of dictionary-making'. In Dictionaries: Journal of the Dictionary Society of North America 14.

42

Elena Barcena/Tim Read

Bärcena, E., T. Read & P. Faber. 1997: 1 diccionario del traductor del mafiana'. Presented at the I Congreso International de Estudios de Traduccion (Universidad de A Corufia) and submited to Revista de la Facultad de Filologia Inglesa. Coseriu, E. 1977: Principios de semantica estructural. Credos. Madrid. Dik, S.C. 1978: Functional Grammar. Foris. Dordrecht. Dixon, R.M.W. 1991: A new approach to English grammar on semantic principles. Clarendon. Oxford. Faber, P. 1994: 'The semantic architecture of the lexicon'. In K. Jensen & H. Pedersen (eds.). Proceedings from the VI International Symposium on Lexicography. Niemeyer. Tübingen. Heslop, B. & L. Budnick. 1996: Publicar con HTML en Internet. Paraninfo. Madrid. Hutchins, W.J. & H.L. Somers. 1992: An Introduction to Machine Translation. Cambridge University Press. Martin Mingorance, L. 1984: 'Lexical Fields and Stepwise Lexical Decomposition in a Contrastive English-Spanish Verb Valency Dictionary'. In R.R.K. Hartmann (ed.). LEX'eter 83 Proceedings: Papers from the International Conference on Lexicography at Exeter. Niemeyer. Tübingen. Read, T., E. Bärcena & P. Faber. 1997: 'Java and its role in Natural Language Processing and Machine Translation'. Proceedings of the Machine Translation Summit VI. University of California, San Diego.

Matthias Kammerer

Kritisches zu Schnelles Applizierung einer »Logischen Semantik« bei Wörterbüchern vom COBUILD-Typ*

l.

Applizierung einer Logischen Semantik auf Wörterbücher vom COBUILD-Typ nach Schnelle 1.1. Darstellung der Argumentation von Schnelle 1.2. Kritik der Korrektheitsthese 1.3. Kritik der Vollständigkeitsthese

l .4. 2. 3. 3.1. 3.2. 3.3.

Bemerkungen zur Axiomenthese Anhang Literatur Corpora Wörterbücher Sonstige Literatur

1. Applizierung einer Logischen Semantik auf Wörterbücher vom COBUILD-Typ nach Schnelle Dieser Aufsatz hat es sich zur Aufgabe gemacht, die verführerischen und auf den ersten Blick hin einleuchtenden Ausführungen in Schnelle (1997),' nach denen die Bedeutungsangaben in Wörterbüchern vom COBUILD-Typ als ein axiomatisches System zu betrachten seien, wörtlich zu nehmen und sie in ihren Konsequenzen zu überdenken sowie auf potentielle Probleme aufmerksam zu machen. Der Vorteil von Bedeutungsangaben als axiomatisches System liegt klar auf der Hand: auf diese Weise könnten Maschinen mit Sprachdaten gefuttert werden, auf deren Grundlage sie Deduktionen vornehmen könnten. Endlich wäre die Semantik der natürlichen Sprache exhaustiv maschinell implementierbar, und »Bedeutung« könnte vom Computer verarbeitet werden. In Schnelle et al. (1996) werden folgende Vorteile genannt („Z" steht für Zitat}: [Z l: Schnelle et al. 1996, 175] [...] wollen wir einen Weg zeigen, der von einer ganz bestimmten theoretischen Konzeption eines natürlichsprachlichen Lexikons über dessen inhärente logische Ordnung zu einer inferentiellen lexikalischen Wissensbasis für sprachverarbeitende Systeme und Online-Wörterbücher führt. Ausgehend vom Collins Cobuild Student 's Dictionary (CCSD) werden wir einen neuen, logischen Ansatz der lexikalischen Semantik und seine Abbildung auf eine integrative unifikationsbasierte A. Furrer danke ich sehr für die Prüfung der formal-logischen Ausführungen. Für konstruktive Kritik danke ich außerdem den Teilnehmerinnen und Teilnehmern des Kolloquiums Capita selecla der germanistischen Linguistik vom WS 1997/98 an der Universität Heidelberg, insbesondere Chr. Bläsi, K.-P. Konerding, K. Mudersbach und H. E. Wiegand. Cf. auch den früheren Aufsatz Schnelle et al. (1996) sowie die Explikationen in Schnelle (1991, 483-567).

•

Matthias Kammerer

Grammatik vorstellen und zeigen, welche Konsequenzen sich für die Praxis der Sprach- und Informationsverarbeitung aus unserer Theorie ergeben.

Schnelle geht dabei von der begrüßenswerten Einsicht aus, daß „[...] es nicht die Aufgabe des Lexikographen sei, die in der Welt existierenden Dinge zu defmierenf,] sondern die Formen des Gebrauchs der Wörter zu klären." (Schnelle 1997, 143.) Trotz dieses Bekenntnisses verwendet Schnelle weiterhin den Begriff der Definition und führt an Hand des Wörterbuches COBUILD vor, wie Wörterbuchartikel so geschrieben und reinterpretiert werden können, daß die Bedeutungsangaben eines Wörterbuchartikels als die Axiome einer Sprache aufgefaßt werden können. (Man bedenke dabei, was Schnelle wahrscheinlich unter einem Axiom versteht: eine Instanz einer nullstelligen Regel in einem Kalkül oder weniger technisch ausgedrückt - „wahres Urteil bzw. wahrer Satz, der bei deduktivem Aufbau einer Theorie, innerhalb einer abgeschlossenen Theorie, ohne Beweis als Ausgangsthese akzeptiert wird und den Beweisen anderer Thesen dieser Theorie zugrunde gelegt wird." [Kondakow 2WBLogik, 60.]) Ausgehend von Huberts Konzeption eines axiomatischen Systems, schlägt Schnelle vor, dies auf die Lexikologie und Lexikographie zu übertragen: [Z 2: Schnelle et al. 1996, 178] In diesem Sinn kann man auch in der Lexikologie versuchen, von dem in einer Sprache ausgedrückten Wissen (dem System der Theoreme, d.h. der „Theorie") auszugehen und das Lexikon als die Teiltheorie aufzufassen, die die Deutung der Wörter ausreichend festlegt. Das Lexikon ist dann keine Kollektion von atomistischen Definitionen von Wörtern, sondern ein axiomatisch vernetztes Geföge von Aussagen, das den Gebrauch der Wörter implizit festlegt. Diese Auffassung leitet unsere Analyse des Lexikons. Das typische Lexikon, wie es dem Benutzer einer Sprache vorgelegt wird (zum Beispiel vom Cobuild-Team), stellt mit den in ihm angegebenen Erklärungen gewissermaßen die Axiome einer solchen Theorie heraus. Es geht davon aus, daß erstens jedem Wort soviele [!] Axiome zugeordnet werden, wie das Wort Lesarten hat, und daß zweitens normalerweise jedes Axiom nur als Angabe genau einer Lesart eines Wortes vorkommt. Die Wörter korrelieren also mit einer Zerlegung der Axiome der Wortlehre in kleine Teilmengen von Axiomen (wobei jede Teilmenge die Menge der Lesarten eines Wortes ist).

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

AC

Eine solche Überlegung hat weitreichende Konsequenzen.2 Diese werden von Schnelle selbst zusammengefaßt:3 [Z 3: Schnelle 1997, 148] Das (logisch transformierte) Wörterbuch [gemeint sind hier eigentlich die logisch transformierten Bedeutungsangaben; M. K.] ist die Menge von Axiomen eines axiomatischen Systems; aus ihr können alle allein aufgrund des sprachlichen Inhalts der Wörter als -wahr verstandenen Sätze allein mithilfe [!] der Regeln der Syntax und der Satzlogik [i.e. Aussagenlogik; M. K.] abgeleitet werden. Diesem Zitat Z 3 können m.E. drei Thesen (= T) entnommen werden, die im Folgenden eingehend diskutiert und problematisiert werden sollen: T,:

T2:

T3:

die Axiomenthese (AT), nach der die Menge der (logisch transformierten) Bedeutungsangaben (IA = {BA„ BA2 BA„} mit n e N \ {0}) in Wörterbüchern vom COBUILD-Typ die Axiome (einer Sprache) sind; die Vollständigkeitsthese (VT), nach der aus dem axiomatischen System der Menge der Bedeutungsangaben in Wörterbüchern vom COBUILD-Typ mittels eines Kalküls (Kai) - hier: den Regeln der Syntax und der Aussagenlogik - „alle allein aufgrund des sprachlichen Inhalts der Wörter als wahr verstandenen Sätze" S einer Sprache L abgeleitet werden können (wobei unter einer Sprache L wohl eine Menge von grammatikalischen Sätzen verstanden wird), und die Korrektheitsthese (KT), nach der aus dem axiomatischen System der Menge der Bedeutungsangaben mittels eines Kalküls „alle allein aufgrund des sprachlichen Inhalts der Wörter als w a h r verstandenen Sätze'1 - und nur diese, wie hier hinzugefugt werden darf abgeleitet werden können.

In Schnelle (1991) ist er teilweise noch vorsichtiger, was die Möglichkeit der logischen Formalisierung der natürlichen Sprache betrifft, und schreibt u.a.: „Es ist von vorneherein [recte: vornherein] nicht klar, wieweit man diese aus den formalen Sprachen der Logik bekannten Mechanismen der Folgerung direkt auf Sätze der natürlichen Sprache wie die [durch die sog. lexikalischen Sätze; M. K.] gegebenen übertragen kann." (ibd. 515) oder: „Die von uns als ideal vorgestellte Form einer semantischen Bedeutungsangabe ist 'Wenn p, dann q', wo p für einen lexikalischen Satz und q für sein Voll-Implikat steht. Der Wenn-dann-Satz ist nicht als materiale Implikation im Sinne der Logik zu verstehen, sondern als Angabe über das, was implizit auch gesagt wird, wenn man etwas anderes explizit sagt." (ibd., 517). Dann aber finden sich auch Aussagen darin wie diese: „Bei der Entwicklung der lexikalisch-implikativen Bestimmungen werden wir uns von den für die formale Logik entwickelten Vorstellungen leiten lassen und überlegen, ob und wie wir die dort geltenden Regeln für die natürliche Sprache übernehmen können." (ibd., 514) oder noch deutlicher wird es, wenn er bei den natürlichsprachlichen »Implikaten« eines lexikalischen Satzes die „Transitivitätseigenschaft der [logischen, materialen] Implikation" konstatiert und darauf die Schlußregel des Modus ponens anwenden möchte (ibd., 505). In Schnelle et al. (1996, 179) findet sich folgende ähnliche, aber weniger pointierte Passage: „Der lexikalisch-semantische Teil eines idealen Lexikons soll nicht als atomistische Angabe von Bedeutungsrepräsentationen zu Wörtern angesehen werden, sondern als indiziertes axiomatisches System, d.h. als eine Menge von Axiomen der Wortsemantik einer Sprache, also als Menge von Sätzen, aus denen Theoreme folgen, die in ihrer Gesamtheit das sprachlich-analytische Wissen der Sprecher einer Sprache wiedergeben und dadurch den Gebrauch der Wörter implizit definieren."

,

Kammerer

Etwas formaler - und für die hiesigen Zwecke dienlicher - lassen sich die drei Thesen wie folgt reformulieren (wobei „=>" eine Abkürzung für die natürlichsprachliche Wenn... dann...-Konstruktion ist):4 Tj: TJ: T3:

AT: VT: KT:

V e B A: L S => BAhKalS =>

0 l-Ka, BA BA h Ka , S L = S

1.1. Darstellung der Argumentation von Schnelle Um nun diese Behauptungen zu illustrieren, sei ein Beispiel im Anschluß an Schnelle (1997) gegeben. Ausgegangen wird vom ersten Satz aus der Bedeutungsangabe im zweiten semantischen Subkommentar des Wörterbuchartikels zu orange in COBUILD („E" steht für Wörterbucheintrag), wo es heißt: [El:COBUILD678 b ] 2 An orange is a round fruit that is juicy and sweet. It has n count or a thick orange skin and is divided into sections inside. See N uncount picture headed FOOD.

Den ersten Satz der Bedeutungsangabe interpretiert Schnelle als eine Definition,5 die aus den Bestandteilen: Definiendum - Deßnitor - Definiens besteht, so daß der erste Satz aus E l wie folgt analysiert werden kann: Definiendum:6 Deßnitor: Definiens:

An orange is a round fruit that is juicy and sweet.

Diese »Definition« kann in ein bedeutungsgleiches Wenn...dann...-Schema transformiert werden, das Schnelle weiterhin »Definition« nennt. Auf diese Weise erhält man das Damit nicht der Verdacht entsteht, daß hier Formalismen benutzt werden, die letztlich von der eigentlichen Sache ablenken oder diese zumindest verschleiern, werden sie - für die Aussagenlogik - im Anhang (vor dem Literaturverzeichnis) näher erläutert. Daß dies keine Definition - zumindest im klassischen Sinne - ist, dazu vgl. man z.B. Wiegand (1989d, insbes. 547-550). Wahrscheinlich, so ist zu vermuten, benutzt Schnelle den Begriff der »Definition« in Anlehnung an die lexikographische Werkstattsprache, in der von lexikographischer Definition bei Bedeutungsparaphrasenangaben die Rede ist. - Nach der mittelalterlichen Schullogik gilt: Definitio fi[a]t per genus proximum et differentiam specificam. Das genus proximum ist dabei der unmittelbar übergeordnete Gattungsbegriff in einem Begriffssystem, das man sich als porphyrischen Baum vorzustellen hat. Bei orange wäre dies nicht fruit, sondern wohl citrus fruit. Außerdem ist zu beachten, daß differentiam specificam Akk. Sg. ist. Demnach wird der nächsthöhere Gattungsbegriff durch genau ein Merkmal differenziert und nicht, wie im COBUILD, durch drei. Tatsächlich gibt es - zumindest für die Botanik - eine streng binäre Nomenklatur, die gemäß dem oben zitierten Schulsatz aufgebaut ist, nämlich das im Jahre 1735 von dem Schweden Carl von Linno entwickelte Linnesche System, das sich auf die Unterscheidung der Sexualorgane stützt. (Cf. Brockhaus l? Enz, XI, 492.) Schnelle (1997, 141) spricht an dieser (und an weiteren) Stelle(n) immer versehentlich vom Definiens statt von Definiendum.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

Schema (oder: den Satz) S„ wobei jetzt die „inhaltstragenden Wörter" (Schnelle 1997, 143) mit den jeweiligen Lesarten aus COBUILD indiziert sind: S,

IF something is an orange2 THEN it is a round, fruit, which is juicy, and sweet,.

Schema S, kann durch die aussagenlogische Formel (=F): F,

A-B

wiedergegeben werden, wobei die atomare Aussage ß, wie der da««...-Teil in S, bereits zeigt, sehr komplex sein kann. In F, entspricht das Symbol „A" dem Definiendum, der Pfeil „—" dem transformierten Definitor und das Symbol „B" dem Defmiens. Auf S, kann nun eine weitere bedeutungserhaltende Transformation durchgeführt werden, die dafür sorgt, daß im then...-Tei\ quasi eine »Normalform« von mit und koordinierten Sätzen vorliegt, wobei versucht wird, möglichst jedes indizierte Wort in einen vollständigen prädikativen Satz zu überführen; auf diese Weise erhält man nach Schnelle (1997) S2: S2

IF something is an orange2 THEN it is round, AND it is a fruit, AND it is juicy, AND it is sweet,.

Eine solche Paraphrase, wie sie im Schema S2 vorliegt, hat den Vorteil, daß sie wie folgt aussagen logisch formalisiert werden kann: A-(B, ... 2 mit n e N \ {0}

F2

)

In F2 entspricht die atomare Aussage A dem Definiendum, der Pfeil „—" wieder dem Definitor und die Konjunktion von atomaren Aussagen B. - ... Bn dem Definiens. Damit hat S2 gegenüber S, (und: F2 gegenüber F,) den Vorteil, daß S2 innerhalb der Aussagenlogik adäquater7 modelliert werden kann, weil F2 quasi »näher« an der Sprache ist und die Struktur des Definiens besser verdeutlicht. Anmerkung: Genau genommen liegt der dann...-Teil in S2 - aussagenlogisch gesehen - in einer konjunktiven Normalform (CNF) vor, nämlich als Konjunktion von Disjunktionen von Literalen L, wobei Liierale Atome (= positive Literale) oder negierte Atome (= negative Literale) sind, so daß man F2 allgemeiner formulieren kann: Fj A - [(L,, v L 1 2 v . . . ) A ( L 2 , v L 2 2 v . . . ) A . . . m i t n e N \ {0}.

(L„ , v L„ 2 v ...)]

Die aussagenlogische Formel F3 kann leicht selbst in CNF gebracht werden; dazu wird sie zuerst einmal auf Grund der Äquivalenzbeziehung (A —· B) (-.A v B) umgeformt zu: Fj.

- v [(L, , v L, 2 v ...)

(L2 , v L 2 2 v ...)

...

(L„ , v L „ 2 v ...)]

und auf Grund der Distributivität von „v" zu:

Der Begriff der Adäquatheit soll in dieser Arbeit verstanden werden als dem untersuchten Gegenstand entsprechend (cf. Kondakow ^WBLogik, 15).

Matthias Kammerer F3b

[(-.A v (L,, v L, 2 v ...)) (-, v (L 2 , v L 22 v ...))

...

(-, v (L„, v L„ 2 v ...))]

Schließlich kann auf Grund der Assoziativität von „v" übergegangen werden zu: FJC

- v L , , v L, 2 v ...)

(-, v L 2 , v L 22 v ...)

...

(-iA v L„, v L„ 2 v ...)]

Lediglich hingewiesen sei noch darauf, daß für Formeln in CNF gilt, daß die Frage nach der Erfüllbarkeit bereits NP-vollständig ist. Nur dann, wenn Horn-Formeln vorliegen (i.e. Formeln in CNF, bei denen in den Disjunktionen höchstens ein positives Literal vorkommt), ist die Erfüllbarkeit in quadratischer Zeit entscheidbar (cf. Menzel 1990, 18-22). Nun soll fruit aus S2 näher betrachtet werden. Dazu wird im COBUILD unter fruit nachgeschlagen, wo man u.a. den folgenden Wörterbucheintrag findet: [E 2: COBUILD 390a] fruit /fruit/, fruits. The plural can be either fruit n count or n uncount or fruits, but it is usually fruit. 1 A fruit is something which grows on a tree or a bush °° and which contains seeds or a stone covered by a substance that you can eat. Apples, oranges, grapes, and bananas are all fruit. See picture headed FOOD. Der in E 2 wiedergegebene erste Teil der Bedeutungsangabe („A fruit is something ... you can eat") kann wiederum auf die bekannte Weise in ein Wenn...dann...-Schema transformiert werden, so daß man zuerst S3 und schließlich S4 mit einer Konjunktion von möglichst einfachen Sätzen im Sukzedenz erhält (die hochgestellten Zahlen kennzeichnen in diesem Fall keine Lesarten, sondern sie identifizieren verschiedene Entitäten): 53 54

IF something is a fruit, THEN it is something which grows, on a tree0 or a bush, and which contains, seeds, or a stone3 covered2 by a substance, which you can, eatQ. IF something1 is a fruit, THEN it is something1 which grows, on something2 AND something2 is a tree0 or a bush, AND something1 contains2 something3 AND something3 are seeds, or something3 is a stone3 AND something3 is covered2 by something4 AND something4 is a substance, AND you can, eat,, something4.

Das Schema S4 läßt sich aussagen logisch wie folgt formalisieren: F4

A-(BACADAEApAGAH).

Fürfruil^ in S, kann nun - quasi, so Schnelle, per »Termersetzung« - der dann...-Teil aus S3 (in dem jedoch „then it is" vorher getilgt wurde) eingesetzt werden, so daß man erhält:

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

49

something which grows,, onatree 0 orabush 1 ( IF something is an orange 2 THEN it is a round,

and which contains2 seeds, or a stone3

which is juicy.

covered2by a substance, which you can,eat0 and sweet.

Desgleichen kann für fruit, in S2, so Schnelle weiter, der dann...-Teil aus S4 eingesetzt werden, wobei hier „then it is a" aus syntaktischen Gründen getilgt werden muß: S6

IF something1 is an orange, THEN something1 is round, 'AND something 1 grows, on something" AND something 2 is a tree,, or something" is a bush, AND something' contains; something"' AND something" 1 are seeds, or something' is a stone·, AND something" 1 is covered; by something" AND something'' is a substance, - A N D you can, eat0 something 1 AND something1 is juicy, AND something1 is sweet,.

Auf Grund der aussagenlogischen Beziehung: F5

{A - (B C)} i-Kal A - B

erhält Schnelle aus S6 schließlich die folgenden Schemata:8 S7l Sn S7c S7d S7e S7f S7g

IF something1 is an orange, THEN something 1 is round,. IF something1 is an orange, THEN something1 grows, on something2. IF something1 is an orange, THEN something2 is a tree0 or something2 is a bush,. IF something1 is an orange, THEN something1 contains2 something3. IF something1 is an orange, THEN something3 are seeds, or something3 is a stone3. IF something1 is an orange, THEN something3 is covered2 by something4. IF something1 is an orange, THEN something4 is a substance,.

S7h S7i S7j

IF something1 is an orange2 THEN you can, eato something4. IF something1 is an orange, THEN something1 is juicy,. IF something1 is an orange, THEN something1 is sweet,.

Bei Fj stellt sich die Frage, ob eine solche - logisch korrekte - syntaktische (!) Deduktion innerhalb der natürlichen Sprache angemessen ist. K. Mudersbach vergleicht eine solche Semzerlegung - denn um nichts anderes handelt es sich unter einem linguistischen Aspekt - treffend mit einem zusammengestürzten Kartenhaus, dessen einzelne Karten (die Inhalte) Genau genommen, geht Schnelle (1997, 146) an dieser Stelle ein klein wenig anders vor, kommt aber letzten Endes zum selben Ergebnis: Aus der Bedeutungsangabe zu orange: 0 — (F ) deduziert er auf Grund der Konjunktions-Regel: 0 — F; zusammen mit der Bedeutungsangabe zu fruit, die die Form F — hat kann er auf Grund der Transitivitäts-Regel dann übergehen zu: 0 - Y.

Matthias Kammerer

durchaus angegeben werden können, die Interrelationen aber, die zwischen den Karten bestanden, damit sie ein Kartenhaus - und nichts anderes! - bilden konnten, fallen jedoch unter den Tisch. Ebenso ist die Bedeutung von orange nicht nur die bloße Summe von S7a bis S7j, sondern mehr, das nach der Zerlegung weggefallen ist. Ganz ähnlich bemerkt z.B. auch L. Tesniere auf der Satzebene, daß ein Satz aus mehr als der Summe der Inhalte seiner Wörter besteht, nämlich aus den Bedeutungen der Wörter und der Konnexion zwischen den Wörtern, die das Bewußtsein wahrnimmt. Ohne Konnexion wäre ein Satz erst gar nicht verständlich. Er schreibt: [Z 4: Tesniere 1980[96], 518] 5. Daraus folgt, daß ein Satz vom Typ Alfred spricht nicht aus zwei Elementen - nämlich erstens Alfred und zweitens spricht - besteht, sondern aus drei Elementen: erstens Alfred, zweitens spricht und drittens der Konnexion, die sie verbindet und ohne die kein Satz bestünde. Wer sagt, daß ein Satz wie Alfred spricht nur zwei Elemente enthalte, der hat oberflächlich und rein morphologisch analysiert und das Wesentliche - die syntaktische Verbindung - übersehen. 6. Genauso verhält es sich in der Chemie, wo die Verbindung von Chlor (Cl) und Natrium (Na) etwas Neues ergibt, das Kochsalz oder Natriumchlorid (NaCl), das ein ganz neuer Stoff mit völlig anderen Eigenschaften ist als Chlor einerseits und Natrium andererseits.

Insofern müßte die Möglichkeit von syntaktischen Zerlegungen, die nach F5 durchgeführt werden können, und ihre Anwendung auf die natürliche Sprache nochmals kritisch überdacht werden.

Wenn man bedenkt, daß die „Aussagenlogik [...] die 'gröbste' oder Oberflächlichste' Spielart der Logik" ist, daß „die Aussagenlogik zur Boole 'sehen Algebra, dem Rechnen mit Wahrheitswerten" wird und daß damit selbst „das (semantikbezogene) Folgern und das (rein syntaktische) Ableiten [...] sich auf dieser besonders elementaren Stufe logischer Betrachtungsweise" befindet (Menzel 1990, 4), ist es u.U. angemessener, die bei Schnelle (1997) ausschließlich in der Aussagenlogik formulierten Zusammenhänge kurz in der Prädikatenlogik 1. Ordnung zu betrachten.9 Hier soll nicht behauptet werden, daß eine Modellierung in der Prädikatenlogik 1. Ordnung hinsichtlich der natürlichen Sprache adäquat sei, sondern lediglich, daß sie adäquater ist, und zwar adäquater ist als in der Aussagenlogik, da sie z.B. die Quantoren und deren Skopi berücksichtigt. Aber selbst die Prädikatenlogik ist nicht in jedem Fall ein adäquates Kalkül für die natürliche Sprache, wie z.B. die über 2000 Jahre alten Probleme mit den Donkey-Sätzen (Jeder Bauer, der einen Esel hat, schlägt ihn bzw.: Wenn ein Bauer einen Esel hat, schlägt er ihn usw. - cf. hierzu z.B. Wunderlich 1991, 41^44, oder hinsichtlich der Schwierigkeiten einer syntaktischen Analyse: von Stechow 1993, 79 f.) Schnelle weist in Schnelle et al. (1996, 181) mit dem Prinzip 4: Prinzip der Subsprachen-Approximation darauf hin, daß das Verständnis einer komplexen Sprache durch schrittweise Annäherung von immer komplexer werdenden Subsprachen möglich sei. Es ist allerdings die Frage, ob gerade die Aussagenlogik zur Verdeutlichung der relativ komplizierten Zusammenhänge dienlich ist oder ob sie nicht wichtige Zusammenhänge auf Grund ihrer »Oberflächlichkeit« unberücksichtigt läßt. (Dasselbe läßt sich natürlich - dessen bin ich mir bewußt - cum grano salis ebenfalls gegen die nun wiedergegebene Darstellung in der Prädikatenlogik 1. Ordnung einwenden. Doch welche Subsprache wäre dann am Ende angemessen, wenn nicht die natürliche Sprache selbst?)

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

*j

zeigen, die bekanntlich nur mit großer Schwierigkeit in der Prädikatenlogik formalisierbar sind (von Heusinger 1997, 173-190, schlägt eine dezidierte Analyse vor). Ein Indiz dafür, daß hier die Prädikatenlogik zugrunde gelegt werden sollte, ist m.E. die Verwendung der Anapher /'/ oder auch die Variablenbindungen, die bei Schnelle unberücksichtigt bleiben und auch unberücksichtigt bleiben können, da er, wie gesagt, ein anderes Darlegungsziel verfolgt. Es ist mir klar, daß es Schnelle zunächst darum geht, die logischen Beziehungen zwischen den einzelnen Aussagen zu verdeutlichen, weshalb unter diesem Aspekt die Aussagenlogik völlig hinreicht. Auf dem Hintergrund des Gesagten ist S, in der Prädikatenlogik 1. Ordnung standardmäßig wie folgt zu formalisieren (daß der natürlichsprachliche Ausdruck ;/ something is an orange mit einem Aliquanter »übersetzt« wird, dessen Skopus über die ganze Formel reicht, soll hier nicht problematisiert werden): F;

Vx (0(x) - Q(x)) mit: ,,O(x)" bedeutet soviel wie ist eine Orange; ,,Q(x)" bedeutet soviel wie ist eine runde Frucht, die saftig ist und süß schmeckt.

S2 ist hingegen zu formalisieren als: FJ

Vx [O(x) - [R(x) F(x) J(x) mit: ,,O(x)" bedeutet soviel wie ,,R(x)" bedeutet soviel wie ,,F(x)" bedeutet soviel wie ,,J(x)" bedeutet soviel wie ,,S(x)" bedeutet soviel wie

S(x)]] ist eine Orange; ist rund; ist eine Frucht; ist saftig; schmeckt süß.

Bei Schema S4 wird es prädikatenlogisch interessanter, denn es ist darauf zu achten, daß die verschieden indizierten something in S4 durch verschiedene gebundene Variablen wiedergegeben werden, damit es keine »Überschneidungen« und sonstige Unstimmigkeiten in den Skopi der Quantoren gibt:10 F4

10

Vx (F(x) -[3y [G(x,y) (T(y) v B(y))] 3z [Con(x,z) (Se(z) v St(z)) 3w [Cov(z,w) mit: ,,F(x)" bedeutet soviel wie ist eine Frucht; ,,G(x,y)" bedeutet soviel wie wächst aufy; ,,T(x)" bedeutet soviel wie ist ein Baum; ,,B(x)" bedeutet soviel wie ist ein Busch; ,,Con(x,y)" bedeutet soviel wie enthält y; ,,Se(x)" bedeutet soviel wie sind Kerne; ,,St(x)" bedeutet soviel wie ist ein Stein; ,,Cov(x,y)" bedeutet soviel wie wird von y umhüllt; ,,Su(x)" bedeutet soviel wie ist eine Substanz; ,,E(x)" bedeutet soviel wie ist eßbar.

Su(w) E(w)]]]}

Es sei angemerkt, daß die geschweiften Klammern „{" und „}" in F^ etc. selbstverständlich keine Mengenklammem sind, sondern nur der Übersichtlichkeit wegen benutzt werden.

^2

Matthias Kammerer

Fi kann beispielsweise in folgende Pr nex-Normalform berf hrt werden:11 [Es gelte: x, y e Var; A, B e Forr und Q e {V, 3}.] Auf Grund der quivalenzbeziehung von: (A Λ QxB) Qx(A Λ B), falls χ ί Frei(A), gilt: p;a Vx (F(x) - 3y 3z 3w [[G(x,y) Λ (T(y) v B(y))] Λ [Con(x,z) Λ (Se(z) v St(z)) Λ [Cov(z,w) Λ Su(w)AE(w)]]]} Auf Grund der quivalenzbeziehung von: (A -» 3xB) 3x(A -» B), falls χ £ Frei(A), gilt: p;b Vx 3y 3z 3w (F(x) - [[G(x,y) Λ (T(y) v B(y))] Λ [Con(x,z) Λ (Se(z) v St(z)) Λ [Cov(z,w) Λ Su(w)AE(w)]]]} Mit der Regel zum Beseitigen von Existenzquantoren sowie nach dem Streichen von berfl ssigen Klammern wird aus Y'Ab: F;C

Vx [F(x) - [G(x,g,(x)) Λ (T(g,(x)) v B(g,(x))) Λ Con(x,g2(x)) Λ (Se(g2(x)) v St(g2(x))) Λ Cov(g2(x),g3(x)) Λ Su(g3(x)) Λ E(g3(x))]J

FJC kann dann seinerseits auf die bekannte Art und Weise in CNF gebracht werden, so da man erh lt: F;„

Vx fhF(x) v G(x,g,(x))] Λ hF(x) v T(g,(x)) v B(g,(x))] Λ hF(x) v Con(x,g2(x))] Λ hF(x) v Se(g2(x)) v St(g2(x))] Λ hF(x) v Cov(g2(x),g3(x))] Λ hF(x) v Su(g3(x))] Λ [^F(x) v E(g3(x))]]

Geht man nun von FJC aus und fuhrt darauf auf Grund der pr dikatenlogischen Beziehung F5 {Vx [A(x) - (B(x) A C(x))]} h Kal Vx [A(x) - B(x)] eine »Semzerlegung« durch, so erh lt man folgende Ergebnisse: F^a F;„ F;C F;d F^ F;f F'6&

Vx [F(x) - G(x,g,(x))] Vx[F(x)-(T( gl (x))vB( g ,(x)))] Vx [F(x) - Con(x,g2(x))] Vx [F(x) - (Se(g2(x)) v St(g2(x)))] Vx [F(x) - Cov(g2(x),g3(x))] Vx [F(x) - Su(g3(x))] Vx [F(x) - E(g3(x))]

Diese Formeln k nnten beispielsweise wie folgt interpretiert und versprachlicht werden: Sft,, Set,

11

Wenn etwas1 eine Frucht ist, dann w chst etwas1 auf etwas2, das funktional von etwas1 abh ngig ist. Wenn etwas1 eine Frucht ist, dann ist etwas2, das funktional von etwas1 abh ngig ist, ein Baum oder ein Busch.

Cf. hierzu: Menzel (1990, 90). Die Gewinnung von Pr nex-Normalformen ist nicht immer eindeutig bestimmt, so da man aus: Vxp(x)-Vyq(y) mit Anwendung von: A -» Vx B Vx (A -> B) und dann: Vx A -* B 3x (A -* B) sowohl: Vy3x(p(x)~ q (y)) als auch mit der Anwendung von: Vx A—-B 3x (A—B) und dann: A—Vx B Vx (A—B): 3xVy(p(x)- q (v)) erh lt! (Menzel 1990, 92.)

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«... Sfo

t?

Wenn etwas1 eine Frucht ist, dann enthält etwas1 etwas2, das funktional von etwas1 abhängig ist.

S6d S^ Sgf SL

Wenn etwas1 eine Frucht ist, dann ist etwas2, das funktional von etwas1 abhängig ist, Kerne oder ein Stein. Wenn etwas1 eine Frucht ist, dann wird etwas1 von etwas2, das funktional von etwas1 abhängig ist, umhüllt. Wenn etwas1 eine Frucht ist, dann ist etwas2, das funktional von etwas1 abhängig ist, eine Substanz. Wenn etwas1 eine Frucht ist, dann ist etwas2, das funktional von etwas1 abhängig ist, eßbar.

Hier wird deutlich, daß der funktionale Zusammenhang, der zwischen dem Antezedenz und dem Sukzedenz besteht und der bei Schnelle nicht berücksichtigt wurde, da er die Aussagenlogik für seine Formalisierung zugrunde gelegt hat, in der Prädikatenlogik klar zu Tage tritt. Wenn eben etwas eine Frucht ist, dann ist etwas anderes ein Baum oder ein Busch, aber dieses „etwas anderes" ist nicht irgend etwas beliebig anderes, sondern steht in einer funktionalen Abhängigkeit von dem Gegenstand, der eine Frucht ist. Damit scheint die Prädikatenlogik den Standpunkt von Schnelle zu stützen, da sie eine Beziehung explizit macht, die bei den Ausführungen Schnelles in der Aussagenlogik zu kurz kamen. Dennoch ist damit m.E. nicht allzuviel gewonnen. Trotz der Explizitmachung des funktionalen Zusammenhangs zwischen etwas2 und etwas1 bleiben S^ bis Sgg auf sonderbare Weise unbestimmt, und zwar in genau dem Maße, in dem die einstelligen Funktionen g,(x), g2(x) und g3(x) unbestimmt sind. Es wird m.a.W. nur ausgesagt, daß etwas2 von etwas1 funktional abhängig ist, aber es wird nicht gesagt, welcher Art diese Abhängigkeit ist. Aber genau das ist es, was für das Verständnis wichtig und grundlegend ist. Es ist eben von fundamentaler Bedeutung, daß, wenn etwas eine Frucht ist, etwas anderes ein Busch oder Baum ist und daß dieses „etwas anderes" nicht nur in irgendeiner funktionalen Abhängigkeit zur Frucht steht, sondern in der spezifizierten funktionalen Abhängigkeit des wächst anyo.H.l Damit bleiben sowohl S^ bis S'6g als auch F'6t bis F^ unbefriedigend.

l .2. Kritik der Korrektheitsthese Als erstes soll untersucht werden, ob die Korrektheitsthese KT tatsächlich mit allen ihren Konsequenzen behauptet werden kann. Dazu werden S7a bis S7J näher betrachtet. Während S7a, S^, S7i und S7j sinnvoll sind, sind zwar S7c bis S7h korrekt, aber in dem Sinne sinnlos, wie z.B. der (logisch korrekte) Satz Sg

Wenn etwas eine Orange ist, dann ist etwas anderes hart.

auf den ersten Blick sinnlos ist (dies ändert sich, wenn man annimmt, daß mit etwas anderes 'Kerne' gemeint sind). Dies liegt zum einen daran, daß einerseits something im Sukzedenz das something aus dem Antezedenz anaphorisch nicht aufnimmt, wie z.B. in S^, was aber in der Aussagenlogik unter den Tisch fällt und nur in der Prädikatenlogik berücksichtigt werden kann (cf, z.B. F'6J, und daß andererseits die Aussagen derart allgemein sind, daß sie (fast) nichts zur Bedeutungserläuterung beitragen, wie z.B. in S7d. Außerdem zeigt sich an S7a bis S7j ein weiteres Problem.

54

Matthias Kammerer

Mit S7d und S7e wird behauptet, daß Orangen einen Stein oder Kerne enthalten, bzw. mit S7b und S7c, daß sie auf Bäumen oder auf Büschen wachsen. Diese Propositionen sind - in der natürlichen Sprache ausgedrückt - falsch, denn Orangen enthalten nicht einen Stein oder Kerne, sondern sie enthalten ausschließlich Kerne und keinen Stein, und sie wachsen nicht auf Bäumen oder auf Büschen, sondern auf Bäumen! Das natürlichsprachliche oder funktioniert mithin nicht immer wie das logische vel, das wahr ist, wenn zumindest eines seiner beiden Argumente wahr ist. (N.B.: Damit wird nicht behauptet, daß das natürlichsprachliche oder nicht einmal wie der logische Operator hinsichtlich der Wahrheitsfunktionalität funktionieren kannl) Hier, und ganz besonders in Texten der Textsorte Wörterbuch- sowie Lexikonartikel, ist es so, daß Alternationen - wie z.B. in der Logik - nicht einfach nach Belieben hinzugefügt werden können, ohne daß sich dadurch der Wahrheitswert ändert. So ist es beispielsweise in einem Schriftstellerlexikon nicht möglich, ja geradezu falsch, wenn dort geschrieben stünde Johann Wolfgang von Goethe starb 1832 oder 1833. Hingegen ist es völlig korrekt zu sagen: Der Maler Hans Holbein d. J. wurde 1497 oder 1498 geboren, da nicht sicher ist, ob er 1497 oder 1498 das Licht der Welt erblickte, wohingegen es sicher ist, daß er nicht vor 1497 und nicht nach 1498 geboren ist. Die Hinzufügung einer weiteren Alternation wie dies in der Logik jederzeit möglich ist, ohne den Wahrheitswert zu ändern - würde eine falsche Proposition versprachlichen, nämlich: Hans Holbein d. J. wurde 1497 oder 1498 oder 1499 geboren, denn Hans Holbein d. J. wurde ganz gewiß nicht 1499 geboren. M.a.W.: In den genannten Fällen der natürlichsprachlichen Alternation gilt für die Argumente, daß der Sprecher, der sie äußert, nicht der Ansicht sein darf, daß sie falsch sind, wenn er die Wahrheit sagen und informativ sein will, kurz: wenn er den Grice'schen Kooperationsprinzipien, nämlich den Maximen der Quantität, der Qualität, der Relation und der Modalität folgt (Grice 1993, 249seq.). Selbst wenn man das natürlichsprachliche oder als exklusive awf-aw/-Alternation interpretieren würde, ändert sich an der hier geäußerten Kritik nichts Wesentliches. Schließlich ist darauf hinzuweisen, daß die Hyperonymie-Relation, die Schnelle wohl als transitiv ansieht, Aussagen ableiten läßt, die m.E. in der natürlichen Sprache nicht gelten. Beispielsweise gilt in unsuellen Benennungskontexten nicht, daß Orangen Dinge sind, 12 wie man über das Verfolgen von Hyperonymieketten über die Transitivitätseigenschaft folgern könnte. Damit dürfte an einem Gegenbeispiel gezeigt sein, daß KT nicht erfüllt ist, denn es sind Sätze S aus der Menge der Bedeutungsangaben BA in Kai (syntaktisch) ableitbar, für die nicht gilt, daß jedes Modell der Sprache L auch ein Modell des Satzes S in der semantischen Folgerung bzgl. ist - oder: Nicht jede wahre Interpretation von L ist dann auch eine wahre Interpretation von S bzgl. . Wenn diese Untersuchungsergebnisse korrekt sind, kann den Ausführungen in Schnelle (1997) nicht zugestimmt werden, in denen es u.a. heißt: 12

Ausgenommen sind davon Philosophie- und Linguistik-Seminare, in denen es selbstverständlich ist, daß Orangen Dinge und Gegenstände sind. - Gemäß meinen Befragungen von Sprechern des Deutschen ist es jedoch nicht sinnvoll (und wahrscheinlich sogar falsch), als Hyperonym zu Orangen in unsuellen Benennungskontexten (!) - Gegenstand oder Ding anzunehmen.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

55

[Z 5: Schnelle 1997, 145] So sichert die Ersetzung, daß die Eigenschaften, die im Defmiens-Satz \onfruit.l eingeführt werden[,] vom Defmiens-Satz von orange.2 ererbt werden. Dies gilt ganz allgemein: Die numerierten Terme oder Wort-Lesarten in Defmiens-Sätzen können immer durch den entsprechenden Teil des Definiendum[-]Satzes [recte: Defmiens-Satzes; M. K.] zu diesem Term ersetzt werden - bzw. durch die entsprechenden Teile in den Implikanda und Implikaten der zugehörigen wenn..dannSätze. Durch diese Ersetzungen können viele neue wenn..dann-Sätze erzeugt werden, die im Wörterbuch nicht explizit formuliert wurden. Die auf diese Weise formal aufgrund der Prinzipien der Logik generierten Sätze sind auch wahr in der Sprache, wenn die explizit im Wörterbuch formulierten Sätze wahr sind. Wenn wir die die [!] Gesamtmenge aller so implizierten Sätze nehmen und sie als Theorie des semantischen Wortwissens verstehen, dann haben die WörterbuchErklärungen, aus denen sie generiert wurden, den Status von Axiomen. In dieser Sicht ist das [!] ein Wörterbuch eine Art axiomatischer Theorie, ein Teil der axiomatischen Theorie des Wortwissens der Sprache.

1.3. Kritik der Vollständigkeitsthese Nun geht es darum zu zeigen, daß auch die Vollständigkeitsthese VT ungelöste Probleme aufwirft. Dies ist relativ einfach, indem wiederum ein Gegenbeispiel gegeben wird. Dazu betrachte man die folgenden beiden Wörterbuchartikel: [E3:COBUILD551 a ] lenvon /'leman/, lemons, l A lemon is a bright yellow fruit with sour juice. See picture headed FOOD. EG...slices ft offllemon. 2 Lemon is a drink that tastes of lemons EG ... glass of lemon [E4:COBUILD 159b] cit-rus fruit /'sitras 'fruit/, citrus fruits. A citrus fruit is a juicy, sharp-tasting fruit such as an orange, lemon, or grapefruit. EG Bananas and citrus fruits are its only cash crops.

N COUNT OR N UNCOUNT

citrus fruit N UNCOUNT

N N

COUNT OR ^01™1"

Mit lemon in der Bedeutung, wie sie in der Bedeutungsangabe im ersten semantischen Subkommentar versprach licht ist, können u.a. folgende wahre Sätze gebildet werden: S9 S ]0

Lemons have Vitamine C. Lemons are healthy.

S9 und S,0, die zu den wahren Sätzen der (natürlichen) englischen Sprache gehören, können nicht aus E 3, E 4 und E 2 syntaktisch abgeleitet werden. Schnelle (1997, 151) würde an dieser Stelle wohl argumentieren, daß S9 und S10 enzyklopädisches Wissen versprachlichen und kein lexikalisches Wissen (Schnelle spricht konkret von „Fakten-Feststellungen", cf. ibd., 151, bzw. auch von „Wortwissen", cf. ibd., 145). Seiner Meinung nach gehört in ein Sprachwörterbuch nur sprachliches Regelwissen. Aber die strikte und intersubjektiv verifizierbare Trennung zwischen enzyklopädischem und lexikalischem Wissen ist bekanntlich

56

Matthias Kammerer

nicht ganz einfach und bisweilen sogar unmöglich, ja es sollte in einem (Sprach) Wörterbuch erst gar nicht durchgeführt werden. 13 Ganz in diesem Sinne wäre nämlich sonst zu fragen, ob S7i oder S7j nicht auch zum enzyklopädischen Wissen gehören und damit aus einem Sprachwörterbuch getilgt werden müßten, oder - anders argumentiert -: Wenn S7i und S7j zum lexikalischen Wissen gehören, was spricht dann dagegen, S9 und S10 nicht auch zum lexikalischen Wissen zu rechnen? - Man betrachte nun den folgenden Satz: S |,

A rotten orange don't taste sweet.

Der Satz S,, ist zweifellos wahr. Aber es stellt sich die Frage, von was hier die Rede ist. Nach Schnelle ist etwas, wenn es eine Orange ist, u.a. genuin süß, d.h.: Süß-sein ist eine notwendige Bedingung für Orange-sein. Orange-sein impliziert demnach süß-sein. In S,, wird jedoch von faulen Orangen ausgesagt, daß sie nicht süß sind. Wird m.a. W. in S n von etwas gesprochen, das alles mögliche ist, aber nur keine Orange, da diese ja angeblich grundsätzlich süß sind? Sicherlich nicht! Auch faule Orangen, die nicht süß sind, sind und bleiben bis zu einem unbestimmten Zustand des Verfaulens Orangen. (Wann eine verfaulende Orange aufhört, eine Orange zu sein, ist genauso schwierig zu beantworten wie die Frage, wann ein Mensch beginnt, ein Mensch zu sein, und wann er wieder aufhört, ein Mensch zu sein.) S u ist also - so nehme ich an - für alle Sprecher des Englischen wahr und unproblematisch. Aber man braucht nicht einmal solche etwas gesuchten Beispiele zu betrachten; selbst ein Satz wie S,2

Navelapfelsinen sind Orangen.

ist nach S7d und S7e falsch, da Orangen, die ja Früchte sind, angeblich Steine oder Kerne enthalten. Navelapfelsinen zeichnen sich allerdings dadurch aus, daß sie keine Kerne (und erst recht keine Steine) enthalten. Mit S,, und S12 sind nun nicht nur zwei Beispiele für einen Satz S aus einer Sprache L gegeben worden, die mittels Kai aus den Bedeutungsangaben von COBUILD nicht nur nicht abgeleitet werden können, sondern die im Kalkül sogar den Wahrheitswert FALSE zugewiesen bekämen. Nun könnte entgegnet werden, daß lediglich vergessen wurde, die propositionalen Gehalte von S9 und S10 in E l zu versprachlichen, was durchaus einmal passieren kann, wenn man bedenkt, wie umfangreich COBUILD ist. Würde man nun diese propositionalen Gehalte in der Bedeutungsangabe von E l berücksichtigen, so die Argumentation weiter, dann wäre das axiomatische System vollständig. Dem ist aber entgegenzuhalten, daß es noch viele weitere propositionale Gehalte gibt, die in der Bedeutungsangabe von E l nicht be13

Cf. hierzu z.B.: Bergenholtz/Kaufmann (1996); Habel (1985, 444 f.); Haiman (1980; 1982) u. dazu: Frawley (1981); Lüdi (1985, 91-94); Rossenbeck (1994) u. Wiegand (1988a; 1989d, insbes. 550-552; 1994). Eco (1985, 77-132) kommt zu dem Ergebnis, daß in einem Wörterbuch sinnvollerweise immer enzyklopädisches Wissen enthalten sein muß. Ein (gutes, hilfreiches, semantisches ...) Wörterbuch ist deshalb immer „eine verkleidete Enzyklopädie" (ibd., 107). -Auch Aitchison (1997, 295), der zwischen semantischem Wissen und konzeptuellem Wissen grundsätzlich unterscheidet, konstatiert, daß eine genaue Grenzziehung unmöglich ist: „Es läßt sich daher nicht sagen, wo die »Bedeutung« eines Wortes endet und das Allgemeinwissen beginnt."

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

57

rücksichtigt wurden, nämlich z.B. daß Orangen neben Vitamin C drei Zuckerarten, fünf Säuren, 13 Mineralstoffe und Spurenelemente sowie sechs Fermente enthalten (cf. Kranz 2 ExotFrGem, 232). Aber selbst dann, wenn man dies alles in der Bedeutungsangabe berücksichtigen würde, würde das axiomatische System der Menge der Bedeutungsangaben nicht vollständig werden, denn es könnten leicht weitere propositionale Gehalte aufgezählt werden, die die Vollständigkeitsthese widerlegen würden. Angesichts dieser Beobachtung ist zu vermuten, daß es wohl nur äußerst schwer, wenn nicht gar unmöglich ist, das axiomatische System durch exhaustive Aufzählung der wahren propositionalen Gehalte innerhalb der Bedeutungsangaben zu vervollständigen. In Schnelle et al. (1996, 193) bemerkt er im Anschluß an Quine, daß für Wenn... dann...-transformierte Sätze gemäß der Gebrauchsregel: Wenn jemand p sagt, meint er damit q. gilt: Wenn (jemand sagt) „etwas ist eine Orange"[,] dann (meint er damit praktisch dasselbe wie) „es ist eine orangefarbene, runde Frucht, die saftig und süß ist". Nun, eine solche Beziehung besteht nur in den seltensten Fällen; man vergleiche hiermit die nachstehenden Zitate aus dem Spiegel sowie aus der F.A.Z. (wobei davon ausgegangen wird, daß das Sagen und damit die Behauptung von Orange gleichbedeutend mit dem obigen Ausdruck etwas ist eine Orange ist): • F. A. Z. vom 27. l. 1993, S. N2: Exotische Früchte wie Orangen, Kiwis und Melonen drängten zu Lasten heimischer Obstsorten auf den Markt. • F. A. Z. vom 3. 2. 1993, S. N l: Beta-Carotin ist vor allem in Fisch, Leber, Milch und Eigelb, in Gemüse wie Karotten, Spinat, Kresse und Broccoli sowie in Tomaten, Paprika, Aprikosen und Orangen enthalten. • F.A.Z. vom 31.3. 1993, S. N3: Man muß kein Chemiker sein, um Zitronen und Orangen an ihrem Geruch zu erkennen. • F. A. Z. vom 19. 5. 1993, S. N2 : Ginseng, Jujube und Orangen zum Beispiel rühmte er mit den Worten: „Diese königlichen Pflanzen gleichen dem Himmel; sie dürfen so lange eingenommen werden, wie man es für gut findet, sie können nicht schaden.''1' • F. A. Z. vom 2. 8. 1995, S. N3: Bei Baumwolle und Orangen verbesserten sich die Erträge um knapp 50 Prozent. • F.A.Z. vom 2. 8. 1995, S. N3: Während Baumwolle, Weizen und Orangen in den Experimenten immer ausreichend gedüngt wurden, begrenzt in Wald- oder Steppenvegetation wahrscheinlich das Fehlen wichtiger Nährstoffe das Wachstum. • Der Spiegel 4/1995, S. 136: Simpson hat auch für Orangensaft im Fernsehen geworben, für Orange Juice. • Der Spiegel 6/1995, S. 222: Cläre O'Sullivan hatte vor einem Jahr eine Diät mit nur WO Kalorien am Tag begonnen, meist ein Apfel und eine Orange. • Der Spiegel 7/1995, S. 172: In der Braunschweiger Farbenfabrik Auro riecht es wie auf einem orientalischen Basar. Ein Hauch von Orangen und Limonen liegt in der Luft. • Der Spiegel 7/1995, S. 172: Verschreckt durch Holzschutzskandale und die Warnung vor benzolhaltigen Farben, greifen umweltbewußte Heimwerker zunehmend zu „Leinöl-Firnis'''' oder Pinselreinigern aus „Orangenschalenöl in Lebensmittelqualität" (Eigenwerbung der Firma Leinos Naturfarben).

58

Matthias Kammerer • Der Spiegel 7/1995, S. 173: Die Firma Volvox wirbt gar in Gourmetmanier für Bereinigte kalte Pressungen aus Orangenschalen". • Der Spiegel 7/1995, S. 173: Gesundheitsängste von Anwohnern in einem Kleingartenverein (Braunschweiger Zeitung: ätherische Orangenöle stinken den Kohlrabizüchtern'') konnte der Mann nur mühsam zerstreuen. • Der Spiegel 7/1995, S. 173: Einer US-Studie zufolge löst Orangenschalenöl nicht nur akute Atemnot und Kopfschmerzen aus, sondern — zumindest bei Ratten — auch Nierenkrebs. • Der Spiegel 9/1995, S. 12: Die sehr sorgfältig erarbeitete US-Studie und andere Recherchen zu dem Inhaltsstoff Limonen resümieren keine gentoxische oder krebserzeugende Wirkung. Sonst wäre wohl auch das vielseitig angewendete Auspressen von Orangensaft aus der ungeschälten Frucht untersagt worden. • Der Spiegel 13/1995, S. 122: Bild ortete ihn [i.e. Schneider; M. K.] im Iran oder auf seiner Jacht im Persischen Golf, unterm Sonnensegel auf dem Achterdeck mit fangfrischen Shrimps und frischem Orangensaft • Der Spiegel 13/1995, S. 166: Abgetrennte Köpfe werden nachts auf Bürgersteigen deponiert oder auf einen Zaun gesteckt, Hände und Arme mit Draht an Bäumen oder Laternen aufgeknüpft. Der dazugehörige Rumpf kann kilometerweit entfernt in einem Orangenhain liegen. • Der Spiegel 13/1995, S. 171: Blida, 80 Kilometer von Algier entfernt, eine Stadt, um die herum paradiesisch üppig Reis, Rosen, Orangen und Wein gedeihen, ist eine Hochburg der Islamisten. • Der Spiegel 17/1995, S. 226: So wie an diesem regnerischen Morgen in einer Hamburger Hotellobby, wo er ziemlich mürrisch und mit einer strengen Studentenbrille im Milchgesicht vor einem Orangensaft sitzt und widerwillig Auskunft gibt über seinen frischen Ruhm: „Uns geht es um Musik, Mann?' • Der Spiegel 19/1995, S. 142: Doch die Bananen und Orangen verrotten am Baum, die Auberginen bleiben auf den Feldern liegen. • Der Spiegel 19/1995, S. 142: Locker könnten unter den duftenden Orangen- und Hibiskusbäumen 250 Gäste gleichzeitig bewirtet werden. • Der Spiegel 21/1995, S. 129: ,fJa,ja,jaa", antworten sie im Chor, bis die Assistenten an jeden ein Glas mit den in Orangensaft aufgelösten 150 Milligramm MDE austeilen. • Der Spiegel 22/1995, S. 115: Irgendwann soll es auch Orangen geben, die niemand mehr schälen muß: Sie bestehen nur noch aus Fruchtfleisch. • Der Spiegel 24/1995, S. 197: Alle Früchte, bis auf die Orange, sind von Isaacson grau bemalt. • Der Spiegel 26/1995, S. 166: Mit klapprigen Pritschenwagen fliehen die „Okies" vor Ausbeutern und Bodenerosion ins gelobte Kalifornien - schon [sind; M. K.] sie auf der Suche nach neuem Lebensstil: „Viel Arbeit und alles schön und grün und kleine weiße Häuser und überall Orangen." • Der Spiegel 26/1995, S. 166: Sie haben Orangen satt. • Der Spiegel 27/1995, S. 175: In dem von Bettina Schäfer präzise übersetzten Roman erscheint die Mutter als eine unberechenbare und amoralische Überlebenskünstlerin: „Bei ihr war die Liebe etwas zum Anfassen, rund und duftend wie eine Orange. [...]" • Der Spiegel 31/1995, S. 130: Kommissar Deraldo Peralha, einst Chef des Polizeireviers in Copacabana, bekämpfte die Sünde mit einer Apfelsine. Männer, die in zu engen Hosen über den Strand schlenderten, zwang Peralha, eine Orange hinter den Hosenbund zu stecken. Fiel die Frucht durch das Beinkleid zu Boden, konnte der Verdächtige des Weges ziehen. Blieb die Orange stecken, mußte er eine Nacht auf dem Revier verbringen. Das obszöne Kleidungsstück ließ der Ordnungshüter dann zerschnippeln.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

59

• Der Spiegel 40/1995, S. 236: In dieser Situation wurde Mein vom SPIEGEL gefragt, was er sich wünschen würde, wenn er drei Wünsche frei hätte. Ein gewöhnlicher Bürger der DDR hätte wahrscheinlich geantwortet: Bananen, Orangen und Schokolade, die diesen Namen verdient. Ein Intellektueller hätte gesagt: Freie Wahlen, freie Presse, freies Reisen. • Der Spiegel 42/1995, S. 251 [i.e. Zitat eines Kochrezeptes]: 25 dkg Zucker wurde an der Schale einer Orange und einer Zitrone abgerieben, den Staubzucker mußte man noch gut absieben, damit keine Klümpchen blieben. • Der Spiegel 52/1995, S. 86: Die Karriere in den deutschen Trümmer-Filmen unmittelbar nach dem Krieg vertauschte die Knefmil der Eheflucht in die USA, die in den Hungerjahren als ein Paradies galten, in dem es schicke Nylons und gute Zigaretten, Orangen und Steaks gab. In den Zitaten ist zwar immer von Orangen die Rede, aber in den seltensten Fällen werden damit orangefarbene, runde Früchte gemeint, die saftig und süß sind. Damit dürfte VT hinreichend widerlegt sein.

Unabhängig von den bisherigen Ausführungen soll noch auf folgende zwei Schwierigkeiten aufmerksam gemacht werden, wobei Anmerkung (1) insofern berechtigt ist, als die Ausführungen in Schnelle (1997) auch Leseanweisungen für menschliche Benutzer darstel-

len: (l)

Mehrere Ersetzungen fuhren schnell zu unübersichtlichen und unverständlichen Satzgebilden.

Dies dürfte unmittelbar einsichtig sein und soll lediglich an Hand eines Beispiels demonstriert werden. Ausgegangen wird von S6. In S6 werden nun round^ grows^ tree^ bush\, can{, eat0 undyw/cy, durch den entsprechenden Teil in den dazugehörigen Bedeutungsangaben ersetzt. Auf diese Weise erhält man das unübersichtliche und nicht mehr verstehbare Schema S13: S 13

IF something' is an orange: THEN something1 is shaped like a ball OR something1 is shaped like a circle 1· OR something1 is curved like a letter J AND something1 increases in size on something2 something; is a tall plant J , something2 is a large plant 1 AND something has a hard trunk I^ ^„^ · sma|]er ^ a ^ AND something; has branches {^ ^^2 has a ,ot of brapchesJ LAND something' has leaves J AND something1 contains? something3 AND something3 are seeds or something^ is a stone3 AND something3 is covered2 by something4 AND something4 is a substance( it is possible for you to do itj something4 into your mouth (

AND °R y°" haVe ?e-SklH ° ab.l,ty to do ,t lOR you are allowed to do it AND something 1 has a lot of juice in it AND something' contains a lot of sugar.

AND you chew something4 M AND you swallow something4 J

Matthias Kammerer

60

Aber nicht nur dann, wenn innerhalb einer Bedeutungsangabe mehrere „inhaltstragende Wörter" ersetzt werden, entstehen schwer verständliche bis unverständliche Sätze, sondern auch dann, wenn mehrmals nacheinander Ersetzungen durchgeführt werden. So läßt sich aus S6 der Satz S,4 ableiten: S 14

IF something 1 is an Orange 2 THEN something 1 is juicy,.

Ersetzt man nun juicy t in S,4 durch den entsprechend transformierten Teilsatz der dazugehörenden Bedeutungsangabe aus dem COBUILD-Wörterbuch, so erhält man: S ,5

IF something 1 is an orange, THEN something' has a lot of juice, in it.

Diese Transformationen und Ersetzungen können mehrere Male nacheinander durchgeführt werden; z.B.: S,6

IF something 1 is an orange,

THEN something' has a lot of something 2 in it AND something 2 is a liquid, AND the liquid, can be obtained from a fruit or a plant

'17

IF something' is an orange 2

THEN something' has a lot of something 2 in it substance, AND the substance, is like water AND something 2 is a AND the substance! is not solid AND the substance, flows LAND the substance, can be pouredJ AND the substance, can be obtained from a fruit or a plant

S, g

IF something' is an orange 2

AND

THEN something' has a lot of something 2 in it

something2 is a solid OR something2 is a powder [ .OR something2 is a liquid f something2 is a solid AND j OR something2 is a powder }· is like water LOR something2 is a liquid something2 is a solid AND OR something2 is a powder f js not solid LOR something2 is a liquid something2 is a solid AND "i OR something2 is a powder }· flows [OR something2 is a liquid something2 is a solid 1 AND j OR something2 is a powder > can be poured LOR something2 is a liquid J is a solid 1 OR is a powder ? can be obtained from a fruit or a plant OR is a liquid J

(

(2)

Nicht jedes Wort des then... - Teils kann sinnvoll ersetzt werden.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

61

Diese Tatsache hat auch Schnelle erkannt, weshalb er nur die „inhaltstragenden Wörter" indiziert und nicht z.B. den unbestimmten Artikel oder eine Konjugationsform des Verbs to be. Wollte man dies dennoch bewerkstelligen, so scheitert man bereits daran, den gegebenen Wörterbuchartikel in eine Wenn...dann...-form zu transformieren, die eingesetzt werden könnte. Man vergleiche hierzu die beiden folgenden Wörterbuchartikel zu dem indefiniten Artikel a und dem Pronomen U: [E5:COBUILD l a ] a /ei, 9/, an /aen, an/. DET 1 You use a or an at the beginning of noun groups when you are referring to someone or something and you do not want to sy which particular person or thing your mean. EG Tom could see a hall-way-.. She wanted to be an actress. 2 You can use a or an instead of the number One'. DET EG... a year or two ago in an hour 's time. 3 When you express rates, prices, and measures- DET ments, you can use a or an to say how many units = Per apply to each of the items being measures. EG He charges 100 dollars an hour 4 You can use a or an in front of uncount nouns DET when they are preceded by adjectives or followed by words that describe the uncount non more fully. EG...a happiness that he couldn quite hide. 5 You use a or an in front of the names of people DET when you are referring to someone who you do not = a certain know personally. EG You dont'/ know a Mrs Burton, do you? [E6:COBUILD518 a ] it /it/ is used as the subject of a verb or as the PKON : SING object of a verb or preposition. 1 You use it 1.1 to refer to an object, animal, or other thing that has already been mentioned or whose identity is known. It can also be used to refer to a child. EG...a tray with glasses on it... The man went up to the cat and started stroking it... The strike went on for a year before it was settled. 1.2 to refer to a situation or fact, or to say something about a situation or fact. EG She was frightened, but tried not to show it... It was very pleasant at the Hochstadts... It took Simon some time to work out what she meant... He found it hard to make friends... It was a pity that her spelling was so bad... It doesn 't matter.

62

Matthias Kammerer 2 You also use it, mainly as the subject of the verb PRON : SING 'be', 2.1 when you are making statements about the weather, the time, the date, or the day of the week. EG It's hot... It's raining here ...It is nearly one o'clock... It's the 6th of April today. 2.2 when you are stating or asking who is speaking on the telephone or who is present. EG It's me - Mary... Who is it? 2.3 when you are emphasizing or drawing attention to something. EG It's my mother I'm worried about.

1.4. Bemerkungen zur Axiomenthese Was ein Axiom ist, ist gar nicht so eindeutig, wie es u.U. auf den ersten Blick erscheint. Szabo (1971, 73&seq.) behauptet, daß ursprünglich mit dem Wort Axiom (gr. ) ein Satz bezeichnet wurde, der nur von einem der beteiligten Kommunikationspartner als wahr angesehen wurde. Damit waren Axiome gerade jene Sätze, deren Gültigkeit nicht fest stand, sondern von den übrigen Kommunikationsteilnehmern bezweifelt oder sogar bestritten wurde. Erst mit Aristoteles, so Szabo weiter, bezeichne der Ausdruck Axiom einen Satz, „[...] der eines Beweises weder fähig noch bedürftig sei [...]" (Szabo 1971, 738). OeingHanhoff (1971, 741 seq.) konstatiert bei Aristoteles folgende zwei Verwendungsweisen des Ausdrucks Axiom: einerseits bedeutet er in der Topik soviel wie 'Annahme' oder 'Forderung', in der Zweiten Analytik hingegen ,,[...] einen Ausgangssatz der beweisenden Wissenschaften, der wahr, ein erster, unvermittelt, einsichtiger und früher als das aus ihm Bewiesene ist und dieses begründet." (Oeing-Hanhoff 1971, 741.) Die Geschichte dieses Begriffs bleibt weiterhin sehr bewegt. In der Mathematik z.B. werden unter Axiomen keine selbstevident wahre Aussagen verstanden, wie das noch bei Aristoteles der Fall war, sondern sie sind als Konventionen konzipiert (Oeing-Hanhoff 1971,747). Hier soll unter Axiom das verstanden werden, was bereits eingangs im Zitat aus Kondakow 2 WBLogik gesagt wurde: ein Axiom sei ein Satz eines Kalküls, der - ohne bewiesen zu sein - als wahr angenommen wird und seinerseits zur Deduktion weiterer wahrer Sätze dient. Die Axiomenthese AT soll hier nicht explizit widerlegt werden, es sei aber Folgendes zu bedenken gegeben: • In Anbetracht der Tatsache, daß es sich bei den Bedeutungsangaben im COBUILDWörterbuch um Sätze der natürlichen Sprache handelt, mutet es zumindest eigenartig an, daß diese axiomatischen Status haben sollen und damit die natürliche Sprache als ein Kalkül angesehen wird (cf. z.B. auch den Titel von Montagues Arbeit English as a formal language, ein Ansatz, der letztlich gescheitert ist). • Außerdem wäre zu klären, ob es im COBUILD-Wörterbuch auch Axiome gibt, die man tilgen könnte, ohne daß das Kalkül unvollständig werden würde. Eine solche Untersuchung wäre darauf ausgelegt, ein Kalkül zu entwerfen, das mit möglichst wenigen Axiomen auskommt und trotzdem vollständig und korrekt ist.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

63

• Schließlich ist noch daraufhinzuweisen, daß es sich - da KT und VT falsifiziert werden konnten - bei den Bedeutungsangaben im COBUILD-Wörterbuch nicht um Axiome der natürlichen Sprache handeln kann. Denn wären sie Axiome der natürlichen Sprache, dann müßte sowohl KT als auch VT gelten.

Am Ende dieser Kritik ist daraufhinzuweisen, daß mit dem durch Schnelle vorgeschlagenen Interpretationsansatz von Bedeutungsangaben in Wörterbüchern vom COBUILD-Typ bereits praktisch erfolgreich gearbeitet wird (cf. Schnelle et al. 1996, insbes. 195-211, oder Hoelter 1996). Insofern wäre zu fragen, ob die Praxis nicht die hier vorgebrachten theoretischen Einwände bereits widerlegt hat. Doch tatsächlich werden in der Praxis »lediglich« 14 die Wörterbuchartikel aus Collins CobuildStudent's Dictionary in ein HPSG-Format überführt, wobei u.a. zwischen dem Antezedenz (als left-hand side) und dem Sukzedenz (als right-hand side) der Implikation unterschieden wird. Deduziert wird jedoch daraus noch nichts, insbesondere noch nicht die Menge der wahren Sätze der englischen Sprache. Vielmehr dient die HPSG-Repräsentation dazu, mittels Unifikation Sätze wie The farmer shot the bull.

als korrekt i.S.d. Bedeutungsangaben in Wörterbüchern vom COBUILD-Typ zu erkennen (also entsprechend der Axiomenthese), und Sätze wie ''The bull shot the farmer.

als inkorrekt zurückzuweisen, weil shoot in der (einzig hier passenden) Lesart 2 in der logischen Subjektsposition ein Nomen verlangt, das das Merkmal [+ person] besitzt. Zwar kann auf der Grundlage der HPSG-Repräsentation bereits einiges getan werden, wovon Hoelter (1996, insbes. 88-94) ein Beispiel gibt, aber diesem System wird es, wie schon gesagt, nicht möglich sein, allein auf Grund der ihm zur Verfügung stehenden Bedeutungsangaben zu entscheiden, ob Lemons do not have vitamine C. oder Lemons have vitamine C.

(und ungezählt viele weitere Sätze) wahr ist. Insofern kann - um dies zu wiederholen - die Menge der Bedeutungsangaben des COBUILD-Wörterbuches nicht die Axiome einer natürlichen Sprache wie dem Englischen sein. Und außerdem ist es nicht möglich, in sinnvoller Weise darauf eine Logik zu applizieren.

14

Es ist klar, daß der Ausdruck lediglich nicht besagen will, daß das Parsen des Wörterbuches und die Überführung in ein HPSG-Format trivial sind.

64

Matthias Kammerer

2. Anhang 15 Eine gewisse Vertrautheit mit den Zeichen: (,), true, false, -,, Λ, ν, —, wird vorausgesetzt. - So kann nun f r die Aussagenlogik festgelegt werden: Symbol Name Σ Signatur

ForOj;

Menge der aussagenlogischen Formeln ber Σ

W, F I val,

M

15

Wahrheitswerte Interpretation Auswertungsfunktion bez glieh I

Modell

Definition Eine (aussagenlogische) Signatur Σ ist eine abz hlbare Menge Σ von Symbolen: Σ={Ρ0....,Ρη}οα6Γ:Σ={Ρ0.Ρ,....}. Zur Signatur Σ ist ForOs, die Menge der Formeln ber Σ (oder: der Aussagen ber Σ), induktiv definiert durch: (l)true e ΡοΚ)Σ fa|se € ForOz ^ ^ n Σ c ForOj; (2) Mit A, B e Σ sind auch (A Λ Β) € ΡθΓθΣ (A v B) e ForOr (A - B) e ForOj; (A *-* B) e ForOj; [Anmerkung: die Null „0" hinter „For" sei ein Kennzeichen f r Aussagenlogik.] W und F stehen fur die Wahrheitswerte -wahr und falsch. Es sei Σ eine aussagenlogische Signatur. Eine Interpretation ber Σ ist eine beliebige Abbildung Ι: Σ - {W, F}. Zur jeder Interpretation I ber Σ wird nun die zugeh rige Auswertung der Formeln ber Σ definiert als die Abbildung val,: ForOj — {W, F} mit: val,(true) = W vali(false) = F val,(P) = I(P)f r j e d e s P e Σ _ |F falls val, (A) = W Va [W falls val,(A) = F [wie bekannt, wird val, entsprechend definiert f r: (A Λ B), (A v B), (A -» B) und (A B).] Ein Modell einer Formel A e ForOz ist eine Interpretation I ber Σ mit val,(A) = W. Zu einer Formelmenge M c ForOz ist ein Modell von M eine Interpretation I, welche Modell von jedem A e M ist.

Die hier wiedergegebenen Definitionen gehen gr

tenteils w rtlich auf Menzel (1990) zur ck.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

Symbol Name

65

Definition

allgemeingü A e ForO£ heißt allgemeingültig gdw.: val](A) = W für jede Interpretation Itig I über . erfüllbar A heißt erfüllbar über gdw.: es gibt eine Interpretation I über mit vat=

( , L)

Relator der semantischen Folgerbarkeit Sprache

*

Stern Regel; Instanz, Prämisse, Conclusio, Axiom

Kai

Kalkül Ableitung

h

Relator der syntaktischen Folgerbarkeit Korrektheit Vollständigkeit

Es seien eine Signatur, M e ForOL, A e ForO^. So sei definiert: M t= x A [lies: aus M folgt A bezüglich ] gdw.: jedes Modell von M ist auch Modell von A. Eine Sprache ist ein Tupel ( , L), wobei ein endliches oder unendliches Alphabet ist und L eine entscheidbare Teilmenge von *. * ist die Menge aller abstrakten Wörter (d,, d2, ..., d n ) mit dj e und n e N* über dem Alphabet . Für ein n e N ist eine n-stellige Regel in ( , L) eine entscheidbare, n+1stellige Relation über L. Ist R c L""1"' eine solche Regel und ist (u h ..., u n , u n+ |) € R, so heißen - (u , , . . . , un, u n+1 ) eine Instanz von R - U], ..., u n die Prämissen dieser Instanz - u n+l die Conclusio dieser Instanz. Die Instanzen von nullstelligen Regeln heißen auch Axiome. Ein Kalkül Kai einer Sprache ( , L) ist eine endliche Menge von Regeln in dieser Sprache. In einer Sprache ( , L) sei Kai ein Kalkül, und es sei M eine Teilmenge von L. Eine Ableitung aus M in Kai ist eine Folge

von Wörtern in L, so daß für jedes i € { 1, ..., m} gilt: - U; ist Axiom; oder - u: e M; oder - es gibt eine Regel R e Kai einer Stelligkeit n > 1 sowie Indices jh ..., jn e {!,..., i-1}, sodaß e R. u e L heißt ableitbar aus M in Kai, kurz: M H Kal u gdw.: es gibt eine Ableitung (u,. ..., u m ) in Kai mit u m = u. Kai ist korrekt, wenn für beliebige M c ForOj gilt: wenn: M (-Ka, A dann: M £= . Kai ist vollständig, wenn für beliebige M c ForOj gilt: wenn: M = A dann: M h-Ka, A.

66

Matthias Kammerer

3. Literatur 3.1. Corpora FAZ-NuW93-95 = Frankfurter Allgemeine Zeitung für Deutschland. Natur und Wissenschaft. Entwicklungen, Entdeckungen und Ergebnisse der Natur- und Geisteswissenschaften von 1993-1995. CD-ROM. Frankfurt/M: Frankfurter Allgemeine Zeitung GmbH 1996. Spiegel95 = Der Spiegel. 52 Ausgaben im Volltext und Register. 1995 CD-ROM Ausgabe. Hamburg: SPIEGEL-Verlag Rudolf Augstein 1996.

3.2. Wörterbücher Brockhaus 17Enz = Brockhaus Enzyklopädie in zwanzig Bänden. Siebzehnte völlig neubearb. Aufl. des Großen Brockhaus. Wiesbaden: F. A. Brockhaus 1966-1976. Bd. \:A-ATE(]966); Bd. 2: ATF-BLIS (1967); Bd. 3: BLIT-CHOC (\961)\ Bd. 4: CHOD-DOL (1968); Bd. 5: DOM-EZ (1968); Bd. 6: F-GEB (1968); Bd. 7: GEC-GZ (1969); Bd. 8: H-IK (1969); Bd. 9: IL-KAS (1970); Bd. 10: KAT-KZ (1970); Bd. 11: L-MAH (1970); Bd. 12: MAlMOS(1971); Bd. 13: MOT-OSS (\91 \)\ Bd. 14: OST-POQ(1972); Bd. 15: POR-RIS(1972); Bd. 16: RIT-SCHO (1973); Bd. 17: SCHR-STAL (1973); Bd. 18: STAM-TR1E (1973); Bd. 19: TRIFWAL (1974); Bd. 20: WAM-ZZ (1974); Bd. 21: Karten (1976); Bd. 22: Ergänzungen. -/(1975); Bd. 23: Ergänzungen. ./-Z (l976); Bd. 24: Bildwörterbuch der deutschen Sprache (1976). COBUILD = PONS Cobuild English Learner's Dictionary. [Ed. in chief: John Sinclair.] Stuttgart: Klett; London, Glasgow: Collins 1989. 'HWP = Historisches Wörterbuch der Philosophie. Unter Mitwirkung von mehr als 700 [Bd. 3: 800; Bd. 4: 900; Bd. 5: 950; Bd. 6: tausend; Bd. 7: 1200] Fachgelehrten. In Verbindung mit Günther Bien et al. Hrsg. v. Joachim Ritter [ab Bd. 4: Hrsg. v. Joachim Ritter t und Karlfried Gründer]. Völlig neubearb. Ausgabe des (Wörterbuchs des philosophischen Begriffe) von Rudolf Eisler. Bd. 1-[13 u. Registerband]. Darmstadt: Wissenschaftliche Buchgesellschaft 1971-1995 [2005]. Bd. \:A-D, 1971; Bd. 2: D-F, 1972; Bd. 3: G-H, 1974; Bd. 4: i-K, 1976; Bd. 5: L-Mn, 1980; Bd. 6: Mo-O, 1984; Bd. 7: P-Q, 1989; Bd. 8: R-Sc. 1992; Bd. 9: Se-Sp, 1995; [geplant: Bd. 10: Su-T, 1998; Bd. 11: U-V, 2001; Bd. 12: W-Z, 2004; Bd. 13: Register, 2005]. Kranz 2ExotFrGem = Brigitte Kranz: Exotische Früchte und Gemüse kennen · zubereiten · genießen. 2. Aufl. München: Südwest 1974. Kondakow 2 WBLogik = N. I. Kondakow: Wörterbuch der Logik. Hrsg. der dt. Ausg. Erhard Albrecht [u.] Günter Asser. 2., neubearb. Aufl. Leipzig: VEB Bibliographisches Institut 1983. Zander I5 HWB = [Robert] Zander: Handwörterbuch der Pflanzennamen. Dr. h.c. Fritz Encke, Dr. Günther Buchheim, Prof. Dr. Siegmund Seybold. 15. Aufl. Korr. Nachdr. d. 14. Aufl. Stuttgart: Eugen Ulmer 1994.

3.3. Sonstige Literatur Aitchison, Jean (1997): Wörter im Kopf. Eine Einführung in das mentale Lexikon. Aus dem Englischen v. Martina Wiese. Tübingen: Max Niemeyer 1997 (Konzepte der Sprach- und Literaturwissenschaft 56). [Titel der engl. Originalausg.: Words in the Mind. An Introduction to the Mental Lexicon. '1987; 21994.]

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

67

Bergenholtz, Henning/Uwe Kaufmann (1996): „Ensyklopädische Informationen in Wörterbüchern." In: Nico Weber (Hrsg.): Semantik, Lexikographie und Computeranwendungen. Tübingen (Sprache und Information 33), 167-182. Deutscher Wortschatz. Lexikologische Studien. Ludwig Erich Schmitt zum 80. Geburtstag von seinen Marburger Schülern. Hrsg. v. Horst Haider Munske, Peter von Polenz, Oskar Reichmann, Reiner Hildebrandt. Berlin, New York: Walter de Gruyter 1988. Dieckmann, Walther (1988): „Man kann und sollte Bedeutungserläuterung und Sachbeschreibung im Wörterbuch trennen. Ein unpraktisches Plädoyer für Sprachwörterbücher." In: f Deutscher Wortschatz[...] 1988,791-812. Eco, Umberto (1985): Semiotik und Philosophie der Sprache. Übers, v. Christiane Trabant-Rommel u. Jürgen Trabant. München: Wilhelm Fink 1985 (Supplemente, hrsg. v. Hans-Horst Henschen, 4). [Titel der ital. Originalausg.: Semiotica efilosofia del linguaggio, 1984.] Eisinger, Norbert/Andreas Nonnengart (1987): „Termersetzungssysteme." In: Deduktionssysteme. Automatisierung des logischen Denkens. Mit 68 Abb. u. 8 Tabellen. Hrsg. v. K.[arl] H. Bläsius u. H.-J. Bürckert. Mit Beiträgen v. K. H. Bläsius [...]. München, Wien: Oldenbourg 1987, 115-133. Frawley, William (1981): „In defense of the dictionary: A response to Haiman." In: Lingua 55. 1981, 53-61. Funk-Kolleg Sprache = Funk-Kolleg Sprache. Eine Einfuhrung in die moderne Linguistik. Hrsg. v. Klaus Baumgärtner et al. Bd. 1-2. 123-127. Taus. [Bd. 1]; 101-105. Taus. [Bd. 2]. Frankfurt/M: Fischer 1982. Grice, H. Paul (1993): „Logik und Konversation." In: Handlung, Kommunikation, Bedeutung. Hrsg. v. Georg Meggle. Mit einem Anhang zur Taschenbuchausgabe 1993. Frankfurt/M: Suhrkamp 1993,243-265. Habel, Chrisopher (1985): „Das Lexikon in der Forschung der Künstlichen Intelligenz." In: t Schwarze 1985,441^74. Haiman, John [M.] (1980): „Dictionaries and Encyclopedias." In: Lingua 50. 1980, 329-357. — (1982): „Dictionaries and Encyclopedias again." In: Lingua 56. 1982, 353-355. Heusinger, Klaus von (1997): Salienz und Referenz. Der Epsilonoperator in der Semantik der Nominalphrase und anaphorischer Pronomen. Berlin: Akademie Verlag 1997 (studiagrammatica43). Hoelter, Martin (1996): „Deriving HPSG Lexical Entries from a Natural Language Dictionary." In: Lexikon und Text. Wiederverwendbare Methoden und Ressourcen zur linguistischen Erschließung des Deutschen. Hrsg. v. Helmut Feldweg u. Erhard W. Hinrichs. Tübingen: Max Niemeyer 1996 (Lexicographica. Series Maior 73), 83—97. HSK 5.1 = Wörterbücher. Dictionaries. Dictionnaires. Ein internationales Handbuch zur Lexikographie. An International Encyclopedia of Lexicography. Encyclopedic internationale de Lexicographie. l. Teilbd. Hrsg. v. Franz Josef Hausmann, Oskar Reichmann, Herbert Ernst Wiegand u. Ladislav Zgusta. Berlin, New York: Walter de Gruyter 1989 (Handbücher zur Sprach- und Kommunikationswissenschaft, 5.1). HSK 6 = Semantik. Semantics. Ein internationales Handbuch der zeitgenössischen Forschung. An International Handbook of Contemporary Research. Hrsg. v. Arnim von Stechow u. Dieter Wunderlich. Berlin, New York: Walter de Gruyter 1991 (Handbücher zur Sprach- und Kommunikationswissenschaft. Hrsg. v. Hugo Steger u. Herbert Ernst Wiegand, 6). HSK 7.1 = Sprachphilosophie. Philosophy of Language. La philosophie du langage. Ein internationales Handbuch der zeitgenössischen Forschung. An international Handbook of Contemporary Research. Manuel international des recherches contemporaines. Hrsg. v. Marcelo Dascal, Dietfried Gerhardus, Kuno Lorenz, Georg Meggle. 1. Halbbd. Berlin, New York: Walter de Gruyter 1992 (Handbücher zur Sprach- und Kommunikationswissenschaft. Hrsg. v. Hugo Steger u. Herbert Ernst Wiegand, 7. l). HSK 7.2 = Sprachphilosophie. Philosophy of Language. La philosophie du langage. Ein internationales Handbuch zeitgenössischer Forschung. An International Handbook of Contemporary Re-

68

Matthias Kammerer

search. Manuel international des recherches contemporaines. Hrsg. v. Marcelo Dascal, Dietfried Gerhardus, Kuno Lorenz, Georg Meggle. 2. Halbbd. Berlin, New York: Walter de Gruyter 1996 (Handbücher zur Sprach- und Kommunikationswissenschaft. Hrsg. v. Hugo Steger u. Herbert Ernst Wiegand, 7.2). Kammerer, Matthias (1995): Bildschirmorientiertes Abfassen von Wörterbuchartikeln. Dargestellt am Beispiel des Frühneuhochdeutschen Wörterbuches. Tübingen: Max Niemeyer 1995 (Lexicographica. Series Maior 68). — (1998): „Die Mediostruktur in Langenscheidts Großwörterbuch Deutsch als Fremdsprache." In: Perspektiven der pädagogischen Lexikographie des Deutschen. Untersuchungen anhand von Langenscheidts Großwörterbuch Deutsch als Fremdsprache. Hrsg. v. Herbert Ernst Wiegand. Tübingen: Max Niemeyer 1998 (Lexicographica. Series Maior 86), 315-330. — (1998a): „Hypertextualisierung gedruckter Wörterbuchtexte: Verweisstrukturen und Hyperlinks. Eine Analyse anhand des Frühneuhochdeutschen Wörterbuches." In: Angelika Storrer/Bettina Harriehausen (Hrsg.): Hypermedia für Lexikon und Grammatik. Tübingen: Gunter Narr 1998 (Studien zur deutschen Sprache. Forschungen des Instituts für deutsche Sprache 12), 144-171. Kammerer, Matthias/Andrea Lehr (1996): „Potentielle Verweise und die Wahrscheinlichkeit ihrer Konstituierung". In: Wörterbücher in der Diskussion H. Vorträge aus dem Heidelberger Lexikographischen Kolloquium. Hrsg. v. Herbert Ernst Wiegand. Tübingen: Max Niemeyer 1996 (Lexicographica. Series Maior, 70), 311-354. Kammerer, Matthias/Herbert Ernst Wiegand (1998): „Ausgewählte Bibliographie zur Pädagogischen Lexikographie und zu Wörterbücher in pädagogischen Kontexten." In: Perspektiven der pädagogischen Lexikographie des Deutschen. Untersuchungen anhand von Langenscheidts Großwörterbuch Deutsch als Fremdsprache. Hrsg. v. Herbert Ernst Wiegand. Tübingen: Max Niemeyer 1998 (Lexicographica. Series Maior 86), 361-385. Keller, Rudi (1996): „Begriff und Bedeutung". In: Joachim Grabowski, Gisela Harras, Theo Herrmann (Hrsg.): Bedeutung. Konzepte. Bedeutungskonzepte. Theorie und Anwendung in Linguistik und Psychologie. Opladen: Westdeutscher Verlag 1996, 47-66. LGL 1980 = Lexikon der Germanistischen Linguistik. Hrsg. v. Hans Peter Althaus, Helmut Henne, Herbert Ernst Wiegand. 2., vollständig neu bearb. u. erw. Aufl. Tübingen: Max Niemeyer 1980. Lüdi, Georges (1985): „Zur Zerlegbarkeit von Wortbedeutungen." In: t Schwarze 1985, 64-102. Lutzeier, Peter Rolf (1985): „Die semantische Struktur des Lexikons." In: Schwarze/Wunderlich 1985, 103-133. Menzel, W. (1990): Vorlesung[sskript]: Formaie Systeme. WS 90/91. Universität Karlsruhe, Institut für Logik, Komplexität und Deduktionssysteme. Karlsruhe: 1990. Oeing-Hanhoff, L. (1971): „Axiom: II. Geschichte." In: i 'HWP, I, 741-748. Partee, Barbara H./Alice Ter Meulen/Robert E. Wall (1990): Mathematical Methods in Linguistics. Dordrecht, Boston, London: Kluwer Academic Publishers 1990 (Studies in Linguistics and Philosophy 30). Porphyrius (1995): „Einleitung in die Kategorien." In: Aristoteles. Philosophische Schriften in sechs Bänden. Bd. 1. Übers, v. Eugen Rolfes. Darmstadt: Wissenschaftliche Buchgesellschaft 1995. [Lizenzausgabe des Felix Meiner Verlags, Hamburg 1995.] Rosenbech, Klaus (1994): „Enzyklopädische Information im zweisprachigen Wörterbuch." In: Burkland, Schaeder/Henning Bergenholtz (Herg.): Fachlexikographie. Gachwissen und seine Repräsentation in Wörterbüchern. Tübingen (Forum für Fachsprachen -Forschung, 23) 137-159. Schnelle, Helmut (1991): Die Natur der Sprache. Die Dynamik der Prozesse des Sprechens und Verstehens. Berlin, New York: Walter de Gruyter 1991 (Grundlagen der Kommunikation und Kognition). — (1992): „Die sprachphilosophischen Annahmen der formalen Semantik." In: i HSK 7.1, 775-784.

Kritisches zu Schnelles Applizierung einer »Logischen Semantik«...

69

— (1997): „Logische Semantik eines Wörterbuchs vom Cobuild Typ." In: Linguistische Theorie und lexikographische Praxis. Symposiumsvorträge. Hrsg. v. Klaus-Peter Konerding u. Andrea Lehr. Tübingen: Max Niemeyer 1997 (Lexicographica. Series Maior 82), 139-153. Schnelle, Helmut/Martin Hoelter/Frank Wegmann/Rolf Wilkens (1996): „Lexika vom Cobuild-Typ." In: Wilfried Hötker/Petra Ludewig (Hgg.): Lexikonimport, Lexikonexport. Studien zur Wiederverwertung lexikalischer Informationen. Tübingen: Max Niemeyer 1996, 175-229 (Sprache und Information. Beiträge zur philologischen und linguistischen Datenverarbeitung, Informatik und Informationswissenschaft 31). [Dreiteiliger Beitrag, bestehend aus: Helmut Schnelle: Zur Logik der lexikalischen Semantik, 177-194; Martin Hoelter/Rolf Wilkens: Die Computerform des Lexikons - Natürlichsprachliche Lexika für natürlichsprachliche Systeme, 195-211; Frank Wegmann/Martin Hoelter/Rolf Wilkens: Die Organisation zukünftiger Lexikographie, 212-229.] Schöning, Uwe (o.J.): Logik für Informatiker. Mannheim, Wien, Zürich: B. I. Wissenschaftsverlag o.J. (Reihe Informatik). Schwarze, Christoph/Dieter Wunderlich (Hrsg.) (1985): Handbuch der Lexikologie. Königstein/Ts.: Athenäum 1985. von Stechow, Armin (1993): „Die Aufgaben der Syntax." In: Syntax. Ein internationales Handbuch zeitgenössischer Forschung. An International Handbook of Contemporary Research. Hrsg. v./Ed. by Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, Theo Vennemann. 1. Halbbd. Berlin, New York: Walter de Gruyter 1993 (Handbücher zur Sprach- und Kommunikationswissenschaft. Hrsg. v. Hugo Steger u. Herbert Ernst Wiegand; Bd. 9.1.), 1-89. Szabo, Ä. (1971): „Axiom: I. Axiom und Postulat." In: T 'HWP, I, 737-741. Tesniere, Lucien (1980[96]): [Auszüge aus:] Grundzüge der strukturalen Syntax. In: Ludger Hoffmann (Hrsg.): Sprachwissenschaft. Ein Reader. Berlin, New York: Walter de Gruyter 1996, 517542. Wiegand, Herbert Ernst (1973): „Einige Grundbegriffe der lexikalischen Semantik." In: t FunkKolleg Sprache, Bd. 2, 23-39. — (1973a): „Lexikalische Strukturen I." In: t Funk-Kolleg Sprache, Bd. 2, 40-54. — (1973b): „Lexikalische Strukturen II." In: t Funk-Kolleg Sprache, Bd. 2, 55-69. — (1977[84]): „Nachdenken über Wörterbücher. Aktuelle Probleme." In: Günther Drosdowski, Helmut Henne, Herbert E. Wiegand: Nachdenken über Wörterbücher. Mannheim, Wien, Zürich: 1977 [Korrigierter Nachdruck 1984], 51-102. — (1985): „Eine neue Auffassung der sog. lexikographischen Definition." In: Symposium on Lexicography H. Proceedings of the Second International Symposium on Lexicography May 16-17, 1984 at the University of Copenhagen. Hrsg. v. Karl Hyldgaard-Jensen u. Arne Zettersten. Tübingen: Max Niemeyer 1985 (Lexicographica. Series Maior, 5), 15-100. — (1988): „Wörterbuchartikel als Text." In: Das Wörterbuch. Artikel und Verweisstrukturen. Jahrbuch 1987 des Instituts für deutsche Sprache. Hrsg. v. Gisela Harras. Düsseldorf: Schwann; Bielefeld: Cornelsen-Velhagen & Klasing 1988, 30-120. — (1988a): „Was eigentlich ist Fachlexikographie? Mit Hinweisen zum Verhältnis von sprachlichem und enzyklopädischem Wissen." In: t Deutscher Wortschatz [...] 1988, 729-790. — (1989): „Der Begriff der MikroStruktur: Geschichte, Probleme, Perspektiven." In: t HSK 5.1, 409-^62. — (1989a): „Arten von MikroStrukturen im allgemeinen einsprachigen Wörterbuch." In: t HSK 5.1, 462-501. — (1989b): „Aspekte der Makrostruktur im allgemeinen einsprachigen Wörterbuch: alphabetische Anordnungsformen und ihre Probleme." In: t HSK 5.1, 371-409. — (1989d): „Die lexikographische Definition im allgemeinen einsprachigen Wörterbuch." In: t HSK 5.1,530-588.

70

Matthias Kammerer

— (1994): „Zur Unterscheidung von semantischen und ensyklopädischen Daten in Fachwörterbüchern." In: Burkland, Schaeder/Henning Bergenholtz (Herg.): Fachlexikographie. Gachwissen und seine Repräsentation in Wörterbüchern. Tübingen (Forum für Fachsprachen -Forschung, 23) 103-132. — (1996): „Textual Condensation in Printed Dictionaries. A Theoretical Draft." In: Lexikos 6 (AFRILEC-reeks / series 6: 1996), 133-158. — (1996a): „Das Konzept der semiintegrierten MikroStrukturen. Ein Beitrag zur Theorie zweisprachiger Printwörterbücher." In: Wörterbücher in der Diskussion II. Vorträge aus dem Heidelberger Lexikographischen Kolloquium. Hrsg. v. Herbert Ernst Wiegand. Tübingen: Max Niemeyer 1996 (Lexicographica. Series Maior 70), 1-82. — (1996b): „Deutsch-Usbekisches Wörterbuch. Einblicke in die Wörterbucharbeit an der Staatlichen Usbekischen Weltsprachen-Universität in Taschkent." In: Lexicographica 12. 1996, 190-254. — (1996c): „Textual Condensation in Printed Dictionaries. A Theoretical Draft." In: Lexikos 6 (AFRILEC-reeks / series 6: 1996), 133-158. — (1996d): „A theory of lexicographic texts: an overview". In: South Africa Journal for Linguistics 14, H. 4. 1996, 135-149. •—(1998): Wörterbuchforschung. Untersuchungen zur Wörterbuchbenutzung, zur Theorie, Geschichte, Kritik und Automatisierung der Lexikographie. 1. Teilbd. Mit 159 Abbildungen im Text. Berlin, New York: Walter de Gruyter 1998. Wiegand, Herbert Ernst/Werner Wolski (1980): „Lexikalische Semantik." In: t LGL, 199-211. Wunderlich, Dieter (1991): „Bedeutung und Gebrauch." In: 1 HSK 6, 32-52.

Jörgen Erik Nielsen Otto Jespersen as a Lexicographer

In his autobiography entitled En sprogmands levned, in the English translation A Linguist's Life, Jespersen informs us that "in between other work, I collaborated on two lexical projects", namely "Brynildsen's large English-Danish dictionary" and "Salmonsen's encyclopaedia", and for the latter "I wrote articles on general linguistics and on several philologists".1 That encyclopaedia, entitled Salmonsens Konversationsleksikon, appeared in 18 volumes over the years 1893-1907 with a supplementary volume in 1911; a second, revised edition of 25 volumes appeared in 1915-28 with a supplementary volume in 1930. A smaller encyclopaedia called Den lille Salmonsen appeared in 1937-40, but that is not an abridged edition of the bigger work, and Jespersen did not have a share in it. The encyclopaedia has its name from the publisher: The first edition was sent out by the publishing house of the Brothers Salmonsen in Copenhagen, established in 1871 by Isaac and Moritz Salmonsen. In 1911 the publishing house of J.H. Schultz took over Salmonsen's business including the encyclopaedia, whose name remained unaltered. Editor of the first edition and of the second one till vol. XXI was the author J. C. Blangstrup, whom Jespersen describes as "amiable", but he had the shortcoming "that he was not sufficiently systematic in allotting space for articles, and that he allowed several contributors to spread themselves disproportionately:" When I called on him to deliver my article on English Language, for which I had rigidly adhered to the allotted length, I was able to point out to him that in the part of the encyclopaedia that had just appeared the article on Pigeon Breeds had been given more space than he had allowed for English Language, which must surely be of more importance to the world than pigeon breeds, the evidence being that many universities had professorships of the former, but not usually of the latter. He was very amenable and gave me back the article with permission to expand it a bit, which was not so easy to do as it would have been if I had been allowed more space from the start.

The article on pigeon breeds precedes that on English, because pigeon is due in Danish. Actually the two articles are in the same volume, but Jespersen here uses the word "hefte": The encyclopaedia appeared in parts. The article on the English language is found under the heading "Sprog" ("Language") with the signature "O.Jsp.", a subdivision of the entry on England, in the 2nd edition 110 pages long, and English language is in that edition (vol. VII, 1918) allotted more than six pages, precisely the same as is allotted to pigeons. The article describes the origin and development of the language, possible influence from other languages, vocabulary etc. At the end of it Jespersen explains and deplores the fact that "English spelling is one of the most unreasonable and capricious in the world", but otherwise, by removing superfluous grammatical difficulties, "the English language has in a higher degree developed to its advantage than any other European language", its vocabulary is "ample and convenient", "it can pride itself on a literature which need not shun comparison with any in the entire world," and it 1

Otto Jespersen, En sprogmands levned (Copenhagen, 1938), pp. 98-99; A Linguist's Life, ed. by Arne Juul, Hans F. Nielsen and J0rgen Erik Nielsen (Odense, 1995), pp. 115-16.

72

J0rgen Erik Nielsen

"has proved and daily proves to be a pliant tool for expressing what agitates the mind of the poet, the thinker, the scholar, the politician, the engineer, the workman, in short everybody." We recognize Jespersen's well-known position on language change, what he called "progress in language;" and in his long article on "Sprog" ("Language") in vol. XXI of the 2nd edition (1927) he enlarges on progressive simplification in languages, not much perceptible in Slavonic languages and Lithuanian, whereas English has "liberated itself from many difficult instances of inflection." 2 Finally Jespersen says that English "has one more merit over its rivals in the world of languages", viz its extension over the globe, which process is "far from ended yet." Indeed, it was not, in the intervening 80 years English has become a second language for most of us. Jespersen's entry on "Anglicisme" ("Anglicism") in vol. I of the 2nd edition (1915) defines that phenomenon as "transfer of some characteristic of the English language into other languages" and says that Anglicisms are common in inferior translations from English and in the everyday speech of non-native settlers in England and the USA. Jespersen did not anticipate the position of English today, when it is commonly used as an international means of communication, the role that he expected an artificial language like Esperanto or Ido to play. He describes the various auxiliary languages in the article "Verdenssprog" ("World Language") in vol. XVIII of the first edition (1907) and again in vol. XXIV of the second one (1928), where he adds information on the schism between Esperantists and Idists following a conference in 1907. Jespersen, who took part in that conference, believed in progress in language and, logically, you might say, in the feasibility of constructing a regular and easily learnt auxiliary language. Jespersen contributed, if I have counted correctly, 294 articles, some very brief, others very long, to Salmonsen, on aspects of linguistics, on linguists and on language teaching. Occasionally one suspects that he had suggested the inclusion of such an article, e.g. the one on "Vovvov -Teorien" ("The Bow-wow Theory") in the supplement to the 1st edition (1911) and in vol. XXV of the 2nd edition (1928). The long entries on "Sprog" ("Language") and on "Sprogvidenskab" ("Linguistics") in vol. 16 (1905), repeated in expanded form in the 2nd edition in vol. XXII (1927) are also by Jespersen, and the ones in the 2nd edition give a clear impression of what Jespersen had to say about a subject that he had been busy with for a lifetime. With some familiarity of his works one easily recognizes his hand, thus French 'aimerai' from 'amare habeo' as an example of a synthetic form having developed on the basis of a periphrastic one is mentioned already in his Progress in Language (London, 1894), p. 332 and can be traced back to the public defence of his doctoral thesis in 1891.3 Salmonsens Konversationsleksikon is a big encyclopaedia, which contains many long and detailed articles on various subjects, the contributors had a reasonable amount of space to express themselves and do more than merely enumerate facts. So in Jespersen's case the encyclopaedia furnishes an excellent opportunity to become acquainted with the linguistics of his days and as he conceived of it. If you are acquainted with some of his works you will often recognize his hand, and time and again the reader has the opportunity to admire his "gift of lucid All passages here quoted from Jespersen's autobiography are from the above-mentioned printed translation, all other translations from Danish are by the present writer. The defence of Studier over engelske Kasus took place on 12 May 1891. One of the examiners, Hermann M011er, had his contribution printed in Nordisk Tidsskrift for Filologi, n.s 10 (Copenhagen, 1892), pp. 292-317, and a rejoinder by Jespersen is in the same periodical, 3d ser. 1 (Copenhagen, 1892-93), pp. 36-41.

Otto Jespersen as a Lexicographer

73

presentation that amounted to genius", to quote C.A. Bodelsen.4. Jespersen was not a scholar who made things more complicated than necessary, and he was an authority on modern linguistics who had himself contributed immensely to its development; of course he was the obvious person to write about "Lydskrift" ("Phonetic transcription") in vol. XVI (1924) of the 2nd edition. That remark brings me on to the other lexical work that Jespersen contributed to, viz Engelsk-dansk-nors k Ordbog, on the English title-page called A Dictionary of the English and Dano-Norwegian Languages, which appeared in instalments from 1900 and came to comprise two volumes, the title-pages bearing the years 1902 and 1907 respectively. An abridged onevolume edition appeared in 1910, like the big edition published by Gyldendal in Copenhagen. The dictionary was composed by John Brynildsen, who was reader of English and French at the naval academy at Horten, Norway, and also a lexicographer who had in 1886 published an English-Norwegian dictionary and in 1892 a Norwegian-English one. The inclusion in the new work of Norwegian as well as Danish equivalents to the English headwords made the market larger of course, indeed the English-Danish dictionary by Svend Rosing which Brynildsen's new work replaced, appears to have sold well in Norway.5 Also the difference between written Norwegian and written Danish was not considerable in those days. In the new dictionary Danisms were supervised by Johannes Magnussen, a well-known language teacher, translator and author of textbooks. Jespersen was entrusted with the phonetic transcription, and he took over the revision of Danisms in the appendix from the word 'dead' after the demise of Magnussen in 1906. Rosing's dictionary, which had run through seven editions between 1853 and 1899, indicated stress at individual words and offered a number of rules for the pronunciation in the introduction, but the lexicographer had been unwilling to use phonetic transcription. In 1855 and in 1874 English dictionaries appeared in Christiania and Copenhagen with phonetic transcription based on John Walker's Critical Pronouncing Dictionary (lst ed. London, 1791), in which the different vowel sounds are indicated by means of ciphers above the letters, thus a notation very different from the one Jespersen used in the new dictionary. One year after the appearance of the last edition of Rosing's dictionary, the first instalment of its successor came out. Volume I of Brynildsen's dictionary contains a preface "Om udtalebetegnelsen" ("On the Phonetic Notation") by Jespersen. He makes it clear that the notation found here is the one he had for a number of years used in various textbooks and in his Fonetik (Phonetics) (1897-99): "It differs very little from the 'international' system which is used in the Maitre Phonetique and is becoming more and more commonly used; the few and slight divergences pertain to points only where the international notation proves in praxis to cause difficulties to Danish and Norwegian readers." Thus man is in Jespersen's notation [man], as [maen] might might mislead Danes to assume the vowel sound to be that ofmcend, the plural of mand; the transcription of'mane' as [mein] would lead to confusion with [main] because of the pronunciation of ei in Danish; and ou in, say, [boun] might lead to "exaggeration of the diphthong"; so for these two diphthongs he chose the notations [e' ] and [o u ] instead.

5

C.A. Bodelsen, "Otto Jespersen, 1860-1943" in his Essays and Papers (Copenhagen, 1964), p. 186. J0rgen Erik Nielsen, "Svend Rosing, his Engelsk-Dansk Ordbog (1853), and Gyldendal", Symposium on Lexicography V, ed. by Karl Hyldgaard-Jensen and Arne Zettersten, Lexicographica: Series Maior 43 (Tübingen, 1992), pp. 217-26.

74

Jörgen Erik Nielsen

The Standard that Jespersen has aimed at describing is that employed by "educated English people", but in some cases he indicates American pronunciation by means of "amr". His transcriptions are based on his own notes, on "the information on pronunciation I have been able to procure in works by real phoneticians from the last decades", and on other dictionaries. Brynildsen has included many dialect words, in which cases Jespersen has once again had recourse to the educated Englishman: How would he pronounce those words? In the case of Scottish, "the best known dialect group", he indicates the dialect pronunciation too, thus his transcription of loch is [läk, sc lax]. He emits a cri de coeur at "the exceptional flair for overlooked matter which marks the author of this dictionary", in other words: Jespersen has had to transcribe a number of words that he had never heard of and not been able to find anywhere. When in spite of all his pains he has not succeeded in solving the problem, he has had recourse to "the [?] which has been the most difficult for me to use during my work." Examples in point are 'ho-gya', an Anglo-Indian word meaning 'in a bad way', and -lan-kwoh', Pidgin-Dutch for 'Dutch people'. Frankly, who among us ever felt the disadvantage of not being familiar with the pronunciations of those words? In his autobiography Jespersen writes that "in contrast to the dictionaries then existing, I managed to indicate the actual weak or reduced sounds in non-accented syllables; English and American dictionaries indicated the same sound as in accented syllables right up to the publication of Daniel Jones's pronouncing dictionary in 1917." Actually, weak or reduced sounds in non-accented syllables are indicated already in Michaelis and Jones, A Phonetic Dictionary of the English Language (Hannover and Berlin, 1913), but that does not alter the fact that thanks to Jespersen, Brynildsen's dictionary was ahead of other dictionaries in this aspect of phonetic transcription. Finally a few words about an entire dictionary by Jespersen. In A Linguist's Life he says that many may regard as the greatest folly of his life the colossal amount of labour he devoted to the auxiliary-language problems; and he adds in brackets: "I am sometimes inclined to agree" (p. 147). However, that did not discourage him from continuing that work alongside the other scholarly work he did after he had retired from the University: In an attempt to overcome what he saw as the shortcomings of Esperanto and Ido he constructed the grammar of a new language, which he called Novial, and introduced it to the world in a book called An International Language (London, 1928), followed by the dictionary NovialLexike "da Otto Jespersen" (Heidelberg, London, Paris, 1930). He spent about a year of work on that dictionary, and he describes it as his "third great flop" (A Linguist's Life, pp. 223-24), as it never sold in large numbers; still he adds: "But, as Gloucester says in Shakespeare, 'there was good sport at its making.'" Novial Lexike contains a preface in Novial, English, German and French explaining the guiding principles behind his new language. As for the vocabulary the chief principle has been "to use, whereever possible, words that are already international, and where these are missing, to take those words which create the least friction." However, many international words, such as "nature", "form" and "romantic", are so far from being precise that they must be avoided. Most of the 251 pages are taken up by a quadrilingual dictionary, translating Novial words into English, German and French. Jespersen does not explain what words have been included, but he is clearly anxious to demonstrate derivation in Novial; as an example may serve this entry:

Otto Jespersen as a Lexicographer

75

klefe F clef, E key, D Schlüssel / -fa F fermer ä clef, E lock, D zuschliessen / -fo (akto). The noun is clearly formed on the basis of the French word, which is here mentioned first among the translations; and from that noun is derived a verb with the ending '-a' and a noun, denoting the action, with the ending '-o'. In some cases no translation whatever is offered by the dictionary, because the meaning is regarded by the lexicographer as obvious, a case in point is album. Personally I have been annoyed by being led astray by a false friend in this dictionary, namely klima, which I understood immediately until I saw that it means 'to climb'. What then is climate, Danish 'klima', in Novial? Oh, I see, it is called klimate, the next entry on the page. So Otto Jespersen, who is not normally thought of as a lexicographer, did do some lexicographical work, and that work is intimately linked up with his other occupations. In his contributions to the encyclopaedia and to the English-Danish dictionary we see him as the scholar who makes his expert knowledge available to a large public, explaining himself lucidly and considering the practical usefulness of his work. He began his work for Salmonsen as a young, newly appointed professor, and when the supplementary volume to the second edition appeared in 1930, in which there is an article by Jespersen on Novial, he was a very active pensioner. Actually, about the same time he contributed long entries, signed "O.J.", on "Grammar", "Language" and "Philology" to the 14th edition of the Encyclopedia Britannica, explaining in the article on "Grammar", among other things, his concept of the three ranks of primary, secondary and tertiary. His Novial Lexike may be a result of his "folly", but at least he had the both noble and useful objective to facilitate international communication; and his work with Novial is closely linked to his studies of general linguistics, of the "philosophy of grammar".

Dorthe Duncker/Hanne Runs

Multi Level Text Representation in an LCB

In the infancy of corpus linguistics in the 1970s, the limits of electronic storage and processing capacities made it infeasible to build a structured database of a medium sized text corpus. The corpus base for frequency studies described in Maegaard and Ruus (1980) is theoretically wellstructured, but turned out to be infeasible in the available main frame computer at the end of the 1970s. The increasing speed and storage capacity of present days' personal computers make it possible for philologists, lexicographers and linguists to work conveniently with large corpora of data. Computational methods and tools for dealing electronically with large corpora have been developed within corpus linguistics (Zernik 1991, Baker et al. 1993); computational lexicography offers tools for producing large dictionaries electronically (The DANLEX-group 1987, Norling-Christensen 1992). The requirements of both disciplines are satisfied in the Lexical Corpus Base, in which the linguistic data are marked and structured in a way that facilitates corpus investigations and offers a lexical view of the data at the same time.

Methodology Many currently used methods for investigating texts with electronic tools have as their primary goal to develop techniques that function fully automatically. Because computers as yet cannot be given a general text understanding competence, the investigators often have to accept a level of correctness in analysis that is far below the complete correctness that is a matter of course in the tradition of philology. This is one source for a certain scepticism from traditional philologists towards computers in textual investigations. In the work described here we have developed computational techniques that comply with the philological requirements of complete correctness and transparency of analysis by taking into account the strengths of both humans and computers. The aim of the work is to make a collection of texts available for philological and linguistic investigations and to develop electronic tools that make it possible for lexicographers to build a dictionary based on the texts with a minimum of tedious, repetitious work and a maximum of accuracy. The treatment of the texts consists of the following steps: electronic registration parsing and SGML-tagging building a multi level representation

This results in a corpus base which is used to make a dictionary of the texts. The dictionary making has the following steps:

Dorthe Duncker/Hanne Ruus

78 building the orthographic dictionary adding semantic field information extracting collocations and formulae extracting raw dictionary entries editing raw dictionary entries The final result is the lexical corpus base shown in figure 1. Dictionary entries

Semantic information

Multi word units

Figure 1: The resulting lexical corpus base. The arrows indicate how the data in one or more boxes contribute to the content of other boxes.

Our method makes an appropriate division of labour between humans and computers by splitting the work to be performed into a set of well-defined processes. The steps of each process are initially described and classified either as tasks that can be performed fully automatically or as tasks that require the assistance of a philologist or a lexicographer because of decisions to be taken that are necessarily based on human text understanding. In the following, we use the term heuristic for the fully automatized processes and the term hermeneutic for the semiautomatized processes that require decisions from a philologist or a lexicographer. In the distribution of tasks between humans and computers it has been useful to recall the strengths of humans and those of computers in the area of textual analysis. Humans are experts in understanding and interpreting texts and in describing texts by applying linguistic categories. Computers' primary strengths in text handling are composed of their ability to process strings of characters and of their practically unlimited capacity for repeating any formally described process a well-defined, often very large number of times with great speed. To minimize the repetitious work and the number of writing errors, we have adopted an accumulative approach that makes use of the results of the analysis of the previously stored texts. This boot-strapping procedure will be illustrated below.

A/w/// Level Text Representation in an LCB

79

The text collection The work described in this paper was done in the research project Dansk Folkevisekullur 15501700 at the University of Copenhagen (Ruus 1996). One of the aims of the project was to examine the oldest manuscripts containing Danish ballads from 1550-1600. Most of these manuscripts were compiled among courtiers at the Danish court and preserved in libraries of the gentry. The manuscripts contain ballads, lyrical and historical poems and songs, and jocular songs recorded by many different persons in a linguistic form representing an abundance of none too stable orthographic norms. Many of the texts in the manuscripts have been published in scholarly editions compiled according to genre: Danmarks gamle Folkeviser (DgF) contains most of the ballads, descriptions of the manuscripts and many useful indexes, but no dictionary; the lyrical and historical poems are published in Danske Viser (DV) with a dictionary for these texts, the jocular songs are published in Danske Skcemteviser (DS). In Danmarks gamle Folkeviser, each ballad type is published in all its known versions comprising handwritten versions dating from mid 16th century to late 18th century, printed versions from 1591 to 1843, and versions recorded from oral tradition in the 19th century. When a handwritten version is recorded in several manuscripts, the text from only one manuscript is printed and variants from the other manuscripts are added as notes. This principle of editing favours the investigation of each ballad type on its own, but complicates the study of the old manuscripts as collections of texts. The texts of the manuscripts are distributed on three scholarly editions and not all texts are published; this makes it extremely difficult to get a coherent view of each of the manuscripts, its content and style. The dating of the ballad genre in Scandinavia has been a recurrent issue in the study of the ballads, but an exhaustive study of the linguistic form of the oldest recorded tradition, that of the Danish manuscripts from the 16th century, is missing. In short, a data base containing the texts from the oldest manuscripts in their entirety will offer a unique opportunity to study and compare the manuscripts and their texts. All the texts from the 9 oldest manuscripts have been registered and stored electronically. The manuscripts record from 50 to 200 texts, in total approximately 1000 ballads, songs, and poems representing about 600 different text types, as many ballads and songs are recorded in different versions. The texts have varying lengths: from 3 to more than 200 stanzas. More than half of the texts have been given a multi level representation and stored in the lexical corpus base.

Making the Lexical Corpus Base In electronic storing of texts for philological purposes, it is important to take care that the texts are properly marked so that they can be moved between computers and processing systems. The best known and widely used mark-up system is SGML (ISO 8879). Among the conventions in SGML is the use of a text header that contains information about the text. In our case, the header for each text contains information about the title of the ballad or poem, the name of

80

Dorthe Duncker/Hanne Ruus

the manuscript, the position in the manuscript, the position in the scholarly edition if published etc. The text itself is marked up giving labels that specify all parts of the text e.g. stanza, verse, word, abbreviation. In electronic investigations of texts it is convenient to have access to the authentic texts and to additional information that facilitates automatic extraction of relevant text sections according to varying criteria. The additional information will depend on the kind of text to be described. It may e.g. comprise phonological and discourse analysis categories in a corpus of spoken texts. In the case of written texts, word form and lemma are useful information categories. To keep track of original source text and added information, a multi level representation is adopted. In the case of the Danish 16th-century texts, the levels are: source level, orthographically neutral level, lemma level. The initial input to the base are the electronically registered texts furnished with headers computed from an electronic index containing the necessary information about the manuscripts and the manuscript texts in the editions. In the mark-up of the text, the hierarchical structure of the text is recorded: the limits of the stanzas and their verses are tagged. The units of the verses, the textual words are also delimited.

The levels

The source level contains the original textual word forms. Information from the edition of the text is implemented as foot notes to the relevant words and sections. Abbreviations, lacunae and emendations are also tagged. The orthographically neutral level contains the neutral word forms corresponding to the source words. An orthographically neutral representation is an absolute prerequisite for searching and finding e.g. different instances of a word in the older Danish texts. In the manuscripts we find as many different orthographical norms as there are scribes. The orthography of modern Danish is fairly conservative and can be used as the basis for defining a notation that neutralizes orthographic variation in the language of the 16th-century song books. This notation may, with a few modifications, be used for all older Danish and modern Danish written texts. The lemma level contains the lemmas of the source words consisting of a lemma form, the relevant word class and - in case of homographical lemma forms with the same word class - a discriminator that makes it possible to distinguish between otherwise identical lemmas. The rules for making the levels and the reasons for building the multi level text representation are found in Ruus (forthcoming).

Multi Level Text Representation in an LCB

81

The accumulative approach

Each text is treated by a series of processes to obtain its multi level representation. When the very first text is processed, the main part of the work must be done by the philologist assisting the computer. The first text is treated by the two hermeneutic processes: Fill-in-unknown-word-forms-and-lemmas Check-multi-level-representation. Then follow the heuristic process Make-dictionary-of-the-words-in-the-text and the hermeneutic process Check-dictionary-of-the-words-in-the-text. The resulting dictionary is the first input to the orthographic dictionary. In the treatment of the first text, the processes numbered 3, 5, 6, 7, and 8 have been executed (cf. figure 3 below). When the second text is treated a new heuristic process is added as the first step Add-known-word-forms-and-lemmas-to-the-parsed-text. The following hermeneutic process deals with the unknown words only Fill-in-unknown-word-forms-and-lemmas It is necessary to check all three levels in each text in the Check-multi-level-representation process because of homographic forms. E.g. a neutral word form like/or occurs in the texts as form of a verb, a preposition, an adverb, and a conjunction. In the checking process all solutions are available, and the philologist chooses the correct solution by clicking on it. The following processes deal only with the new words: Make-dictionary-of-new-words-in-the-text Check-dictionary-of-new-words-in-the-text Update-orthographic-dictionary-with-the-new-words-from-the-text. The same series of processes (no. 1,3, 5,6,7, and 8 in figure 3 below) treat the next texts. As soon as the orthographic dictionary comprises the most common function words, a heuristic process that suggests solutions to unknown words is added (no. 2 in figure 3). This process uses orthographic variant rules deduced from the texts so far treated to Forecast-new-word-forms-of-known-lemmas. A small extract of the orthographic dictionary based on one, two, 250, and 450 texts respectively, illustrates the accumulation of information, cf. figure 2. Extract of dictionary based on 1 text vinde (vb) vunden vonden. vinter (sb) v inter winter. Extract of dictionary based on 2 texts vinde (vb) vunden vonden, vindue (sb) vindu vyndu. vinter (sb) vinter vyntter, winter. Extract of dictionary based on 250 texts vinde (vb) af vinde affWinde; vandt vandt, wandtt, wanndt, wanndtt, wannt, want; vinde Winde, wende, wennde, wunde, winde, wynde, wynnde, wynne; vinder Wender, uender, viinder, vinner,

82

Dorthe Duncker/Hanne Ruus wender, wennder, wenner, winder, wynder, wynner; vindes venndis, windis, winndis; vunde vonde, vone, wonde, wondne, wonnde, wonne, wunde; vunden Wanden, vonden, vonnden, wanden; vundet vonditt, vonnditt, wondil, wonndit, wonnditt, wundeth. vindue (sb) vindu vyndu, winnuff, wynne, wynnu; vinduet wenued, weuod, wienduett, windqffiied, winduffuid, winnuedt; vindvet vindvet. vinge (sb) vinge vinge, winge; vinger wenger, wynger. vinke (vb) v inker vinncker, wencker. Vinstermand (pers) Vinstermand Vinstermand, Winnstermand, Winnslermannd, Winstermand, Winstermannd. vinter (sb) vintcr vinnther, vinter, vyntter, wennther, \venter, wenther, wientter, winter, winther, wynter; vinters wynters.

Extract of dictionary based on 450 texts vinde (vb) af vinde qffWinde; vandt vand, vandt, vandtt, vanndt, vanntth, vant, wand, wandtt, wanndl, wanndtl, wannl, wanntt, want; vind vinndt; vinde Winde, vinde, vinnde, vunde, vyndhe, wende, wennde, wunde, winde, winnde, winne, winnhe, wynde, wynnde, wynne; vinder Wender, uender, viinder, vinder, vinnder, vinner, wender, wennder, wenner, wiendder, winder, winnder, wynder, wynner; vindes venndis, vindis, wendis, windis, winndes, winndis, wyndis; vunde Vunde, vonde, vone, vonen, vunde, wonde, wondne, wonnde, wonne, wunde; vunden Wonden, vonden, vonnden, wanden; vundet vonditt, vonnditt, vunditt, wondit, wonditt, wonndit, wonnditt, wundeth, wundit, wunditt, wunett, wuonditt. vinding (sb) vinding Winding. vindue (sb) vindu vyndu, winnuff, wynne, wynnu; vindue vindve, wiindu; vinduet venduett, vinduet, vinduett, vindvet, wenued, weuod, wienduett, windoffued, winduet, winduffuid, winnuedt. vinge (sb) vinge vinge, winge; vinger wenger, wynger. vinke (vb) vinker vinncker, wencker. Vinstermand (pers)Vinstemiand Vinstermand, Winnstermand, Winnstermannd, Winster-mand, Winstermannd. vinter (sb) vinter vennther, vinnther, vinter, vinther, vyntter, wennther, wenter, wenther, wentter, wientter, wiinter, winnther, winntter, winter, winther, wintter, wynter, wynther; vinteren winteren; vinters winntthers, wintters, wintthers, wynnthers, wynters. Figure 2: The accumulating orthographic dictionary. Four extracts of the orthographic dictionary from the entry vinde 'winn' to the entry vinter ''winter'. The source text forms are printed in italics, the neutral word forms in plain typeface. The extracts show the growth from a version based on one text to the version based on 450.

The poetic texts in our LCB are characterized by a repetitive style. This makes it possible to use recurring patterns of word forms and lemmas at a fairly early stage in the base building work. A new heuristic process is added that compares the patterns of the possible solutions for a homograph with the current context and puts the most likely solution first. As the possible solutions are stored in a ring, we call this process Adjust-homograph-wheels (no. 4 in figure 3). When this process is added, each text is treated by eight processes, five fully automatized, heuristic processes and three semi-automatic, hermeneutic processes. In figure 3 the order of the processes is shown. When a fair amount of text has been furnished with its multi level representation, the tasks of the philologist when treating the next texts, are reduced to checking, choosing, and approv-

Multi Level Text Representation in an LCB

83

ing suggestions supplied by the computer. The following example will demonstrate the benefits of the accumulative approach. Orthographic dictionary

8

Multi level text representation

Parsed source text 1. Add-known-word-forms-and-lemmas-to-the-parsed-text 2. Forecast-new-word-forms-of-known-lemmas 3. Fill-in-unknown-words 4. Adjust-homograph-wheels 5. Check-multi-level-representation 6. Make-dictionary-of-new-words-in-the-text 7. Check-dictionary-of-new-words-in-the-text 8. Update-orthographic-dictionary-with-the-new-words-in-the-text Figure 3: The eight processes used to build the multi level text representation and the orthographic dictionary, in order of application. The processes no. 3, 5, and 7 are hermeneutic: the computer is assisted by the philologist. The other five processes are performed fully automatically by the computer.

An example stanza

The following stanza from a ballad (no. 36 in Skj0nne Fru Seiverlad, Anne Munk's manuscript, DgF IV, p. 546) has been processed with a dictionary based on two texts, with a dictionary based on 40 texts, and with a dictionary based on 450 texts. The stanza is first quoted in the source form with a fairly literal English translation:

84

Dorthe Duncker/Hanne Ruus De thog poser paa dieris bag they took bags on their back och pillegrims-stauff y hennde: and pilgrims' staff in hand: saa gick dy saa lanngtt aff lee, then went they so far a-way, som inngenn mand dennom kiende. as no man them kenned.}

In figure 4 the stanza is given in three versions with word forms added from versions of the dictionary based on an increasing number of texts. The added word forms are italized, unidentified word forms are marked by stars. This is the presentation for the philologist after the application of the heuristic process Add-known-word-forms-and-lemmas-to-the-parsed-text. The two levels, source level and neutral word form level, are displayed, the lemma level is available in pop-up windows. The increasing number of known words is the result of the accumulative approach. In the case of the homographs, the alphabetically first form is displayed; the other possible solutions are available in pop-up windows. Note that the source form>> in the first version of the stanza has the suggested neutral word form /, in the second version /, in the third e. This reflects that the largest dictionary has at least three possible solutions of which the e-solution is alphabetically first. word forms added from dictionary based on two texts De thog poser paa dieris bag ^e **** * * * * * * ****** *** och pillegrims-stauff y hennde: o„ ***************** j ******. saa gick dy saa lanngtt aff lee, sa

**** ** sa ******* fly ***

som inngenn mand dennom kiende. *** ******* **** ****** ******

word forms added from dictionary based on 40 texts De thog poser paa dieris bag de tog *****/>cj deres bag och pillegrims-stauff y hennde: og ***************** i hen(je: saa gick dy saa lanngtt aff lee, sa gik de sä ******* afle, som inngenn mand denwom kiende.

Scottish 'knew'.

Multi Level Text Representation in an LCB

85

word forms added from dictionary based on 450 texts De thog poser paa dieris bag de dog ***** pa deres bag och pillegrims-stauff y hennde: og ,»„»**»*„»»,«*, ehende. saa gick dy saa lanngtt aff lee, säe gik de säe langt ad led, som inngenn mand dennom kiende. sam ingen mand dennum kende. Figure 4: The three versions of the same stanza after the process Add-known-word-forms-and-lemmasto-the-parsed-text has been applied to the parsed source text using increasing versions of the orthographic dictionary. The suggestions for neutral word forms are given in italics. For the homographic source forms y, aff, saa, and som several suggestions are available, as can be see n from the different forms suggested in the three versions. The alphabetically first form is displayed, the other possible solutions are available in pop-up windows.

When the dictionary based on 450 texts suggests neutral word forms, only two source words miss suggestions. When the process Forecast-new-word-forms-of-known-lemmas is applied, these two words get suggestions too: De thog poser paa dieris bag de dog passer pä deres bag och pillegrims-stauff y hennde: og pilegrimsstav e hende: saa gick dy saa lanngtt aff lee, säe gik de säe langt ad led, som inngenn mand denwom kiende. sam ingen mand dennum kende. Before the philologist checks the neutral word form level and the lemma level, the process Adjust-homograph-wheels is applied. This yields the following result (the adjusted homograph solutions are in bold face): De thog poser paa dieris bag de tog passer pa deres bag och pillegrims-stauff y hennde: og pilegrimsstav i hcende: saa gick dy saa lanngtt äff lee, sä gik de sa langt afled, som inngenn mand dennom kiende. som ingen mand dennum kendte.

86

Dorthe Duncker/Hanne Ruus

As the result of three heuristic processes, we now have a multi level version of the stanza where the philologist can accept all suggestions except one: in the case of poser, he must choose a solution which is available, but not as the first one.

Extracting orthographic dictionary entries from the marked-up texts The description of the corpus base so far has been centred on the philological aspects, and the orthographical dictionary has been viewed as a lever to obtain the multi level text representation. The different stages of accumulating the corpus base have been presented in a procedural manner. However, when a lexical view is adopted in order to produce the final dictionary entries, it is necessary to address the mark-up structure of the corpus base directly. Underneath the previously described procedures, resulting in the fully developed multi level text represen-

lemma.form part.of .speech I HP gloss 1] word.unit-

word .form

text.word

Figure 5: Entry structure of the orthographic dictionary.

tation, the initial mark-up structure of the texts gradually becomes more complex and refined. Below, it will be shown how the mark-up structure can be exploited to achieve further enrichment of the corpus base, and how it consequently is turned into the LCB proper. When one text is furnished with neutral word forms and lemmas, the new information hereby gained enters the global orthographic dictionary. While the multi level text representation is built up, the dictionary entries mainly contain orthographic information. The information types and the structure of the entries appear from the diagram in figure 5.2 The lemma.form and the part. of. speech elements both occur just once in each entry. The discr element contains information to discriminate homographs; the gloss element contains glosses. Both elements are optional. The wordunit element holds the neutral wordform and one or more text, word elements. The wordunit is a compulsory element that can occur one or more times. The orthographic dictionary entry structure constitutes the head element of the dictionary entries in the overall corpus base. When this view is adopted, it also appears clearly how the entry content is retrieved directly from the multi level text representation of the individual texts in the sample elements (cf. figure 6).

2

The diagrams are drawn with Near & Far®, Microstar Software Ltd.; an application for designing SGML document type definitions (DTD).

Multi Level Text Representation in an LCB

87 —llemma.forml —I part, of. speech I

~H dictionary ffl horn TO entry

Figure 6: Retrieval of orthographic information from the multi level text representation.

The text.wrd.unit element in the sample element holds the text.word, the neutral wordform, and the lemma. This substructure matches the elements of the head element structure. When the orthographic dictionary is updated, frequency attributes in the head and the word, unit elements keep track of identical element contents (indicated here by a swung dash). The lemma element can occur zero or more times: It can be omitted when the source text is illegible for some reason, and it can occur more than once in connection with enclitic word forms.

ocA og og og itort/ ogsfl og ac/v v/ oj pron f/»7/ //7ye

tilje tilje ift f/7 fliA //7 JtoM/ //7 /;7

oc/i og og jto«y og50 og atA vi oi pro/i /Ai// //7 /»7 prap //7ye ///ye ji» /i/ fi/ a£n //7

Dorthe Duncker/Hanne Runs

88 prar/> ifewj

/tow/ ifewi

den

cf. H. Reszeg (, mint a) diszno (MSZK) (literally: 'drunk pig, or drunk as a pig'; used with reference to the way a drunk walks). Having fed itself it immediately goes to sleep. Because of this it is thought to be a m) 12

Note that an ink blot on a piece of paper used to be called diszno or malac in Hungarian motivated perhaps by the oval shape of a pig. Since this use of these lexemes has become rather obsolete, they will be ignored here.

On Animal Frames in English and Hungarian

121

To lazy pigs the ground is always frozen (CODP) H. Rest, mint a diszno (MSZK) (literally: 'He is as lazy as a pig') and n) animal. (By extension pig is used to indicate an unpleasant person and/or thing in British English, or, similarly hog and swine, an inconsiderate, selfish, illmannered and lazy person) Cf. What a pig! He refused to help, even though he could see that we were having trouble. (CIDE) a chauvinist pig Some drivers are real pigs. (OALD) He realised what a swine he was. (COBUILD) That tune ispig to play. (CIDE) Stop hogging the biscuits and pass them round. (OALD) ... a huge lorry hogging the middle of the road... (COBUILD) What can you expect from a pig/hog/sow but a grunt? (CODP). In Hungarian the semantic feature is supported by secondary lexemes as well. The derivatives disznosag (literally:'pig-ness') and malacsag (literally: 'piglet-ness') denote obscene or filthy talk and smut, respectively. In addition, the nouns disznolkodas (literally: 'pigging') and malackodas (literally: 'pigletting') denoting swinish behaviour, or dirty talk, and the corresponding verbs disznolkodik (literally: 'to pig') with the meaning 'behave like a swine' or 'to tell obscene things' and malackodik (literally: 'to piglet') with the meanings 'to tell indecent things' or 'to eat messily' are further examples for the saliency of this semantic feature. Finally, mention must be made of the attributive use of diszno and malac in phrases like diszno/malac beszed 'filthy talk, smut', diszno viselkedes 'swinish behaviour', diszno alak/ember 'filthy/inconsiderate/rude person'. On the basis of the above example semantic features k) and m) are found linguistically relevant in English; in Hungarian, on the other hand, the respective attributes are k), 1) and m). These features, similarly to those in APPEARANCE, also underlie the figurative uses of the lexemes in question.13

3.4. Thematic part 4: RELATION TO PEOPLE The semantic features identified in the above thematic part and the linguistic evidence supporting them implicitly have demonstrated that BEHAVIOUR and RELATION TO PEOPLE are firmly related. This can be put down to the fact that the behavioural patterns animals display, or more precisely, the way people conceive of these behavioural patterns greatly affect the relationship between people and animals. Consider, for instance, the semantic feature n) above: the description of it and the examples justifying it clearly indicate that a pig is perceived as unpleasant by people, because it displays such forms of behaviour that they find strongly rejectable. In this section, however, another aspect of the people-pig relationship will be discussed. Despite the mostly negative connotations triggered by their appearance and behaviour, pigs, being indispensable sources of meat and leather, are held to be basically o) . The examples below are intended to prove this economical usefulness. MoreFor more information on behavioural patterns related to animal frames see M. Neagu (1994).

122

SandorMartsa

over, these examples are also meant to confirm that pigs, at least diachronically, are associated with wealth, husbandry and, especially in Hungary, with luck.14 Cf. Feed the pig and you will have a hog. (CODP) One pig pays for his brother. (CODP) You know I'm not going to buy that kind of pig in a poke. (COBUILD)

I think piggybank, a the corresponding Hungarian malacpersely, indicating a small container, usually shaped like a pig, and used mostly by children to save money in may also belong here. In Hungarian a lot of secondary lexemes, primary compounds, as a matter of fact, are used as names for meals that can be made from pigs. Annual pig-killings, especially in country households, are still important events with a relatively great number of words and phrases related to them. (Note that in most names of meals and in words related to pig keeping series or malac can also be used as determiners.) Cf. disznoaprolek 'chitterlings', disznohizlalas 'fattening of pigs', disznohus 'pork', disznokaraj 'pork-chop', disznoöles 'pig-killing, pigsticking', disznotor 'dinner on pig-killing day', malacpecsenye 'roast pig/pork', sertosorja 'chine/fillet of pork'.

The phrases disznaja van or malaca van (literally: 'to have a pig/piglet') is used with reference to a lucky person. Moreover, in Hungarian cards aces are also called diszno. Cf. makkdiszno (literally: 'pigs of clubs; i.e. aces of clubs'). (In this connection see also footnote 13.) Finally, the description of the pig frame would not be complete if we did not mention that pigs (piglets, piggies) can also be viewed as p) and r) animals in both languages. Consider for example tales in which pigs are witty and smart animals, or the first line of the poem for children said while counting their toes This little piggy went to market... (LDELC) The English expression

Pigs might fly and analogous (though, already dated) Hungarian saying Ha repülni tudna, a legjobb szarnyas allot volna a diszno. (MSZK.) (literally: 'If a pig could fly, it would be the best bird')

are used to emphasise that a particular thing is very unlikely indeed to happen. Semantic features specified in this thematic part also confirm what has been said about the ambivalent nature of animal frames in section 2. Indeed, we have seen that on the one hand a pig is usually conceived of as a dirty, smelly and lazy animal of formidable appearance and rejectable forms of behaviour. On the other hand, we have also seen that from another perspective pigs are useful, witty and smart animals.

14

In Hungary, the piglet is believed to bring good luck if its tail is pulled on New Year's Eve. It is also a mascot animal symbolising luck.

On Animal Frames in English and Hungarian

123

4. Concluding remarks The aim of this article was to investigate animal frames as they reveal themselves through linguistic evidence. In the first part, it was demonstrated that due to taxonomic and attitudinal uncertainties concerning the status of animal in ethnobiological classifications and dictionary descriptions the generic animal frame is conceptually ambivalent. As the crosslinguistic analysis of the pig frame have shown, the ambivalent nature of the generic animal frame is partly inherited by individual animal frames in the sense that they often contain contradictory semantic features. It has been found that the frame or frame knowledge conjured up by pig is shared to a large extent by the speakers of the two languages. In a broader context, this probably can be put down to the fact that both cultures basically belong to the same European tradition. Some aspects of animal frames that are not or not easily accessible by using the methods of linguistic analysis could not be discussed in this paper. Thus we could not deal with individual varieties of the pig frame and the possible changes of this frame in a diachronic perspective. That is to say, we could not examine how the social, cultural and national identities of individual speakers determine or influence the actual content of frame knowledge; we could not examine either how the content of the pig frame, traditionally connected with rural life, have changed over centuries, or at least in this century. In this connection, for example, mention must be made of G. Orwell's Animal Farm which, being an enormous allegory of Russian Revolution, a piece of "modern animal bestiary " l5 and one of the best ever examples of discourse metaphor (cf. R. Dirven 1985), undoubtedly has contributed to the formation of the content of the pig frame. Finally, the problem of conceptual ambivalence has not been explored properly either. If domestic animals are really indispensable as they are from the point of view of the maintenance of human life, then how is it that their names, especially when they are applied to people, almost always generate negative emotions? By contrast, how is it that names for wild animals, with some obvious exceptions (e.g. names for reptiles), evoke as a rule positive feelings? Is it because in the long process of domestication people developed the feeling of superiority and dominance over cows, dogs, pigs, etc., and consequently it became a (verbal) assault if people were ascribed the features of creatures that were inferior to them? On the other hand, can the fact that wild animals have not (yet) been fully dominated by people explain why calling someone a tiger or a lion is the expression of appreciation and respect? To give adequate answers to these and other questions raised in the previous paragraph requires further research into the nature of animal frames.

5. Bibliography Apresjan, Jurij (1974): Leksioeskaja Semantika. Moscow: Nauka Beaugrande, R.de, Dressler, W. (1981): Introduction to Text Linguistics. London: Longman 15

This expression is taken from the the title of H. Neisi's article (1985).

124

Sandor Martsa

Berlin, B., Breedlove, D.E., Raven, P.H. (1973): "General Principles of Classification and Nomenclature in Folk Biology." In: American Anthropologist 75, 214-42. Berlin, Brent (1981): "The Concept of Rank in Ethobiololgical Classifications: Some evidence from Aguaruna Folk Botany." In: R. Casson (ed.) Language, Culture and Cognition. New York: Macmillan, 92-113. — (1992): Ethobiological classification: Principles of Categorization of Plants and Animals in Traditional Societies. Princeton, NJ: Princeton University Press. Brown, Cecil (1979): "Folk Zoological Life Forms : Their Universality and Growth." In: American Anthropologist 81, 791-817. Cruse, D.A. (1986): Lexical Semantics. Cambridge: Cambridge University Press. Crystal, David (1995): The Cambridge Encyclopedia of Language. Cambridge: Cambridge University Press. Fillmore, Charles., Atkins, Beryl. (1992): "Towards a Frame-based Lexicon: The Case of Risk." In: A. Lehrer, E.Feder (eds.) Frames, Fields and Contrasts: New Essays in Semantic and Lexical Organization. Hillsdale, NJ: Lawrence Erlbaum, 75-102. Hunn, E.S. (1976): "Toward a Perceptual Model of Folk Biological Classification." In: American Ethnologist 3, 508-24. Isaöenco, A.V. (1972): "Figurative Meaning, Derivation, Semantic Features." In: The Slavic Word. Proceedings of the International Colloquium at UCLA. The Hague-Paris: Mouton, 76-91. Kripke, Samuel (1972): "Naming and Necessity." In: D. Davidson, G. Harman (eds.) Semantics of Natural Language. Dordrecht: Reidel, 253-355. Lehrer, Adrienne (1990): "Polysemy, Conventionality, and the Structure of the Lexicon." In: Cognitive Linguistics 1-2, 207-46. Martsa, Sändor (1997): "Animal Verbs." In: M. Kurdi, J. Horväth (eds.) HUSSE Papers 1997, Proceedings of the Third Biennial Conference. Pocs: University Press, 314-22. Neagu, Mariana (1994): "Behavioral Properties of Animal Frames" In: Revue Roumaine de Linguistique Tom XXXIX , No 3-4. Bucarest: Editura Akademiei Romane, 403-7. Neisi, Hilary (1995): "A Modern Bestiary: A Contrastive Study of the Figurative Meanings of Animal Terms." In: ELT Journal 49/3, 272-8. Rosch, Eleonor (1975): "Universal and Cultural Specifics in Human Categorization." In: R. Brislin, S. Bochner, W. Lonner (eds.) Crosscultural perspectives on Learning. New York: Wiley, 177-206. Taylor, Tohn.R. (1989/1995): Linguistic Categorization. Prototypes in Linguistic Theory. Oxford: Claredon Press. Wierzbicka, Anna (1985): Lexicography and Conceptual Analysis. Ann Arbor: Karoma. — (1996): Semantics: Primes and Primitives. Oxford/New York: Oxford University Press.

The examples were taken from the following dictionaries: CIDE 1995 - Cambridge International Dictionary of English. Cambridge University Press. COBUILD 1990 - Collins Cobuild English Language Dictionary. Collins Publishers. CODP 1996 - The Concise Oxford Dictionary of Proverbs. Oxford University Press. LDELC 1992 - Dictionary of English Language and Culture. Longman. MESZ 1966 - A Magyar Nyelv Ertelemzo Szotära. Budapest: Akadomiai Kiado. MKSZ 1976 - O. Nagy, Gabor: Magyar Szoläsok es Közmondäsok. Budapest: Gondolat. OALD 1990 - Oxford Advanced Learner's Dictionary. Oxford University Press. WEBSTER 1985 - Webster's II New Riverside Dictionary. New York: Berkley Books.

Geart van der Meer

Further Ways to Improve the Active Dictionary: Collocations, Non-morphological 'Derivations', Grammar

1. Introduction

In this article I intend to deal with three seemingly unrelated subjects. Yet, they all often cause insuperable problems to the foreign learner, in the present case the learner of English as a foreign language, and even more in particular, the Dutch learner of English. I will try to show that the present active bilingual Dutch-English dictionaries - and by implication then of course any other similar active bilingual dictionary - should be improved in such ways that they offer the user more guidance with respect to common combinations to be translated into the target language (i.e. mainly the so-called collocations, but there are many other common word groups that are not really collocations) and grammar (i.e. the dictionary should not only translate, but give information about grammatical properties of the target words when necessary). In addition, the well-known English learner's dictionaries' would in my view profit from giving more information on what I have here called 'nonmorphological derivations'. By the latter I mean cases like maternal and maternity, which in a non-morphological but yet very significant way are the adjective and noun 'derivations' belonging to mother, in addition to what are normally considered the genuine derivations motherly and motherhood. As will be seen, most of these derivations use Latin or Greek suffixes. In the following pages I will deal with each subject in turn, but will already here say a little more about my decision to argue for the inclusion of non-morphological derivations 2 in learner's dictionaries. It is one of the great advantages of dictionaries that they are alphabetically arranged, according to form, i.e. according to the sequence of the letters of the alphabet. This in principle makes for a unique find spot for each item. Yet, this arrangement is totally arbitrary: none of the possible connections (morphological, semantic or oterwise) between words are systematically retrievable. Thus, the noun to the adjective opaque, i.e. opacity, will in dictionaries based on the alphabetical principle precede the adjectives by a number of entries, though in another case, e.g. profundity as the noun to profound, the noun will follow the adjective by a number of entries. And matters are obviously much worse in cases like maternal and urban (adjectives to resp. mother and city/town, but separated from them by a huge number of other entries). There has, therefore, throughout the history of 'wordbooks' been a strong pull in another direction: order-

1 2

LDOCE, OALD, CC, CIDE, HEED. Obviously, for learner's dictionaries, only the more frequently occurring ones.

126

Geart van der Meer

ing according to certain content principles. A thesaurus is a case in point, as might be the synonym dictionaries, though these are usually at the same time alphabetical.3 Of the major learner's dictionaries it is CIDE which has chosen to do the unusual thing by letting derivation - and also composition - take (a certain amount of) precedence over the strict alphabetical ordering principle. This dictionary in fact includes opacity in the opaque entry, and the same goes for profundity, which is found under profound.4 One of the reasons is apparently the desire to save space (cf. CIDE's Introduction p. ix), for obviously the sense definition for profundity may (largely) take for granted the definition for profound as given just a few lines higher up in the same entry. This is the only reason given in CIDE's front matter. There may, in addition, also have been the pedagogical lexicographer's desire to teach the learner about sets of words belonging together, but this has not been given as an official reason.5 Yet, however this may be, even CIDE has not taken the next - even bolder - step of cross-referencing from e.g. mother to maternal or from city to urban, let alone giving these 'derivations' in nested entries. In this contribution I will argue in favour of drawing attention to the major non-morphological derivational connections in learner's dictionaries. Dictionaries claiming to be learner's dictionaries - and pretending to be usable as active dictionaries as well - could usefully provide at least some degree of guidance to foreign learners in this difficult field. The highly composite and hybrid character of the English vocabulary - i.e. partly of Germanic origin but with a huge number of Romance (Latin & French) and Greek derivations - poses numerous vocabulary problems to the foreign learner. Learner's dictionaries would be the ideal instrument for solving at least a fair number of these problems.

2. Collocations and phraseology Over the past few decades an increasing amount of attention has been lavished on habitually occurring word combinations that are felt to be neither idioms - which are often said to be not

See Tom McArthur's wonderful Worlds of Reference. Lexicography, Learning and language from the Clay Tablet to the Computer (1986) for a history and survey of such reference works. Though the latest 1995 edition of the OALD still nests derivations and compounds under the headword, it does not go so far as CIDE here: both opacity and profundity are here in their proper alphabetical slots. See Van der Meer 1996 for the related problem of listing compounds under one headword, and cf. the quote from the OALD there: '[i]n this dictionary, compounds do not normally appear as separate entries but are included in the entry for the first element in each case (thus check-list appears at check and doorknob at door). This is helpful because it brings close together items which are often related in meaning' [emphasis mine, GvdM].

Further Ways to Improve the Active Dictionary

127

fully semantically transparent6 - nor totally free combinations. 7 1 am here referring to combinations like commit murder, follow somebody's example or take a decision. These are readymade linguistic building blocks larger than words that are used (by the native speaker) as units of form and meaning. These are, in Sinclair's words (1991: 110) a propos of the "idiom principle" "semi-preconstructed phrases that constitute single choices, even though they might appear to be analysable into segments" [emphasis mine]. As may be expected, collocations - as many linguists would call at least a number of the fixed combinations referred to - are another case of gradience: in the same way as not all idioms are equally opaque, not all examples of what are called collocations are equally transparent. In this light it is no wonder that opinions on the demarcation of this concept also vary somewhat. Thus, for Hausmann (1985: 118) collocations are typical, specific and characteristic word combinations, which he calls prefabricated or at least semi-prefabricated ("Fertigprodukte" and "Halbfertigprodukte"), and he adds that obviously not all combinations qualify for the term collocation. Thus, he thinks that a "banal" combination like ein Buch kaufen / buy a book is not really a collocation, presumably because it is not 'typical, specific and characteristic' enough. In Benson, Benson, llson (1986b: 253) we find another interesting discussion of the concept of collocations. They define collocations as "loosely fixed combinations" of the type to commit murder, taking up a position between "idioms, on one hand, and free combinations, on the other". Collocations are "psychologically salient"... "fixed phrases" (idem) differing from idioms in being semantically transparent, and from free combinations in being frequent and not freely variable by means of synonyms. This latter point (the impossibility of synonyms replacing the word in question) seems to imply that there is a certain amount of arbitrariness in the choice of collocators: out of a number of potential ("synonymous") candidates only a subset is normally selected.8 Thus, the authors assert, the only acceptable synonym, if at all, for commit in our example seems to be perpetrate, which does not normally occur with murder (idem: 253). Incidentally, it seems to me, the crucial word is here "synonym". It is often said that complete synonyms are very rare indeed. One may legitimately believe that the meaning of a word is also reflected in the company it keeps. In that case commit and perpetrate are certainly not full synonyms, for their sets of collocators only partly overlap (cf. be-

I should point out here that this cannot be taken at its face value, for - as has been amply demonstrated in Gibbs 1990 and Verstraten 1992 - what we feel are idioms are quite often not so much unanalysable as non-literal. Quite frequently idiomatic expressions are easy to analyse as metaphors, especially when they are instantiations of underlying conceptual metaphors like ANGER IS HEAT or ARGUMENT IS WAR (also cf. the groundbreaking Lakoff & Johnson 1980). So the point seems rather to be that collocations are non-metaphorical. Obviously, words will always have to fit together in a meaningful way, so that strictly speaking total freedom should not be taken too literally. The word 'arbitrariness' is explicitly used by Mel'chuk & Wanner 1994: 325. They also hint at the unpredictability or lack of motivation of the collocations, though they concede that there may be 'partial motivation' (idem). Yet, idiosyncracies abound in collocations (idem: p. 326), 'so that many of them have to be simply listed'. Also cf. Benson's own definition in 1989, p. 3: collocations are 'arbitrary, recurrent word combinations'. Benson also points out that the arbitrariness of 'collocations' is especially clear when languages are compared in this respect, something every learner will confirm.

128

Geart van der Meer

low). Saying that perpetrate and commit are synonyms therefore begs the question and leads to circularity. Collocations (the prototypical ones), therefore, are to be regarded as preconstructed combinations of words, where the constituent words are semantically literal (i.e. non-metaphorical). Because they are preconstructed they are conventional, i.e. frequent and psychologically salient. Due to their semantic literalness the meanings of collocations can be arrived at through normal analysis. We would then have the three following possibilities in the case of word combinations: FREE COMBINATIONS are: not preconstructed, and semantically literal (i.e. the words have retained their conventional literal meanings) COLLOCATIONS are: preconstructed, and semantically literal (i.e. the words have retained their conventional literal meanings) IDIOMS are: preconstructed, not semantically literal (i.e. the words, or at least one of them, have NOT retained their conventional literal meanings, or at least cannot be analysed as such)

In some more detail this becomes (as worked out in Van der Meer 1998b): 1. two or more lexical units, with meanings also occurring independently elsewhere (in other combinations); 2. which are used non-metaphorically, 3. which combination occurs repeatedly and normally in a language (cf. Carter 1987: 47), as a conventional building block, 4. which the language user has available as a whole, to express conventional established concepts, 5. whose constituent words are typically in a grammatical modifier - modified relation (including that of verb-object), 6. whose constituent words (in spite of point 1!) naturally select each other because the sense definition of the modifier includes the modified (and sometimes vice versa) in a non-banal way (semantic motivation) 7. which typically functions as part of a larger group and not as a complete utterance (sentence) itself.

This, however, is too restricted for the practical purpose I have in mind, the dictionary-supported help for using natural word combination in a foreign language, for there are many more kinds of standard word combinations. Carter (1987: 60) lists the following "types of fixed expressions" (after Alexander 1984): idioms, which are either unanalysable into separately usable words or in which at least one word is used metaphorically (spick and span, smell a rat), proverbs, as complete utterances (which may also frequently be metaphorical), in which the words moreover do not really necessarily select each other in a meaningful way (A watched pot never boils}, stock phrases, i.a. because the words do not really select each other in a meaningful way (When all is said and done, unaccustomed as I am to ..), catchphrases, being complete utterances or because the words do not really select each other in a meaningful way (That's another fine mess you got me into), allusions and quotations, idem (You've never had it so good), idiomatic similes, being unmotivated (as sober as a judge), and discoursal expressions, being complete utterances or because the words do not really select each other in

Further Ways to Improve the Active Dictionary

129

a meaningful way (How do you do, mark my words, we'll now take questions from the floor, Ladies and gentlemen, I thought you 'd never ask).9 The expression commit perjury is a collocation, for the meanings of both the verb and the noun are independently definable and used non-metaphorically. However, the expression (idiom) hoist with one 's own petard is not, for obvious reasons. A cover term for all these prefabricated word combinations might conveniently be phraseology (cf. Howarth 1997: 6). As a convenient definition we might use a rewording of Sinclair's definition of the "idiom principle": in phraseology we are concerned with "preconstructed or semi-preconstructed10 phrases that constitute single choices, even though they are analysable into segments". It is precisely these phraseological units that cause the learner no end of trouble. They are extremely numerous, and are hardly treated in the general bilingual dictionaries. I would here like to argue for either a more extensive treatment of these phrases in general bilingual dictionaries or else, which would even be preferable because there would be more space available, for a separate bilingual volume giving only phrases." To give an indication of the large number of difficulties encountered by the learner I will here print the outcome of my research on English equivalents of combinations ('collocations') with the Dutch word angst 'fear, etc.'. As will be seen, not all combinations are necessarily technically genuine collocations, but the point was that the translations of even the rather non-collocational ('free') combinations were not always unproblematic: the freedom of word combinations may not always extend across languages. ANGST (zie ook vrees) * fear (of), anxiety (about, over), dread, terror (of), * (psychologische «) anguish (at), » (« voor wat gaat geheuren, he t onbekende) dread (of), * (neurotische ») angst; D in verbindingen (zie ook vrees) (= fear tenzij anders aangegeven) (iem.) a. »aanjagen frighten; terrify; put the fear of God into sb, to inspire fear (of. terror), put (of. strike) fear in the hearts of (bijv youngsters); *aanwakkeren add to; *achtervolgen haunt; '•q/leggen, shrug off; *afnemen: diminish, subside; (van je) "afzetten shrug off; "algemene: common, public, prevailing (ook de a. is algemeen dat fears are rife that); *beheersen control; >bekruipen creep over (s.b.); '•bestaan (bij) (people) suffer

9

10

11

fear, have a fear of; met a. en *beven (iets tegemoet zien) (bijv view (of. await) sth.) with terror in one's heart, .. with fear and trembling, (bijv anticipate sth.) with (of. in) fear and trembling; >bevorderen promote; »bewaarheid worden be confirmed, realised; ••bezweren allay; ••bezwijken van bijv I almost died (of. passed out) with fear; >bij (de mensen): fear among (bijv the people); '•blijven bestaan persist; de '•boventoon voeren (pre)dominate; >diepe profound, deep, grave; >diepgewortelde deep-seated; *dodelijke deadly, mortal; *doen afnemen reduce, diminish, relieve; *doen herleven (of. opleven) renew, revive; *doen verdwijnen dissolve; >doen toenemen add to, increase, heighten; >door (bijv paralysed) with (of. by) fear, shaken with pain and

Respectively subcategorised as 'social formulae/cliches', 'connectives; structuring devices', 'conversational gambits', 'stylistic formulae' and 'stereotypes'. This is on purpose: phrases are often not completely fixed, allowing some degree of variation.

To date, there seems to be only one example of such a separate bilingual dictionary, i.e. Ilgenfritz etal 1989.

130 fear; »doorstaan suffer fears, be terrified, be anxious; »duizend angsten, zie uitstaan; »enorme: tremendous; »houden (de a. erin ») perpetuate; »gegronde reasonable, wellfounded, well-grounded; (zander ook maar de) »geringste: without the slightest fear; een »golf van a wave of fear; »groeien grow, increase; »groeiende ook mounting; »groot zijn: bijv fears are running high; »grote big, great, grave; iem. om het »hart slaan: to be gripped by; met a. in het »hart with fear in his heart; »hebben (voor) be afraid, scared (of), be (of. stand in) dread (of. fear) of; »hebben ook experience, have (got) a fear (bijv of failure), be in fear of; »heersen rule (bijv »er heerst a. dat there are widespread fears that); »hevige (of. heftige) strong, tremendous; »herleven: revive; »in a. verkeren, zijn, zitten e.d be afraid (of. scared), be (go, live) in fear (and trembling), in terror, be in a fright (of. panic), be under fear; »inboezemen strike fear (terror) in the hearts of, inspire fear, instil s.b. with fear, instil fear, inject s.b. with fear, geen a. inboezemen (bijv their surroundings) hold no fear (no terrors); »ingegeven door a. prompted by fear; een »klimaat van: a climate of fear; »koesteren stand in fear (of); »krijgen begin to fear, begin to be afraid; »leiden (zieh laten leiden door) be impelled by; »leven (bij) there are fears (bijv among employees) that; met a. in zijn »·////with fear in his heart; het »loon der a. the Wages of Fear; »met bijv people living with the fear of death; »niet-aflatende nagging, niggling, lingering; de a. »om fears for (bijv your (of. people's) livelihood); »om te fear of (bijv failure); »on(der)bewuste subliminal; »onberedeneerde irrational; »onderdrukken suppress; »ongegronde idle, groundless, unfounded; »onterechte unjustified; »ontstaan arise; »overheersen (pre)dominate; »opkomende rising; »opleven revive; »oproepen cause, provoke, give rise to, spark, evoke, raise, generate, engender, arouse, create, prompt, trigger; »over about, over (bijv the Aids virus); »overdreven over-anxiety; »overmand door be gripped, overwhelmed by; »overwinnen overcome; »permanente constant; a. is een siechte »raadgever: fear is a bad counsellor; »schuilen achter (bijv behind these comments) lies the fear that, (bijv

Geart van der Meer the fear is always there,) lurking behind those crocodile tears; »sluimerende dormant; »sterker worden zie »toenemen; »sterven van die (of. pass out) with fear; »stijf (van) be rigid (of. petrified) with fear; »stijgende growing fear(s) (of: anxiety); »stille unspoken; »sussen soothe, allay; met enige a. »tegemoet zien: anticipate with some trepidation; »teweegbrengen zie oproepen; »toenemen increase, deepen, grow; »toeslaan bij iem. bijv I was gripped by fear; »uit a. (voor) for fear (of), out of fear, from fear (of), on the fear that, (bijv the words were uttered) through fear of (the supposed killer); »uiting geven aan express, voice; »uitspreken express, voice; »uitstaan suffer fears, be terrified, be anxious, be terror-stricken, suffer agonies (of fear), duizend angsten uitstaan endure a thousand terrors); »van (bijv sidderen van a.} (shake, tremble) with fear; »verdrijven dispel, expel; »verdwijnen disappear, vanish, evaporate; »vergroten add to, increase, heighten; in a. »verkeren be afraid, have fear(s) that; »verlamd van numb with fear, frightened out of one's wits, rooted to the ground with; »verliezen lose; »veroorzaken zie »oproepen; »verspreiden spread (of: pass on) fear, »versterken heighten, reinforce, exacerbate; »verstijfd (van) rigid (of: frozen) with fear (of: terror, dread); »vervuld (van): imbued with fear; »vervullen met a. fill a p. with fear (of. dread); »voeden feed, fuel, give fuel to; »voor (fear) of (zie verder vertaalprofiel); »voornaamste main; »voortdurende continued, perpetual, chronic, constant, nagging; »in a. en vreze in fear and trembling, in dread and fear; »wakker maken (bij) activate fear (among); »waren (door) fear is stalking (bijv the streets of Belfast); »wegens a. voor because of, for, out of, on fear(s) of; »wegnemen drive out, expel, quell, remove, defuse; »\vekken zie oproepen; »\vijdverbreide widespread, rampant, (ook er heerst een wijdverbreide a. dat: fears are rife that); »würgende strangling; »zaaien sow, spread, unleash, whip up; er »zijn bijv »er is grote a. dat; de a. is dat the (big) fear is that; in a. »zitten zie verkeren; erin »zitten bijv de a. zit er goed in fears are running high; »zander without (bijv the slightest fear that)

Further Ways to Improve the Active Dictionary

131

3. Non-morphological derivations As already stated in the introduction, the hybrid and opaque character of the English vocabulary presents extraordinary problems to the learner and, one may surmise in the light of the evidence, the native speaker as well. 12 The mixed origin of English has caused countless words to be formally totally unrelated, despite an obvious semantic link, a link in fact which in languages like Dutch and German is often much more clearly present by means of morphological derivation and composition. Thus, the adjective to Dutch koning 'king' is koninklijk 'royal, regal', a derived form which is clearly semantically motivated and nonopaque,13 whereas the English equivalents are not. They are opaque morphologically, though the -al suffix to some extent does at least hint that the words are adjectives. The fact that English, in addition, also has kingly shows the extraordinary richness of the English vocabulary, and with it its complexity, for the three words are certainly not synonymous. For the general meaning Of, or pertaining to kings/queens' Dutch has just the one word koninklijk.™ The mainly French, Latin and Greek foreign heritage of the English language has made huge swathes of its vocabulary unmotivated, or one might also say, opaque. As a further example, consider the words opaque and transparent themselves. The 'derived' nouns are respectively opacity and transparency, somewhat less than productive formations.15 Also the words themselves are totally unmotivated, apart maybe from the fact that the -ent of transparent may be intuited as an adjectival ending. Perhaps the 'prefix' trans- also adds some modest amount of semantic motivation. Now consider the Dutch equivalents, which are respectively doorzichtig and ondoorzichtig, meaning 'through-see-able' and 'unthrough-see-able'.16 The derived nouns are doorzichtigheid and ondoorzichtigheid 'through-see-able-ness' and 'un-through-see-able-ness', by means of adding the nounforming suffix -held. German, whose vocabulary is often even more systematic and hence morphologically transparent than that of Dutch in this respect, offers the wonderful example of the four words Jahr, Jahrzehnt, Jahrhundert and Jahrtausend, for which English has year, decade, century and millennium, which in a nutshell sums up the origins of English (from Germanic, Greek, and Latin (twice) respectively).17 For a quite original study of English word transparency compared with German I refer the reader to Alwin Fill's Wortdurchsichtigkeit im Englischen (Fill 1980), which to once again illustrate the point I am making translates literally as 'Word-through-see-able-ness etc.'.

12

13 14

15

16

17

Cf. the frequent error of mitigate against instead of militate against in English papers, caused by the phonetic similarity of two otherwise formally opaque, i.e. unmotivated, words. Though the [k] in koninklijk is due to historical conditions, and is no longer really productive. Dutch does have the word rojaal, but this only means 'generous', and there is therefore no link with koning. Despite the mind-boggling absurdities in Chomsky and Halle's Sound Pattern of English (1968) and similar works. Strictly speaking, -zieht- is a noun meaning 'vision', so that the gloss might instead be 'througsight-ish(-ness), etc. See Fill 1980: 136 for this wonderful example.

132

Geart van der Meer

Though speakers of English may marvel at the ponderousness of such words they are certainly not felt to be such by speakers of these languages, who will often wonder why on earth the adjective to sun is solar, as in solar eclipse, which to them is simply 'sun-darkening' or 'sun-darkness', cf. Dutch zonsverduistering or German Sonnenfinsternis. Now sun and solar still have some initial sounds in common, but the situation is dramatically more difficult in most other cases, as for example m filial devotion, where there is phonetically little resemblance between son/daughter and filial. Of the four best-known learners' dictionaries OALD and CIDE frequently nest morphologically derived forms under the headword, but almost nowhere do they refer to the non-morphological derivations like filial by means of a cross-reference. The same holds for CC and LDOCE, which give the derivations in their strictly alphabetical slots. Since learners' dictionaries are intended for active as well as for passive use, it comes as something of a surprise that the user of such reference works cannot possibly make the step from e.g. son to filial. I therefore suggest that at least in the case of the not too infrequent words this information be given by means of cross-references, as from king to ROYAL, REGALE In itself such cross-referencing would be nothing new in these dictionaries. It would take up some space, but not exorbitantly so, for the fact is that the number of words concerned is relatively small, as can be seen from the list below. Though the number of non-morphological derivations in the language as a whole is huge, quite a lot of them belong to specialised parts of the vocabulary, which would not be included in a learners' dictionary. I here print a list of the less uncommon words I have been able to find. Some readers may think some of them are rather too specialised or infrequent for inclusion in a learners' dictionary, yet I give the entire list just to give an impression of the extent of English morphological opacity. Most of the word are nouns, with adjectival 'derivations', but there are also some verbs and adjectives with their own 'derivations'.

3.1. List of words and their non-morphological 'derivations' air - aerial (assault, photography) angle - angular (face) ape - simian (forehead) back - dorsal (fins) bad - deteriorate (health), worsen (political situation) ball - spherical (object) beast - bestial (savagery), beastly beginning - initial (stages) believe - credible (alternative), believable (explanation) bend - flexible (rules), bendable bird - avian (malaria) bishop- episcopal (duties) blood - sanguinary (wars), bloody body - physical (defect), corporeal (needs), corporal (punishment), somatic (cells) bone - ossify brain - cerebral (haemorrhage) break - fragile (china, peace), breakable In addition, it would be wise for ROYAL and REGAL to cross-refer too.

Further Ways to improve the Active Dictionary

133

brother - fraternal (feelings), brotherly calf-veal carry - portable (phone) cat - feline (grace) century - centennial, centenary (celebrations)19 chalk - calcareous (shells), chalky chest - pectoral (muscles) child - infantile (diseases, behaviour), puerile Gokes. behaviour), childish, childlike church - ecclesial (teachings), ecclesiastic(al) (history) citizen - civil (strife, disobedience) city - urban (development) coast, shore - littoral (forests) compel - compulsive (liar, eating), compelling (novel), compulsory (education, subject, test) country(side) - rural (community), rustic (farmhouse, idyll) cow - bovine (diseases, stupidity, cf. BSE); beef daughter - filial (duties) day - diurnal (rhythms, clock), daily death - lethal (weapons, blow), mortal (wound, enemy), deathly, deadly deceive - deceptive (appearances) dog - canine (tooth, behaviour) donkey, ass - asinine (remark) ear - aural (test) earth - terrestrial (TV channels), earthly, earthy east - oriental (art), easterly, eastern eat - edible (snails, fungi), eatable OALD: > edible; CIDE idem20 eighty - octogenarian21 enemy - hostile (attitude), inimical (climate, conditions, influence); OALD inimical: > hostile error - erroneous (conclusion) example - exemplary (behaviour) eye - ocular (defects) fast - accelerate (rate of growth), speed up father - fatherly, paternal (authority), fatherhood - paternity fever - febrile (activity), feverish flesh - carnal (desires, knowledge), fleshy flower - floral (pattern), florid (style), flowery; CIDE > floral friend - amicable (settlement), friendly god - divine (wisdom, inspiration), godlike, godly good - improve (health), ameliorate (working conditions) hair - hirsute (animals), hairy heart - cordial (smile), cardiac (arrest), hearty heat - thermal (energy) heaven - celestial (bodies, beauty), heavenly heavy - gravity, heaviness 19

20

21

Though strictly speaking this is Latin from Latin, I have included a couple of such cases because these derivations, too, are not so easy for foreign learners. In this, and in some other cases, OALD and/or CIDE, though very infrequently and unsystematically, have done exactly what I am arguing for here: giving cross-references to latinate 'derivations'. Also cf. hexagenarian, septuagenarian and of course nonagenarian.

134 hell - infernal (heat, machine), hellish iron - ferrous (metals), iron joke - jocular (remarks), jokey king - regal (splendour), royal (family), kingly (bearing) lie - recumbent (figure), lying life - vital (functions, organs), lively light - luminous (paint), illuminate lion - leonine (head) lips - labial (sounds) love - amorous (looks, advances), loving, lovely man - male (child, voice) - masculine (face, word) - virile (body), manly many - multitude, mass; multiply, proliferate marriage - nuptial (bliss, ceremony) merchant - mercantile (interests, ships) mind - mental (image, process) mistake - erroneous (beliefs, conclusions), mistaken(ly) money - pecuniary (gains, advantage), financial (gains) monkey (cf. ape) - simian (forehead) moon - lunar (eclipse) mother - motherly - maternal (grandfather, love) - motherhood - maternity mouth - oral (exam, sex) move - mobile (phone, home), movable (goods, items) new - novelty, newness night - nocturnal (visit, animal), nightly north - boreal (forests), northerly (winds), northern nose - nasal (voice, sounds) old - senescent,senescence owner - (proprietor) proprietorial (air, rights) peace - pacific (community), peaceful; pacify people - popular (vote, misconception, wisdom) picture - pictorial (encyclopedia, history), picturesque (cottage, village) pig - porcine (aroma), piggish; pork place - local (authorities, hero) rain - pluvial (weather conditions), rainy read - legible (signature, hand), readable; OALD: > legible receive - recipient (n.), receptive (audience, market) river - fluvial (deposits) river bank - riparian (wildlife) rule - regular (routine, intervals) salt - saline, salty (solution), salty (taste) sea - marine (biologist, pollution, habitat) sheep - ovine (offal), sheepish; mutton side - lateral (branch, pressure) sight, see - visual (field, arts, image) sister - sisterly, sisterhood, sorority sit - sedentary (life), sitting skull - (cranium) cranial (nerve injury) sleep - dormant (inflation, giant), somnolent (cat, village), sleepy slow - decelerate, slow down snake - serpentine (course of the river), snaky son - filial (duty, love)

Geart van der Meer

Further Ways to Improve the Active Dictionary

135

sound - phonetic (alphabet, elements), phonic (skills), acoustic (properties, possibilities), sonic (waves, boom) south - austral (winter), southern, southerly space - spatial (awareness), spacious (room, kitchen); both in CIDE under space; spider - arachnoid, spidery spring - vernal (equinox, grove) star - stellar (constellations, performance, career), sidereal (day, zodiac), astral (bodies, navigation), starry sun - solar (eclipse, temple), sunny table - tabular (form) three - treblepredet. (the usual charges), treble adj. (line), triple (murder, jump) time - temporary (accommodation), temporal (perspective, distance), timely tooth - dental (surgeon, treatment), toothy tree - arboreal (birds) two - double (figures) uncle - avuncular (manner) understand - intelligible (English), understandable voice - vocal (chords, performance) wall - mural (design) war - martial (arts, law), warlike water - aquatic (sports, environment), aqueous (lotions), watery west - occidental, westerly, western whale - cetacean (anatomy) winter - hibernate wolf- lupine (pack), wolfish woman - womanly - female (animal) - feminine (voice, figure), effeminate (manners), womanhood, femininity, womanliness word - verbal (abuse, skills), verbose (style), wordy world - global (warming, problems), mundane (life, matters), worldly year - annual (salary, report), yearly young - rejuvenate youth -juvenile (delinquency, offender), youthful

4. Grammar Everyone using an active bilingual dictionary will occasionally have wondered why such dictionaries take the user only part of the way. Thus, after having found the word research the user is usually not warned that *a research would be incorrect English. Likewise, having found the verb prevent it may well be that no hint is provided that the continuation should not be * (prevent) that but (prevent) from (we must prevent them from making the same mistake twice; this should be prevented from happening; cf. Van der Meer 1997 for this and similar problems). Though there may be respectable reasons for this, such as space (always a major concern for lexicographers) or an assumption that the grammar should take account of this (again cf. Van der Meer 1997), I see no real reason why active bilingual dictionaries should not be made to resemble learner's dictionaries, which do in fact provide assistance here, a little more in this respect. It is, at any rate, extremely misleading and un-

136

Geart van der Meer

satisfactory to suggest the target word research without attaching the warning that it cannot - like the source word - be preceded by the indefinite article.22 It is, indeed, "rather difficult to accept that the dictionary user is given less information on the foreign language, which is of primary concern in the active bilingual dictionary, than on his native language. This goes for the phonological, morphological and syntactic levels" (Maria Toporowska-Gronostaj 1988 : 346). Ϊ will here list a number of cases, based on Van der Meer 1997, where translating dictionaries should make sure that the necessary information is given to the foreign user, in this case of English. The suggestions fall under a number of headings, presented in random order. Since the source language in the 1997 article was Dutch, it may well be that problems encountered in other language pairs will be overlooked, but the following is meant as suggestive and not intended to be exhaustive.

4.1. Adjectives For well-known pairs like ill/sick and drunk/drunken the translating dictionary should make it quite clear that there are distribution restrictions: only the latter are eligible for the attributive positions.23 Likewise, if the word afraid is given as a target word, there should be some indication that it does not occur before its head noun. The comparison of adjectives might be given some attention, as in the case οι politer and more polite, though this is perhaps not the most pressing problem. More importantly, the formation of adverbs from adjectives should be dealt with in a systematic way. Thus, one should like to hear when slow is to be used and when slowly is preferred, or whether the adverb to friendly is in facifhendlily or perhaps better avoided. All this need not be done in an abstract way, but could simply be pointed out in a short example, or if there is space enough, by means of a source language example and.a target language translation.

4.2. The articles (definite and indefinite) A grammatical difference between English and other languages like Dutch, which is responsible for many errors, is the presence/absence of articles. English, in words like society, literature, culture and history, often has no indefinite article, where other language frequently do. Again, no general rules are needed here, a simple example plus translation will nicely do the job. Such simple examples could also be used for examples like go by train/boat and go to church/university, where other languages often use definite articles. Attention should also focus on the use of the indefinite article a in English in cases like he is a teacher, where a language like Dutch, for example, normally does without the article (hij is leraar). It would perhaps be feasible to solve this particular problem by giving the information, by means of examples, under the source language equivalents of English to be: he is a 22

23

Assuming, of course, that from a number of possible candidates (examination, investigation, inquiry, survey, poll, study...) the correct word has been chosen with the help of the dictionary. It is true that ill is certainly not excluded from attributive positions, yet it seems unwise not to say what most modern monolingual dictionaries say, e.g. LDOCE 'not usually before noun'.

Further Ways to Improve the Active Dictionary

\ 37

lawyer, they are lawyers.2* The latter example (with lawyers) incidentally illustrates the related problem of the plural, where other languages would have singular words.

4.3. Verb complementation Another hard nut to crack for many learners is verb complementation. This would be easy to handle in a dictionary, since this phenomenon is essentially word-based (cf. the by now wellknown verb-patterns in the English learner's dictionaries). Thus, Dutch willen dot + finite clause ('want that') corresponds to the English object with infinitive (/ want you to arrive at six), but Dutch voorkomen dat 'prevent that' is in English prevent Xfrom +-ing: we must prevent this from happening again. Yet, as I have shown elsewhere (Van der Meer, 1997), bilingual dictionaries do not always manage to give this information consistently. There are, in this same area, other cases where English has a choice between a clause complement and a gerund complement, as in he admitted that he., and he admitted having.. Though it is not strictly necessary, it would still be useful for users to be told about this choice in bilingual dictionaries. English constructions like / would prefer you not to go or gerundial constructions like / prefer walking to cycling could also easily be treated with more consistency in active translating dictionaries.

4.4. Inversion Unlike Dutch and other Germanic languages, English has little inversion of subject and finite verb in non-interrogative sentences, yet in those restricted cases where this inversion should occur foreign users of English often do not observe it, as in sentences beginning with never, rarely, nowhere, or only.25 For the first three the two dictionaries I examined, for example, do not provide any information at all about this phenomenon. Presumably, this state of affairs will obtain in many other dictionaries as well.

4.5. Nouns: uncountability, plurals, singulars In the field of the (un)countability of English nouns there is a great need for user guidance. Translations like *a research or *an advice (in the intended sense of'a recommendation') are to be prevented. Similar well-known cases like (pair oß glasses/trousers are potential danger areas in that it is not always made absolutely clear that *a trousers instead of a pair of trousers is wrong.

24

25

I realise that a word like 'be' is not usually consulted in bilingual dictionaries. Perhaps the most frequent nouns which may be expected to present this problem should be treated in this way as well. The number could easily be multiplied with expressions with no, like on no account, under no circumstances, in no case, by no means, on no account, not on any account.

138

Geart van der Meer

Foreign plurals such as phenomena, fungi, amoebae should be given in the target part of the entries, and even quite regular English plurals like museums, where for example Dutch has the Latin plural musea, might be given, just to avoid any misunderstandings. English differs from Dutch and other languages in that after a number of (Dutch een aantal) the verb should normally be plural: a number of persons were late. Dutch usually uses a singular verb here. Another case involving an English plural noun where other languages frequently have singulars is the above-mentioned case of they are teachers (not: *they are teacher). Another case involving the plural/singular problem which may be troublesome to learners is English it is .. who/that, where Dutch has a plural: het zijn.. die (= plural form of 'to be') and where *it are ... -who/that is a common error. With regard to the well-known much/many problem it should be remarked that translating dictionaries must be expected to clearly point out the differences here between much + sg. noun and many + pi. noun. Other cases involving the plural/singular problem are police (the police are), 26 per cent is/are, six new staff (\.Q. staff members), The Netherlands is. the United States is (not usually are), my family is/are, the majority is/are to mention just a few cases.

4.6. Gender An area where users of English often make mistakes as well is gender as reflected in pronoun use. English may have lost the gender system for its noun and adjective inflection, a development we should indeed be very grateful for considering the problems posed by languages like German and Russian,26 but this does not mean there are no remaining problems, such as the fact that things, but also animals, babies etc. are referred to as it, and that she may be used used for e.g. ships, trains, machines and contraptions, countries, big cities etc. Perhaps dictionaries could in some way or other take the trouble to tell us - in the pronoun entries - that the pronoun reference to the target word car may also be she and that, on the other hand, a target word like meeting or government must never be referred to as *she, as for example Dutch people often do. A related problem is gender-free generic references, in order to avoid the today increasingly unacceptable generic he when women are also meant. The dictionary should show under source words for 'he' that there are various options available, such as they, following for example non-gender words like a person, someone, somebody, etc. (... someone / a person ....they...).

4.7. Pronouns In the area of pronouns improvements are possible as well. One of the notorious pronouns is Dutch men (also cf. German man, French on) which indeed can be people, they, we, you, a man or be equivalent to a passive. But this Dutch word has various meanings and hence various different equivalents. A search through two major Dutch dictionaries reveals that the user is given plenty of opportunity to choose the incorrect equivalent. The relative pronoun, like26

Even Dutch and French, with their two-gender 'system' and much less complicated adjectival inflection cause no end of problems, due to the arbitrariness of what is in fact no system at all.

Further Ways to Improve the Active Dictionary

139

wise, is a source of errors: it will simply not do, as one dictionary does, to simply list, without any comment, who, which, that. Combinations like it's me (for Dutch ik ben het' (Ί am it') should also be given the attention they deserve, but do not always get.

4.8. Concluding remarks and suggestions for improvement It is quite obvious that grammatical tense is extremely difficult to treat in a normal dictionary, since it is not word-bound. Apart from some isolated cases, this will have to remain an area which the (bilingual) dictionary will not be able to cover. This applies of course also to problems like word order. Yet, as demonstrated above and indeed in learner's dictionaries, there do remain many opportunities for improvement. I will now make a number of general suggestions to improve active bilingual dictionaries in the field of grammatical information about the target language equivalents. These suggestions will of course of necessity be somewhat general. To start with, active bilingual dictionaries would benefit from a greater awareness on the part of the compilers of the grammatical difficulties their suggested equivalents, once chosen, could cause the user. The compilers' responsibility does not stop at simply giving, for example, research, hopefully after suggesting the proper choice from various alternatives (examination, inquiry, etc.}. Since, without any further support, an error like *a research is quite predictable this should be forestalled, for instance, by explicitly telling the user that research is uncountable: research (U).27 This would be the most space-saving method, but it should be realised that users often overlook abbreviations or simply do not bother to try to understand them. If space is not the overriding factor, more concrete instructions could be given, such as explicitly translating into anything but *a research, and/or by explicitly indicating 'not *a research'. Obviously, this choice between telling the user rather abstractly about grammatical properties by means of (abbreviated) grammatical terms or through concrete examples and/or warnings has to be made in many other cases. Thus, no bilingual dictionary can really shirk its responsibility of keeping apart sick and ///, or drunk and drunken. Here, simple relevant example sentences would easily do the job. Mistakes like *the western culture, *he is / they are teacher, *we should avoid that this happens again, *Never I have.., *it are the Dutch who.., *much people think that.., *the government in her latest meeting.., *the police is.., *in the Victorian age one still believed that.., *the government Major are all easily prevented by properly alerting foreign users to the dangers, again most profitably by providing relevant examples (plus cautionary remarks, if still necessary). It is the job of the lexicographer to choose the most appropriate word(s) under which to provide the required data. Thus, it is under the source language words for 'to be' or 'to become' that the he is a teacher / they are teachers example should be given, as otherwise the same thing would have to be said again and again under each separate noun (which of course, in itself, would be fine if space allowed it).28

27

28

As I realise full well that the plural researches is perfect English, a way should be found to point this out too. I realise that dictionary users do not normally expect such information under such general words, but I trust that sooner or later they will discover what the dictionary has on offer.

140

Geart van der Meer

Likewise, the more morphological trouble areas, like irregular plurals or verb forms, should be covered either directly, by for example telling that, for instance, Dutch nijlpaardis English hippopotamus (plur. -uses or -mi)29 - which does not require too much space - or by referring the user to a list of irregular verbs, or irregularities in general, at the back of the book, e.g. stride"3, the superscript number Ί33' leading to the required information in a list at the end. This latter method has its drawbacks in that it requires the users to look elsewhere, which they usually resent, but they will at least have been alerted to the danger. The Dutch bilingual dictionaries I studied do pay a certain amount of attention to grammatical problems connected with the translation equivalents they offer, but this attention does not seem to be based on a really consistent or well thought out policy. The time has now arrived for the revolution of the learner's dictionaries over the past few decades to be widened to the active bilingual dictionaries. Attention should no longer be restricted to choosing the correct equivalent, which in fact places the main focus on semantic equivalence - which aspect is itself still in need of improvement as well -, but there should be a stronger spotlight on the grammatical properties of the equivalents themselves. This means taking a step away from the direct interface between source and target language towards a stronger focus on the target language itself, as in monolingual dictionaries. It is now time to emulate the great strides forward taken in the monolingual learner's dictionaries by incorporating the grammatical information found there in active bilingual dictionaries, at least as far as it is word-based. Considerations of space will of course place restraints on this in paper dictionaries, but not really in the case of electronic versions. It should never be lost sight ofthat every active bilingual dictionary is also in its own way a learner's dictionary, learners who must be presumed to be "imperfectly familiar with the target language, and therefore to have incomplete or no intuition as regards the construction" (Svensen 1993: 87).

Bibliography Dictionaries referred to CC = Collins Cobuild English Dictionary, 2nd ed (London, 1995). CIDE = Cambridge International Dictionary of English (Cambridge, 1995). HEED = Harrap's Essential English Dictionary. (Edinburgh, 1995) LDOCE = Longman Dictionary of Contemporary English, 3rd ed. (Burnt Mill, Harlow, 1995). OALD = Oxford Advanced Learner's Dictionary, New Edition, 5th ed. (Oxford, 1995).

Other references Alexander, R. J. (1984). "Fixed Expressions in English: Reference books and the teacher"; English Language Teaching Journal, 38: 127-34.

29

Obviously, the metalanguage of an active Dutch-English dictionary would be Dutch, so that 'plur' really should be 'meerv(oud)'.

Further Ways to Improve the Active Dictionary

\ 41

Benson, Morton (1989). "The structure of the collocational dictionary"; International Journal of Lexicography 2: 1-14. Benson, Morton, Evelyn Benson & Robert Ilson (1986). Lexicographic Description of English; Amsterdam. — (1998). The BBI Dictionary of English Word Combinations; Amsterdam/Philadelphia: John Benjamins. Carter, Ronald (1987). Vocabulary. Applied Linguistic Perspectives; London: Allen & Unwin. Cowie, A P, R Mackin, I R MaCaig (1983). Oxford Dictionary of Current Idiomatic English. Volume 2: Phrase, Clause & Sentence Idioms; Oxford: Oxford University Press. Fill, Alwin 1980: Wortdurchsichtigkeit im Englischen. Eine nicht-generative Studie morphosemantischer Strukturen. Mit einer kontrastiven Untersuchung der Rolle durchsichtiger Wörter im Englischen und Deutschen der Gegenwart. Innsbruck. Fontenelle, Thierry (1997). Turning a Bilingual Dictionary into a Lexical-Semantic Database; Tübingen: Max Niemeyer Verlag (Lexicographica Series Major 79) Gibbs, Raymond W (1990). "Psycholinguistic studies on the conceptual basis of idiomaticity"; Cognitive Linguistics 1-4:417-51. Hausmann, F.J. (1985). "Kollokationen im deutschen Wörterbuch. Ein Beitrag zur Theorie des Lexikographischen Beispiels"; in Bergenholtz/Mugdan (eds.), Lexikographie und Grammatik 118-129. Herbst, Thomas (1996). "What Are Collocations: Sandy Beaches or False Teeth?"; English studies. Vol. 77 (1996), no. 4, pag. 379-393. Howarth, Peter Andrew (1996). Phraseology in English Academic Writing. Some Implications for Language Learning and Dictionary Making; Tübingen: Max Niemeyer Verlag (Lexicographica Series Major 75) Ilgenfritz, P., N. Stephan-Gabinel & G. Schneider (1989). Langenscheidts Kontext-Wörterbuch Französisch-Deutsch', Berlin-München: Langenscheidt. Lakoff, George & Mark Johnson (1980). Metaphors we live by; Chicago: Chicago University Press. McArthur, Tom 1986: Worlds of Reference. Lexicography, Learning and language from the Clay Tablet to the Computer; Cambridge etc. Mclntosh, Angus (1966). "Patterns and ranges"; in Mclntosh, Angus, Halliday, M.A.K.. (1966). Patterns of language : papers in general, descriptive and applied linguistics; London: Longmans. Meer, Geart van der (1996): 'How alphabetical should a dictionary be? (the case of HIGH and its combinations in some dictionaries)'; SYMPOSIUM ON LEXICOGRAPHY VII. Proceedings of the Seventh Symposium on LexicographyMay 5-6, 1994 at the University of Copenhagen (Lexicographica, Series Major 76), edited by Arne Zettersten and Viggo Hjernager Pedersen. [pp. 183-197] — (1997): 'Grammar, construction information, and collocations in two active bilingual dictionaries. A comparison and assessment of Wolters' Nederlands-Engels and Van Dale Handwoordenboek Nederlands-Engels'; in Leuvense Bijdragen, vol. 86, pp: 109-123. — (1998a). "Het actieve vertaalwoordenboek en collocaties" ; Trefwoord 12: 134-142. — (1998b). 'Collocations as one particular type of conventional word combinations.Their definition and character' in: Euralex 1998 Proceedings /-//, Papers submitted to the Eighth EURALEX International Congress on Lexicograhpy in Liege, Belgium. Vol. I, pp. 313-322. Mel'chuk, Igor & Leo Wanner (1994). "Towards an Efficient Representation of Restricted Lexical Cooccurrence". In Euralex 1994 Proceedings. Papers submitted to the 6th EURALEX International Congress on Lexicography in Amsterdam, The Netherlands, pp. 325-338. Sinclair, John (1991). Corpus, Concordance, Collocation; Oxford. Svensen, Bo (1993). Practical Lexicography. Principles and Methods of Dictionary-Making; Oxford. Toporowska-Gronostaj, Maria 1988: 'Contrastive valency in the bilingual dictionary', in Studies in Computer-Aided Lexicology; ed. by Martin Gellerstam et al.; Göteborg. Verstraten, L P (992). Vaste verbindingen; Utrecht: Led

Stefan J. Schierholz Governed Prepositions A Database for German, English and Portuguese Nouns

1. Introduction and Outline of the Problem The Erfiirt research project, "Pr positionsvalente Substantive - Trilinguale Grundlagenforschung zum Deutschen, Englischen und Portugiesischen" (Prepositional valencies of nouns - basic trilingual research in German, English and Portuguese)1 (abbr.: SUBVAL), investigates nouns which can govern one or more prepositions, and which occur as the head of a complex nominal phrase. Such nominal phrases (cf. (1) and (2)) have recently begun to be referred to as prepositional attribute constructions (= ΡΑ-constructions) in the specialist lexicographic literature.2 (1) die Jagd NACH den Dieben (2) # die Jagd zu den Dieben ( l a) the hunt FOR the thieves (2a) # the hunt to the thieves (1 b) α οαςα A OS (= A os) ladr es (2b) # α οαςα para os ladrdes3

Selecting the appropriate preposition to formulate a correct ΡΑ-construction poses a problem, particularly for the non-native speaker. This can be illustrated by the contrast between governed and non-governed prepositions.4 (3) (3 a) (3b) (4) (4a) (4b)

die Jagd NA CH den Dieben the hunt FOR the thieves α οαςα AOS (= A os) ladr es die Jagd nach dem Mittagessen the hunt after the dinner α οαςα depots ο αίπιοςο

In (3), the preposition "nach" is governed by "Jagd", the predecessor nounPA, whereas (4) contains a lexical preposition, which forms part of an attributive adverbial phrase of time. The governed preposition "nach" (in (3)) is translated differently in the English (3a) and 1

2 3 4

This is a cooperative project involving staff at the PH Erfurt, the University of G ttingen and the Universidade Nova de Lisboa. The Thuringian Ministerium f r Wissenschaft und Kunst (Ministry of Science and Art) supports the project in Erfurt by funding two members of staff. Cf. Schierholz (1995), (1996a:150f.), (1998a), (1998b), Wiegand (1996). In the illustrative examples, governed prepositions are shown in capitals. Non-governed prepositions are also termed "lexical prepositions", as in Rauh (1995), amongst others, who established that there are a total of five types of prepositions in English, but for the purposes of the project only the two types named here are relevant.

144

Stefan J. Schierholz

Portuguese (3b) than the lexical preposition "nach" (cf. (4a) and (4b)).5 Whereas in German the different circumstances are only distinguishable by the semantics of the postposition nouns (here: "Dieb" and "Mittagessen"), but not by the preposition, English and Portuguese have two markers available: the meaning of the postposition noun and the different prepositions used. When translating into German, English and Portuguese native-speakers probably need confirmation that German uses the preposition "nach" on both these occasions, whereas the German native speaker has to know that in English and Portuguese the different circumstances are not only marked by the noun but also by the selection of different prepositions. Complex nominal phrases, in which a PP occurs with a lexical preposition (as in (4) - (4b)), are termed AA-constructions - AA stands for Attributive Adverbial.6 Whereas, in most cases, one can translate the lexical prepositions without any great difficulty, because the preposition occurs in a concrete situation, most often with its original meaning, the same cannot be said, without reservations, about governed prepositions because of both their abstract meaning and the idiosyncratic bond it has with the predecessor nounPA. One has to learn which governed preposition has to be selected because the predecessor nounPA and the preposition have to be regarded as a unit. The SUBVAL project workers are compiling a database in which almost all the predecessor nounsPA will be listed. Under each noun, the governed prepositions and the description of the syntactic and semantic attributes required for the ΡΑ-construction will be collected and entered. This will create an important tool, which will contain basic linguistic data for various areas of lexicographical application.

2. The State of Research

The above description and illustrations may have given the impression that the topic of "ΡΑ-constructions" primarily plays a role in foreign language teaching, and that the need for the basic research can be justified primarily from the perspective of foreign language didactics. This has, however, proven to be of only secondary importance, simply because the linguistic foundations for foreign language teaching have not yet been laid down with sufficient thoroughness to be able to set up clear objectives in the content of a specific teaching methodology. The analyses of the topic complex "ΡΑ-construction" show that the problems which confront non-native speakers in learning or remembering the unit "noun plus governed preposition" can hardly be solved with the aid of didactic and methodological strategies, because the range of information available is insufficient, at least in the languages concerned: German, English and Portuguese.

The ambiguity in (4) will only be dealt with briefly here: the temporal variant in which the lexical preposition occurs means "the hunt, which took place immediately after the dinner or somewhat later", and the variant in which the governed preposition occurs would be based on the same syntactic functions as in (1) and would mean "the hunt for a dinner which someone has hidden somewhere". Cf. Schierholz (1996a: 151).

Governed Prepositions. A Database for German, English and Portuguese Nouns

145

Most grammars only deal with a fraction of the types of prepositional usages, and specifically those, which are adverbials or occur in connection with verbs.7 The topic "PAconstruction" is treated in a similarly unsatisfactory manner in the specialist literature.8 In addition, in most cases they fail to note that, in the ΡΑ-construction, the selection of the preposition and the characteristics which identify the ΡΑ-construction are idiosyncratic phenomena of the governing nouns concerned that can rarely be explained by grammatical rules. The recognition of this fact, might, of course, in one way explain the inadequate handling of this construction in the grammars and the specialist literature on grammar, if the treatment were not also by and large very superficial and imprecise, and based upon too few demonstrative examples, which can be contradictory and sometimes incorrect.9 One can conclude from this that the ΡΑ-construction, on the one hand, needs to be treated with more care in the grammars and the specialist literature and, on the other hand, it is primarily a topic for theoretical and practical lexicography. Whereas some work has recently been done in theoretical lexicography on this topic,10 the analyses of grammar information in German, English and Portuguese dictionaries, and in bilingual dictionaries (German-English, German-Portuguese), reveal that the dictionaries - even those which have been published in the nineties - do not contain the expected, and, for the non-native speaker, necessary information." This can be demonstrated by the example of the selected nouns in tables Ito3. Table 1 has been compiled from data from the German corpus used by the SUBVAL research project. This corpus contains over 40 million words and primarily consists of newspaper texts.12 Table 1 provides an overview of those prepositions, which are governed by the nouns. For example, "Verh ltnis" occurs with the governed prepositions "bei", "mit", "unter", "zu" and "zwischen"; whereas "Umfrage" just occurs with the governed prepositions "unter" and "zu", with the prepositions "bei" and "in" it only occurs in AA-constructions such as, "eine Umfrage bei 20 Polizeibezirken" or "eine Umfrage in Bremen".13 The frequency values entered cannot be quantitatively evaluated and give only an approximate indication whether the occurrence of a predecessor nounPA with a particular governed preposition is infrequent (e.g. "Kontroverse in") or frequent (e.g. "Kontroverse mit").

7

8

9 10 11

12

13

Cf. amongst others for German Duden (1984), Duden (1995), Eisenberg (1994), Helbig/Buscha (1988), for Portuguese Cunha/Cintra (1984), Mateus/Brito/Duarte/Faria (1989). Cf. amongst others Droop (1977), Lauterbach (1993), Teubert (1979), Sommerfeldt/Schreiber (1983). Cf. amongst others the critique in Schierholz (1996a), Wiegand (1996). Cf. amongst others Schierholz (1996a), Wiegand (1996). Cf. amongst others Figge (1994), Ghitescu (1992), Schierholz (1995), (1996a), (1996b), (1997), (1998b), (1998c), Wiegand (1985), (1996). It largely consists of "Die Tageszeitung" (TAZ 1995), plus some hundreds of record examples from the "Frankfurter Allgemeine Zeitung" (FAZ 1994). AA-constructions have only been indicated in table 1 if it could be supposed that there could also be occurrences in which the preposition concerned was governed by the noun.

Stefan J. Schierholz

146 Table l: Prepositions governed by selected nouns Predecessor nounPA Kontroverse Auseinandersetzung Unterschied Zusammenhang Verhältnis Umfrage

bei

mit über >10 >10 >10 >10 >10 AA >10 AA >5

in 10

1 >5

AA

2 3 >10 -

um unter >10 5 >10 >10 1 >5 >10

zu 1 5 >10 >10 >10 10

zwischen

>10 >10 >10 >10 >10 -

Key to table l: AA = only in AA-constructions; >5 = more than five corpus records; >10 = more than ten corpus records.

Table 2 indicates whether three selected German dictionaries contain information about the governed prepositions. The shaded areas indicate the governed prepositions, which ought to be in the dictionaries on the basis of the results of the corpus searches (cf. table 1). The items in the LGwDaF are differentiated as to whether a governed preposition is illustrated by a structural formula or only by an example item. The purpose of the structural formula, amongst others, is to state which preposition a noun governs.14 If only example items Table 2: Items with their governed prepositions in monolingual German dictionaries

Auseinandersetzung Unterschied

Key to table 2: L = LGwDaF 1994; D = DDUW 1989; W = DW 1994; ! = structural formula; example phrase; - = item missing.

=

are provided, the dictionary user needs to have a considerable amount of interpretative ability in order to be able to recognise correctly the syntactic relationships concerning the governed preposition. This is always the case in the DDUW and DW because the grammatical items in regard to governed prepositions consist solely of the example items. It can be seen from table 2 that under "Kontroverse" the LGwDaF gives structural formulas for "mit", "zwischen" and "über" together with example items for "zwischen" and "über", whereas no such items are given for the prepositions "in", "um", "unter" or "zu". In contrast, the DDUW only contains example items for "urn" and "über", whereas the DW gives no items on governed prepositions for "Kontroverse" whatsoever. A comparison with table 1 shows clearly that each dictionary contains items about, at most, one third of the possible governed prepositions. This proportion would hardly improve if one removed the nouns with low frequencies (e.g. < 3) from the comparison, be14

Cf. Schierholz (1998b).

Governed Prepositions. A Database for German, English and Portuguese Nouns

147

cause, by the same criteria, the grammatical items which are lacking under the entries for those nouns which occur particularly frequently in -constructions (e.g. "die Auseinandersetzung über", "der Unterschied zu"), would be regarded as particularly serious omissions. If one also takes the quality of the individual items of grammatical information into consideration, it can be seen that a proportion of the items of information is more like a guessing game for the dictionary user than a clearly structured information base.15 Table 3 contains the information on governed prepositions found in bilingual GermanEnglish and German-Portuguese dictionaries. There is a lemmatic gap in LGSchw-E-D and LTw-D-P concerning the noun "Kontroverse", whereas LEW-D-E provides lexicographical information for "Kontroverse mit" and "Kontroverse über" which is comparable to the structural formulas used in the LGwDaF. Example items are also given: "mit j-m eine Kontroverse über (acc) etwas haben: to have a controversy with s.o. on (od. about) s.th.; ich hatte eine Kontroverse mit meinem Vater: I had a dispute [...] with my father". LGSchw-D-E only provides useful items for "Verhältnis zu" and "Auseinandersetzung zwischen". In the case of "Verhältnis zu", the translation of the preposition is given as "with" and it provides two example items: "in freundlichem Verhältnis mit: on friendly terms with", and "er hat kein inneres Verhältnis zu seiner Arbeit: his heart is not in his work". However, there is no reference to the phraseology in the last example, although this is necessary in order to be able to understand the use of the preposition "in". As the LGSchw-D-E also contains the formulation: "er hat ein gestörtes Verhältnis zur Demokratie; he has an odd idea of democracy", which is described as ironic, it is difficult for the dictionary user to interpret the individual variants correctly. In the case of the lemma "Unterschied", "between" is given as the translation for the preposition "zwischen" in connection with "ein Unterschied machen", so that one can regard this as an aid to the translation of a -construction with "Unterschied zwischen". Table 3: Items with governed prepositions in bilingual dictionaries

Key to table 3: S = LGSchw-D-E; E = LEW-D-E; P = LTw-D-P; · = noun not lemmatised; item missing; ? = unclear item; = item; ! = translation of the preposition.

15

Cf. the critique in Schierholz (1996a:175ff.) which can also be applied in many respects to the examples mentioned here; e.g. in the DDUW, which poses the user the (unwanted) riddle, whether, in respect of the potential -constructions, there is a difference between the item giving the competence example "die Umfrage [zur/über die Atomkraft] machen, veranstalten" and "eine Kontroverse über, um etw. austragen".

148

Stefan J. Schierholz

In the LEW-D-E the governed prepositions are supposed to be stated in the dictionary entries.16 However, it is only in the case of "Auseinandersetzung mit", "Unterschied in", "Unterschied zwischen", "Zusammenhang zwischen", "Verhältnis mit" and "Verhältnis zu" that the German governed preposition is translated and an example item is also provided. It is not always clear in what connection a preposition is to be translated e.g. "Verhältnis mit" and "Verhältnis zu" with "with", but "Verhältnis zu" also with "to". In the case of "Umfrage", for "bei" the preposition "with" is given and, as a user one has the impression that this is a governed preposition, but it is not, at least not according to the definition of governed prepositions used here. A translation is given for "eine Umfrage unter der Bevölkerung [...]" ("a public opinion poll [...]"), but it does not contain a translation of the governed preposition "unter". Even this large, bilingual dictionary only contains information about 32% of the governed prepositions, and only 25% of the items can be regarded as satisfactory. The LTw-D-P does not provide any items about governed prepositions for any of the nouns, it merely contains the translation equivalents for the nouns (exception: "Kontroverse"). None of the dictionary entries contain any information about the possible serialisations of prepositional attributes. The syntactically correct constructions with "Auseinandersetzung" plus prepositional attribute-serialisation, which are possible in German, are shown by the following examples:17 (5a) (5a' (5a'') (5a'") (5b) (5b') (5b") (5b'") (5c) (5c') (5c")

die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung die Auseinandersetzung

ZWISCHEN den Zwillingen ZU dem Vorfall ZWISCHEN den Zwillingen UM den Entwurf ZWISCHEN den Zwillingen ÜBER den Vorfall ZWISCHEN den Zwillingen IN dieser Sache UNTER den Teilnehmern ZU dem Vorfall UNTER den Teilnehmern UM den Entwurf UNTER den Teilnehmern ÜBER den Vorfall UNTER den Teilnehmern IN dieser Sache MIT dem Vorsitzenden ZU dem Vorfall MIT dem Vorsitzenden UM den Entwurf MIT dem Vorsitzenden ÜBER den Vorfall

(5c''') die Auseinandersetzung MIT dem Vorsitzenden IN dieser Sache

The listing from ((5 a) to (5c)) does not take either the frequency of the individual formulations or the order of the prepositional attributes into consideration. Dictionaries need to provide information about all the prepositions governed by a predecessor nounPA together with the serialisation possibilities of these prepositions. Those serialisations, which are regular, can also be described in a dictionary grammar or in grammar handbooks. However, establishing this regularity should be done on the basis of extensive linguistic data and not on the basis of a few demonstration examples, because, in some cases, changing the order of the prepositional attributes produces fundamental changes in the meaning of the PAconstruction. 16 17

Cf. LEW-D-E (1974:XXIX). Additional serialisations (e.g. "die Auseinandersetzung mit dem Vorsitzenden um den Entwurf zu der Planung über die Weiterentwicklung in dieser Sache") have not been considered, because not all the prepositional attributes are dependent on "Auseinandersetzung".

Governed Prepositions. A Database for German, English and Portuguese Nouns (6) (6a)

149

die Entscheidung F R die Zustimmung ZU dem Plan die Entscheidung ZU dem Plan F R die Zustimmung

These few examples show that ΡΑ-constructions are far more complex than is generally accepted. The SUBVAL project attempts to research the existing ambiguities in respect of the ΡΑ-construction, which are demonstrated in detail elsewhere,18 for the German, English and Portuguese languages more precisely than has been done before.

3. Results Obtained In the SUBVAL project, the linguistic processing of the ΡΑ-construction is being carried out with the German language first. There are 18 governed prepositions (an, auf, aus, bei, durch, f r, gegen, in, mit, nach, (ohne), ber, um, unter, von, vor, zu, zwischen) to investigate, some of which also occur in portmanteau forms (am/ans, im/ins, vom, zum/zur) which must also be included in the analysis. The preposition "ohne" is not generally regarded as being amongst the governed prepositions; however, it should be included in the investigation as a complement to "mit". The percentage proportion that each individual preposition comprises of the total number of all the prepositions varies from 0.5% ("zwischen") to 20.25% ("in").19 The number of different lemmata which can function as predecessor nouns PA is less than 1000, but most of these can govern more than one preposition (cf. table 1), so that the number of units, predecessor noun PA plus governed preposition, is ca. 3000. However, one finds that many predecessor nouns PA have two or more variants that always occur if a predecessor noun PA can govern different prepositions which mutually exclude each other in the serialisation. This means that four variants have to be generated for "Auseinandersetzung zwischen" (cf. (5a)) because the prepositions governed by "Auseinandersetzung", i.e. "zu", " ber", "um" and "in" cannot be combined with one another. If each data set is restricted to those prepositions which occur in combination with one another, then, including all the variants, the total could amount to over 5000 database candidates (i.e. 5000 data sets). It is not possible to compile a list of all the nouns which govern prepositions on the basis of the information contained in the existing dictionaries and grammars. Neither can one copy a list of these nouns from some source or other, nor work through a dictionary from A to Z noting down the nouns on the way. For this reason, basic linguistic research is needed, which requires considerable knowledge of linguistics and high competence in the evaluation of the language examples. The linguistic results in SUBVAL are checked empirically and in such a way that it is not the subjective linguistic competence of the investigators which is decisive for the determination of the database candidates, but rather that they are the result of the application of an intersubjective checkable test. Together with the corpus,

18

19

Cf. amongst others Bergenholtz/Mugdan (1985), Figge (1994), Schierholz (1995), (1996a), (1996b), Wiegand(1985). Cf. Schierholz (1992); cf. Meier (1978), in which Meier does not distinguish between governed and lexical prepositions.

150

Stefan J. Schierholz

thesauruses and onomasiological dictionaries are also used as aids to determine additional predecessor nounsPA, because it has been established that, in many cases, semantically similar nouns govern the same preposition, e.g., in German: "Angriff, " berfall", "Attacke", "Attentat" and "Vorsto " all govern the preposition "auf.20

3.1. Collecting the Predecessor NounsPA The procedure for compiling the database for the nouns in the German language can be divided into three working steps: training the project workers, determining the database candidates and implementing the data in the database. It must be emphasised that the second and the third working step must be carried out separately, because, before a decision can be made whether or not a noun can be accepted as a database candidate, a comparative linguistic analysis of the corpus records must be undertaken. This is the only way to ensure that nouns of similar meaning, which occur in similar textual contexts, are also processed in the same way for the database. Working steps two and three are carried out preposition by preposition, i.e., in each case, the syntactic conditions for one preposition are checked, the nouns searched for, and the results of the process are then entered in the database.

3.1.1. The First Phase: Theory The project workers received initial general introductory instruction on the complex topic of the "ΡΑ-construction". Subsequently, specific intensive training is required for each preposition. This is necessary because in no case can the predecessor nounsPA be identified automatically. On the contrary, intellectual decisions are always required, which may well show similarities from preposition to preposition, but are not always the same. It is, however, relatively difficult to devise instructions for the project workers which lay down when a noun can function as a predecessor noun PA because - as shown above - there is no consensus of opinion concerning the prepositional attribute in the specialist literature, the grammars or dictionaries and because very few reliable test criteria for the determination of ΡΑ-constructions have been developed.21 Nevertheless, in very many cases, a native speaker can say with great confidence whether a given preposition is governed or not - although this stands almost in contradiction to the absence of a test procedure.

3.1.2. The Second Phase: Evaluation This phase of the work begins with the processing of a list of nouns which govern the same preposition. This list comprises the so-called initial set. It has been compiled, on the one 20 21

Cf. Schierholz (1998a). I have already referred to the Erfurt project workers, Claudia Hardut and Andrew Kerr, as "PAspecialists" (cf. Schierholz 1998a: 75). I repeat this here and must also give them the recognition for probably having checked more German corpus records than has ever been done before for the presence of ΡΑ-constructions and checking whether the test procedures developed are valid.

Governed Prepositions. A Database for German, English and Portuguese Nouns

151

hand, from years of searching through the research literature and, on the other hand, from computerised searches for the noun-preposition unit. This search was based on a text corpus, which is held at the Institut für Maschinelle Sprachverarbeitung (Institute for Computerised Language Processing) in Stuttgart.22 A search is made in the Erfurt SUBVAL-corpus for all the nouns in this initial set, including the inflected forms, plus the governed preposition concerned. For example: the search for "Anschlag auf will also include "Anschlags auf, "Anschläge auf and "Anschlägen auf. So far, the predecessor nounsPA have been recorded for nine prepositions. Table 4 shows the most important data. Table 4: Corpus Evaluation Data Preposition

(1)

an auf durch

mit nach unter

von zu zwischen

Number of corpus records Number of nouns Total QuantitaQualitative Predecessor Low fre- Doubtful Discards Compound tive frefrequency nounPA quencies cases quency (6) (7) (2) (3) (4) (5) 31902 15287 3144 247 303 31086 13145 3721 200 68 6940 88 44 539 6 33000 25194 3764 273 132 6862 6008 8 4269 39 3980 350 813 553 68 95300 4037 68 62 739 48913 37767 541 7190 467 244 13600 9038 3280 248 271583 111377 26704 1680 2183

nouns (9)

(8)

100 24 146 2 0 5 174 33 17 501

650 601 320 840 193 400 403 157

219 71 166 225 0 67 26 204 134

5411

1112

1847

Column (1) contains a list of the governed prepositions for which searches have been made for the predecessor nounsPA. Column (2) contains the number of corpus records which were obtained during the first corpus search in the SUBVAL-corpus. The term "first corpus search" is actually not absolutely accurate, because new nouns are found again and again during the process of recording the database candidates. However, the searches for these new nouns are not performed until the first search, with the initial set, has been completed. This means that the value for each preposition (e.g. for "an" 31,902) consists of the sum of many individual searches.23 The values in column (2) include not only corpus records for the predecessor nounsPA (column 5), but also records for those nouns which only occur infrequently (column 6), doubtful cases (column 7), nouns in other syntactical constructions24 22

23

24

I would particularly like to thank Judith Eckle and Ulrich Heid, whose assistance made the initial approach to analysing the corpus in the SUBVAL-Projekt considerably easier. No statements can be made about the conditions which were applied in the search algorithm. However, the results show that the initial set used in SUBVAL provided only a proportion of the total number of predecessor nounspA collected at the end of the second work phase, because many of the nouns cannot function as predecessor nounsp^ and because a large number of additional nouns are found during the work in SUBVAL. Faulty corpus records, which can be quickly discarded with the aid of simple sorting algorithms, are not included in the statistics. These include, amongst others, AA-constructions (Sonntag nach Pfingsten, Haus an der Leine, Feier am Jahrestag), prepositional objects (er bietet Beteiligungen an) and phraseologisms (er denkt über die Forschung nach).

l52

Stefan J. Schierholz

(column 8) and compound nouns (column 9). The total number of corpus records of complex nominal phrases present in the corpus is certainly noteworthy (271 583 records). The procedure for determining a predecessor noun PA is very complex, but it depends upon the individual noun and the quantity and the variability of the corpus records concerned. If there is no possibility of the noun which is to stand in front of the preposition functioning as a predecessor nounPA, e.g. nouns such as "Sonntag" or "Haus", the corpus rec??ords or the noun concerned can be relatively easily discarded. However, it is frequently the case that the unit, noun plus preposition, occurs both in ΡΑ-constructions and in other syntactical constructions. (7) die Reduktion A UF das Wesentliche (7a) die Reduktion auf zwei Milliliter (8) die Ausgabe AN die neu angekommenen Kinder (8a) Sie hat mit dem Dieter eine leicht verbesserte Ausgabe an Land gezogen. Whereas (7) and (8) contain ΡΑ-constructions, in which the PAs can be substituted with "worauf (7) and "an wen" (8) respectively, (7a) contains an AA-construction which can be substituted by "auf wieviel" and (8a) a phraseologism ("an Land ziehen"). The nouns "Reduktion" and "Ausgabe" can, therefore, function as predecessor nounsPA, but not all the corpus records extracted for these nouns contain ΡΑ-constructions. If at least two records are found in which a given unit, noun plus preposition, functions as a PA, then the predecessor noun PA is added to the list of database candidates. The number of database candidates which have been obtained after processing is contained in column 5 and currently stands at 1680 predecessor nounsPA. The quantities of corpus records found for the nouns listed in column 5 are differentiated into two types. The quantitative frequency (column 3) includes all the records, which are found with the unit, noun plus preposition, without taking into consideration the syntactical functions in which the nouns occur. Thus, the quantitative frequency includes only those nouns which can sometimes or always function as predecessor nounsPA. The qualitative frequency (column 4) includes only those PA-constructions in which the nouns are found to occur as predecessor nounsPA (column 5). If one took examples (7) and (8), then "Reduktion" and "Ausgabe" would each have a qualitative frequency of "1" and a quantitative frequency of "2". Columns (3), (4) and (5) are shaded because the values in these columns can be related to one another. Because a very large number of corpus records are found for some predecessor nounsPA, of which over 90% are ΡΑ-constructions (e.g. in the case of "Frage nach" there were 966 [= qualitative frequency] ΡΑ-constructions from a total of 1005 records [= quantitative frequency]), quantitative restrictions have been placed on the evaluation procedure in order to enable the corpus rec??ords to be processed efficiently (cf. table 5). These restrictions were not decided upon until after the first preposition ("nach") had been processed in order to be sure that this would not cause any significant distortion of the data. When the search for a noun plus preposition unit produces a result lying between 1 and 20 all the corpus records are checked and the actual number of corpus records found to contain a ΡΑ-construction is recorded. If the corpus search yields between 21 and 50 examples, numbers 10 to 29 are evaluated, and the number of ΡΑ-constructions occurring in this sample is recorded. If the result lies between 51 and 100, the last 30 records are checked (e.g. in the case of 78 records, nos. 49 to 78 would be checked), and the number of

Governed Prepositions. A Database for German, English and Portuguese Nouns

153

PAs found in this group is recorded. If more than 100 corpus records are found, the first 40 records are evaluated.25 Table 5: Restrictions on the corpus analysis Frequency noun plus preposition

Checking procedure

1 to 20 21 to 50 51 to 100

all records record numbers 10 to 29 the last 30 records record numbers 1 to 40

more than 100

Recordable qualitative frequency per predecessor nounpA max. 20 max. 20 max. 30 max. 40

The evaluation restrictions shown in table 5 enable the work to be carried out efficiently and do not influence the identification of the database candidates because the SUBVAL project is only concerned with ascertaining whether or not a specific noun can function as a predecessor noun PA . The frequency values cannot therefore be interpreted as comparable, quantitative analyses of the corpus records for individual prepositions because not all prepositions will have been evaluated on the same quantitative basis and because the number of records for the individual candidates which have been included in the evaluation cannot be reconstructed on the basis of the data in table 4. Only in the case of the preposition "nach" (cf. table 4) do the values in columns (3) and (4) demonstrate the relationship between the values in columns (3) and (4), which is, that a total of 6008 corpus records were found for the 39 predecessor nouns PA (column 5), of which 4269 were found to contain a ΡΑ-construction. The values in column 3 of table 4 show that there are a total of 111,377 records for the predecessor nouns, and the values in column 4 show how many records have been examined, tested, intellectually processed and identified as PA-constructions by the project workers, namely 26704. For the preposition "an", (table 4, line 1) 247 predecessor nouns PA have been recorded, of which at least 3144 corpus records have been found to contain ΡΑ-constructions. Because not all the records for high-frequency predecessor nouns PA have been investigated, the actual number of corpus records of PAconstructions with "an" is higher. In all, there are 15,287 records containing one of the 247 nouns from column 5 plus "an", however, there is no guarantee that these are all PA-constructions. If one compares the number of records obtained in the first search (31,902 rec??ords in the case of "an") with the quantitative frequency (15,287 records), there remain 16,615 corpus records which can be allocated to columns (6) to (9), amongst others. The number of nouns for which less than two corpus records are found (e.g. "ein Geschmack nach Leder"), is entered in column (6), whereby it must be noted that not all these low frequency records will have been checked for the presence of a PA-construction. Column (7) "Doubtful cases" contains the number of nouns which have not yet been able to be allocated with certainty to the category predecessor noun PA because either the validity of the tests is not clear-cut or it is considered there may be connections with other governed

25

In exceptional cases, if it is absolutely certain on the basis of the subjective assessment of a noun that it can be a predecessor nounpA, and if there are a large number of corpus records for it, but the evaluation of a proportion of the records (e.g. the first 40 records for a quantitative frequency of 134) does not yield a single ΡΑ-construction, then the remainder of the records will also be investigated.

154

Stefan J. Schierholz

prepositions which have not yet been processed. The numbers of doubtful cases are very high for the prepositions "an", "durch" and "von" because the relations in these cases are particularly difficult to resolve. For example, in the case of the preposition "an", a substantial proportion of the constructions can be designated as "indefinite statements of quantity". (9) (10) (11) (12) (13) (14)

der Absatz AN Verhütungsmitteln die Fülle AN Informationen der Überschuß AN Kriegsmaterial der Höhepunkt AN sportlicher Brisanz den ganzen Katalog AN Folterinstrumenten die Vielfalt AN Naturräumen

Whereas in (9) the prepositional attribute specifies what is "abgesetzt" (note: Absatz: sales, abgesetzt: sold), in (10) the quantity of information is given by the predecessor noun. The preposition "an" can be omitted in (10) but not in (9). In (11) it is stated, on the one hand, that a quantity (note: Überschuß: surplus) exists, on the other hand, it is specified what this surplus consists of. In (12) the predecessor noun indicates a point on a scale, whereas in (13) the complete assortment of instruments of torture (FolterInstrumente) is meant, without this quantity being more precisely determinable. (14) can be interpreted in a similar way, but whereas in (13) one could do without the preposition "an" in certain contexts, this is not possible in (14). In all the examples except (12), the preposition "an" can be replaced by "von", and in (9) and (13) this would then form an analytical genitive in which the preposition "von" is not regarded as being governed. Examples (12) and (14) can certainly also be paraphrased with genitive constructions ("der Höhepunkt sportlicher Brisanz") and ("die Vielfalt der Naturräume"), even if this results in a distinct shift in meaning in (14). A satisfactory solution for the doubtful cases with "an" has not yet been found. This short discussion of some of the nouns occurring with the preposition "an" which have been assigned to the "doubtful cases", is merely intended to make it clear that the identification of a predecessor noun PA is, in many cases, bound up with very complex linguistic considerations. The quantities of nouns which have been discarded because they do not form PA-constructions or because no corpus records were found for them is entered in column 8. Column 9 contains the number of nouns which have been analysed as compound nouns and rejected. The criteria for rejection are that, firstly, the basic element has been recorded in the database as a predecessor nounPA and, secondly, that the meaning of the basic element is contained in the meaning of the compound noun (e.g. "Handelsabkommen mit"). Various test frames are available for identifying a predecessor noun PA (e.g. case marking, meaning of the preposition, pronominal adverb substitution, use of interrogative pronouns for the prepositional phrase, insertion of a specificator, basis verb or adjective).26 In the case of the first two prepositions to be processed, "nach" and "auf, most of the tests were applied to the corpus records in order to establish which test would enable a predecessor noun PA to be identified most quickly and reliably. It has been found that, as a rule, it is hardly necessary to apply any special test to those nouns derived from verbs and adjectives in which the base verb or adjective governs the same preposition. Derivatives of verbs which govern a case can also be relatively easily identified as predecessor nounsPA ("die 26

Cf. Schierholz (1996a:167ff.).

Governed Prepositions. A Database for German, English and Portuguese Nouns

155

hnlichkeit mit der Marionette" "die Suche danach" or "die Suche nach der erfolgreichen Beendigung eines Mittagessens" -» "die Suche danach", even if with different stress). In the case of the preposition "mit", the pronominal adverb "damit" could not be used because it can also function as a substitute for instrumental AAs ("der Mann mit dem Fahrrad" -> "der Mann damit"). If the postposition-N PPA is animate or an institution, then it cannot be substituted by a pronominal adverb, instead one has to use the preposition plus a personal pronoun ("der Anschlag auf den Vorsitzenden" -> "der Anschlag auf ihn"; "das politische Abenteuer mit der Bundeswehr" -» "das politische Abenteuer mit ihr"). Although some difficulties with the application of the tests have been indicated, in the project procedure it has been laid down that, in principle, only those nouns for which there is an intersubjective, checkable test for their identification as predecessor noun PA are to be accepted as database candidates. For the time being the remainder are being retained as doubtful cases. During the corpus analyses, not only are the test procedures applied and modified, but also additional characteristics of the ΡΑ-constructions are noted. These can be of a syntactic nature (e.g. governed case, requirement of an article, numerus restrictions, possible insertion of adjective) or semantic (e.g. meaning of the governing noun, meaning of the noun in the attribute) and are required for further processing of the data sets. For the preposition "an", it is relevant whether it governs the dative ("die Arbeit AN der Statue") or the accusative ("seine Erinnerung AN den Unfall") because it cannot be directly ascertained from a general statement about the ΡΑ-construction ("seine Erinnerung AN etwas" bzw. "die Arbeit AN etwas") which the governed preposition is to take. In some ΡΑ-constructions, the postposition-NPPA occurs without an article ("der Bedarf AN Wohnungen") unless it refers to specific "Wohnungen". Although, in such a case, it would be necessary to add a relative clause ("der Bedarf an den Wohnungen, die es seit kurzem zu vermieten gibt, [...]"). Some predecessor nounsPA only occur in the singular, this type of noun is called a Singulariatantum, for example, Einigkeit in "Einigkeit MIT den Nachbarl ndern". But there are also cases in which a lemma is not subject to any restrictions on number; however, in its function as a predecessor noun PA it only occurs in one number, e.g. "Unstimmigkeit" plus "mit" only occurs in the plural. The meaning of a predecessor noun PA is noted, because nouns with a similar meaning often govern the same preposition. In some cases, the meaning of the noun standing in the postposition-NPPA determines whether the construction is a ΡΑ-construction or an AA-construction: "die Grenze ZWISCHEN Ironie und Zynismus" vs. "die Grenze zwischen Hamburg und Niedersachsen". In addition, a note is made of other grammatical and semantic features which come to attention during the course of processing so that a linguistic evaluation can be made at a later time.

156

Stefan J. Schier holz

3.1.3 The Third Phase: Implementing the Database In this work phase, the recorded data on the predecessor nounsPA are entered in the database27. The information in the database could serve as the prerequisite for compiling a special dictionary on governed prepositions which contains an index containing all the predecessor nounSp A plus their translations in Portuguese and English. For this purpose, not all the processes from the second work phase are needed (this concerns, e.g., the test procedures, whose areas of application are not transferred into the database). On the other hand, additional information is entered in some of the fields of the data set which were not determined in the second work phase. To this extent the transfer of the data takes place semiautomatically. Figure 1 shows the most important fields which have been defined in the database for the data sets. The noun, which can function as a predecessor noun PA , is entered in the field "Candidate" in the nominative singular form. Pluraliatantums are entered in the nominative plural form. If there are restrictions concerning particular forms in its function as a predecessor noun PA , these are not noted because the noun entered in the field "Candidate" is always the lexeme which has been sought and it should be subject to the sequence rules of a strictly alphabetically ordered lemma index of a dictionary. The governed preposition is entered in the field "preposition", the quantitative frequency ("F 1") and the qualitative frequency ("F2") are entered in the frequency fields. These values can later be used as a basis for formulating information about relative frequencies. In figure 1, the frequency values which have been entered are relatively high because no restrictions were placed on the evaluation of the preposition "nach" in respect of the corpus records.

27

This is an MS-Access database.

Governed Prepositions. A Database for German, English and Portuguese Nouns Candidate

Bedürfnis

Port-candidate

necessidade

Preposition

nach

Port-prep

de

Fl

193

Grammar

meist Sg

Construction

das Bedürfnis nach jmdm/etw

Portexample phrase

a necessidade de confirma?äo

F2

191

Comment on the construction Example phrase

das Bedürfnis nach Bestätigung

Das Bedürfnis nach Rausch ist Example sentence (original) normalerweise bei jedem vorhanden.

Semantic restrictions Comment on restrictions Preposition 2 Construction PI +P2 Comment on PI + P2 Preposition 3 Construction P1 + P3 Comment to PI +P3 Preposition 4 Construction PI + P4 Comment on PI + P4 Theoretical semantics Figure 1: Layout of the database

Comment on the translation

157

158

Stefan J. Schierholz

The "Grammar" field contains statements about the predecessor nounPA, relating to the number usage. The possible entries are: "nur Sg", "meist Sg", "nur PI", "meist PI" or no entry. The decision upon which entry to make is taken after the corpus records have been evaluated and after consulting various monolingual dictionaries.28 In the "Construction" field, a general (relatively abstract) formulation is entered which indicates the possible forms which the noun or predecessor-NPPA plus the governed preposition plus the postposition-NPPA can take. The entry in the "Construction" field is similar to the dictionary item "structural formula" as found in the LGwDaF.29 The constructional information should indicate which general syntactic and semantic characteristics exist regarding the PAconstruction. There are six variants available for designating the postposition-NPPA in German: "jmdn/etw", "jmdm/etw", "jmdn", "jmdm", "etw (Akk)", "etw (Dat)". The entries in the field "Comment on the construction" are initially, tentatively determined and refer to other possible constituents e.g. "meist/immer Adjektiv vor Vorgänger" (mostly/always adjective before predecessor), "meist/immer Genitivattribut nach Vorgänger" (mostly/always genitive attribute after predecessor), determination possibilities of the postposition nouns, if they are marked, "obligatorische Adjektiveinfügung" (obligatory inclusion of adjective), "obligatorisches Genitivattribut" (obligatory genitive attribute). In addition, any remarks can be entered to indicate special features. Complete systematisation will not be instituted until the results for all the prepositions can be taken into consideration. A PA-construction is entered in the "example phrase" which demonstrates the use of the noun in the "Candidate" field and the preposition entered in the "preposition" field. A -construction is selected from the corpus records which is typical for all the records, syntactically simple and not ambiguous. Syntactically simple means that it does not contain any additional adjectives, genitive attributes or verbs. In the field "Example sentence", one or two sentences are entered which are largely originals from the SUBVAL-corpus. The example sentences should be selected so that they contain a minimum of world knowledge needed to understand the sentence. Its validity should, as far as possible, be independent of time, politics and economics. Sentences containing phrases such as "Lösung von Rußland", "DDR-BRDVergleich" should not be entered in the database. However, this requirement can not always be fulfilled in the case of predecessor nounsPA, with a very low, qualitative frequency. For the predecessor noun PA "Erinnerung", which has the construction "die Erinnerung an jmdn/etw", the illustrative phrase could be "die Erinnerung an ihren Vater", and the example sentences could be: "Noch ist die Erinnerung an kalte Zehen und Eisblumen hinter Glas taufrisch." and "Er schenkte jedem eine goldene Uhr zur Erinnerung an einen der größten Tage der Vereinsgeschichte". So far, only a few entries have been made in the "semantic restrictions" field because the presence of any semantic restrictions on the PA-construction cannot be ascertained until all the data has been subject to a linguistic evaluation. The statements mostly concern the postposition nounPA, but can also refer to the predecessor nounPA. Currently, the following characteristics are available: "meist/nur Lebewesen" (mostly/only animate), "meist/nur Personen" (mostly/only persons), "meist/nur Institutionen" (mostly/only institutions), "meist/nur Abstrakta" (mostly/only abstract), "meist/nur Konkreta" (mostly/only concrete). Combinations (e.g. "mostly animate or institutions") are also possible. Informal comments are entered in the "Comments on restrictions" field, in 28

DDUW (1989) and LGwDaF (1994) are normally used.

29

Cf. LGwDaF (1994:XXIII).

Governed Prepositions. A Database for German, English and Portuguese Nouns

159

which reference is made too specific, semantically determined, noteworthy features. If a noun can govern more than one preposition, these are entered in the fields "preposition 2", "preposition 3" and "preposition 4", and this is followed by an entry in the corresponding fields: "Construction PI + P2", "Construction PI + P3", "Construction PI + P4". In the fields "Comment on PI + P2", "Comment on PI + P3", "Comment on PI + P4", comments are entered in the same manner as provided for in the field "Comment on the construction" (cf. above). In addition, notes can be made about the serialisation. This particularly concerns the sequential order of the prepositions if this serialisation affects the meaning of a ΡΑ-construction or the serialisation of prepositional attributes produces syntactic or semantic changes. No entries are being made at the present time in the field "Theoretical semantics". The fields "Port-Candidate", "Port-Prep", "Port-example phrase" and "Comment on the translation" are to be completed by the Portuguese partners. So far, the necessary entries have been made in the German-Portuguese database for the predecessor nouns PA , which govern the prepositions "nach" and "auf. In addition, the prepositions ("an", "von", and "zwischen") have been prepared for automatic transfer into the database.

3.2 The Translations Once the entries in the German language have been made, the translations from German into English and from German into Portuguese can be entered. An individual database is to be compiled for each language pair and for each translation direction, that is a total of six databases in all. In principle, these are all to have the same structure so that they can be combined with one another at a later time. After the information has been entered in the "German" fields of the "German-Portuguese" database (cf. figure 1), the fields for the Portuguese information are to be completed. The translation of the German noun in the "Candidate" field is entered in the "PortCandidate" field if there is a Portuguese equivalent. If there is no equivalent, the field remains empty and a remark to this effect is entered in the field "Comment on the translation". The translation of the German preposition is entered in the "Port-Prep" field, if there is a Portuguese equivalent which can occur with the predecessor noun PA in a PA-construction. The Portuguese translation of the German example phrase is entered in the "Port-Example phrase" field if there is a Portuguese equivalent for the candidate. If it is totally impossible to form a ΡΑ-construction in the Portuguese translation, then a syntactically different translation is entered. Even if a translation with a ΡΑ-construction is rare or unusual, it is still entered. The "Comment on the translation" field contains a comment on the Portuguese translation of the illustrative phrase if there are any special features in the translation. This would be the case, if there were no Portuguese equivalent to the ΡΑ-construction or if a translation with a ΡΑ-construction were rare or unusual. If the latter should be the case, the usual (or better) translation is also entered. Following these activities, the Portuguese ΡΑ-constructions are processed in detail in the "Portuguese-German" database. For the German-Portuguese database, the Portuguese partner only receives the information which has been entered in the database from the German side, but no information from the tables which were compiled in the second work phase (cf. above). In this way, an automatic check is made during the translation process into Portuguese as to whether the German entries in

160

Stefan J. Schierholz

the database are complete, sufficiently comprehensive and understandable for the non-native speaker. There are hardly any problems with any of the translation processes if there are translation equivalents in the target language for the German ΡΑ-constructions and these PA-constructions are in normal use in both languages. Naturally, this also applies to the translations into German. However, in those cases in which a German ΡΑ-construction is formulated with a different syntactic construction in the foreign language, exact analyses are required and appropriate notes have to be entered in the database. Table 6 contains selected problem cases in translating German predecessor nouns PA into Portuguese. A special note must be made if a language does not contain a translation equivalent (e.g. "Ausschau") and alternative translations have to be considered. If there are more than one possible translations for a lemma (e.g. "Bitte", "Drang") then, after consulting monolingual and bilingual dictionaries, searching the corpus and discussion amongst the project workers, a decision has to be taken which translation is to be preferred. The same applies to ΡΑ-constructions in which the meaning can only be demonstrated with additional explanation, although they are usually unambiguous because of their context. For example, "Duft" usually has a positive connotation, the exact meaning of "Geruch" is given by the context, whereas "Gestank" is always meant pejoratively. Connotations pertaining to the Portuguese ("emboscada") need not apply in the same way to the German equivalent ("Suchaktion"). Because there are no comprehensive, electronically readable, Portuguese corpora available, problem cases can only be resolved through specialist discussions amongst the project workers. For processing English, an English native-speaker is available and additional

Governed Prepositions. A Database for German, English and Portuguese Nouns

161

Table 6: Translation problems with predecessor nounspA Port Port Comment on the Prep Example translation

German Can- Prep Gram- Construcmar didate tion

Example phrase Port Candidate

Ausschau

nach nur Sg die Ausschau nach jmdm/etw nach meist die Bitte nach Sg jmdm/etw nach meist der Drang nach etw Sg (Dat)

die Ausschau nach anderen Einkommensm glichkeiten die Bitte nach Mitteln

requenmento, pedido der Drang nach desejo. Unabh ngigkeit impeto

de

nach meist Sg

der Duft nach Weihrauch

aroma, cheiro, perfume der Geruch nach cheiro Brot

a

der Gestank nach fauligen Eiern

a

phrase

Bitte

Drang

Duft

Geruch

Gestank

Such-aktion

der Duft nach etw (Dat) nach meist der Geruch nach etw Sg (Dat) nach nur Sg der Gestank nach etw (Dat)

nach

no translation of the noun possible

pwete. fedor

die Such- die Suchaktion emboscada aktion nach nach den T tern jmdm/etw

de

a

a

o pedido de recursos o desejo de independencia

requerimento - elevated speech (authorities, officialese)

ο desejo is the desire (not so strong an urge o impeto is the impulse (to) stronger urge. If the PA is abstract impeto is preferred o cheiro is more neuperfume a tral than perfume

incenso o cheiro a po o pivete a fedor is a little ovos stronger but is used podres less often because it is "phonetically awkward", it does not sound so good a embos- The meaning of cada aos emboscada implies a criminoso well organised action

assistance can be expected from the excellent, comprehensive text corpora of the English language available.

4. Possible Applications of the Results The features of the ΡΑ-construction which have been ascertained, are held in the database, as far as possible, in a modular form which is designed to enable it to be used for further research in as many ways as possible. It can be safely assumed that there is a multitude of possible applications because the results entered in the database have not previously been available in a comparable form in the field of linguistic research. Therefore, the data can be related directly to the "ΡΑ-construction" topic area because an empirically based syntactic

162

Stefan J. Schierholz

and semantic evaluation of the interior structure of the -construction has been carried out. Furthermore, the results can also be used to improve presentations in grammar books. This particularly applies to the selection of representative examples and to a more accurate description of the area of validity of the individual test procedures. The governed prepositions are also important for computerised translation because they need to be treated as a syntactic feature in the interplay between lexicon and parser in a similar way to that of case government of the lexemes individually affected. Further applications primarily affect theoretical and practical lexicography. Alongside the possibilities for improving existing monolingual and bilingual dictionaries, special dictionaries can be compiled for predecessor nounsPA. These dictionaries could be monolingual or bilingual, or could even be published as trilingual dictionaries, which could then be regarded as derivatives of the database. The production of such dictionaries is primarily conceived of as an object for future research projects, in which one can conceive of these dictionaries being published both in printed and CD-ROM versions. However, it can be seen that further consideration will have to be given to the lexicographical presentation for this purpose. The database concept is however currently organised so that it does not predetermine the microstructure of potential dictionary articles. The dictionary items must be presented in a user-friendly form and must also be orientated to the needs of the non-native speaker, offer the quickest possible access time to the required dictionary information and guarantee that the language in the dictionary entries is correct and easily understandable. The linguistic analyses carried out in the SUBVAL project will form a basis for such applications.

5. Literature

Bergenholtz, Henning/Mugdan, Joachim (1985) (Hrsg.): Lexikographie und Grammatik. - Tübingen (= Lexicographica. Series Maior 3). Cunha, Celso/Cintra, Lindley Luis F. (1984): Nova Gramatica do Portugue's Contemporäneo. Lisboa. Droop, Helmut Günter (1977): Das präpositionale Attribut. - Tübingen (= Forschungsberichte des Instituts für deutsche Sprache Bd. 34). Duden (1984): Grammatik der deutschen Gegenwartssprache. 4., völlig neu bearbeitete und erweiterte Aufl. Hrsg. und bearb. von Günter (sie!) Drosdowski in Zusammenarbeit mit Gerhard Äugst, Hermann Gelhaus, Helmut Gipper, Max Mangold, Horst Sitta, Hans Wellmann und Christian Winkler. - Mannheim/Wien/Zürich (= Duden Bd. 4). — (1995): Grammatik der deutschen Gegenwartssprache. 5., völlig neu bearb. und erw. Aufl. Herausgegeben und bearbeitet von Günther Drosdowski in Zusammenarbeit mit Peter Eisenberg, Hermann Gelhaus, Helmut Henne, Horst Sitta und Hans Wellmann. - Mannheim/Leipzig/Wien/ Zürich (= Duden Bd. 4). Eisenberg, Peter (1994): Grundriß der deutschen Grammatik. 3., überarbeitete und erweiterte Aufl. Stuttgart/Weimar. Figge, Udo F. (1994) (Hrsg.): Portugiesische und portugiesisch-deutsche Lexikographie. - Tübingen (= Lexicographica. Series Maior 56). Ghitescu, Micaela (1992): Dicionärio prätico de substantivos e adjectives com os regimes preposicionais. - Lisboa. Heibig, Gerhard/Buscha, Joachim (1988): Deutsche Grammatik. Ein Handbuch fur den Ausländerunterricht. 11. Aufl. - Leipzig.

Governed Prepositions. A Database for German, English and Portuguese Nouns

\ 63

Lauterbach, Stefan (1993): Genitiv, Komposition und Präpositionalattribut - zum System nominaler Relationen im Deutschen. - München. Mateus, Maria Helena Miva/Brito, Ana Maria/Duarte, Ines/Faria, lsabel Hub (1989): Gramätica da Lingua Portuguesa. 2.' edifäo revista e aumentada. - Lisboa. Meier, Helmut (1978): Deutsche Sprachstatistik Bd. 1-2. 2., erweiterte und verbesserte Aufl. Hildesheim. Rauh, Gisa (1995): Präpositionen und Rollen. In: Sprachwissenschaft Bd. 20. H. 2, 123-167. Schierholz, Stefan J. (1992): Syntaktische Merkmale deutscher Substantive. - IBM-Deutschland (= IWBS-Report2I4). — (1995): Präpositionalattribute: Wieviel Syntax - wieviel Semantik? Wieviel Grammatik - wieviel Lexik? In: GAL-Bulletin 22/1995. Herausgegeben im Auftrag der Gesellschaft für Angewandte Linguistik von Dieter Wolff. - Wuppertal, 19-26. — (1996a): Grammatische Informationen zu Substantiven in einsprachigen deutschen Wörterbüchern. In: Wiegand, Herbert E. (Hrsg.): Wörterbücher in der Diskussion II. - Tübingen, 140203 (= Lexicographica. Series Maior 70), — (1996b): Grammatik im Wörterbuch. Zur Wörterbuchbenutzung aus fremdsprachiger Perspektive. In: Deutsch als Fremdsprache 33. Jg. H. 4, 223-232. — (1997): Prepositional Attributes: Syntax or Semantics? Grammar or Lexicon? In: EESE (Erfurt Electronic Studies in English), http://webdoc.sub.gwdg.de/edoc/ia/eese/ eese.html. - Erfurt. — (1998a): Zur Semantik der Präposition "auf in komplexen Nominalphrasen sowie notwendigen und möglichen Darstellungen im Wörterbuch. In: Wiegand, Herbert E. (Hrsg.): Wörterbücher in der Diskussion III. - Tübingen, 55-105 (= Lexicographica. Series Maior 84). — (1998b): Die Grammatik der Substantive in Langenscheidts Großwörterbuch Deutsch als Fremdsprache. In: Wiegand, Herbert E. (Hrsg.): Perspektiven der pädagogischen Lexikographie des Deutschen. Untersuchungen anhand von "Langenscheidts Großwörterbuch Deutsch als Fremdsprache". - Tübingen, 88-103 (= Lexicographica. Series Maior 86). — (1998c): A Trilingual Dictionary of Nouns which govern Prepositions. In: Zettersten, Arne/Pedersen, Viggo H./Mogensen, Jens Erik (eds.): Symposium on Lexicography VIII. Tübingen, 257-265 (= Lexicographica. Series Maior 90). Sommerfeldt, Karl-Ernst/Schreiber, Herbert (1983): Wörterbuch zur Valenz und Distribution der Substantive. 3. Aufl. - Tübingen. Teubert, Wolfgang (1979): Valenz des Substantivs. Attributive Ergänzungen und Angaben. - Düsseldorf (= Sprache der Gegenwart 49). Wiegand, Herbert E. (1985): Fragen zur Grammatik in Wörterbuchbenutzungsprotokollen. In: Bergenholtz, Henning/Mugdan, Joachim (Hrsg.): Lexikographie und Grammatik. - Tübingen, 2098 (= Lexicographica. Series Maior 3). — (1996): Über primäre, von Substantiven "regierte" Präpositionen in Präpositionalattributkonstruktionen. In: Harras, Gisela / Bierwisch, Manfred (Hrsg.): Wenn die Semantik arbeitet. Klaus Baumgärtner zum 65. Geburtstag. - Tübingen, 109-147.

6. Dictionaries and Corpora Used DDUW = Duden (1989). Deutsches Universalwörterbuch. Hrsg. und bearb. vom Wissenschaftlichen Rat und den Mitarbeitern der Dudenredaktion unter der Leitung von Günther Drosdowski. 2., völlig neu bearbeitete und stark erweiterte Aufl. - Mannheim/Wien/Zürich. DW = Wahrig, Gerhard (1994): Deutsches Wörterbuch. Neu herausgegeben von Dr. Renate WahrigBurfeind. Mit einem "Lexikon der deutschen Sprachlehre". - Gütersloh.

164

Stefan J. Schierholz

f.A.Z. (1994): Die F.A.Z. (= Frankfurter Allgemeine Zeitung) auf CD-ROM. Jahrgang 1993. Frankfurt a.M.. LEW-D-E = Springer, Otto (1974) (Hrsg.): Langenscheidts Enzyklopädisches Wörterbuch der englischen und deutschen Sprache. Teil II. Deutsch-Englisch. 2 Bde. Völlige Neubearbeitung. Berlin/München/Wien/Zürich. LGwDaF = Götz, Dieter/Haensch, Günther/ Wellmann, Hans (1994) (Hrsg.): Langenscheidts Großwörterbuch Deutsch als Fremdsprache. Das neue einsprachige Wörterbuch für Deutschlernende. 3. Aufl. - Berlin/München. LGSchw-D-E = Messinger, Heinz (1989 (1977)): Langenscheidts Großes Schulwörterbuch. DeutschEnglisch. 19. Aufl. - Berlin/München/Wien/Zürich/New York. LTw-D-P = Beau, Albin Eduard (1982 (1969)): Langenscheidts Taschenwörterbuch der portugiesischen und deutschen Sprache. Zweiter Teil. Deutsch-Portugiesisch. 14. Aufl. Berlin/München/Wien/Zürich. TAZ = tazCompact (1995): Sieben lange Jahre, Wort für Wort: taz, die tageszeitung. 1.7.1988 bis 30.6.1995.-Berlin.

Ulrich Busse

Recent English Learners' Dictionaries and their Treatment of Political Correctness

1. Introduction: political correctness, the language learner, and the dictionary "We live in interesting lexicographical times. There is now an unprecedented range of general-language learner dictionaries available." With this statement Pätzold (1997: 184) begins his comprehensive review of "the new generation of learner dictionaries." The section of his review that concentrates on the four dictionaries for advanced learners that appeared in new editions in 1995 (cf. 1.1 of the bibliography)', is restricted to the points of how they handle meanings (especially of selected Americanisms) presumably unknown to students, unknown uses of well-known items, body collocations and, finally, the treatment of examples. Herbst (1996: 328) in his detailed review article on the four dictionaries casts a short sidelong glance at sexism in connection with illustrative examples. All of these points are important features, yet for various reasons the topic of political correctness (PC)2 is also highly relevant for language learners. The concept of PC itself is highly controversial and even the term is a matter of contention for those avowedly in favour of it and those contemporaries who vehemently oppose censorship. For one group of people PC is the embodiment of a totalitarian Newspeak thinly disguised under the fig leaf of the protection of minorities in order to implement political and cultural objectives, e.g. a change in the canon of texts to be studied; for the other group, PC is regarded as the effort to prevent racism, sexism, etc. by avoiding linguistic aggression and Stigmatisation. Although it is not possible to outline the socio-political history of PC within the framework of this article a few notes are necessary to understand why the discussion has become so heated, at least in the United States. During the 1980s, awareness of linguistic prejudice was largely publicised under the keyword of PC, which arose on American university campuses, then spread to Great Britain and is currently gaining some ground in Germany. The original idea was to establish binding speech codes in Universities, publishing houses etc. so to abolish linguistic discrimination of minorities, "and especially about groups of people whom it [society] perceives to be disadvantaged or oppressed. The most sensitive domains are to do with race, gender, sexual affinity, ecology, and (physical or mental) personal development" (Crystal The American Heritage English as a Second Language Dictionary (1998) [1008 pp.], to my knowledge the first ever American Learners' Dictionary, came to my attention too recently for the inclusion in this article; and the COBUILD English Learner's Dictionary (1994) [1154 pp.], originally intended for comparison with the other dictionaries, proved to be too small in scope and coverage. Although the two terms political correctness and its abbreviation PC may differ in their connotations, the latter will for the sake of brevity be employed throughout this article.

166

Ulrich Busse

1995: 177). The origins of this movement go back to the Civil Rights Movement of the 1960s, Feminist criticism and the philosophy of post-structuralism of Barthes, Derrida, Foucault, Habermas, Lacan and others. The underlying idea is that a change in language is a starting point for the reformation of socio-political power structures in society. The criticism voiced against this maxim from the early 1990s onwards is its inherent intolerance infringing on the freedom of speech and the trend towards a terminological absurdity. Although proponents of PC acknowledge this reproach they justify their policy by arguing that these restrictions are necessary for the sake of preventing discrimination against marginalised groups (cf. for example Berman 1992, Dunant 1994, Hoffmann 1996, Greil 1998). In the controversy about PC the exchange of arguments has often taken place on the level of essayistic journalism without truly taking into account the problem of effecting language change through active language policy, and what is more, the public debate has seldom been helpful in this respect because often the discussion about the desirability of PC gets bogged down in an exchange of polemics (cf, however, Cameron 1995). No matter how one thinks about this issue, it can hardly be avoided. On the one hand, the lexicographers have to take a stance towards this change of language in progress by making decisions of what terms to include, how to define, and, moreover, with what usage labels and illustrative examples to supply them. Even if all the dictionary makers would subscribe to the point of view that they only register usage without passing value judgements, it is also a universal truth that descriptive dictionaries (can) have a prescriptive effect, and thus contribute to linguistic change. The language learners and dictionary users, on the other hand, need to be aware of the problem, both in terms of passive language skills, e.g. to understand the impact of a Hemingway short story like "I can tell Wops miles off and, what is even more important, to mind their own language use so as not to cause offence by involuntarily stepping into a linguistic minefield because, as the late Dwight Bolinger put it as early as 1980, language is a "loaded weapon." It is arguable whether exchanging one term for a new one does indeed lead to more social equality or not. Some scholars have their doubts about this, e.g. Zimmer (1997: 136f.) states that domino-effect-like chain reactions as exemplified by the subsequent changes from Negroes, Coloured, Blacks, Afro-Americans to African-Americans have no lasting effects. As long as social, ethnic, etc. prejudice persists the new terms will in turn undergo the process of pejoration and have to be replaced by new ones. PC-critics often ridicule this process by giving excessive lists often with a plethora of pathetic examples of terms in "Oldspeak" and "Newspeak", thus trying to ridicule the movement. The language learners or their teachers, moreover, have to develop a linguistic awareness and consciousness of the problem that speech acts, as Austin and Searle put it, are doing things with words. This means, among other things, that one can insult or offend other people with inappropriate vocabulary. For these reasons it is imperative that dictionaries give guidance on usage and that the dictionary user is made aware of this problem. Since the Collins COBUILD Dictionary (COBUILD 2), the Longman Dictionary of Contemporary English (LDOCE 3) and the Oxford Advanced Learner's Dictionary (OALD 5) came out in major new editions in 1995, and with the Cambridge International Dictionary of English (CIDE) — a completely new dictionary — making its appearance on the

Recent English Learners' Dictionaries...

167

scene in the same year, it seems all the more appropriate to investigate how the four learners' dictionaries treat this important issue. In 1992 Tom McArthur commented rather defensively on this matter. In a newspaper article entitled: "Enter the era of caring words — PC has made it past the critics into the dictionaries" he says that "inevitably, dictionaries are drawn into the discussion" (Independent 23.8.1992: 22). Therefore the four dictionaries will be investigated to find out to what extent the new "caring terms" have found access into the dictionaries. Although it is rather commonplace to state that the lexicon of a living language is always in a state of flux, it is interesting to take a snapshot view of this rapidly changing field of PC. In a broader diachronic study it would be compelling to compare the present generation of learners' dictionaries to their predecessors and also to the new editions to come in order to verify whether the substitutions have caught on and have become part and parcel of everyday usage, or whether it is much more difficult to change the linguistic codes of a speech community than some advocates of PC would like to see. In this context though, one impressive example of the pejoration of a term, i.e. nigger must suffice: OALD 1 (1948, 3rd impr. 1952) nigger (coloq.) a negro; a black man of any race. OALD 2 (1963, 17th impr. 1972) !3 (impolite word for) Negro; member of any dark-skinned race OALD 3 (1974)! [=taboo] (impolite and offensive word for) negro OALD 4 (1989, 2nd impr. 1990) ( ! [=taboo] derog offensive) black person; negro. OALD 5 (1995) ( ! [=taboo] derog offensive) a black person. In the following I shall be looking at the way in which ethnicity, gender, sexual affinity, physical/mental achievement and old age4 have been dealt with in the dictionaries under scrutiny in terms of: - the inclusion/exclusion of headwords - bias in the definitions - usage notes or labels - bias in the examples.5 From these five areas gender will we given the most emphasis. As dictionaries are the object of this study and its objective being meta-lexicography, the procedure of the investigation will neither be pro or anti-PC but rather strictly empirical.

The sign "is [...] a warning of danger and the need for caution. Any word or phrase marked by this sign should also [as the taboo words] be avoided" (xi). For reasons of space the treatment of terms concerning ecologically motivated choices and constraints has to be omitted. The dictionary of Beard/Cerf (1992) draws attention to ecological correctness by providing entries such as lumberjack (tree butcher), fishing (raping the oceans), etc. For an introduction to the topic of ecolinguistics cf. Fill (1993, 1996, 1998). Because pictures and large-scale illustrations are not provided in all of the dictionaries COBUILD 2 doing completely without them - they will not be part of the investigation.

168

Ulrich Busse

2. Investigation of the four dictionaries 2.1. Political Correctness as a headword in the respective dictionaries Although the above discussion has proved that PC is considered a dogma or political programme rather than a precise technical term in linguistics, it is interesting to see how the dictionaries under investigation define the term, because the tenor of the definition can provide important clues as to the positions of the lexicographers in this issue. For many people in the course of the 1990s, the term "PC" itself has acquired pejorative connotations. CIDE: politically correct (s.v. politically, s.v. politics): If someone is politically correct (abbreviation PC), they believe that language and actions which could be offensive to others, esp. those relating to sex and race, should be avoided. · A word or expression which is politically correct (abbreviation PC) is used instead of another word or expression to avoid being offensive: Some people think that 'fireman' is a sexist term, and prefer the politically correct term 'firefighter'. [reference to language portrait on Sexist language] COBUILD 2: political correctness. Political correctness is behaviour and beliefs that reflect the attitudes and language that are typical of people who hold left-wing or liberal views; often used showing disapproval. In some sectors, where political correctness is a powerful force, I've been criticized for that. politically correct. If you say that someone or something '^politically correct, you mean that they reflect the attitudes and language that are typical of people who hold left-wing or liberal views; often used showing disapproval. politically incorrect. ... you mean that they reflect old-fashioned ideas and beliefs about equality. Gershwin's lyrics would today probably be deemed politically incorrect. LDOCE 3: politically correct: language, behaviour, and attitudes that are politically correct are regarded as right and acceptable because they are careful to avoid offending women, black people, DISABLED people etc: It's not politically correct to say 'handicapped' anymore. — see also PC2 — political correctness OALD 5: (s.v. political) politically correct adj (abbr PC) showing political correctness. political correctness: (sometimes derog[atory]) the principle of avoiding language or behaviour that may offend particular groups of people; an excessive display of or belief in this. When compared to each other it becomes apparent that two of the dictionaries, namely COBUILD 2 and OALD 5 point to the fact that the term may be used to show disapproval or scorn. CIDE and OALD 5 mainly stress the idea of avoiding offensive vocabulary; so does LDOCE 3, but this dictionary also refers to PC's claim of righteousness. Whereas CIDE gives a fairly neutral definition, OALD 5 adds the comment "an excessive display or belief of this." The definitions to be found in COBUILD 2 approach the term from a completely different perspective. This dictionary does not mention that PC involves avoiding potentially offensive words, but points to the underlying ideological background, which in my opinion is not very helpful for the learner because in this way awareness of offensive vocabulary is not raised.

Recent English Learners' Dictionaries...

169

2.2. Political Correctness as a topic in the general introductions of the dictionaries None of the four dictionaries has a general note on the treatment of PC, as does e.g. the COD (1995: 1662) in its style guide under the heading "offensive language and sexism." COBUILD 2 is the only dictionary that does completely without extra usage notes or language portraits, yet words that express a certain speaker attitude have been assigned the label PRAGMATICS in the extra column. Besides the inclusion of pragmatic information in the definitions the introductory section on pragmatic functions (xxxiv-xxxvii) has been substantially extended in comparison to COBUILD 1. CIDE has a one page language portrait6 on sexist language (p. 1305) arranged in two columns of "old-fashioned sexist language" and its counterpart "modern non-sexist language" including information on third person singular personal pronouns and possessives (he, his, him), the concord between the indefinite pronouns (anybody/-one, somebody/-one etc) as antecedents and the corresponding personal pronoun, words formed from -man, jobs or activities strongly connected with a particular sex, job names and "irreversible" binominals such as "men and women" and, finally, forms of address (Mr, Miss, Ms, Mrs). LDOCE 3 provides a number of usage notes interspersed among the alphabet, some of which are of special importance in relation to PC, e.g.: BEAUTIFUL, FAT, THIN; HE, MAN, MR; NEGRO and OLD. OALD 5 pursues a similar strategy, however, this dictionary provides only usage notes on FAT and THIN, and one larger usage note on GENDER, which includes the use of man and mankind, the suffix -ess in names and occupations; neutral words for names of jobs, social gender for jobs that have been traditionally held by one sex, and information that especially women find the indication of sex in expressions like woman doctor offensive. The pragmatic impact of address forms like Miss and Mrs has to be looked up at the respective headwords and not in a single usage note. To ease this process the entries are cross-referenced. The above comparison has revealed that the four dictionaries pursue different approaches ranging from the rather short reference to PRAGMATICS in COBUILD 2 to one extensive language portrait in CIDE to the usage notes in LDOCE 3 and OALD 5. These usage notes describe the neighbours in the respective semantic field and provide valuable information both on semantics and pragmatics as well as on the register of an item, viz. whether the term is euphemistic, friendly, offensive, informal, etc. The entries included in these notes have a cross-reference at the proper alphabetical slot in the dictionary referring back to the usage note. Thus the usage notes constitute an onomasiological position in a semasiological dictionary that helps learners in their proficiency in the foreign language by providing appropriate choices. As already mentioned at the beginning, none of the dictionaries has centralised information on - call it PC or offensive language - at any point in the dictionaries. Although OALD 5 has a study page (A 8) on polite expressions such as excuse me, sorry, I'm afraid, there is no information to be found on PC there. Nonetheless, OALD provides a complete inventory of usage notes (pp. 1350), as does CIDE for its "Pictures, language portraits and false CIDE's language portraits do not have exactly the same function as the usage notes in LDOCE 3 and OALD 5. The type of information provided and their length likens them more to OALD's study pages.

170

Ulrich Busse

friends (pp. 1772). Unfortunately, in LDOCE 3 such a table of contents for the usage notes is missing. In my opinion, however, it would be more user-friendly to provide one central column on offensive language and to give further hints where to find the respective usage notes, because a language learner is not in the same position as a dictionary reviewer who can sample the relevant data from the whole range of entries.

2.3. The labelling systems of the dictionaries The treatment of PC is largely a problem of usage. For this reason the pragmatic labels provided in the dictionaries need investigation so that they can be compared to each other. However, the amount of labelling as well as the actual labels used in the four dictionaries are different.

2.3.1.CIDE CIDE uses the following five categories of labels. The inside front cover features 28 mostly abbreviated labels in alphabetical order, which are explained and categorised in a language note s.v. label (p.790). 1) labels showing formality and informality: 2) labels showing the speaker's feelings or opinion: formal (fml) approving informal (infml) disapproving slang humorous taboo 3) Labels referring to the origin of a word, 4) Labels referring to different varieties of or to the situation in which the word is used: English7 specialized 5) Other labels law abbreviation medical figurative (fig) literary female, male poetic not standard saying trademark dated old use

These labels have been left out because they are not relevant for this study.

Recent English Learners' Dictionaries...

171

2.3.2. COBUILD2 In contrast to the five categories of labels used in CIDE, COBUILD 2 operates with only two classes, i.e. geographical labels8 and style labels (pp. xx-xxi), the reason being that COBUILD 2 integrates information on usage (register, variety, etc.) into the meaning gloss given in form of a whole sentence, e.g.: elderly "You use elderly as a polite way of saying that someone is old." Style labels: formal informal journalism legal

literary medical offensive old-fashioned

spoken technical written

2.3.3.LDOCE3 32 labels are explained on the inside front cover. They comprise the following categories:9 3) Words which are used in a particular situation, or show a particular attitude: approving humorous formal informal 4) Words which are used in a particular context or type of language: biblical old-fashioned spoken dialect old use taboo law poetic technical literary slang trademark not technical

2.3.4. OALD 5

OALD 5 also explains its labels (23 alltogether) on the inside front cover. Mention is made that "there are also many labels for expressions that are normally restricted to specialized use in specific fields, e.g. (anatomy), (computing), (grammar), (law)" which makes it impossible to establish how many labels are really being used in this dictionary.

See footnote 7. The labels 1) Words which are used only or mainly in one region or country [11 all in all] 2) Words which are used in English but which come from another language ... [5] have been omitted.

172

Ulrich Busse

1) The following labels indicate words that express a particular attitude or are appropriate in a particular context: approving (approv) informal (infml) rhetorical (rhet) derogatory (derog) ironic sexist euphemistic (euph) jocular (joe) slang (si) figurative (fig) offensive taboo (!) formal (fml) 2) The following labels indicate other restrictions on the use of words: archaic (arch) dialect Scottish (Scot) British (Brit) proprietary name (propr) technical (techn) catchphrase saying US dated This synopsis of the labelling systems shows that the labels (cf. appendix) are not immediately comparable across dictionaries, because they are part of differently organised systems and hierarchies, with LDOCE 3 and CIDE providing more categories than the other two dictionaries (cf. also Herbst 1996: 342).

2.4. The treatment of terms concerning ethnicity None of the four dictionaries has a usage label on the term race itself. However, some people tend to avoid the term, except in its strictly anthropological sense, in favour of words like community, ethnic group, nation, people, etc.10 For limitations of space only the prime example of Negro shall be discussed in detail. The sequence of terms ranging from Negro to African American illustrates the continuing pejoration of the successive terms. Because COBUILD 2 does not use taboo as a label, items such as coon and nigger are labelled as "extremely offensive", whereas the other dictionaries characterise them as taboo, slang (CIDE), taboo, very/extremely offensive (LDOCE 3) and taboo, slang, derogatory offensive in OALD 5. Whereas the dictionaries despite differences in individual labels unanimously agree in matter that coon and nigger are extremely offensive, consensus already ends at the labelling of Negro. While all dictionaries note its offensive potential and mention that the word is dated or old-fashioned, OALD 5 is the only dictionary not providing a diachronic label dated. Black has no usage label in CIDE. LDOCE 3 says in the usage note s.v. NEGRO: "People of African origin (recent or long ago) usually prefer to be called black, or in the US, often African-American. In the UK the term Afro-Caribbean is often used." COBUILD 2 and OALD 5, on the other hand, label the term black as sometimes offensive. While Coloured is again characterised as offensive by all dictionaries, only COBUILD 2 and OALD 5 indicate that the term is dated. Interestingly, Afro-American appears in none of the four dictionaries as an entry, only in LDOCE 3 as an example s.v. Afro-. The latest term, African-American is missing in CIDE and COBUILD 2. The PC-term person ofcolo(u)r for any non-white person is not entered in any of the dictionaries. 10

The sample of test words on ethnicity and the other domains is to be found in the appendix.

Recent English Learners' Dictionaries...

173

This example illustrates that it is difficult for the learner to come by accurate and up-todate information. In this respect, LDOCE 3 with its usage notes provides, in my opinion, the most vital information on lexical choices and their pragmatic implications at a single glance. For reasons of limited language proficiency and lack of socio-historical background, learners cannot be expected to work their ways from one entry to the next, especially when the entries are not consistently cross-referenced. The awareness that certain expressions can indicate racial prejudice is, however, developed to different degrees. While the lexicographers indicate that Native(s) and Negro are dated and offensive terms, the sensitivity towards people traditionally called Indians and Eskimos is not yet so highly developed. The dictionaries agree in their treatment of Red Indian as old-fashioned or dated and now considered to be offensive. Indian in the sense of Native American is regarded as offensive in CIDE, but no restrictive usage label is given in the other three dictionaries. All dictionaries list the term Native American without a usage label. The term First Nations Peoples chosen as a term for self-reference among Native Americans is not registered in any of the dictionaries, and that the word Eskimo is potentially offensive is indicated only by COBUILD 2 and LDOCE 3. In contrast to the comprehensive treatment of minorities as exemplified by the above examples, the coverage of upbeat PC-expressions is patchy, as the following examples illustrate: DWEM (for Dead White European Males), a term voicing criticism against the cultural dominance of this group is not included in any of the dictionaries, neither is Afrocentrism, however, Eurocentrism, another keyword in the PC debate, has found entry to COBUILD 2: "Eurocentric, ... you disapprove of it. PRAGMATICS."

2.4.1. Insulting abbreviations for members of various nationalities Abbreviated disparaging terms like Chink(y), Frog, Jap, Kike, Kraut, Limey, Paki, Polack, Yank and Κ/α? are labelled accordingly in all dictionaries. The preceding ten of these depreciatory stereotypes were used as testwords. CIDE lemmatises seven out of the ten items, COBUILD 2 only three, LDOCE 3 has eight, and OALD 5 includes five. The labels indicating register range from taboo to informal; that is to say that some of the terms are perceived to be less insulting or disparaging than others. The advice on slang terms already given in OALD 3 (1974: xi) is, in my opinion, still true today: "Some words and phrases in this dictionary have the sign !. These are taboo words. [...] The foreign learner of English may need to understand them but will do well to avoid using them, especially in conversation." Since all of the dictionaries under scrutiny are corpus-based the non-inclusion of terms may be due to infrequency. If this, however, should not be the case, there is no reason to exclude them, even if they are racist, because students might encounter them in works of literature and thus need to know their connotations, as the quotations for kike from the OED( 1992) illustrate: 7932 J. Dos Passes 1919 164 The little kike behind the desk had never been to sea. 1940 R. Stout Over my Dead Body vi. 84, I don't care if the background is wop or mick or kike.. so long as it's American.

174

Ulrich Busse

1956 D. Karp All Honorable Men 74 If you repeat that lie, I'll wring that skinny kike neck of yours with my own hands! 1963 V. Nabokov Gift iii. 179 My better half.was for twenty years the wife of a kike and got mixed up with a whole rabble of Jew in-laws.

2.5. The treatment of terms concerning gender

2.5.1. The third person personal and possessive pronouns he, his, him As the English language does not provide a neutral third person personal pronoun or possessive adjective that applies to both sexes, the treatment of he, his, him, which according to older conventions were used with respect for both sexes, seems to be a good starting point for the investigation of gender because this usage is considered offensive by some women." CIDE and COBUILD 2 in their articles on he mention that generic he is disliked by some people. LDOCE 3 and OALD 5, on the other hand, say that he is used for unknown persons whose sex is not known: "Everyone should do what he considers best." (LDOCE 3). Yet both dictionaries make a cross-reference to the respective usage notes [LDOCE 3 s.v. he, OALD 5 s.v. GENDER)] where more information, and furthermore, alternatives are provided.

2.5.2. Indefinite pronouns All dictionaries point out that indefinite pronouns like anybody/-one, somebody/-one, etc. are now usually followed by they, if one wants to avoid saying 'he or she'. However, CIDE labels this usage as informal and COBUILD 2 adds that some people find this use of they incorrect. The indefinite pronouns do not have usage notes in any of the dictionaries, except the OALD 5. In the example sentences in cases where two or more people are concerned they, them and their are used consistently. The OALD 5 provides cross-references from the individual entries to the usage note s.v. GENDER, but has not done so for anybody'/-one. The usage note s.v. gender reads like this: ''''Everybody, everyone, anybody, anyone, somebody, someone, etc are usually followed by the plural pronouns they, them and their ... You can also use he or she, his or her, or him or her, especially in more formal speech or writing ... Alternatively, the sentence can be rephrased, using a plural noun."

11

On the history of prescriptivism and descriptivism in the use of generic he see Wales (1996: 110133).

Recent English Learners' Dictionaries...

175

2.5.3. The use of man and mankind The use of man and mankind in the sense of 'all human beings, humankind' has been assigned as sexist or old-fashioned language. Womankind, on the other hand, is classified as dated or formal. The neutral term humankind is registered in all dictionaries. CIDE in its definition draws particular attention to the fact that it includes men and women: "the whole of the human race, including both men and women." LDOCE 3 only says "people in general", and OALD 5 refers to the usage note s.v. GENDER. 2.5.4. Job designations ending in -man or -woman The next major point in my investigation are words for professions in which pairs of terms ending in -man and-woman co-exist, and where to avoid sex-specific reference new, neutral terms have in many cases been coined. In most of these cases the dictionaries provide entries for all of these forms. It is interesting though to observe that they implement a different practice to lemmatise these items. With the only exception of policeman and policewoman, which are two separate entries, CIDE normally groups all terms under just one joint entry, e.g. chairperson, chair, chairman, chairwoman, all in bold face type and in the same font. The definitions for such entries begin with: 'a person who ...' or, to take another example: salesperson, salesman, saleswoman, dated also salesgirl 'a person who ...' These examples show that the sequence is not alphabetical but ranging from neutral to marked for gender to restricted in usage (dated in the above example). In contrast to this practice, COBUILD 2 keeps the entries for -man and -woman apart, e.g.: craftsman: '...a man who makes things skilfully with his hands' craftswoman: '...a woman who makes things skilfully with her hands'. This is also true for LDOCE 3: salesman: 'a man whose job is to persuade people to buy his company's products.' salesperson: 'someone whose job is selling things.' saleswoman: 'a woman whose job is selling things.'

The definitions, however, do not match, but this is not necessarily a sign of sexism as the job profiles could not be identical. OALD 5 usually (except policeman/'-woman and salesman/'-woman) also has joint entries for both forms, yet in alphabetical order, which means that e.g. businessman is the headword followed by femfiminej businesswoman in a smaller font, and the definition reads: 'a person who ...'. Depending on the form of lemmatisation (joint entry) or whether or not a feminine counterpart is listed, the second constituent -man has a changing meaning, varying between generic and male exclusive reference, e.g. COBUILD 2, LDOCE 3 and OALD 5 only register cameraman. In the three dictionaries -man has a generic meaning 'someone/person who ...'. CIDE, on the other hand, has an entry for: cameraman, -woman, -operator. 'With fireman the case is more complicated: CIDE again has a joint entry for firefighter, fireman, firewoman. COBUILD 2: '... a person, usually a man, whose job is to put out fires.'

176

Ulrich Busse

LDOCE 3: 'a man whose job is to stop fires burning; FIREFIGHTERS'. OALD 5: 'a member of the firebrigade.' Again, entries for the feminine formßrewoman are not provided. Craftsman and spokesman are also interesting cases in point because in COBUILD 2 they are defined as 'a man who ...', 'a male spokesperson'. In LDOCE 3 the definitions for these terms have generic reference: 'someone who ...'. This practice, however, is not consistently adhered to in LDOCE 3, because spaceman: 1 informal 'a man who travels into space; ASTRONAUT' 2 'someone in stories who visits the Earth from another world' and space-woman: informal 'a woman who travels into space; ASTRONAUT' provide counter evidence. OALD 5 offers two different senses for craftsman: 1 ' a skilled man, esp one who makes beautiful things by hand ... 2 'a person who does sth very well and with great attention to detail ...'. Whereas in these two cases the female counterpart is given as an entry in all of the dictionaries, postman and mailman are examples in which with the exception of CIDE the female terms are missing, but nonetheless the definitions in the remaining three dictionaries waver between generic and male exclusive reference: CIDE: postman, postwoman, infml postie, Am usually mailman, letter carrier 'a person whose job is to deliver and collect letters ...' COBUILD: postman British English 'a man who ...'; mailman American English 'a man who ...', LDOCE 3: postman especially BrE 'someone whose job is ...'; mailman AmE 'a man who delivers mail...' OALD 5: postman 'a person employed to collect and deliver letters'; mailman (US) = POSTMAN Thus, the treatment of the -manf-woman compounds seems to depend on whether both forms are being treated as separate entries, but apart from this, the lexicalisation of -man has obviously reached different degrees, i.e. it can imply 'male only' or 'person in general', or moreover, the above results could also be interpreted as pointing to a lack of consistency in lemmatisation, because at least to my mind the reasons for the different treatment of individual compounds are not fully apparent. If, on the one hand, one acknowledges that dictionaries are to reflect social reality and to portray it realistically, and if it is still true that the number of women in firebrigades is marginal, this could explain why three out of the four dictionaries do not see a reason to provide an entry firewoman, but if, on the other hand, all the dictionaries under investigation have usage notes on -man and -person, then it is not clear why in the LDOCE 3 craftsman and spokesman have generic reference. It is arguable indeed whether dictionaries should try, e.g. by means of parallel entries, to help initiate social change - although I have doubts that social inequalities can be changed in this way. Nonetheless, it is desirable to have consistency as a guiding line. In this case, CIDE, with its joint entries, is the dictionary most advanced in this respect, because questions of generic or male exclusive reference and definitions that do not match do not arise. However, the explanation of the label female/male (p. 790) is open to misunderstanding: "The form which can refer to both females and males, e.g. lion, is not labelled. Other examples: fire fighter, fireman (male), fire-woman (female)". Here it is not clear whether male pertains to fireman only or to both expressions, and what is more, in the entry

Recent English Learners' Dictionaries...

Ill

under the headword firefighter, fireman, fir ewoman ... 'a person who ...' none of the above labels is used. At least in respect to the designation of professions the label needs to be removed, because none of the terms ending in -man has the label male, or -woman (female) respectively.

2.5.5. Compounds ending in -woman and in -girl -Woman as a second constituent of compounds with reference to sexist language is included as a lemma only in CIDE s.v. sexism: "job names are used which refer unnecessarily to the sex of the person ... Jane is a camerawoman" are regarded as "old-fashioned sexist language" and OALD 5 s.v. GENDER "It is also usually possible to use a neutral word ... instead of man/woman in the names of jobs. Policeman/policewoman = police officer" LDOCE 3 s.v. woman: "6 businesswoman/spokeswoman etc. a woman who has a particular kind of job: a spokeswoman for the charity" does not give a usage note. Compounds like womanhour, womanyear, womanpower and phrases like one's own woman, woman in the street are still largely missing from the dictionaries. The use of -girl in compounds like salesgirl, shopgirl is marked on by all dictionaries except OALD 5 as offensive. OALD 5 does not provide a pragmatic label: "3 (usu[ally] in compounds) a female worker: an office girl/a shop-girl/a telephone-girl.

2.5.6. -Person compounds as an alternative to -man/-woman compounds The substitution of -woman for -man in job titles etc. draws attention to the fact that the holder of the respective position is female. In constrast to this gender specification the substitution of both terms through -person is an alternative leading to gender abstraction. The combining form -person is granted the status of a separate entry with extensive usage notes in all of the dictionaries, e.g.: "-person is used by people who do not want to use a term which indicates whether someone is a man or a woman" (COBUILD 2). In addition to this, several single items are registered, such as anchorperson, chairperson, salesperson and spokesperson.

2.5.7. The coinage of neutral terms Another alternative abstracting from the gender of the jobholder is the creation of new neutral words or of reducing the old compounds by means of back-clipping of their -man/woman constituent. Here the dictionaries have included these terms to varying degrees: - chair is listed in all dictionaries. - firefighter features also as an entry in all dictionaries; in OALD 5 as a subentry s.v.flre. - policeqfficer as well (although in CIDE and in OALD 5 only s.v. police and not as a headword such as policeman and policewoman) - flight attendant is registered in three dictionaries (not in the OALD 5 and in CIDE s.v. flight).

178

Ulrich Busse

- anchor is lemmatised only in COBUILD 2 and in LDOCE 3.

- camera operator only in CIDE. - business executive in none. The way in which these terms are perceived differs for many speakers. According to Allan/Burridge (1991: 44) "the -man locutions are not dysphemistic when used of a male denotatum, and the neutral locution is primarily intended to name the office (job) itself so as to acknowledge that women may hold such an office." They also note that neutral terms are usually "preferred in nonreferential contexts." In addition to that they say that a number of women do not find the term chairman dysphemistic and "are quite happy to be called Madam Chairman, because they understand the word chairman as an idiom denoting the office of chairperson" (45).

2.5.8. Gender specification vs. gender abstraction — the use of woman- and lady- as first constituents in job names As far as present-day English word-formation is concerned, gender specification may be realised by using lady- or woman- as first parts in compounds of the type lady doctor or woman doctor to show that the post is held by a female. Information on the status of such formations in the four dictionaries is scarce. Lady- is an entry only in CIDE and LCOCE 3 and the information given is not exactly equal. CIDE says that the addition of lady- to the names of particular jobs is dated, and LDOCE 3 describes lady doctor/councillor etc as "a polite word, which many women find offensive ...". CIDE is the only dictionary to include a meaning woman- "Woman also means female: She is Ireland's first woman (a female) president." but does not provide any guidance on usage.

2.5.9. Morphological means for gender specification In comparison to German, for example, morphological means to derive feminine forms from masculine forms have never been very productive in English (cf. Marchand 1969: pp. 286, Hellinger 1990: pp. 72; Weite 1996: pp. 200). Historically, the feminine forms have undergone the process of pejoration as pairs like governor/governess, major/majorette, master/mistress and many others demonstrate. Owing to this, feminine forms like authoress, poetess etc. are felt by some to convey negative connotations and to point unnecessarily to the fact that the person in question is a woman. For these reasons the lemmatisation of the four suffixes -ess, -ette, -ine, -trix and their combinations was tested. The four suffixes vary greatly in their morphological productivity, and are accordingly not all listed as entries in their own right. CIDE and COBUILD 2 do not register any of the above suffixes as headwords, and LDOCE 3 and OALD 5 lemmatise only -ess and -ette. Of the two dictionaries the OALD 5 is the only one to point out its usage (s.v. GENDER): "The suffix -ess in names of occupations such as actress, authoress and hostess shows that the person doing the job is a woman. Many people now use the same word for both sexes because it is unnecessary to make a distinction between men and women doing the same job: actor, author, host."

Recent English Learners' Dictionaries...

179

Accordingly, the number of entries with these suffixes is limited in all dictionaries under investigation. The only pairs to appear in all dictionaries are: actor/actress, host/hostess, murderer/murderess, steward/stewardess, usher/usherette and waiter/waitress. As regards actor/actress, COBUILD 2 gives the most concise information: "'Actor' in the singular usually refers to a man, but some women who act prefer to be called 'actors' rather than 'actresses'." Author is another interesting case in point. CIDE does not provide an entry for the feminine form. COBUILD 2 points to the fact that "some female writers object to this word [authoress], preferring to be called authors." LDOCE 3 has entries for both forms without usage labels: author: 'someone who writes books: Jeffrey Archer, successful author and politician. It's clear that the author is a woman.' authoress: 'a woman who writes books.' OALD also features two entries, but adds that the feminine form is dated. author: 'The writer of a book, play, etc.: Dickens is my favourite author. She is the author of numerous scientific articles and books.' authoress: '(dated) a woman author.' Hero is a further interesting example of ongoing semantic change which is treated differently in the four dictionaries. All of them enter hero and heroine as headwords. CIDE has again a joint entry: hero, female also heroine which points to the fact that hero can be used to refer to both sexes. COBUILD 2 has three senses for hero: 1 'male character...'; 2 'someone, especially a man ...'. 3 'If you describe someone as your hero you mean that you admire them a great deal, usually because of a particular quality or skill that they have.' LDOCE 3 hero: 1 'a man who is admired for doing something extremely brave.' 2 'the man or boy who is the main character in a book, film, play, etc.' 3 'someone you admire very much for their intelligence, skill, etc.' OALD 5 hero: 1 'a person who is admired by many for her or his noble qualities or courage'. 2 'the chief male character in a story, poem, play, etc. Thus in the sense of 'main character in a book, film, etc' hero is used with reference to men, whereas in the more general sense of 'someone who is admired' it seems to have lost the semantic property [+ MALE].

2.5.10. Chef vs. cook and tailor vs. seamstress In some cases, gender specific designations for professions formerly held predominantly either by men or by women stem from different etymological sources as the examples in the heading illustrate and in which the male term has had a higher social prestige than the female term. For tailor CIDE gives the definition: 'someone who ... usually men ...', LDOCE 3 and OALD 5 on the other hand stress that a tailor is someone (or a person) whose profession it is to make men's clothes to measure. For seamstress CIDE and LDOCE 3 point out that the term is dated or old-fashioned, respectively. Whereas in older editions of the dictionaries chef\s explained as 'the male head cook in a restaurant etc.' the definitions are now neutral, and CIDE has a gender balanced

180

Ulrich Busse

set of examples: He is one of the top chefs in Britain. She is head-chef at the Waldorf-Astoria. LDOCE 3 s.v. chef refers to a picture for restaurant [p. 918, not 917!] where a male chef is shown.

2.5.11. Social gender Apart from morphological or lexical means to designate gender there is also what Hellinger (1990: pp. 61) calls social gender, i.e. due to stereotypical images of gender roles in society there are a number of nouns which are grammatically and/or morphologically unmarked for gender but where the referents can be said to be prototypically male as in lawyer or female as in nurse. For this reason, a number of job designations were analysed for supposed sexist bias. Because the nouns do not overtly indicate gender it was investigated whether they are pronominalised by he or she or whether job names co-occur with male or female Christian names. Luise Pusch (1983) once wrote an article "Sie sah zu ihm auf wie zu einem Gott.' Das Duden-Bedeutungswörterbuch als Trivialroman'" where she analyses the gender roles in the example sentences and comes to the conclusion that the men are doing business and the women do the cooking and rear children. In comparison to these findings the example sentences given in the four learners' dictionaries give evidence that social reality has changed or, to put it more carefully, that at least most of the stereotypes have been removed from the dictionaries. Cowie (1995) compares the treatment of gender in OALD 1 to that in OALD 4 and focuses on the two aspects: "first, whether nouns or pronouns with male or female reference are given the more prominent place in dictionary examples; and second, whether in definitions and examples, women are portrayed in positions of responsibility and authority or are simply assigned to supporting domestic roles (286). The result for OALD 1 is: masculine items (38) are double that of feminine ones (19), and, furthermore, some of the examples which at this time (i.e. 1948) were invented and not corpus-based show male prerogatives. "OALD 4, by contrast, sees the application of a conscious decision by the editorial team (several of whom were women) to give greater prominence to females overall, and especially, to reflect present-day patterns of employment by featuring women in a wider range of professional roles" (291). Thus in OALD 4 the relationship between nouns, pronouns and determiners of male and female reference is in the ratio of 7 : 6. Regarding the illustrative examples, Cowie finds that in OALD 4 "women are portrayed in a number of responsible and prestigious jobs (not in all cases in the entries which define the relevant nouns)" (292). Herbst (1996: 328) compares the number of female/male persons referred to in examples in the section need to nervy in the four 1995 learners' dictionaries, among which "COBUILD 2 displays a very clear majority of male persons." Female references Male references

OALD 5 24 59% 17 41%

LDOCE 3 17 41% 24 59%

CIDE 26 39% 40 61%

COBUILD 2

33 34% 64 65%

Recent English Learners' Dictionaries...

181

My own investigation of pronominalisation and the presence of men and women in the illustrative examples shows that, on the whole, the dictionaries no longer reflect old social stereotypes. On the one hand, this can be attributed to the phrasing of the example sentences where third person singular examples are avoided in favour of first or second person singular examples, surname only or pluralisation: - First person: "/ want to be a ballet dancer when I grow up" (LDOCE 3).' 2 - Second Person: "Don't sign anything until you've consulted a lawyer." (OALD 5) - Surname only: "Nurse Jones (LDOCE 3) - Pluralisation: "Criminals should be reeducated as well as punished." (CIDE)

On the other hand, apart from these strategies the dictionaries often provide two example sentences showing a man doing something, and a woman doing something. In some cases they even try to be ostensibly non-sexist: (1) "He worked as a nurse in a psychiatric hospital." (CIDE) (2) "Jane Coker, lawyer ..." (3) ''She is a civil/mechanical/electrical/software engineer." (CIDE)

For secretary and typist the dictionaries are cautious in that they do not give an example at all, or one which does not show sexist bias. The only example to fit the stereotype is to be found in LDOCE 3: (4) "Julie works as a secretary in a lawyer's office."

On the other hand, the examples for drinker and criminal show that these still seem to be male domains: (5) "He's a heavy/light drinker." (CIDE) (6) "Dave has always been a bit of a drinker r (LDOCE 3) The examples given under professor aptly illustrate social reality and change (?). (7) Professor Stephen Hawking .... She was lecturer at the university for 20 years before she became a professor. (CIDE) (8)... Professor Cameron .... In 1979, only 2% of British professors were female. (COBUILD 2) (9) Thank you, Professor Barkley, for your comments. | my science professor, a professor of history. Ted's a college professor. | Professor, can I ask you a question? (LDOCE 3) (10) Professor (James) Brown. · She was made (a) professor at the age of 40. (OALD 5)

2.6. The treatment of sexual orientation Terms denoting sexual orientation very often have strong connotations. Apart from the neutral terms homosexual, lesbian, transvestite, etc. there exist a number of terms that belong to the register of slang, especially for male homosexuality. Owing to this, when talking about others whose orientation or practices one does not approve of, speakers often resort to the registers of slang or taboo terms. 12

Here and in the following examples my emphasis.

182

Ulrich Busse

According to this, the dictionaries provide entries for a number of such terms and assign usage labels to them that range from informal over derogatory to offensive and slang. On the whole, the pragmatic classification of the terms is comparable in the different dictionaries; all of them give warnings that using these words may cause offense. In my opinion, the most interesting term in this set is queer because the connotations of the word seem to depend on whether it is being used as an in-group solidarity marker among homosexuals or disparagingly by heterosexuals. The only dictionary to mention this is CIDE: "Queer is sometimes used disapprovingly. Queer is also used in a way that is not offensive, esp. by homosexuals." In addition to this, CIDE is also the only dictionary to include compounds like gay activist, gay liberation and gay liberation movement.

2.7. The treatment of physical/mental achievement All of the dictionaries are well aware of the fact that many terms denoting physical or mental handicap can cause offence. If e.g. the terms cripple(d), disabled and handicapped are scrutinised, the dictionaries describe cripple as dated (CIDE), or as (sometimes) offensive. Handicapped is also undergoing the process of pej oration. Whereas COBUILD 2 and LDOCE 3 do not give usage labels, CIDE treats the term as dated, and OALD as sometimes offensive. Disabled is the word that appears in all of the dictionaries without a usage label. The term advocated by the proponents of PC, -challenged, for those with a particular physical or mental disability, is to be found only in COBUILD 2 and LDOCE 3. The description of the term is, however, slightly different. LDOCE 3 treats it as an American euphemism: "visually/physically/mentally challenged AmE an expression for describing someone who has difficulty doing things because of blindness etc, used when you want to avoid saying this directly." COBUILD 2, on the other hand, does not restrict the term to any national variety of English and indicates that the term is "often combined with inappropriate words for humorous effect" as in vertically challenged or intellectually challenged. This shows that a word originally coined to avoid terms laden with negative connotations can easily be converted from a euphemism to a dysphemism, especially by people who do not sympathise with the movement of political correctness and combine -challenged with adjectives that sound rather ridiculous or pathetic such as vertically challenged for 'small in stature'. Limited intellectual capacity can also be referred to in various ways. Retarded has no usage label, except in CIDE (dated) and COBUILD 2, which treats the term as old-fashioned. The treatment of more recent euphemistic terms such as mental handicap and learning difficulty/disability shows different degrees of PC in the dictionaries. CIDE says for mental handicap that this is dated for learning disabilities and s.v. learning difficulties (!) the definition tells us that it is a euphemism: "you mean that they have low mental abilities ...". LDOCE 3, on the other hand, lemmatises both terms in a neutral way without any notes on pragmatics. Surprisingly, COBUILD 2 and OALD 5 do not provide entries for these terms. Another case where the treatment in the dictionaries diverges is spastic, ranging from taboo to medical:

Recent English Learners' Dictionaries...

183

CIDE: taboo '(a person who is) suffering from cerebral palsy ... (dated) My sister is (a) spastic. The Spastics Society in Britain changed their name to SCOPE in 1994.· Children sometimes describe other children as spastic, to mean that they are foolish or not very skil-

fid:

COBUILD: ... 'Most people now refer to someone with this disability as having cerebral palsy: LDOCE 3: 1 'slang an offensive word meaning stupid or lacking in skill, used especially by children' 2 Old-fashioned having CEREBRAL PALSY' OALD 5: (medical) suffering from cerebral palsy ...' The terms denoting insanity are also an interesting point for observation. Allan/Burridge (1991: pp. 186) note that "in the twentieth century, mental illness remains an area of abundant euphemism [...]. One reason for this is that mental illness still has a great deal of mystery to it." As in the other cases the labels assigned to the individual words across the dictionaries do not fully match. What is striking, nonetheless, is that the whole paradigm either in the strictly medical sense and the colloquial use seems to be either old-fashioned or offensive. As illustration the labels provided in CIDE will be presented as representative for this set of terms: -

demented: informal or dated insane 'mentally ill': dated or informal lunatic 'mentally ill': dated, now taboo mad 'mentally ill': Some people consider this use of 'mad' to be old-fashioned and offensive.

The same basically holds true for terms denoting mental hospital: asylum and lunatic asylum are treated as dated or old-fashioned. Whereas COBUILD 2 and OALD 5 have no pragmatic information on the use of mental hospital, LDOCE 3 gives psychiatric hospital as a synonym after the meaning gloss, and CIDE goes so far as to characterise the word as dated or informal. The coverage of humorous and informal terms differs quite considerably from dictionary to dictionary as the entries for loony bin illustrate: taboo slang humorous (CIDE), humorous (LDOCE 3), no entry in COBUILD 2, and slang, jocular offensive in the OALD 5. Thus the treatment of terms for mental illness and the corresponding institutions does not leave open many lexical choices for the learner. The terms ableism/ableist are not registered in any of the dictionaries.

2.8. The treatment of terms concerning old age In contrast to the missing term ableism/abelist for discrimination of the able-bodied against the disabled, the term ag(e)ism is listed in all dictionaries. COBUILD 2 gives a pragmatic notice saying that the term is "used showing disapproval", OALD 5 has the usage label derogatory attached to it. The other two dictionaries do not provide labels. All of the dictionaries have usage notes for elderly. In comparison to the section of mental/physical disability for which all the dictionaries show great concern, no corresponding "caring labels" have been introduced for senile, which can be as hurting as lunatic, mad etc. Geriatric, however, outside its strictly medical sense of Of or for old people', is treated differently in the dictionaries:

l84

Ulrich Busse

CIDE: (disapproving) 'too old and weak' COBUILD 2: 'If you describe someone as a geriatric, you are being disrespectful' PRAGMATICS. LDOCE 3: informal 'too old' OALD 5: no usage label [only the medical sense is given] In contrast to mental hospital, Old peoples' home does not have any label saying demeaning or offensive. OALD 5 is the only dictionary providing a label (informal). If registered at all Old folks' home is classified as informal (OALD 5). Old age pensioner, OAP, pensioner, retiree do not have any usage labels either. For senior citizen the treatment is again different in the dictionaries: CIDE and OALD 5 describe it as a euphemism: "... is used to avoid saying Old person'." (CIDE). Whereas LDOCE 3 gives the following advice on usage s.v. OLD: "Most people however now prefer to be called senior citizens, and this is the most common, polite, and acceptable expression to use." Most of the combinations with old, e.g. oldie, oldtimer, old boy, old girl, etc. with the only exception of old maid, are labelled as informal or dated.

3. Conclusion Cowie (1995: 294) writes that the learners' dictionary "can serve as the mirror both of the conscious (or unconscious) social values and judgements of its compilers, and of the chief social and ideological concerns of the period of compilation."13 The treatment of PC gives evidence for this, ranging from the inclusion or non-inclusion of a term, to the attitude that is signalled by usage labels, the style of the definition or the avoidance of bias in the illustrative examples. None of the four dictionaries under investigation devotes a section of the general introduction to the topic of PC. As far as usage notes are concerned, the domains of my investigation are covered to very different extents. All of the dictionaries under scrutiny, except COBUILD, have notes on sex or gender. Both LDOCE 3 and OALD 5 provide usage notes on appearance (s.v. FAT, THIN). LDOCE 3 is the only dictionary to give one on race (s.v. NEGRO) and age (s.v. OLD). Notes on sexual orientation and mental/physical achievement are missing in all of the dictionaries. What Bolinger (1980: 104) said in respect to sexism — that the greatest danger of biased language is unawareness — is also relevant to the treatment of PC, as is his statement that "the rehabilitation of names of things and of creatures comes in part automatically as the things and creatures are rehabilitated, but it is a factor too in their rehabilitation." The treatment of the terms concerning ethnicity has shown that usage labels for offensive words are generally provided. The comparison of the words denoting Negro, Indian and Eskimo has revealed, however, that awareness of potentially offensive language is still developed to different degrees. 13

My colleague Oliver Grannis (Osnabrück) has suggested the term sociolexicography for studies that investigate the reflexes of social processes in dictionaries.

Recent English Learners' Dictionaries...

185

The treatment of gender in the dictionaries has shown that the topic can be handled in various lexicographical ways: 1) to provide joint entries for male, female and neutral forms with a uniform definition of the kind 'person who ...'as does CIDE. 2) to provide specific commentaries on usage, either directly in connection with the respective headword, or in greater detail in language portraits or usage notes. 3) to give a balanced set of examples, usually under the "unmarked" form for men and women. 4) to attach a pragmatic label, especially a diachronic one, to indicate that an item is dated or oldfashioned.

In 1985, Cannon and Roberson in their investigation of American dictionaries of neologisms asked the question "Sexism in present-day English: is it diminishing?" For a synchronic study it is well-nigh impossible to answer this question, but by comparing their examples to my data-base, and by incorporating the evidence given in Cowie (1995) and in Herbst (1996), I would say yes, it is diminishing. All of the learners' dictionaries are keenly aware of the issue. As far as terms denoting sexual orientation and mental or physical achievement are concerned, the four dictionaries, while showing individual differences in usage labels, generally show great concern to point out offensive usage. In comparison to the previous domains, where all of the dictionaries have entered into "the era of caring words", the treatment of terms concerning old age still lags behind, as most of the terms are not marked for register, or if they are marked they are treated as informal or dated. In my opinion, Nida (1995: 69) sums up neatly the dangers involved in euphemisms and dysphemisms in relation to old age: "We clearly prefer Leisure World and Golden Years Manor to Old folks Home; but when brochures describe small apartments for people over fifty as Retirement Estates, the use of lexical cosmetics becomes absurd." Concerning coverage and lexicographical treatment of the terms I investigated, all of the four dictionaries have achieved an admirable standard and could be said to be a step or two further along "on the way to the perfect learners' dictionary" (Herbst 1996), although on the micro-level of investigation it is surprising how many discrepancies in lemmatisation, register assignment, etc. are still to be found. But all of this becomes apparent only when the data are directly comparable in form of tables (see appendix). This situation is unlikely ever to materialise for the average language learner. As I only looked at the most obvious entries, it might be revealing or indeed might lead to partly different results to check those above immediate suspicion. By means of the newly available OALD 5 on CD-ROM hidden bias may easily be retrieved.14 On the basis of the data I selected, it is difficult to pass a value judgement on the respective merits of the dictionaries and to recommend a particular one. Impressionistically, from the type of joint entries, the usage labels and the gender ratio in the illustrative examples I would say that the CIDE seems to be the most "politically correct" dictionary. COBUILD 2 is concise and pragmatically informative in its entries but does not pay special attention to the topic in form of usage notes. LDOCE 3 and OALD 5, on the other hand, provide a number of invaluable notes on usage. 14

For further information on hidden lexicographical information see Lemberg (this volume) and Goebel/Lemberg/Reichmann (1995).

186

Ulrich Busse

4. References 4.1. Dictionaries 4.1.1. Learners' Dictionaries (Corpus) CIDE = Cambridge International Dictionary of English. Ed. P. Procter. 1995. - Cambridge: Cambridge University Press. COBUILD 2 = Collins COBUILD English Dictionary. Ed. J. Sinclair. 2nd ed. 1995. - London: HarperCollins. LDOCE 3 = Longman Dictionary of Contemporary English. Ed. D. Summers. 3rd ed. 1995. München: Langenscheidt-Longman. OALD = The Advanced Learner's Dictionary of Current English: OALD 1 (1948) and OALD 2 (1963) Eds. A. S. Hornby, E. V. Gatenby and H. Wakefield - London: Oxford University Press; Oxford Advanced Learner's Dictionary of Current English: OALD 3 (1974) Ed. A. S. Hornby with the assistance of A. P. Cowie and J. Windsor Lewis; OALD 4 (1989) Ed. A. P. Cowie; OALD 5 (1995) Ed. J. Crowther - Oxford: Oxford University Press.

4.1.2. Other Dictionaries COD = The Concise Oxford Dictionary. Ed. D. Thompson. 9th ed. 1995. - Oxford: Clarendon Press. OED = The Oxford English Dictionary on CD-ROM. Eds. E. S. C. Weiner and J. Simpson. 1992. Oxford: Oxford University Press.

4.2. Secondary Sources Allan, Keith and Kate Burridge (1991): Euphemism and dysphemism: language used as shield and weapon. New York: Oxford University Press. Beard, Henry and Christopher Cerf (1992): The official politically correct dictionary and handbook. New York: Villard Books. Berman, Paul (ed.) (1992): Debating p.c.: the controversy over political correctness on college campuses. New York, NY: Delta Books. Bolinger, Dwight (1980): Language — the loaded weapon: the use and abuse of language today. London: Longman. Cameron, Deborah 2 (1992): Feminism & linguistic theory. Houndmills: Macmillan. — (1995): Verbal hygiene. London: Routledge. Cannon, Garland and Susan Roberson (1985): "Sexism in present-day English: is it diminishing?", in: Word. Journal of the International Linguistic Association 36 (1), 23-35. Cowie, Anthony P. (1995): "The Learner's dictionary in a changing cultural perspective", in: Kachru, Braj B. and Henry Kahane, eds. Cultures, ideologies, and the dictionary. Studies in honor of Ladislav Zgusta. Tübingen: Niemeyer, 283-295 (= Lexicographica. Series Maior 64). Crystal, David (1995): The Cambridge Encyclopedia of the English language. Cambridge: Cambridge University Press.

Recent English Learners' Dictionaries...

187

Dunant, Sarah (ed.) (1994): The war of the -words: the political correctness debate. Reading: Cox & Wyman. Fill, Alwin (1993): Ökolinguistik. Eine Einführung. Tübingen: Narr. — (ed.) (1996): Sprachökologie und Ökolinguistik. Tübingen: Stauffenburg. — (1998): "Ecolinguistics: state of the art 1998", in: Arbeiten aus Anglistik und Amerikanistik 23(1), 1-16. Greil, Tanja (1998): Political Correctness und die englische Sprache: Studien zu (nicht-) diskriminierendem Sprachgebrauch unter besonderer Berücksichtigung des Social Labeling. Hamburg: Kovac. Goebel, Ulrich, Ingrid Lemberg and Oskar Reichmann (1995): Versteckte lexikographische Information. Möglichkeiten ihrer Erschließung dargestellt am Beispiel des Frühneuhochdeutschen Wörterbuchs. Tübingen: Niemeyer (= Lexicographica. Series Maior 65). Hellinger, Marlis (1990): Kontrastive feministische Linguistik. Mechanismen sprachlicher Diskriminierung im Englischen und Deutschen. Ismaning: Hueber. Herbst, Thomas (1996): "On the way to the perfect learners' dictionary: a first comparison of OALD 5, LDOCE 3, COBUILD 2" and CIDE", in: Internationaljournal of Lexicography 9, 321-357. Hoffmann, Arne (1996): Political Correctness. Zwischen Sprachzensur und Minderheitenschutz. Marburg: Tectum. Marchand, Hans (1969): The types and categories of present-day English word-formation. München: Beck. 2nd ed. McArthur, Tom (1992): "Enter the era of caring words", in: The Independent 23.8.1992: 22. Nida, Eugene A. (1995): "Lexical cosmetics", in: Kachru, Braj B. and Henry Kahane, eds. Cultures, ideologies, and the dictionary. Studies in honor of Ladislav Zgusta. Tübingen: Niemeyer, 69-72 (= Lexicographica. Series Maior 64). Pätzold, Kurt-Michael (1997): "The new generation of learner dictionaries (Part II of Words, Words, Words)", in: Fremdsprachen Lehren und Lernen 26, 184-218. Pusch, Luise F. (1983): '"Sie sah zu ihm auf wie zu einem Gott.' Das Duden-Bedeutungswörterbuch als Trivialroman", in: Der Sprachdienst 27 (9/10), 135-142. Sieburg, Heinz (ed.) (1997): Sprache - Genus/Sexus. Frankfurt/M.: Lang. Wales, Katie (1996): Personal pronouns in present-day English. Cambridge: Cambridge University Press. Weite, Werner (1996): Englische Morphologie und Wortbildung. Frankfurt/M.: Lang. 2nd ed. Whitcut, Janet (1984): "Sexism in dictionaries", in: R.R.K. Hartmann, ed. LEXeter '83 Proceedings. Papers from the International Conference on Lexicography at Exeter, 9-12 September 1983. Tübingen: Niemeyer, 141-144 (= Lexicographica. Series Maior 1). Zimmer, Dieter E. (1997): "Die Berichtigung. Über die Sprachreform im Zeichen der Politischen Korrektheit", in: Zimmer, Dieter E., Deutsch und anders. Die Sprache im Modernisierungsfieber. Reinbek: Rowohlt, 105-180.

Ulrich Busse

188

Appendix: synopsis of dictionary entries investigated for the treatment of ethnicity, gender, sexual orientation, physical/mental achievement and ageism15 Ethnicity (I) coon

nigger

Negro

Black

Coloured

CIDE taboo slang. This word is generally considered to be very offensive. taboo slang. A very offensive word

COBUILD 2 is an extremely offensive word for a black person

LDÜCE 3 ÜALD5 taboo a very offen- (! si derog offensive) sive word for a black a black person person

taboo an extremely offensive word for a black person; see NEGRO (USAGE) dated now consid- an old-iashioned old-fashioned ... now ered offensive by word which some usually considered many people people find offensive offensive USAGE NOTE no usage label 3 Black people are see NEGRO sometimes referred to (USAGE) as blacks, though some people find this use offensive. this word is consid- 3 an old-fashioned an offensive word ered offensive by use which many peo- see NEGRO many people. ple find offensive (USAGE)

AfroAmerican

is an extremely otfensive word for a black person

as an example s. v. Afro- very wellknown AfroAmerican family

AfricanAmerican

"

""

Person ot Colo(u)r

"

Native American Indian

no usage label

taboo dated see at NATIVE

(sometimes

offensive)

(sometimes

offensive)

1 (a) (dated often offensive) 2 (b) (S African) offensive

An Afro-American (=a black American person) s. v. Afrosee, NEGRO (USAGE)

an American Black

"

™

"

no usage label

no usage label

(also dated Amerindian) 2 = NATIVE AMERICAN

= Native American. an old-fashioned use no usage label This is considered offensive by many people. Red Indian taboo dated see at an old-fashioned now considered ofNATIVE word which is now fensive considered offensive redskin

(! derog offensive) a black person

(also redskin) (! dated offensive) a Native American, an American Indian old use ... now con- (.' dated offensive) = sidered to be offen- RED INDIAN sive

The abbreviations for the usage labels are to be found in chapter 2.3

189

Recent English Learners' Dictionaries...

Ethnicity (II) Eskimo

Native

CIDE In Canada, Eskimos are called INUIT

CUBU1LÜ 2 Ihese people now usually call themselves Inuits and Aleuts, and the term Eskimo is sometimes considered offensive. 3 Some European people use native to refer to a person who was born or lives in a non- Western country ... some people consider this use offensive

also used ot a person who lives in a simple and traditional way in a country considered to be less developed than the speaker's own country. This usage is increasingly considered offensive. DWEM WASP esp. Am disapproving EurocenEurocentric, you distrism approve of it. => Pragmatics ~ Atrocen- ~ trism race no usage label no usage label

LDUCE 3 a word ... that many of them consider offensive compare INUIT

ÜALD5 no usage label

a word that is now considered offensive, in former times used by Europeans ...

2 (dated usu offensive) a person born in a place, esp one who is not white and considered ... to be inferior

esp. Am L

(esp t/i' usu derog)

~

~

USAGE NOTE

no usage label

Ulrich Busse

190 Ethnicity (III): Insulting abbreviations for members of various nationalities Chink(y) Frog 'French'

Jap Kike

CIDE taboo slang (an otfensive word for) a Chinese person ß r and Aus slang a French person. This word is considered offensive by most people. Am taboo ( a very offensive word for) a Jew taboo ( a very offensive word for) a Jew

an informal use which some people find offensive

OALD 5 LDOCt 3 (! si offensive) a Chia very offensive word for a Chinese nese person person informal an oifensive (infml offensive) a French person word for a French person

-

-

CÜBUILÜ 2

-

taboo an extremely (! si offensive) a Jew offensive word for a Jewish person Am and Aus dated Some Americans Limey AmE old-Jashioned a slang a British per- refer to British peo- slightly insulting son ple as limeys; an word for a British informal word person Kraut slang an insulting (! si offensive) a Uer'German' word for someone man from Germany Paki taboo disapproving a BrE taboo a very person from Pakioffensive word for a stan; This word is person from Pakistan considered offensive Polack AmE an insulting word for someone from Poland Yank Yank, Yankee ...an informal use informal 1 ... some- {Brit infml sometimes infml ... (dispprov- which many people times used in an in- aerog) = YANKEE find offensive sulting way by => note at AMERIing) someone who is not CAN American

Yid

Recent English Learners' Dictionaries...

191

Gender (I) C1DE 1 ) pro- He is sometimes used nouns: he to refer to a person whose sex is not known. Some people do not like this use because it is offensive to women, and prefer he/she, s/he or they => Sexist language they

(infml) 1 hey can be used to avoid saying 'he or she' => Sexist language

anyone / example sentences: everyone they/them/their anybody/ no usage note everybody ...

COBU1LD 2 LDOCE 3 2 In written English, 2 used to talk about writers sometimes use anyone, everyone, he to refer to a person or an unknown perwithout saying son who may be whether that person is either male or fea man or a woman. male: Everyone Some people dislike should do what he this use and prefer to considers best. use 'he or she' or USAGE NOTE 'they'. 2 You use they in4 used to avoid saystead of 'he or she' to ing 'he' or 'she' refer to a person after words like 'anyone', 'no one', without saying whether that person is 'everyone' etc. a man or a woman. see EVERYONE Some people think (USAGE), this use is incorrect. HE (USAGE) example sentences: no usage note! [exthey/them/their amples containing no usage note them: Has everyone finished their drinks'?] usage note: Each and Every

no usage note

USAGE NOTE

2 sometimes used to refer to all human beings, including bot males and females. Some people dislike this use.

3 'human being' a) old fashioned: a person either male or female b) people as a group USAGE NOTE

mankind => Sexist language

some people dislike this use

womankind

a formal word

compare WOMANKIND -MAN (USAGE) compare MANKIND

2) (-) man combining form Some people find this use offensive ... They prefer ... other words such as police officer, business executive ... => Sexist language man 'human race' ... Some people find this use of man offensive... They prefer ... humanity, humankind and people => Sexist language

humankind

(dated) Compare MANKIND; HUMANKIND no usage note

no usage note

no usage note

OALÜ5 2 a person, male or female, whose sex is not stated or known,.... => note at GENDER

2 (used to refer to a person without specify ing the sex, eg after someone or nobody) ... => note at GENDER

any body /-one: no usage note, no cross reference everybody/-one: note at GENDER somebody/-one: note at GENDER note at GENDER

2 a) a human being ot either sex => note at FEMALE 3 a) the human race b) the people of a particular period of history => note at GENDER See also HUMANKIND => note at GENDER (/ml) (/ml) = MANKIND => note at GENDER

Ulrich Busse

192

Gender (II): Man-IWoman: Compounds and natural terms anchor -man

χ ^l joint entry

CUBUIL 2 x a person who ... x a person who ...

-woman

X

x

-person businessman

χ

CI E

| joint entry: x a person who... x χ ) -

-woman business executive camera man x -woman χ camera operator χ chairman X

Γr joint· entry: a

x

LDOCE 3 x AmL· x AmE a male anchorperson a woman who ... x AmE a female anchorperson x AmE s.o. who ... separate entries: a x separate entries: a man who... man who... a woman who... x a woman who . . . [business person/ -people] a person who ... x s.o. who ...

1 person who... )

x [2 meanings] 1. head of ... r joint entry un2. person in charge -woman x l der -person', a x woman in charge... chairperson x person who ... x person who ... chair x person who... X ) chairmanship X x -womanship craftsman -woman craftsmanship -womanship tireman -woman firefighter milkman -woman policeman -woman

x

1.r joint entry: a

x ) person who... x

x a man who... x a woman who... x

'") x ) (subentry) a person who... x

| joint entry: a person who... x ) (subentry) x a person who ...

x x 12 meanings] 1 . someone, esp a man /· joint entry: 2. BrEs.o. who... ) person who ... x woman who ... x (subentry) x someone who ... x => note at GENDER x x x

Γ·

x 's.o. who...' x a man who... [generic reference] x 'a woman who ... ' x x

x a person, usually a x a man who ... · entry man, who.. x fI-joint x J under -fighter: x s.o. who ... a person who ... x people who ... x a person who ... x a person who ... x s. o. who ... X

x a male x x a female separate x entries! aolice officer s. v. police a male x or female of the police force

UALD5

x a member ot a tire brigade x s. \.fire a person who ... x a man who ...

a man who ... a woman who...

x a male ... separate x a male ... separate x a female ... entries x a female ... entries

a member of the police force

x a member of the police force

x s. v. police a member of the police force

Recent English Learners' Dictionaries...

193

Gender (III): Man-Woman - compounds; social gender UDH CUBUIL 2 χ Br. English: a man χ | joint entry: a χ ) person who... χ Am for POSTMAN χ Am. English: a man χ Am for POSTMAN man who ... χχ Jr joint entry under χχ aa woman who ... χ ) -person: a person χ a person who ... who ... χ a male astronaut, r joint entry: a used mainly by children J person who...

L CE 3 χ esp BrE S. 0. who

UAL 5 χ a person who

χ AmEa man who ...

χ (US) = POSTMAN

-woman spokesman

X

χ χ

-woman

χ f joint entry under 1 -person: a person χ J who ... combiningform. It is often used instead of -man or -woman => Sexist language

postman -woman mailman lettercarrier salesman -woman -person spaceman

-person 3) -person

4) woman instead of man one's own woman Woman to woman woman in the street womanhour/ -year womanpower 5) social gender bricklayer criminal

dancer

*T · x 1

χ a male spokesperson χ a female spokesperson χ a person who ... 2 is used by people who do not want to use a term which indicates whether someone is a man or a woman. -people can also be used in this way.

χ a man who ... χ a woman who... sep. entries χ a person who ... 1 a man who ... χ 1 2 s. o. in stories }· joint entry: a I person who... a woman who ... χ J (subentry) 's.o. who" (generic] χ ] joint entry: a J person who... 'a woman who' χ (subentry)

χ a man who ... χ a woman who ... χ s o. who ... χ

χ

χ usage note χ a person who ... usage note under MAN (forming compound ns) a person concerned with: salesperson, spokesperson => note at GENDER

'

X

be your own woman

""

~

~

"

~

~

~

~

~

~

~

"

~

-

-

-

-

no example Criminals should be reeducated as well as punished.

no example no example no example A group of gunmen Ihe man is a criminal. petty criminals / a war attacked a prison and How could the people criminal set free nine criminals elect him to office? in Moroto. s.o. who ... Ex.: He ... aerson who ... 1 profes- 1 want to be a ballet a) a person who dances: sional: His previous dancer when 1 grow up. He's a good dancer, *irl-friend was a dancer b) a person whose occupawith the Royal Ballet tion is dancing: She's a 2 He was the best dancer. dancer in L.A.

Ulrich Busse

194

Gender (IV): Social gender doctor: 1) medicine

CIDfc He decided to become ...

2) education no example

COBU1LD 2 LDÜCL 3 ÜALD 5 Doctor Paige will be 1 think you'd better You'd better see a go to the doctor ... doctor about that cut. here ... Doctor of Philosophy a Doctor of Law He is a doctor of philosophy. a heavy drinker no indication ot sex Dave has always in the examples been a bit of a drinker. no indication of sex Mike's an electrical no indication oi social in the examples engineer. sex

drinker

He's a heavy/light drinker.

engineer

She is a civil/mechanical/ electrical/software engineer. Jane Coker, lawyer example in the plu- no example ral

lawyer nurse plumber professor

scholar 'academic' scientist secretary Office worker'

She had spent 29 He worked as a nurse in a psychiat- years as a nurse. ric hospital. no example no indication oi social sex in ex. Protessor Stephen Professor Cameron Hawking ... She was a lecturer at In 1979, only 2% of the university for 20 British professors were female. years before she became a professor. Plural example; ... no indication ot social sex in ex. an influential Islamic scholar Plural example no indication oi social sex in ex. no example no indication ot social sex in ex.

teacher

My father was a listory teacher.

typist

no example

Don't sign anything until you've consulted a lawyer. Nurse Jones / a male no indication of social nurse sex no example

no example

Ted's a col lege pro- Professor (James) fessor. Brown. She was made (a) professor at the age of 40. ... a Latin scholar

a Greek/history scholar

no example

no indication ot social sex no indication oi social sex

Julie works as a secretary in a lawyer's office. You can ring my secretary to make an apaointment. I'm a teacher ... / ... Miss lindale's my my English teacher ier chemistry favourite teacher. teacher no indication ot no example no indication oi social social sex sex

Recent English Learners' Dictionaries...

195

Gender (V): Morphology, suffixation C BU1LD 2

C1DE

LD CE 3

ALD5

6) morphology sutlixation a) -ess actor-tress

"

χ | joint entry x )

"

a temale: an actress temale: lioness, actress ( = a female actor) => note at GENDER χ s.o. who ... 'Actor in χ s. o. who ... χ a person who ... χ the singular usually χ a woman who ... χ a woman who... refers to a man, but some women who act prefer to be called 'actors' rather than 'actresses'. χ Jill Philipps, author χ Jettrey Archer, χ Dickens is my tavourof the book ... successful author ite author. and politician ... She is the author of... It's clear that the author is a woman

author

\ Barry Black is the author She's the author

-ess

χ Some female writers χ a woman who ... object to this word, preferring to be called authors. χ a person who χ s.o. who ... χ s. o. who ...

doctor -toress/-tress host -ess

χ

hosljemale also hostess

χ a person who ... χ a woman who ...

χ (dated) a woman author χ a person who ...

χ s. o. who ... χ a woman who ...

χ a person who ... χ a woman who...

χ s. o. who ... χ old-fashioned a woman who ... χ 1-3 a man who ... 4 r£s. o. who ...

χ a person who ... χ a female murderer

X

murderer -deress

murderer, female χ s. o. who ... dated also mur- χ a woman who ... deress steward steward (female χ 1 a man who... las also stewardess) opposite to -ess] On an aircraft 2 s.o. who ... such a person is 3 a man or woman now more usually who ... called a flight attendant. -ess χ a woman who ... flight attendant s. v. flight X waiter waiter male, χ a man who ... female waitress -tress χ a woman who ...

χ

I (jem stewardess)

\ joint entry 1 a person who ... χ a woman who ... χ s. o. who ... χ a man who ... χ a woman who ...

χ ) χ

| joint entry (Jem waitress) χ ι a person who ...

Ulrich Busse

196

Gender (VI): Morphology, suffixation, combining forms C1 E

b) -ette

C BUIL

2

AL 5 LDUCE 3 a woman who is doing a female: usherette particular job: an usher- => note at GENDER ette ( = female USHER) χ an officer ... χ an officer who ...

χ an officer who ... purser · X -ette χ 1-2 s. o. who ... χ a person who ... usher usher: a person who ... χ a person who ... χ esp ErE a woman χ a girl or woman who ... -ette usherette female, male χ a woman who ... , an old-fashioned who ... separate entries usher word c) -ine χ 1 a person who ... hero, jemale also hero- χ 1 male character... ; χ 1-2 a man who ... hero 2 someone, especially 2 chief male character ine: a person who ... 3 s. o. who ... a man ... X χ woman or girl ... -ine χ [cross-reference] d) -trix aviator χ dated Amy Johnson χ an old-fashioned χ old-fashioned χ (dated) a person who was a pioneering word aviator -trix X educator χ esp. Am. χ a person who ... used χ formal especially AmE a teacher mainly in formal -trix - British English e) -girl adult women consider 3 Young women are a word meaning woman, 2 (a) a young woman, usu it offensive to be cal- often referred to as girls. which is sometimes one who is not married considered offensive by (b) a woman of the speciled girls by other peo- Some people find this ple, esp. men, although use offensive. women: the office girls fied type: (infmf) Old girl', 'new girl' this was common in the past, but they 3 (usu in compounds) a female worker might call themselves or their friends girls => Sexist language t) lady- (dated) lady is somelady doctor/councillor times added to the etc ( = a polite word, names of particular which many women jobs: find offensive, for a a lady doctor woman doctor, councillor, etc) combining Jorm => combines with numbers (iormmg compound «5) g) - woman Sexist language to indicate that s. th. Compare -MAN involves the number of women mentioned Woman also means t) woman- female: She is Ireland's first woman ( = female) president => Sexist language

Recent English Learners' Dictionaries...

197

Sexual Orientation (I) CIDE

CÜBULLD2 Ciays are homosexual people, especially homosexual men. tag infml esp. Am slang In informal American a homosexual man English, a fag is a homosexual; an offensive use faggot infml fag esp. Am In very informal Amerislang. This word is can English, a faggot is considered offensive a homosexual man; an when it is used by offensive use people who are not homosexual. -queen slang esp. disapa male homosexual ... , an informal use proving a homosexual man

gay

fairy

infml

infml disapproving a homosexual man. Many people find this word offensive.

LDOCE 3 OALD5 a gay ( 1 ) person someone who is HOMOSEXUAL, especially a man 2 AmL slang an (esp US) = FAUÜOT 3 offensive word for a HOMOSEXUAL 2 AmL slang an of- 3 (also esp US fag) fensive word for a (dated infml offensive) a male HOMOSEXUAL HOMOSEXUAL

informal an insulting word for a male HOMOSEXUAL, especially one who behaves like a woman It someone describes a an offensive word man as a fairy, they for a HOMOSEXmean that he is a homo- UAL man sexual and they disapprove of this; an offensive use. no usage label no usage label an informal use which is 2 an offensive word considered offensive for a LESBIAN

lesbian no usage label infml les slang a homosexual dike/ woman. Many peodyke ple consider this word offensive. butch no usage note It you describe a woman informal as a butch, you mean that you think she behaves or dresses in a masculine way; an offensive use. kinky infml It you describe someinformal thing, usually a sexual practice or preference, a kinky, you mean that it is unusual and would be considered strange by most people; an informa use.

5 (si derog) a male homosexual who behaves in a feminine way

2 (si derog) a male HOMOSEXUAL

no usage label (si usu offensive) a LESBIAN I (often derog) (of a woman) having a MASCULINE appearance and way of behaving (infml usu derog) involving or suggesting unusual sexual behaviour

Ulrich Busse

198 Sexual Orientation (II) tramp

CIDE esp. Am disapproving

COBUILD 2 an offensive word used mainly in American English

tag hag

no usage label

"

queer

straight

LDOCE 3 3 old-Jashioned especially AmE

slang especially AmE an offensive word slang. Queer is some A man who is queer informal ... , contimes used disapis homosexual; an sidered offensive provingly... . Queer informal use which when used by people is also used in a way some people find who are not hothat is not offensive, offensive. mosexual esp. by homosexuals. in/ml not homoan informal use slang HETEROsexual SEXUAL

OALD5

" 2 (si derog offensive) (esp. of men) HOMOSEXUAL

9 (b) (injml) not HOMOSEXUAL

Recent English Learners' Dictionaries.

199

Physical/Mental Achievement (I) ableism cripple

C1DE dated

disabled no usage label handicapped

dated

(-) challenged

demented in/ml or dated mentally ill

deranged no usage label dated or in/ml insane mentally ill lunatic

(dated, now taboo)

spastic

taboo (a person who is) suffering from cerebral palsy dated

retarded

LDOCE 3 COBUILD 2 1 a use which some people 1 an offensive word find offensive compare CRIPPLE, people who are disabled are sometimes referred to HANDICAPPED no usage note as the disabled. You can refer to people no usage label who are handicapped as the compare DISABLED handicapped It you say that someone is visually/physically/ challenged in a particular mentally challenged way, you mean that they AmE an expression for have a disability in that describing someone area. Challenged is often who has difficulty combined with inappropri- doing things because ate words for humorous of blindness etc, used effect: when you want to 'vertically challenged' avoid saying this di'intellectually challenged' rectly 1 ... a severe mental illness, 1 'crazy' no usage label especially Alzheimer's disease; an old-fashioned 2 'insane' oldor medical term. fashioned suffering 2 ... strange, foolish uncon- from a mental illness. trolled; an informal use no usage label no usage label 1 'mentally ill' no usage I 'stupid, crazy' inlabel formal 2 'very foolish' => Prag- 2 'mentally ill' espematics cially law 4 People who were men1 'crazy, stupid' no tally ill used to be called usage label lunatics; an old-fashioned 2 'mentally ill' olduse which some people fashioned find offensive [for the other meanings => PRAGMATICS] Most people now reter to 1 'stupid' slang someone with this disabil- 2 old-fashioned havity as having cerebral ing CEREBRAL palsy. PALSY an old-fashioned use no usage label

ÜALD5

• (sometimes offensive) Compare DISABLED Compare CRIPPLED (sometimes

offensive)

(a) mad (b) extremely upset no usage notes

no usage label 1 'not sane' no usage label 2 (infml) 'very foolish' 'wildly foolish' n 1 a wildly foolish person 2 (dated now offensive) mentally ill person' (medical) suffering from cerebral palsy no usage label

Ulrich Busse

200 Physical/Mental Achievement (II) COBUILD 2 C1ÜE dated for learn- ~ ing disabilities -; learning diffi- ~ culties dated a mental mental hospital no usage label hospital tunny farm humorous mental handicap learning disability asylum

lunatic asylum

(dated, now taboo)

loony bin

(taboo slang humorous) dated or in/ml mental hospital for psychiatric hospital psychiatric no usage label hospital

LÜÜCE 3 mentally handicapped no usage label no usage label

OALD5 ~

~

2 old use a MENTAL 2 (dated) a lunatic asyHOSPITAL lum informal humorous an expression meaning a PSYCHIATRIC HOSPITAL that some people consider offensive a place where men- old-fashioned (dated) ... a mental tally disturbed peohospital ple used to be locked up no usage label ™ humorous (si joe offensive] no usage note

PSYCHIATRIC HOSPITAL

-; only as an example s. v. mental

no usage label

-; psychiatric clinic as an example s. v. psychiatric

Ageism (I)

C1DE ag(e)ism Am and Aus usually agism treating people unfairly because of their age

COBU1LÜ 2 the behaviour that occurs as a result of the belief that older people are of less value than younger people; used showing disapproval elderly (used to avoid saying) 1 you use elderly as a old => Age polite way of saying that someone is old geriatric I specialized ot or for 1 a medical use old people 3 If you describe some2 (disapproving) too one as geriatric you are old and weak being disrespectful ...=> PRAGMATICS senile no usage label no usage label granny-

no usage label

LDOCb 3 BrE unfair treatment of people because they are old; AGE DISCRIMINATION see OLD (USAGE)

OALU5 (derog) the practice of treating people unfairly because of their age, esp because they are considered too old (often euph) => note at OLD

1 (medicalj [medical] 2 informal too old no usage label

not technical men- no usage label tally confused BrEoi a style typically used by old women

Recent English Learners' Dictionaries...

201

Ageism (II) (J1UE

OAF

Br abbreviation for old age pensioner Am => Age

LÜÜCE 3 (JUBU1LÜ2 British English = sen- BrE Old Age Penior citizen sioner

used mainly in Ameri- ~A~inE can English old people's no usage label => no usage label no usage label Age home " " old folks' injormal home old bat an intormal offensive expression => PRAGMATICS (dated) old boy informal British Eng- BrE spoken an old lish disrespectful man old butter (infml) old girl (esp. Br in/ml) informal British Eng- spoken an old woman lish, disrespectful infml an old per- miormal English; oldie injormal someone or son affectionately something that is old old maid (dated esp. disap- an offensive expres- 1 an offensive exproving) sion pression old-timer (infml) an old 1 'veteran' informal 1 old hand man, or someone use 2 especially AmE an who has been or 2 Old man' informal old man worked in a place American English = for a long time old man senior citizen (Am also senior, = pensioner an old person.... see Br and Aus also OLD (USAGE) Most old age pensioner) people however now ...is used to avoid prefer to be called saying Old persenior citizens, and son' this is the most common, polite, and acceptable expression to use. retiree

UALÜ5 (Brit) old-age pensioner ~

(also old folks' home) (infml) (infml)

(infml) an old person (infml) an old person (infml) old person or thing (derog) old hand no usage label

(euph) an old or retired person

Bernhard Diensberg

Towards a Revision of the Etymologies in the Oxford English Dictionary (OED 3rd edition, in progress)

1. Prologue Nearly six years ago I read a paper entitled "Towards a revision of the Oxford Dictionary of English Etymology (ODEE)" at the Sixth International Symposium on Lexicography May 7-9, 1992, at the University of Copenhagen (Diensberg 1994: 207-233). Now that the revision of the parent dictionary (OED2) is well under way a few remarks on the aims and methods may be called for. Since the OED etymologies regarding Old French loanwords basically reflect 19th century scholarship, a thorough revision of the respective entries is an absolute necessity. - After the completion of this huge task (or even earlier) the revision of the ODEE will be commissioned with T.F. Hoad as its most likely editor. More than a decade ago the CODEE (1986) was published under the editorship of Hoad. Apart from the OED2 and the ODEE, reference will be made to the BDE and the recently published KDEE.

2. Introduction Two phonological systems in contact - conclusions regarding the phonological reception/integration of Anglo-French and Old French word material into Middle English.

a) Middle English syllable structure: It is well known that French was in its earlier stages (and still is nowadays) what may be called an accent/stress-timed language which duly reflected the accent/stress pattern of the Latin mother tongue (as do all the other Romance languages and dialects). That means that vowel quantity essentially depended on this factor: accented vowels being longer than the ones outside the main stress (e.g. pretonic, countertonic vowels). The only exception to that rule was Old French e, from Latin ä in open syllables (not followed by a nasal consonant), e.g. OF pre -

—

^

··—

»—

*- -r-

*-

-—

·τ- »—

,

>

r- ,

»—

*—

·—

-r-

ϊ^ «D ϊ^ «^

OOOOOCM OO CD OO CM

ls (O

CO

^

SCM

^

^-

^-^^-^-^

^ - ^ ,

.-

^^^

^ ^ .

^

^

.

^-^^

-

^CD^CM

-

CD Γ~- CM

'•S

-8 tr> CVJ

S O ° -

^

—C M c o - w i n i o r - o o c r > 2 ^ 2 ; i 2 i i ^ i 2 C : i 2 ^ S c M f y c M i M C M

248

Lars Holm

CO OJ

Ο*) ^> τ- 10

»— 53

LTt

co o co

8 8

8

ο ο ο ··— r-.

53 β

CO

^O

OJ

" co ^ co

O

U> CM

- CM

"S g>

C» CM ·.—

e\j u> ^ CM

55 ~ S 3S

^r o ·« τ- ^ *

«a ,_ ^^. «» co «o -is «a

ssfllttltt

·

·»— oj co -^r ir>

249

The bishop who strove for completeness...

1/1

CO CO CM

58*

CM CM O

^ cS o>

j5 ^O

tn *τ ™

ΟΪ CO

PL »— 10

O

^ ^— CM

ef-\

j: *— S ^ r-*· ^ Lf>

CO ^ S

i^s

& ._

26 16 68 88 98 98 98 co

cn

Si cn

25 words in 100 dictionaries - in chart form

57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 81

σ σ»

I

Mi

-*«-»·

•ι

—ι —ι —ι — 3g 3ΓΤ

OZ 89 89

j

-^

99

S

09

ι

99

sa Ξ

-ss SB s

mm P

•1

Ι·

·

' · · · ·

^•is

TS; S3 ^3 s5

™ΤΓ1

•i •i

3 X n

ΞΞ

•u H

53 £,°"

o

tn

spJOM jo jaqiunN

o

ZS

es es es ZS OS

a

ssκ -^!^?ί

B aa asLs Ot

Ί

^ in

09

Lf SE =3 6C •13E 3E Si ^ψ 9G •i •i •i SMS K

κ

α c

^

^

•i

i.Ji

0)

n «n

•

· ·

ι i. •1

-1 1

09

TC3 ZSS3

1 +

LZ

J

•six sgs

£

ι

£ /

-Q

cz >

·· J

•i

O

1 t 1 53=· pJΞΞΞΞ ^Μ,,ρ.ι.!....*^^.!^,.

---

•I •i •i

|

28 29 08 08 U LL Sl frZ

•il.i-Kh

1

ij

in

H-

H

o

C2 6LI

1 He claimed thai his parents /Kid abandoned htm... (VB-peri picl The mad ix strewn with abandoned vehicles. ANTONYM: stay with verbum = opgive; skrinlaegge • If you abandon an activity or piece of work, you stop doing it before it is finished. D (VB nt The authorities have abandoned nur attempt l distribute food in an orderly fashion... (VB n) The scheme's investors, fearful of bankrupted, decided to abandon the project. SYNONYM, guv up ANTONYM: amunue verbum = fravige; opgive • It you abandon an idea or way of thinking:, you stop having that idea or thinking in that way. D (VB n) Logic had prevailed und he hail abandoned the idea. SYNONYM: give up verbum • hengive sig til; käste sig ud i • If you abandon yourself to on emotion, you think about u a lot and feel it strongly. D (VB refl pron HI ni We are scared lo abandon ourselves to our feelings in case M f seem weak or out of control. abandoned adjektw = lossluppen; haemningslos Substantiv, ulxlleligt: (sstdv w«n Ni = lossluppenhed; ubekymrethed • If you say that someone does something with abandon, you mean that they behave in a wild, uncontrolled way and do not think or care about how they should behave. He has splashed dollars around with gay abandon... Their permissiveness toward their children reflects the wild abandon of their own lives. NB: with abandon = lessluppent; ubehersket; hiemningslest BRUC: mishilligende Se ogsä abandoned. fräse: (vb bej.) = forlade et synkende skib • If people abandon ship, they get off a ship because u ts sinking. ü The officers and crew prepared in abandon ship in an orderly fashion.

256

Hanne Lauridsen/Arne Zettersten

One example of a word with new meanings within the areas of electronics and business, benchmark, will also be given here: benchmark ·«"»«=»* [bentjmarlc] sb: benchmarks; vb: benchmarks, benchmarked, benchmarks. BEMERK: Staves ogsä bench mark. l Substantiv, tjelleligt: (sadv sg; ofte N for n> = norm; referencepunkt; standard; kriterium; sammenligningsgrundlag • A benchmark is something whose quality, quantity, or capability is known and which can therefore be used as a standard with which other things can be compared. The truck industry is a benchmark for the economy. NB: benchmark opr = referencepunkt i landmäling: hejdefikspunkt SYNONYM: yardstick ~L. Substantiv, tslleligt = [standardised program til testning og sammenligning af computere]; testprogram • A benchmark is a standard program used to compare different makes of computer. BRLIG: fagudiryk inden for edb

3. verbum = [analysere konkurrerende virksomhed-

er] • If you benchmark a company's performance, you evaluate it by comparison with a standard. BRUC: iseer amerihunsk engelsk: fagudiryk inden far forreiningsverdenen 4. verbum = teste; afpreve • If you benchmark a computer system, you evaluate or test il using a benchmark or standard program. BRUC: fagudtryk inden for edb

First of all we should like to point out that all definitions are in English. One reason for keeping the COBUILD definitions in English in our new English-Danish Dictionary is that the level of English proficiency in Denmark is generally rather high, especially for understanding typical COBUILD-definitions which are usually composed in a defining style that is rather easy to follow. The defining vocabulary is normally about 2,500 words. Another reason for keeping the definitions in English is the above-mentioned fact that important collocations can be studied by the user both in the whole-sentence definitions and in the complete examples, which are usually two or three in number. Thirdly, since a complete set of Danish equivalents will be given as well as a detailed analysis in Danish of the syntactic construction of about 100,000 examples, we believe that maximum information on all Danish equivalents of all English entries will be obtained. This will be even more obvious from the fact that ample indications of style, usage and pragmatics will be included in Danish. Bearing in mind that some typologists discuss dictionaries in terms of'active' and 'passive' ones, meaning that the former are 'encoding' (writing and speaking) and the latter 'decoding' (reading and listening), our new bilingual English-Danish Dictionary is certainly more 'active' than most standard bilingual dictionaries. The present project in Denmark falls in the category that Henri Bejoint in his Tradition and Innovation in Modern Dictionaries calls "monolingual-bilingual dictionaries" (p. 87).

The New Politiken English-Danish Dictionary...

257

"Among the dictionaries for foreign students, one interesting recent development is the monolingual-bilingual dictionary with its source-language definitions and its equivalents in the target language added to each meaning. These dictionaries are important, because they og beyond the traditional distinction between monolingual and bilingual dictionaries." In the English monolingual COBUILD 1995, the grammatical notations and the information on synonyms, antonyms and pragmatics were placed in an extra column in the margins. We have kept all such information in the articles under each headword but translated all the 100,000 grammatical codes into Danish, so that all the 100,000 grammatical patterns are explained in Danish under each meaning or submeaning of a word. This was quite a gigantic enterprise but it certainly provided a rather full practical grammar of authentic written and spoken English within the framework of a bilingual dictionary. Some examples of the codes will be given below. DET DET-NEG DET-POSS DET-POSS-QUEST DET-QUEST EXCLAM FRACTION MODAL N-COUNT N-COUNT-COLL N-FAMILY N-IN-NAMES N-MASS N-PLURAL N-PROPER N-PROPER-COLL N-PROPER PLURAL N-SING N-SING-COLL N-TITLE N-UNCOUNT N-UNCOUNT-COLL N-VAR N-VAR-COLL N-VOC

DETERMINATIV NEGATIVT DETERMINATIV POSSESSIVT DETERMINATIV POSSESSIVT, SP0RGENDE DETERMINATIV SP0RGENDE DETERMINATIV UDRÄBSORD BR0K MODALVERBUM SUBSTANTIV, T^LLELIGT SUBSTANTIV, T^ELLELIGT, KOLLEKTIVT SUBSTANTIV, FAMILIEMEDLEM SUBSTANTIV l NAVNE SUBSTANTIV, MASSEORD SUBSTANTIV, PLURALIS PROPRIUM PROPRIUM, KOLLEKTIVT PROPRIUM, PLURALIS SUBSTANTIV, SINGULARIS SUBSTANTIV, SINGULARIS, KOLLEKTIVT SUBSTANTIV. TITEL SUBSTANTIV, UT^ELLELIGT SUBSTANTIV, UT/ELLELIGT, KOLLEKTIVT SUBSTANTIV, T^LLELIGT/UT^LLELIGT SUBSTANTIV, T^LLELIGT/UT/ELLELIGT, KOLLEKTIVT SUBSTANTIV, TILTALEORD

With such wealth of information on definitions, meanings, examples, collocations, grammar patterns, synonyms, antonyms and pragmatics, the user will be in a good position to select the most satisfactory equivalent in Danish. In addition, we trust that the user's knowledge of English for productive purposes will provided he or she reads and contemplates all the information given under each entry - increase considerably. Finally, we definitely claim that our wish to reach maximum information on all aspects of the choice of equivalents, is in agreement with recent discussions of linguistic awareness. We therefore claim that the only way for a foreign learner of reaching absolute linguistic awareness, namely exact knowledge of the precise meanings of an English word, is by using a bilingual dictionary produced along the lines indicated above.

258

Hanne Lauridsen/Arne Zetters ten

We further claim that it is very important in a bilingual dictionary like this - which will be used by the Danish user to decode, or translate English texts as -well as to produce texts in English - to give detailed and clear information about style, pragmatics, illocutionary meanings, varieties of English. It is notoriously difficult - even to advanced learners of foreign languages - to recognize, and reproduce stylistic differences, to use the correct pragmatic strategies and so on. In order to understand and use a foreign language effectively it is necessary to know what the pragmatic conventions are for that particular language, and to know about stylistic variation. Moreover - and this is particularly important, when the foreign language in question is English - it is relevant to a foreign learner to know about geographical varieties of the language so that he or she will be able to recognize i.e. an American English term versus a British English term. We decided that this information should be placed in a conspicuous way in the entry (and not just be part of a definition, often indicated in a rather indirect way). We have carefully studied the very detailed definitions (given in full English sentences) in Collins Cobuild English Dictionary and extracted all information on style, pragmatics, illocutionary meanings, varieties of English found here and adapted it for our purpose (e.g. adapted it for a bilingual dictionary, adding information in some cases, and - in very few cases - leaving out information). We have placed the results of this editing of information on style, pragmatics etc. in a special paragraph called "brug" (= use) at the end of every entry. In order to arrange this paragraph "Brug" in a systematic way, thus systematizing the various types of information found here, we have arranged the paragraph as follows with four possible categories to be filled in, which will always appear in the same order: Category (1) Relevant information about the variety of English which the entry in question represents. (examples: amerikansk engelsk amerikansk engelsk eller australsk engelsk britisk engelsk isaer britisk engelsk) Category (2) Relevant information aboutßelds of technical terms. (examples: fagudtryk inden for sprogvidenskab (= a technical term within linguistics) fagudtryk inden for samfundsvidenskab (= sociology) fagudtryk inden for 0konomi (= economics) fagudtryk inden for medicin (= medicine) fagudtryk inden for sofart (= shipping)) Category (3) Relevant information about style (examples: skriftsprog talesprog formel uformel litterasr gammeldags (= old fashioned) journalistisk)

The New Politiken English-Danish Dictionary...

259

Category (4) Relevant information about pragmatics. (examples: nedsaettende (= derogatory) fremhaevende ( = emphasizing) rosende ( = complimentary) misbilligende ( = critical, disapproving) stedende ( = offensive) billigende ( = approving))

This paragraph ("brug") may thus contain information about maximum 4 categories, normally, however, only one or two of these categories are relevant. On the other hand, in quite a number of cases a category is represented by more than one piece of information, as the following examples will demonstrate. Examples: (categories 1 & 2): amerikansk engelsk; fagudtryk inden for medicin (1 & 3 & 4 & 4): britisk engelsk; gammeldags; fremhaevende; st0dende (3 & 4 & 4): formel; fremhaevende; misbilligende (3 & 3 & 4): talesprog; formel; h0flig

In the following we shall give some examples of headwords from the dictionary and the information given under "brug"; Headword: boy racer

Original definition: British journalists sometimes refer to young men who drive very fast, especially in expensive and powerful cars, as boy racers; used showing disapproval. We leave out "used showing disapproval", and add (at the end of the entry): Brug: britisk engelsk; journalistisk; misbilligende Headword: wannabe

Original definition: If you call someone a wannabe, you are saying in an unkind way that they are trying very hard to be like another person or group of people; an informal word. We leave out "an informal word" and add (at the end of the entry): Brug: uformel; misbilligende Headword: wandering

Original definition: Wandering is used to describe people who travel around rather than staying in one place for a long time; a literary use. We leave out "a literary use" and add (at the end of the entry): Brug: litteraer Headword: (6) wind up

Original definition: If you wind someone up, you say untrue things in order to trick them; used in informal British English.

260

Hanne Lauridsen/Arne Zettersten

We leave out "Used in informal British English" and add (at the end of the entry): Brug: britisk engelsk; uformel Headword: (2) coda Original definition: In music, a coda is the final part of a fairly long piece of music, which is added in order to finish it off in a pleasant way. We leave out "in music" and add (at the end of the entry): Brug: fagudtryk inden for musik Headword: (2) topside

Original definition: On a ship, if you go topside, you og up onto the top deck; a technical use. We leave out "A technical use" and add (at the end of the entry): Brug: fagudtryk inden for sefart Headword: (I) corpus

Original definition: A corpus is a large collection of written or spoken texts that is used for language research; a technical use. We leave out "a technical use" and add (at the end of the entry): Brug: fagudtryk inden for sprogvidenskab Headword: (2) compiler

Original definition: A compiler is a computer program which converts language that people can use into a code that the computer can understand; a technical use. We leave out "technical use" and add (at the end of the entry): Brug: fagudtryk inden for edb

Finally we would like to display the elements of an entry: POLITIKENS ENGELSK/DANSK med betydningsforklaringer. (1) Headword (from Cobuild + additions) (2) Frequency (3) Pronunciation (4) Inflections (5) Grammar (in Danish) (6) Danish equivalents (7) Definitions (from Cobuild, but adapted, see (12 below) and with extra information on etymology) (8) Examples (from Cobuild, but: grammar in Danish) (9) NB (extra paragraph for special equivalents, idioms etc.) (10) Synonyms (11) Antonyms (12) Use (the paragraph, discussed in detail earlier in this paper)

The New Politiken English-Danish Dictionary...

261

The italic elements: (1), (5), (6), (7), (8), (9) and (12) are either new (6, 9, 12) or edited, adapted, translated or expanded by us. Illustration from the final product: abandon •••4-4 fctaendan] vb: abandons. abandoned, abandoning. 1. verbum

= forlade; lade i stikken; lebe fra; efterlade; svigte • If you abandon a place, thing, or person, you leave the place, thing, or person permanently or for 3 long time, especially when you should not do so. O (VB n) Hf claimed thai Ins parents Imil abandoned hint... (VB-perf pic) Tlie mud is siren η \vith abandoned vehicles. ANTONYM: stay with 1. verbum

= opgive; skrinlaegge • If you abandon an aaivity or piece of work, you slop doing it before it is finished. Q (VB n) The authorities have abandoned tiny attempt In distribute food in an iirderlv fashion... (VB n) Tlie scheme's investors, fearful of btinkniplrv. derided in abandon the pmject. SYNONYM: give up ANTONYM: continue 3. verbum = fravige; opgive • If you abandon an idea or way of thinking. you stop having that idea or thinking in thai way. Ο (VB ni Logic had prevailed und lie had abandoned the idea. SYNONYM: give up 4. verbum = hengive sig til; k ste sig ud i • If you abandon yourself to an emotion, you think about it a lot and feel it strongly. O (VB red pron iu n) We are scared to abandon ourselves 10 our feelings in case we seem weak or out of conrroi

0 abandoned adjeknv

= lessluppen; haemningsles 5. Substantiv, uLcllelign (satdv miH N>

= lessluppenhed; ubekymrethed • If you say chat someone does something with abandon, you mean that they behave in a wild. uncontrolled way and do not think or care about how they should behave. Ο He has splashed dollars around with gay abandon... Tlieir permissiveness toward their children reflects the wild abandon of their own lives. NB: with abandon = lessluppent; ubehersket: hxmningslesl BRUC: mi billigende

6. Se ogs abandoned. 7 frase: anon untydras ealle onwocon eotenas and ylfe. clOOO Sax. Leechd. II. 296 wiö aelfe and wip uncupum fidsan gniö myrran on win. c!205 LAY. 19256 Sone swa he com an eoroe aluen hine iuengen. c!386 CHAUCER Man Lawes T. 656 The mooder was an elf by aventure. 1426 AUDELAY Poems 77 (Matz.) Alfe Rofyn be-gon to rug. c!460 Towneley Myst. (Matz.) He was takyn with an elfe. 1513 DOUGLAS ALmis viii. vi. 7 Wyth Nymphis and Favnis apoun euery syde, Quhik fairfolkis, or than elvis, clepyng we. 1579 E.K. in Spenser's Sheph. Cal. June 25 Gloss., For Guelfes and Gibelines, we say Elfes & Goblins. 1610 SHAKES Temp. v.i. 33 Ye Elves of hils, books, standing lakes and groues. 1635 HERRICK Hesper. (1869) II. App. 477 Come follow, follow me You fairie elves that be. 1700 DRYDEN Wife Bath's T. 3 The King of elfs. Gamboll'd on heaths. 1712-4 POPE Rape Lock i. 33 Airy elves by moonlight shadow seen. 1866 KINGSLEY Herew. xv. 193 You are an elf and a goddess. 1875 B. TAYLOR Faust i. i, Then the craft of elves propitious Hastes to help where help it can. t 'eten, 'ettin. Obs. Forms: 1 eoten, eten, 3 eatand(e, -ante, eotand, -end, -ind, Seten, 4 eten(e, -in, yhoten, 4-5 etayn(e, 6 Sc. eitin, etin, eyttyn, 7 ettin. [OE. eoten, eten = ON. iojunn (Sw. Jätte, Da.ye/fe):-Oteut. *ituno-z.] A giant. c!205 LAY. 1801 Heo funden i bon londe twenti eotandes [c!275 eatantes] stronge. Ibid. 17275 Hit hatte pere Eotinde King. c!250 Gen. & Ex. 545 Of hem woren oe aetenes borne, Miati men, and fisti. a!300 E.E. Psalter xviii. 6 [xix. 5] He gladed als yhoten to renne his wai. c!325 Leg. Rood (1871) 118 Quen dauid fast againe pat etin has he noat his staf for-3etin. c!340 Gaw. & Gr. Knt. 723 He werrez. Bobe wyth bullez & berez, & borez oper-quyle, & etaynez. c!380 WYCLIF Serm. Sei. Wks. II. I l l No man is an etene to fede him pus bodili of Crist. c!440 MS. Lincoln A. 1. 17.f. 128 (Halliw.) Fy, he said, thou foule! thou etayne! Alle my knyghtes thou garte be slayne. 1549 Compl Scot. 63 The taiyl of the reyde eythyn vitht the thre heydis 1611 BEAUM. & FL. Knt. Burning Pestle i. ii, They say the King of Portugal cannot sit at his meate but the Giants and the Ettins will come and snatch it from him.

t 'etenish. Obs Forms: I eoten-, etonisc, 3 aeteniss. [f. eten + isc. -ish.] Gigantic. alOOO Beowulf 1559 He..^eseah on sear-vum si?;e eadi^ bit eald sweord etonisc. Ibid. 2616 And his majum atbaer brun fa^ne helm hringde byrnan eald sweord etonisc. c!250 Gen. & Ex. 3715 3etenisse men ben in ebron. orken, an adaptation of *orcen assumed as sing, of orcneas in Beowulf, (?) monsters or seamonsters. alOOO Beowulf 112 Eotenas and ylfe and orcneas. 1855 THORPE tr. Beowulf 225 Thence monstrous births all sprang forth, eotens, and elves, and orkens. Here we see two problems at once: that Thorpe borrowed and transplanted the Old English noun orkneas into his translation, which fact does not prove the continuity of its usage from Old English, and that the quotation from Thorpe contains the noun eotens as well, which, if quoted by Bradley, would have changed the status symbol of the headword eten. In order to achieve some unity in the treatment of Old English words, especially those found in the same source, we could either add the two citations for orken to the headword eten as the

Quotations from Beowulf and Other Old English Poems...

267

first and the last examples, removing the dagger and the label Obsolete, or perhaps change the status of the headword orken to Obsolete. But if they had consistently included archaism of certain renowned translators as part of the current usage of a certain word, many headwords with the label Obsolete might have been labeled current and it would have produced a totally different picture of the English Language. William Morris, fifty years later than Thorpe, translated the poem and published a gorgeous book called The Tale of Beowulf. This book is full of Old English words transplanted from the poem and among hundreds of them the OED cites only nine instances, the selection of which seems quite arbitrary.9 The editors, of course, did not generally collect examples themselves but worked on the quotation slips sent by voluntary readers, who have sometimes been "unreliable". James Murray, had he read Morris' translation himself, would undoubtedly have quoted, for example, the adjective brim for the headword breme, which he was working on, and accordingly, he would definitely have made the adjective current as he did with the headword orken. But the present labels to the word brim, Obsolete except Poetical and Dialectal, seem safer and better grounded in linguistic facts. This kind of sudden revival of an old word which has long been obsolete does not prove its survival or continued use through the thousand years after its first appearance. Preference by the editors for contemporary works, which is repeatedly pointed out by Professor Osselton and Professor Willinsky, has resulted in considerable unevenness and chronological imbalance in quotation columns. One of those is what I would like to call the hollowing out of examples from intermediate periods between Old or Early Middle English and the nineteenth century. This is clearly seen in the headwords ash n4 and scop. In the quotation column of ash the two examples from Old English are followed by a number of instances from the nineteenth and the twentieth centuries. And for the headword scop three examples from Old English and Early Middle English are followed by thirteen examples taken from the periods not earlier than 1774. With these modern instances the two headwords stand as current, without a dagger symbol meaning obsolete. But these instances are all from grammarians or philologists and they are direct quotations of Old English nouns. I would also like you to look at the quotations gathered for the noun mere, which is supposed to mean sea or lake. Let me invite your attention to the second definition of the noun, 'a sheet of standing water', which stands as current without a dagger, with many modern quotations. But six out of eight modern quotations are taken from philologists translating or discussing Beowulf and a note put here defines that the noun mere is, as far as twentieth-century quotations go, always used to refer to Grendel's mere. Grendel is the chief enemy-monster in the poem. And the fourth, fifth, and sixth quotations are from

On pages 16 and 17 of Morris' translation there are fifteen compounds, which are coined by Morris, apparently after Old English counterparts: wan-heed, man-lord, hand-grip, war-hall, blood-mark, hand-work, hand-bane, war-dread, youth-tide, hoard-burg, hate-wiles, hall-floor, war-heap, ale-stoup, and beer-hall. The OED took hand-work and hoard-burg for the headwords and hate-wiles for a subordinate headword. It quoted the line 965 of Beowulf, where the compound hand-gripe appears, for the headwod handgrip but it did not pay any attention to the instance on page 16.

268

Hideki Watanabe

Middle English alliterative poems. So the label Poetical, appropriately put here, shows a restricted use of the noun in the tradition of alliterative poetry. The scholarly use and direct quotation of these nouns, despite their high frequency, does not qualify currency in present-day English in a general sense. It is comparable to the academic use of such Greek or Latin words as Thanatos or comitatus: the OED classifies most of them as "not naturalized" words.10 I have discussed the OED mainly with reference to quotations from Old English poems. I have noted various titles and their abbreviations, and different dates of composition given for a single poem, and double citations of a single item for different headwords. I also discussed different criteria for the evaluation of the currency of Anglo-Saxon words. I just touched on some problems caused by multiple editions of a single work searched for quotations when I discussed the quotations from The Owl and the Nightingale. Let me give you just one more example here. Everybody knows that the poem Beowulf has 3,182 lines, but this poem was once edited in the 3,184-line format by some editors who supposed that two lines were missing. And it used to be printed in the half-line format as against the present long-line format, with over six thousand lines, in the nineteenth century. Thus, there are, at least, three kinds of numbering system employed for Beowulf.11 We cannot make the best use of the dictionary's invaluable information, which is unobtainable anywhere else, until we fully and properly appreciate those inconsistencies. The OED, the largest printed English dictionary on historical principles, has now become an electronic corpus equipped with an extensive search menu. We have yet to explore the versatility of the information therein with well over a million quotations. In order to pick up the data most needed, we have to devise new retrieval techniques. This is a dictionary which deserves our untiring and systematic exploration, for it has always been the first and the last authority for the English Language.

Works Consulted Barnbrook, Geoff (1996): Language and Computers: A Practical introduction to the Computer Analysis of Language, Edinburgh University Press. (On the OED see p. 82; p. 134; p. 137; pp. 163-5.) Berg, Donna Lee (1993): A Guide to The Oxford English Dictionary, Oxford University Press.

10

11

|| comitatus [L., collective deriv. of comes, comit-em, companion, count.] 1. A body ofcomites or companions; a retinue of warriors or nobles attached to the person of a king or chieftain, b. The status or relationship of such a body to their chief. 1875 Stubbs Const. Hist. I. vi. 152 The development of the comitatus into a territorial nobility seems to be a feature peculiar to English History. Ibid. I. vi. 160 The Danish jarl..seems to have been more certainly connected by the tie of comitatus with his king than the Anglo-Saxon ealdorman. 1875 Maine Hist. Inst. v. 138 The first aristocracy springing from kingly favour consisted of the Comitatus or Companions of the King. Simpson notes; "The editions of works cited in the OED are often not those we would expect to use today." (op. cit. p.3)

Quotations from Beowulf and Other Old English Poems...

269

Chase, Collin (ed.) (1981): The Dating of Beowulf, University of Toronto Press. Girvan, Ritchie '(1971): Beowulf and the Seventh Century, reprint with a new chapter by Rupert Bruce-Mitford, Methuen. Murray, Elizabeth (1977): Caught in the Web of Words: James Murray and the Oxford English Dictionary, Yale University Press. Kasmann, Hans (1997): (Review Article) John Willinsky, "Empire of Words: The Reign of the OED". Anglia, 113, pp.217-20. Newton, Sam (1993): The Origins of Beowulf and the Pre-Viking Kingdom of East Anglia, D.S. Brewer. Osselton, Noel (1993): 'The Oxford English Dictionary, Second Edition, prepared by J.A. Simpson and E.S. Weiner.' (Review Article) Internationaljournal of Lexicography, Vol. 6, No. 2, pp. 12431. Simpson, John (1997): 'Lexicographical Work in Progress, OED3: progress details,' Euralex Newsletter, Internationaljournal of Lexicography, Vol. 10, No. 4. p.4. Watanabe, Hideki (1996): 'The Quotations from Otto Jespersen in OED2' in Sententiae: Festschrift for the Sixtieth Birthday of Prof Yoshitaka Mizutori, Hokuto Shobo, pp. 160-9. (In Japanese) — (1998) (Review Article): "Jane Roberts and Christian Kay with Lynne Grundy: A Thesaurus of Old English, King's College London Medieval Studies XI, 1995." Studies in English Literature. The English Literary Society of Japan, pp. 132-8. — (1998): The Quotations from Linguists and Philologists in OED2''. Lexicographica Series Maior, Max Niemeyer, pp. 299-306. Wilks, Y.A., Slator, B.M., and Guthrie, L.M. (eds.) (1996): Electric Words: Dictionaries, Computers, and Meanings. MIT Press. (On the OED see p.32; p.64; pp.77-8; pp. 101-2; p. 243.) Willinsky, John (1994): Empire of Words: The Reign of OED, Princeton University Press.

Andreas Gröger A Semantic Analysis of Old English Mental Verbs, Based on the Material of the Helsinki Corpus

0. Introduction This paper presents a summary of my MA thesis (Gröger 1997) submitted to the Faculty of Language and Literature of Bayreuth University in April 1997. In it I analyzed the lexical field of 15 so-called mental verbs in Old English, using an approach first developed by Edgar W. Schneider in his Habilitationsschrift (Schneider 1988a/b). Schneider's work dealt with the same lexical field in Present-Day English, which for obvious reasons is much larger and includes about 140 lexemes. His method has since been successfully applied by Anke Wildenhain (1995) to the field of verbs of suffering in Old, Middle and Early Modern English; thus her Ph.D. dissertation provided another model for my MA thesis (cf. also Wildenhain 1994).

1. Sources: Dictionaries and the Helsinki Corpus Any study in an older stage of the English language must rely on written texts as the basic linguistic evidence. The corpus chosen for my project was the diachronic part of the Helsinki Corpus of English Texts, more precisely its Old English sub-corpus (henceforth abbreviated as HCO).1 The HCO contains 413,250 Old English word-forms and is further divided into four parts with a time-span of about a century each (cf. Table I). For the purposes of the present study, however, I have used the three-file version of the Helsinki Corpus rather than its 11 -file version, so that diachronic variation within Old English was largely neglected. In view of the relative scarcity of Old English textual evidence (as compared to later stages of English) this seems a reasonable policy. KWIC concordances for the mental verbs were compiled with the help of the concordance program WordCruncher.

For information about its principles and content cf. Kytö (1991).

Andreas Gröger

272 Subperiod Old English 1 Old English 2 Old English 3 Old English 4 Old English Middle English Early Modern English HELSINKI CORPUS

Time Span Beginnings to 850

Word Forms

850 to 950 950 to 1050 1050 to 1150

92,050 251,630 67,380 413,250 608,570 551,000 1,572,820

Beginnings to 1150 1150 to 1500 1500 to 1700

Beginnings to 1700

2,190

Table 1: The Helsinki Corpus of English Texts Corpora provide a necessary complement to dictionaries, which usually are the first source of information for investigations of lexical semantics of an earlier stage of English, but do not provide a scholarly semantic analysis and often suffer from vagueness and/or circularity of the definitions. The lexicographic information was mainly taken from Bosworth-Toller-Campbell's Anglo-Saxon Dictionary with Supplement and Addenda (henceforth abbreviated as BT (S)). The Dictionary of Old English (DOE) has so far only been published up to the letter E; thus, I have used it for the verbs bepencan, gecnawan, cunnan and demon only. In addition, the Oxford English Dictionary and the Concise Anglo-Saxon Dictionary (Clark Hall 1960) were consulted.

2. Delimiting the Lexical Field of Old English Mental Verbs The term mental verbs refers to verbs denoting mental activities and states, such as thinking, believing, remembering, supposing etc. The group of verbs in question is characterized by two semantic roles that all verbs share, though they are not necessarily expressed syntactically: • the person who 'thinks' (henceforth referred to as THINKER (TH), nearly always a human being);

• that which is 'thought', often expressed in the form of a proposition (henceforth referred to as INFORMATION (INF)). A first set of candidates for analysis came from searching Clark Hall's (1960) Modern English equivalents for verbs included in the comprehensive lexical index of Schneider (1988b: 400-402). It soon became apparent that the considerable variety of Old English verbs formed with prefixes (particularly the notorious ge-]2) called for somehow arbitrary but necessary limitations in order to keep the number of lexemes down to a reasonable size. Moreover, pre-

The prefix, not the past participle marker (= ge-2).

A Semantic Analysis of Old English Mental Verbs...

273

fixed verbs are semantically transparent and can be disregarded in a study whose primary aim is to analyze semantics rather than word-formation (Schneider 1988b: 36). I therefore decided to include only those prefix verbs which show a considerable difference of meaning compared to the corresponding simplex verb. This is the case in gefbepencan (vs. simplex pencari). Gemunan and gelifan are almost always prefixed in Old English. Ge-/oncna\van were included as well, since although the simplex cnawan exists in Late Old English 3 , only the prefixed forms are encountered in the corpus. - The very frequent impersonal verb fiyncan, which requires that the THINKER be expressed as a dative object, was likewise excluded; its ModE dictionary equivalent seem is not mentioned in Schneider's index. Lexical fields typically have fuzzy boundaries; no cut-and-dried rules can be given for ascertaining which verbs do belong to the field of mental verbs, since so many activities of human beings involve the mind. The field under scrutiny can therefore be delimited only 'ex negative' by mentioning some types of marginally mental verbs, among them the following (cf. Schneider 1988b: 40 f.): • verbs of sensory perception (here some borderline cases occur, namely, sememes of Old English -cnawan and ongitan); • verbs primarily expressing emotions or having a strong emotive component (e.g. Old English hopian, hyhtan 'hope', hatian 'hate'); • verbs expressing social attitudes founded on opinions (positive - e.g. Old English trywan 'trust, confide in'; negative - e.g. Old English forgieman, ofergitan 'neglect'); • verbs primarily denoting physical actions, particularly speech acts (sememes of Old English smeagan meaning 'to deliberate' and 'to scrutinize').

Altogether, 15 Old English mental verbs remained to be analyzed, while six others were excluded on grounds of low frequency, following Leisi's (1985: 146) criterion that a lexeme should occur at least about 10 times in a corpus in order to permit meaningful semantic analysis4. The following Table 2 lists first the retained and then the excluded lexemes together with their frequencies in the HCO; an "x" in the dictionary columns indicates that a dictionary gives semantic and (more or less) syntactic information about them.

Cf. the DOE entry s.v. cnawan, which gives a total of 21 occurrences for the verb. One may nevertheless argue that they should better be considered Early Middle English instances of cnawen; Gutch (1979: 132, 232) considers Old English *cnawan a 'ghost word'. This limit of significance ensures that one does not spend too much time on describing lexemes for which there is little corpus evidence anyway, in which case one would have to reproduce the information already found in the dictionaries.

Andreas Gröger

274 No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Lexeme witan6 ongitan gelifan wenan cunnan pencan gepencan gemunan understandan oncnawan gecnawan smeagan teohhian hycgan7 bebencan

*1 *2 *3 *4 *5 *6

myntan undergitan (ge)myndgian beahtian deman eahtian

Frequency BT ca. 460 X

256 221 200 154 103 100 100 55 49 39 30 25 21 17 7 7 5 5 4 3

BTS5 -

X

-

DOE -

X

X

X

X

-

X

X

X

X

X X

-

-

X

X

X

X X

-

X

X

-

X

X

X

X

X

X

-

-

X

X

X

X

X

X

X

-

-

X

X

X

X

X

X

Table 2: List of Old English mental verbs

3. The Theoretical Framework 3.1. Semantic Analysis8 The semantic analysis was carried out within a modified feature semantics framework which allows for a more adequate treatment of phenomena of "variability, polysemy and fuzziness of word meaning" (cf. Schneider's book title). Prototype semantics appears to yield more insights when applied to concrete objects. (Lipka 1992 and other publications frequently recur to the time-honoured examples of birds, drinking vessels or furniture; but cf. Geeraerts' (1997) book on diachronic prototype semantics, published after the comple-

Only entries in BTS which contain additional Modern English equivalents, not mere additional quotations, are counted. Including the negated form, which is usually contracted: nytan (infinitive) < ne + witan, not (1st person singular present indicative) < ne + \vat, etc. Including the Late Old English variant hogian; hycgan and hogian are treated separately in BT(S), but there is no semantic reason for doing so. The abbreviations used in the semantic formulas are explained in the Appendix.

A Semantic Analysis of Old English Mental Verbs...

275

tion of my thesis.) Wierzbicka's (1972, 1996) approach has been heavily criticized for being very cumbersome, as it involves lengthy recursive paraphrases, and overly optimistic in assuming a remarkably small set of some 50 universal semantic primitives, among them the mental predicates THINK and KNOW. The fact, however, that THINK and KNOW occur in Wierzbicka's set of primitives, shows the starting-point for an analysis of mental verbs. The members of a lexical field must belong to the same word-class and have at least one semantic feature - or a seme, in Schneider's terminology - in common (Lipka 1992: 152). This shared seme is called the archi-seme (an archi-lexeme, or cover term for the field members, can exist). The Old English mental verbs all share one of the two archi-semes of the dimension PREDICATION TYPE, namely PREa [have stored (the INFORMATION) in brain] or PREb [process (the INFORMATION) in brain].9 THINK as the archi-lexeme of mental processes contains PREb, whereas KNOW (as well as ModE believe, suppose, ...) contains PREa. We can further distinguish several kinds of INFORMATION storage or processing, which are formally expressed by semes TYMa to TYMf of the dimension TYM 'Type of Mental Activity', for instance: • reception of an INFORMATION in the (short-term memory, i.e.) mind (= TYMc; cf. Modern English perceive, realize etc.), • storage of an INFORMATION in the (long-term) memory (= TYMb; cf. Modem English know, believe etc.), • creation of a hitherto non-existent mental picture of an INFORMATION (= TYMf; cf. Modern English imagine, contrive etc.).

Other possible distinctions which step by step help to structure the lexical field are, for instance: • the 'Temporal Situation of INFORMATION' (TSI). The INFORMATION can belong to the past (= TSIa; cf. Old English gemunan, Modern English remember, recall) or to the future (= TSIb, 'verbs of expectation and intention'; cf. Modern English plan, intend, expect etc.); • whether or not a 'GOAL [is] aspired by the THINKER' (GOA) - 1 do not distinguish between a consideration of a future action and the fixed intention to perform it, subsuming both under the seme GOAa; • the degree of truth subjectively assigned to a stored INFORMATION (semes of the dimension TSO).

Semes are established by examining and comparing with each other the occurrences of the verbs in their respective contexts, trying to find minimal pairs. The dictionary entries together with the respective chapters in Schneider (1988b) provide a basic semantic classification which must be tested against the concrete data in the corpus. Each lexeme has at least one, usually several sememe(s), i.e. individual meanings which stand in paradigmatic opposition to each other. Each sememe can formally be described as a concatenation of a finite number of semes, some of which are obligatory and (optionally) others which are not. These latter variable semes are a key feature of Schneider's approach and especially useful for coping with the semantic variability of parole; Lipka's (1992: 109

I prefer the term "brain" as a cover term embracing the mind (= short-term memory) and the memory proper (= long-term memory).

276

Andreas Gröger

ff.) inferential features are similar to them. - The sememes can sometimes be further subdivided, yielding SI ->S1.1 (SI.2 etc.)-» S 1.1.1 (S 1.1.2 etc.) and so on.

3.2. Syntactic Patterns, Collocations and Textual Distribution Any semantic analysis of verbs is incomplete without a description of their syntagmatic behaviour, i.e. their syntactic patterns as observed in the corpus as well as typical collocations. It is well-known that (especially verb) semantics and syntax are closely interrelated. The syntactic behaviour of all Old English mental verbs has therefore been investigated and frequencies of all observable constructions have been given, using the apparatus of the Longman Dictionary of Contemporary English (2nd edition, 1987).10 In several cases strong or even full correlations can be observed, i.e. it is possible to predict the semantics of a given syntactic pattern. Frequent realizations of the thematic roles of THINKER and INFORMATION were also noted. The INFORMATION can, for instance, be characterized by a thematic group (such as 'religious' or 'concrete object'); the semes of the dimension THE 'Thematic Kind' may occur in a semantic formula. - Finally, particular attention was paid to collocational tendencies and formulaic usage." The Helsinki Corpus broke new ground by making extensive use of sociolinguistic parameters in its coding conventions. Many of these are unfortunately not applicable to the Old English subcorpus; however, the textual distribution of lexemes in the corpus (in different genres) has been investigated.12

4. Analyzing the Individual Verbs: the Case of Pencan t>encan occurs 106 times in the HCO. Only 3 occurrences are outside the lexical field of mental verbs. Comparing the semantic structure of pencan with that of its Modern English continuation think, we can first note that the overall frequency of the Opinion' use of pencan (= my sememe S3, cf. example 6) is only about 15%, much lower than in Modern English. This is 10

Schneider delved deeper into the matter - his investigations draw on a modified Fillmorean 'case grammar' framework and employ a complex notation which is very suitable for expressing the close connections between semantics and syntax. However, Old English syntax is much freer than Modern English syntax, and most investigations of Old English syntactic phenomena use a traditional terminology anyway; the LDCE notation serves its purpose quite well, is easy to understand, and was also employed by Wildenhain (1995). " Cf. Koskenniemi (1968); the DOE is the first Old English dictionary to treat collocations thoroughly and systematically. 12 The relevant HCO parameters are encan is a

A Semantic Analysis of Old English Mental Verbs...

279

syntactically and semantically flexible lexeme (unlike its rival wenan, whose inflexibility and relatively narrow semantic range may have joined forces with the Early Modern English ween/wean homophony to speed up its eventual decline). The verb occurs in all textual categories of the HCO, though most frequently in poetic texts (labelled "XX XX", a wastepaper basket category).

5. Structuring the Lexical Field

After analyzing the individual mental verbs, a synthesis was presented in the form of an overview field diagram and five quantitative diagrams devoted to one sub-field (or lexical paradigm) each. However, it should be borne in mind that graphic representations of lexical fields always simplify the complex semantic interrelationships between their members to a certain extent and may suffer from cross-classifications. Table 4 shows the basic structure of the field as a whole. The dimensions DYN 'Dynamicity', TYM 'Type of Mental Activity' - and several others referring to the difference between objective knowledge and subjective opinion, as well as the notion of 'understanding' an INFORMATION - divide the field into several larger sub-paradigms. The following five sub-paradigms emerge: • • • •

the verbs of opinion and belief; the verbs of knowledge; the verbs of understanding; the verbs for mental processes whose INFORMATION is situated in the past or future, i.e. the verbs of remembering and intention/expectation; • the verbs for mental processes whose INFORMATION is situated in the present, i.e. the verbs of (mental and sometimes sensory) perception and consideration.

Andreas Gr ger

280 Table 4: The Structure of the Lexical Field of Old English 'Mental' Verbs Q

'b': future

FORMATION

L)

Ζ

s O

^ ω

«. K J H < >

C3

ο r-

2 >-

£ H

—'

i

H

S

tu H1 ·«? r

· H

c.

u- p

S^ l

+^

irrelevant

DYNAMICITY/PREDICATION

gd

ΓΊ

ζ D S

Q Ci< Z C < ro f-

Λ

ί

objective

i>

X

g^

i

*

S > H U >

l

subjective

C

C

.5ω

έ5 §>

C

1

5 3

υ g^^ ·&

2

S

«•SS ^ S c S

M

3

U

O

ll__i__|j 1 M

g 1 I

is

3

'· *

2" ^ S

'S)

1 -§iH

s

Si I ftS 1 Q, >·

jn

ls 3i s= s? 3*

πι τ

ω

a

j,| cd i: 2

m o

Λ o . cd rv cd C U "g" CJ

11 3 X

c

2

ου

Λ g

30

SUBJECTIVITY

'

§ S (Λ

Έ. δ °

υ

W

-1

r ι ~

J3

•Sl

- &.·=

U O Q W

£

Λ 3

C

J3

static

^X

a

-

'

ssC Si1 &Ca £esg JLcds 15«·§§ M ja o §

c W

"ea

Ξ

£

? 0

D

e

|| f1 i| l li 1 ll

3o. ιί.

i «-?

τ, cdJ

'S

s 1

0 QJ ft, > H

g § e 3 ^J3 3 υ) c

o -!ΛΞ

Ο

oo

s .2 s

8s 1

ο

>^ -a

00 >,

=^ 1 «

τ

c cd > ~ t

>

ς« 5

uj

r-1 < % t~

=

u ofl p—

c3

u Λ 3 c ca tn

Λ

C

1

«

I lit

ι ι 4 >

S

ja.

r-

8

?

A Semantic Analysis of Old English Mental Verbs...

281

The sub-paradigm of the 'verbs of opinion and belief is depicted in Table 5 as an example of a finer classification. Note that the verbs are grouped according to the THINKER'S evaluation of subjective truth of the stored INFORMATION (the semes of the dimension TSO Type of Subjective Opinion' are an example of a scalar, or gradient, feature type). Gelyfan 'to believe', which mainly occurs in religious (con-)texts, is clearly differentiated by its subjective evaluation TSOa = "INFORMATION is certainly true" (without a necessary factual basis!) from other verbs such as wenan, frencan and teohhian. (The latter is more frequent as a 'verb of intention'.) t>encan plays a minor role in this sub-paradigm. For each sememe the following statistical information is provided in parentheses: • the number of the sememe belonging to the given sub-paradigm, • its absolute frequency, • its share of all 'mental' occurrences of this verb in the HCO. DYNAMICITY/ PREDICATION TYPE static: INFORMATION stored SUBJECTIVITY subjective TYPE OF SUBJECTIVE OPINION

I

'c': low Lexeme (sememe I number percentage)

'b': medium

'a': high

wenan (Sl.l I 112 | 56%) bencan (S3 | 161 16%) wenan (SI.21 cf. TSOb)

wenan (SI.2 | 70 I 35%) teohhian (S1.2UI 16%) gebencan (S613 I 3%) wenan: cf. also TSIb

gelifan (SI 1221 | 100%) teohhian (Sl.l | 2 | 8%) understandan (S2 | 1 | 2%)

Table 5: Sub-Paradigm of the "Verbs of Opinion"

6. An Outlook on Middle English Mental Verbs

A Ph.D. dissertation project "Semantic Analysis of Middle English Mental Verbs" was begun in November 1997. Its principles are similar to the project discussed in this article. However, because of the considerable influx of predominantly Romance loanwords, the total number of lexemes to be investigated increases (from 15 to 27 relevant verbs), and so

282

Andreas Gröger

does the complexity of the semantic apparatus. The lexicographic information concerning Middle English mental verbs is taken from the Middle English Dictionary and the CDROM version of the OED, while the four Middle English sub-corpora of the Helsinki Corpus (in its 11-file version) serve as textual basis and permit a finer diachronic partition. Thus, it should be possible to highlight important lexical and semantic changes at the transition from Early to Late Middle English, which was a period of far-reaching linguistic change on all levels.

Appendix: Semantic Dimensions and Semes in the Description of Old English Mental Verbs

ANI = Type of Animate Being /- COEa/13 a [human] b [non-human] COE = Type of Concrete Entity /- COMb/ a [animate being] b [inanimate object] COM = Type of Complement a [abstract entity] b [concrete entity] DEI = Degree of Intention /- GOAa/ a [strong, decided] b [not (particularly) strong]14 DYN = Dynamicity a [static] b [dynamic] GOA = Goal Aspired by THINKER a [future activity or decision] b [knowledge] c [focussing mind on INFORMATION] d [changed disposition to act in a specific way on a specific occasion]

13

14

The remark /- COEa/ (etc.) indicates that there is a hierarchical combination of two or more semes. For instance, in the "verbs of subjective opinion" such as wenan, the semes TSOa/b/c all presuppose SUBb. This seme is the "default" seme applying unless otherwise specified. - It is of course possible to postulate further semes situated between the endpoints of the scale, but this should be avoided in dealing with an earlier language stage such as Old English as it may lead to a pseudo-exactitude which cannot be supported by linguistic evidence.

A Semantic Analysis of Old English Mental Verbs...

283

IMP = Intensity of Mental Processing /- PREb/ a [intensive] b [not (particularly) intensive] KNO = Type of Knowledge /- SUBb/ a [familiarity of THINKER with INFORMATION (due to previous contact) emphasized] b [no emphasis on familiarity] c [THINKER is familiar with sth. that can be learned] PKN = Previous Knowledge of INFORMATION a [new: INFORMATION previously unknown to THINKER] b [old: INFORMATION previously known to THINKER] PRE = Predication Type a [have stored in brain] b [process in brain] SOU = Source of INFORMATION a [visual perception] b [auditive perception] c [INFORMATION is received as a linguistic item] d [tactile perception] SUB = Subjectivity/Objectivity a [subjective opinion] b [objective knowledge] THE = Thematic Kind of INFORMATION a [religious] b [philosophical] c [sth. which can be learned: a craft, skill, text,...] d [language or linguistic item] e [concrete object]15 f [other abstract object, esp. an event or situation] TSI = Temporal Situation of INFORMATION a [past] b [future] TSO = Type of Subjective Opinion /- SUBa/ a [INFORMATION is certainly true] b [INFORMATION is probably true] c [INFORMATION is probably or possibly true]

15

In some borderline cases of verbs which have sememes denoting sensory perception (-cnawan and ongitan), the seme THEe includes objects which can be perceived by ears or eyes (cf. the semes SOUa/b).

284

Andreas Gröger

TYM = Type of Mental Activity a [process or have INFORMATION in mind] b [have INFORMATION stored in memory] c [receive INFORMATION in mind] d [(be able to) transfer INFORMATION from memory to mind] e [receive INFORMATION in mind, process it and finally store the result in memory] f [create (a picture of) INFORMATION in mind] UND = Type of Understanding /- TYMb/c/e/ a [THINKER knows the meaning of (a) linguistic item(s)] b [THINKER knows INFORMATION and its ADDITIONAL COMPLEX PROPERTY (= logical structure, reason(s), etc.)]16 c [THINKER interprets INFORMATION in a certain way]

References

a) Text Corpora The Helsinki Corpus of English Texts: Diachronie Part, 3-File Version, Old English Subcorpus. (1991) Helsinki: Department of English, University of Helsinki. — 11-File Version, Middle English Subcorpus. (1991) Helsinki: Department of English, University of Helsinki.

b) Dictionaries An Anglo-Saxon Dictionary Based on the Manuscript Collections of the Late Joseph Bos-worth. Edited and Enlarged by T. Northcote Toller. Supplement and Addenda Enlarged by Alistair Campbell. (1972-1976) 2 vols. Oxford: Oxford University Press. Cameron, Angus et al. (eds.) (1986ff.): Dictionary of Old English. Toronto: Pontifical Institute for Mediaeval Studies. Clark Hall, John R. (ed.) (1960): A Concise Anglo-Saxon Dictionary. 4th Edition with a Supplement by Herbert D. Meritt. Cambridge: Cambridge University Press. Kurath, Hans/Sherman M. Kühn/Robert E. Lewis et al. (eds.) (1952ff.): Middle English Dictionary. Ann Arbor: University of Michigan Press. Simpson, John/Edmund S.C. Weiner (eds.) (1989): The Oxford English Dictionary, Second Edition. 20 vols. Oxford: Oxford University Press. The Oxford English Dictionary, Second Edition, on Compact Disc for the IBM PC. (1992) Oxford: Oxford University Press. 16

Cf. Kay & Samuels's (1977: 59) definition of understand as "COGNIZE (PARTS) + THINK + COGNIZE (WHOLE)". A paraphrase of understand along the lines of "to know something and the reason(s) for it" or following Kay's & Samuels's suggestion seems a workable alternative to Schneider's (1988b: 154 f.) rather complex and technical definition of the verb's meaning; however, I am aware that the seme UNDb may need to be redefined more exactly in the course of further investigations.

A Semantic Analysis of Old English Mental Verbs...

285

Summers, Delia et al. (eds.) (1987): Longman Dictionary of Contemporary English. 2nd edition. Harlow: Longman.

c) Other Works Bradley, S.A.J. (ed. & tr.) (1982): Anglo-Saxon Poetry. An Anthology of Old English Poems in Prose Translation with Introduction and Headnotes (...). London: J.M. Dent. Geeraerts, Dirk (1997): Diachronie Prototype Semantics: A Contribution to Historical Lexicology. Oxford: Clarendon Press. Gröger, Andreas (1997): Die semantische Beschreibung des Wortfelds der mentalen Verben des Altenglischen auf der Basis des HELSINKI CORPUS. Unpublished MA thesis, University of Bayreuth. Gutch, Ulrike (1979): Altenglisch '-cnawan', 'cunnan', 'witan' - neuenglisch 'know'. Ph.D. thesis, Free University of Berlin. Kay, C./M.L. Samuels (1977): "Componential Analysis in Semantics: Its Validity and Applications". In: Transactions of the Philological Society (1975) [1977], 49-81. Koskenniemi, Inna (1968): Repetitive Word Pairs in Old and Early Middle English Prose. Turku: Turun Yliopisto. Kytö, Merja (ed.) (1991): Manual to the Diachronie Part of the Helsinki Corpus of English Texts: Coding Conventions and Lists of Source Texts. Helsinki: Department of English, University of Helsinki. Leisi, Ernst (1985): Praxis der englischen Semantik 2nd edition. Heidelberg: Carl Winter. Lipka, Leonhard (1992): An Outline of English Lexicology: Lexical Structure, Word Semantics and Word-Formation. 2nd edition. Tübingen: Niemeyer. Schneider, Edgar W. (1988a): Variabilität, Polysemie und Unscharfe der Wortbedeutung. Band l: Theoretische und methodische Grundlagen. Tübingen: Niemeyer. — (1988b): Variabilität, Polysemie und Unscharfe der Wortbedeutung. Band 2: Studien zur lexikalischen Semantik der mentalen Verben des Englischen. Tübingen: Niemeyer. Swanton, Michael J. (ed. & tr.) (1993): Anglo-Saxon Prose. 2nd edition. London: J.M. Dent. Wierzbicka, Anna (1972): Semantic Primitives. Frankfurt/Main: Athenäum. — (1996): Semantics: Primes and Universals. Oxford: Oxford University Press. Wildenhain, Anke (1994): "On 'Verbs of Suffering' in Old, Middle and Early Modern English". In: Hyldgaard-Jensen, Karl/Viggo Hj0rnager Pedersen (eds.). Symposium on Lexicography VI. Proceedings of the Sixth Symposium on Lexicography, May 7-9, (1992): at the University of Copenhagen. 93-102. Tübingen: Niemeyer. — (1995): Ausgewählte englische 'Verben des Leidens'. Eine synchronisch- diachronische Bedeutungsbeschreibung. Ph.D. thesis, University of Leipzig. WordCruncher Index 4.5, WordCruncher View 4.5. Text Indexing and Retrieval Software. (1992) Provo: Brigham Young University & Electronic Text Corporation.

Ljubima Jordanowa

Die Zweite Wende (1996-1997) und die neuen Probleme für die Lexikographie in Bulgarien

Die Sprache der Wende hat für Osteuropa seit geraumer Zeit ein neues wissenschaftliches Gebiet ins Leben gerufen, das sich unter dem Begriff "Wendezeit" zusammenfassen lässt. Jedes Land der mittel- und osteuropäischer Region hat seinen eigenartigen Entwicklungsweg. Für Deutschland umfaßt z. B. diese Periode die Zeit von 1989/90 - die Zeit der Flucht aus der DDR, der Demos, der Wiedervereinigung.

l. Zur Periodisierung der Wendezeit in Bulgarien Die Zeit der ersten Wende umfaßt die Jahre 1989-1991, als gegen den Sozialismus demonstriert wurde, als die ersten demokratischen Wahlen durchgeführt wurden, bei denen die UDK (Union der demokratischen Kräfte) den Sieg errungen hatte und die erste UDKRegierung 1991 gefallen war. Die zweite Periode, 1992-1996, zeichnet sich durch den Rückgang der bereits gestarteten sozialen Prozesse aus, die Wiederkehr der Sozialisten ehemaliger Kommunisten - an die Macht nach den parlamentarischen Wahlen 1994. Die dritte Periode, 1996- 1997, lässt sich durch den Sieg von Peter Stojanov, dem Kandidaten der UDK in den Präsidentenwahlen vom November 96, den friedlichen Protest vom Januar 97 und die vorfristigen parlamentarischen Wahlen im April 97 charakterisieren, als die Union der demokratischen Kräfte mit ausschlaggebender Mehrheit den Sieg errungen hatte. Die letzte Periode (1996 - 97) nenne ich die Zweite Wende in Bulgarien: der Höhepunkt wurde durch den friedlichen Bürgerprotest vom Januar (3.01.-5.02.) erreicht, als die WENDE von den Bürgern Bulgariens vollzogen wurde; Friedliche Demos in den Städten des Landes und die Barrikaden vom 3. und 4. Februar brachten die Sozialisten, ehemalige Kommunisten dazu, auf eine nächste sozialistische Regierung zu verzichten und in den vorfristigen parlamentarischen Wahlen einzuwilligen. Im April 97 fanden die Wahlen statt. Die erzielten Ergebnisse wurden bereits im Januar 97 entschieden. Meiner Meinung nach ist die letzte Periode die bedeutendste Zeit der neuesten bulgarischen Geschichte, soweit die erste Wende 1989 von der Bulgarischen Kommunistischer Partei realisiert, kontrolliert, eigentlich - und dies ist sehr wichtig - nachgeahmt wurde: Nachdem am 9.11.1989 in Berlin die Mauer gefallen war, entschied sich die BKP auf ihrer Tagung vom 10.11.1989, Veränderungen in ihrer Nomenklatur vorzunehmen. An den Demos gegen die sozialistische Regierung vom Januar 97 nahmen Bürger jedes Standes teil: Ärzte und Lehrer, Akademiker und Studenten, die Gewerkschaften, Fahrer, Bergleute, ..., Leute unterschiedlichen Alters Schüler und Rentner; Man streikte im ganzen Land und aus der politischen und wirtschaftlichen Krise konnte nur die vorfristigen parlamentarischen Wahlen helfen. Januar 97 haben die Bulgaren ihre zivilisierte Wahl getroffen. Nach den Wahlen strebt nun das Land nach demokratischen Veränderungen und dem Bau der bürgerlichen Gesellschaft.

288

Ljubima Jordanowa

2. Zur Fachliteratur Ich möchte gleich in diesem Zusammenhang auf das "Wörterbuch in Texten der Wendezeit 1989/90. Ein Wörterbuch zur lexikalischen Erschließung des 'Wendekorpus'" von Manfred Hellmann vom Institut für deutsche Sprache in Mannheim hinweisen. Ein zweites Wörterbuch wurde von Autoren aus demselben Institut verfasst und ist bereits (1997) erschienen: Dieter Herberg, Doris Steffens, Elke Tellenbach. Schlüsselwörter der Wendezeit. Wörter-Buch zum öffentlichen Sprachgebrauch 1989/90. Das Wörter-Buch, wie die Autoren wortspielerisch ihre Untersuchung genannt haben, verzeichnet monographisch Schlüsselwörter. Es baut auf der Grundlage des 'Wendekorpus', dem neuesten Archiv des Instituts für deutsche Sprache, auf. Die Untersuchung ist ein Teil des Forschungsprojekts unter der Leitung von Wolfgang Teubert "Sprachliche Varietäten der Wendezeit". Das Buch stellt die Art lexikologischen und lexikographischen Know-hows dar, durch die Gegenwart untersucht werden kann, die sich ja am schwierigsten einer Periodisierung bzw. Rezeption durch das stete Einfließen neuerer überraschenderer Information über die historischen Ereignisse unterziehen lässt. Es muss auch auf den vielseitigen und ausgiebigen Umfang der Belegsammlung, das bewusste objektive Sammeln von historischen Fakten in beiden deutschen Staaten, einschließlich der Periode 1989/90, als der Gedanke an die Wiedervereinigung noch keinem Deutschen vorschwebte, hingewiesen wird. Vier Millionen Wortformen, die innerhalb von anderthalb Jahren während der dramatischen Ereignisse in der DDR gesammelt wurden, tragen die Bilanz einer gewissenhaft verrichteten Arbeit. In diesem Sinne verdient es die Arbeit der deutschen Kollegen, der Wendezeit-Fachleute, dass man an sie Beispiel für die Registrierung einer historischen Periode, für die berufliche Verantwortung der Geschichte gegenüber nehmen sollte. Die Sprache der ersten Wende in Bulgarien und der folgenden Periode 1992-1996 wurde mehr oder weniger anhand der Texte von Losungen aus den Demos auf den Straßen von Sofia, von Programmen und Werbetexten unterschiedlicher Parteien, sowie aus Pressetexten untersucht. Die gesammelten Belege reichen durchaus nicht aus, da fehlen z. B. Belege aus den Stenogrammen der Sitzungen in der Volksversammlung; doch ist der Grundstein für den Aufbau von einem 'Wendekorpus' für das Bulgarische gelegt. Manche Rezeption der Ereignisse, vor allem aber manches über die Sprache der Wende ist bereits erschienen (Siehe Jordanowa 1993a,b; Jordanowa 1994; Bulgarien 1994). Die Studie "Sprache der Wendezeit" war nur der Anfang. Das Computerwörterbuch der Neologismen im Bulgarischen (Ende der 80er und Anfang der 90er Jahre) belegt den Korpus der ersten Wendezeit und der Periode danach bis 1993. Seit 1994 leite ich die wissenschaftlichen Reihe "Soziolinguistik", die Sprache der Wende in Osteuropa nach 1989 zum Gegenstand hat. 1997 wurde die erste Dissertation über die Sprache der Presse (1989-1996) verteidigt (Siehe Stojanov 97). Nach dieser Promotion gilt die Aufmerksamkeit der Linguisten bereits in größerem Maße der Problematik der sprachlichen Entwicklung in neuester Zeit.

Die Zweite Wende (1996-1997)...

289

3. Zum Korpus der Zweiten Wende in Bulgarien (1996-1997) Die Zweite Wende umfaßt folgende Phasen: 1. Phase: März 1996 - 1. Juni 1996. Peter Stojanov wurde zum Kandidaten von der Union der demokratischen Kräfte für den Sitz des Präsidenten auf der 8. nationalen Konferenz der DDK gewählt; er hat in den Grundwahlen am 1. Juni gesiegt, somit wurde er Vertreter der Vereinigten Opposition aufgrund der Vereinbarung unter der UDK, dem Volksverband (dem Demokratischen Bauernverband und der Demokratischen Partei) und der Bewegung für Rechte und Freiheiten. 2. Phase: 1. Juni - 3. November 1996. Die s.g. heiße Phase - Präsidentenwahl; Peter Stojanov errang den Wahlsieg. 3. Phase: (19. Dezember 1996) 3. Januar - 5. Februar 1997. Höhepunkt. Sie zeichnet sich durch friedliche Proteste aus, die am 19. Dezember von der Vereinigung "Wende" organisiert wurden und nach Weihnachten und Neujahr lawinenartig in der Zeit vom 3. Januar bis 4. Februar ausbrachen. Der politische Konsultationsrat, den der Präsident der Republik Peter Stojanov einberufen hatte, erreichte die Vereinbarung zum Verzichten der Bulgarischen Sozialistischen Partei auf ein zweites Regierungsmandat. Der 5. Februar blieb für die Geschichte erhalten durch den letzten Zug der Studenten in den Straßen von Sofia und den 'Victory-Meeting' auf dem Platz vor der Kathedrale "St. Alexander Newski". 4. Phase: vorfristige parlamentarische Wahlen vom 19. April 1997. Eine Periodisierung der Gegenwart ist ein außerordentlich schwieriges Vorhaben: Ereignisse sind nicht vorauszusagen, sie kommen oft unerwartet. Höchstwahrscheinlich muss in der Zukunft manches in der o.a. Periodisierung korrigiert werden. Aus der o.a. Periodisierung ist ersichtlich, dass der Korpus der Zweiten Wende in Bulgarien die öffentliche und die politische Kommunikation, die Rezeption der sozialen Situation seitens der politischen Parteien mit unterschiedlicher Akzentuierung und Prioritätensetzung ihrer Tätigkeit vorstellen sollte. Die Grundlage vom Korpus bilden also: 1. Texte von Losungen, rhythmische Ausrufe, Aufschriften aus der Zeit der Januardemos 97. Sie bilden den Fokus der Soziums, der unterschiedlichen sozialen Gruppen. Sie wurden bereits im Buch "Vom Lumpen zum Staatsbürger" veröffentlicht (Jordanowa 1997: 177301). 2. Texte von Flugblättern, Wahltransparenten regierender und oppositioneller Parteien. Eine Analyse aus der Sicht manipulativer Strategien und deren sprachlicher Ausdruck liegt bereits vor (Jordanowa 1997a: 50-79; 1997b: 88-103). Die politische Botschaft unterschiedlicher Parteien an den Wählern wird auf der Grundlage des Transparententextes, des Textes von Losungen, der Programme, der Symbolik und des Farbenspiels betrachtet. Die Vielseitigkeit und Vieldeutigkeit der politischen Information, des politischen Textes, wird durch den Hinweis auf die Vermengung einiger manipulativer Strategien, einiger Signale der Darbietung politischer Information verdeutlicht, z.B. der Vereinigungstrategie. Als konkretes Beispiel sei hier auf das Wahltransparent der UDK vom April 97 hingewiesen. Der Text lautet: Ooegunenu JleMokpamuvHu CUAU ... sä > . OffC 97. Hue . /Die Vereinigten Demokratischen Kräfte ... für Bulgarien. VDK 97. Wir können es./ Die Vereinigungsstrategie wird vielfach signalisiert: durch den Namen Vereinigte Demokratische Kräfte, dessen Abkürzung VDK, durch den Satz: einigende erste Person Plural des Subjekts «we/wir und Verb M oofce.M/können; durch das in Großbuchstaben ausgeführte Bulgarien, das dokumentarisch im Text als Ausschnitt von Aufnahmen friedlich demonstrierender Menschen zusammengesetzt wurde. Da ist auch

290

Ljubima Jordanowa

die semantisch positiv geladene Präposition 3a /für. Die Vereinigungstrategie wird weiterhin durch die protestierenden Leute unterstützt: im Vordergrund treten durch die Montage drei junge Leute auf. Sie treten zwar hervor, doch sie bilden eine Einheit mit dem friedlichen Zug. Dies auf der einen Seite. Zum anderen bilden die drei jungen Leute eine Gruppe, zusammengehalten von der nationalen Flagge. Die Menschen sind im Januar protestierend auf den Straßen gegangen, jetzt im April während der Wahlkampagne sollen sie wieder Mut haben. Es wird also an das bereits Erreichte erinnert: die Erinnerung ist ein außerordentlich wichtiges Element der politischen Propaganda. Die Menschenmenge wird durch Leute unterschiedlichen Alters gezeigt, was ebenfalls in der 'Vereinigung' als Grund für den Protest überzeugen soll. Die lachenden Gesichter muten die Information auch positiv an. Die Vereinigung Union der demokratischen Kräfte - Vereinigte demokratische Kräfte - Bulgarien wird durch einige Modi signalisiert. Nebeneinander treten die blaue und die nationale Flagge auf, dokumentarisch dargestellt durch die Leute, die während des Protestes Fahnen tragen. Die nationale Flagge wird von drei jungen Leuten getragen, die auf ihrer Kleidung das UDK-Abzeichen mit dem kleinen Löwen bzw. ein blaues Band auf ihren Ärmeln und eine im Victory-Zeichen gehobene Hand haben. Selbst die Abkürzung UDK hat dieselbe Strategie inne: auf blauem Hintergrund stehen in Weiß die Buchstaben und daneben in Grün und Rot die Ziffern 9 und 7. - Es wären an die zehn Varietäten, durch die eine Vereinigungstrategie signalisiert wurde. Unter den Strategien der politischen Parteien in den Propagandatexten wurden festgestellt: Strategie der Hyperbolisierung Lobpreisung und Verniedlichung des politischen Gegners, nicht überwundener Stil des Ausdrucks aus der Zeit des Sozialismus; das Umschweifen von Problemen; Konkretheit; Kontrast!vität usw. Von Interesse seien die Texte der BBB-Partei (des Bulgarischen Business Blocks) mit ihren Mythen von Überparteilichkeit, Schuld des anderen, des Erfolgs aus dem Nichts - der Arbeitslosigkeit, Mühelosigkeit. 3. Texte aus Zeitungsartikel, Interviews, Kommentare u.a.m. Das Archiv von Krassimir Stojanov belegt Vieles der vorangehenden Periode, bis 1996. 1998 soll das Sammeln von Belegen fortgesetzt werden. 4. Texte der Stenogramme aus den Sitzungen der Volksversammlung. Dies sollte zwecks der Fülle von Belegen erfolgen. 5. Texte von Songs.

4. Zur Rezeption von neuem und zeittypischem Wortschatz Als zeittypischen Wortschatz nenne ich hier: npoMana/Wende/, cno/\yka/Erfo\g/ als Schlüsselwort der BSP-Propaganda, vepBenu ßoi/iyifw/roter Mist/, vepBemi/Rote/, cwww/Blaue/, ,wa0iia/Mafia/ - Wörter und Wendungen, die auch in der Zeit der ersten Wende nach 1989 da waren. Die Zeit zwischen 1996-97 ließ Wörter durch ihre aktuelle Bedeutung hervortreten: «a/iÄM/Gummiknüppel/ anlässlich des Prügels, den friedliche Bürger in der Nacht vom 10. auf den 11. Januar 1997 erleiden sollten; B-bAzapux, ÖTiAzapuH/Bulgarien, Bulgare/ trug einen besonderen Akzent während des Protestes, um die Sorge um das Land und die Angst vor dem Rückgang gesellschaftlicher Prozesse zurück in die Geschichte zu betonen; Buratino - Spitzname des Politikers Shorsh Gantschev, der die Strategie der Vemiedlichung in Gang zu setzen vermag. Slangwörter,

Die Zweite Wende (l996-1997)...

291

wie z. B. /ty,v