Words: Structure, Meaning, Function: A Festschrift for Dieter Kastovsky 9783110809169, 9783110167931

Table of contents :
Tabula gratulatoria
Publications
Introduction
Old Irish dúilem ‘maker, creator’
The importance of being Leofwine Horn
The desemanticisation of existential there in a synchronic-diachronic perspective
Anti- and pro- in Indian English
The meaning of naked and a theory of metaphor
On contrastive word-formation: German and Hungarian denominal adjective formation
The greening of greening
Middle English beck in the Midlands: the place-name evidence
Conceptual and semantic change in the history of English
Genitive case or determinative phrase? The status of ’s forms in Modern English
Direction and location in Modern Irish
John Locke, semanticist
Thou in the history of English: a case for historical semantics or pragmatics?
Diminutives: an interface of word-formation, semantics and pragmatics
Which end is your head on? Typological remarks on Modern Persian
Word-formation and (proper) names: a neglected field
Text condensation in the press: the case of compound adjectival premodifiers
On the semantics of English double modals
The third-person singular -(E)S and -(E)TH revisited: the morphophonemic hypothesis
Paths of loan-word grammaticalisation: the case of according to
The third person present plural in Shakespeare’s First Folio: A case of interaction of morphology and syntax?
The function of word-formation and the case of English -cum-
What happened to Old English clitic pronouns and why?
Tall As He Was: On the meaning of complement-as-constructions
Eyes and ears
Subject index

Citation preview

Words: Structure, Meaning, Function

W DE G

Trends in Linguistics Studies and Monographs 130

Editor

Werner Winter

Mouton de Gruyter Berlin · New York

Words: Structure, Meaning, Function A Festschrift for Dieter Kastovsky

edited by

Christiane Dalton-Puffer Nikolaus Ritt

Mouton de Gruyter Berlin · New York

2000

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Die Deutsche Bibliothek -

CIP-Einheitsaufnahme

Words: structure, meaning, function : a festschrift for Dieter Kastovsky / ed. by Christiane Dalton-Puffer ; Nikolaus Ritt. Berlin ; New York : Mouton de Gruyter, 2000 (Trends in linguistics : Studies and monographs ; 130) ISBN 3-11-016793-X

© Copyright 2000 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Printing & Binding: Hubert & Co, Göttingen. Cover design: Christopher Schneider, Berlin. Printed in Germany.

AUSTRALIJSKIKRAJOBRAZ JEZYKOWY Ludwika Amber Gdziei pomi?dzy wielkimi miastami Sydney i Melbourne przystaj? na drodze poSröd ska} i zrzucaj^cych kor? eukaliptusöw Ptak przylatuje albo dwa kr?c^ si? w pobli£u zwabione czlowiekiem W zielniku slöw mam j u i coraz wi?cej ptasich imion i zapisz? w moim pierwszym j?zyku te drzewa te skafy te ptaki ze przygl^damy si? sobie w ruchu w kolorze i dostrzegamy si? w buszu A potem - ju1 p6£niej b?dziemy to wszystko (wszystko?) thumaczyd na drugi j?zyk i po raz drugi pröbowad dotkn^d diugiego wahania pasma kory na gal?zi australijskiego powietrza i po raz drugi b?dziemy cierpliwie odkrywad pod palcami na tych samych wypuktoSciach i p?kni?ciach skat nowe diwi?ki i slowa w krajobrazie AUSTRALISCHE SPRACHLANDSCHAFT Translated by Dorothea Müller-Ott Irgendwo zwischen den großen Städten Sydney und Melbourne halte ich inne inmitten von Felsen und des die Rinde abwerfenden Eukalyptus ein Vogel kommt geflogen oder zwei angelockt von der Nähe des Menschen verweilen sie Im Herbarium der Worte habe ich immer mehr Vogelnamen und schreibe sie in meiner ersten Sprache auf die Bäume die Felsen die Vögel wie wir uns gegenseitig ansehen in Bewegung in Farbe und uns wahrnehmen im Busch

Dedicated to Dieter Kastovsky

with gratitude for his support of the Literary Translation Circle of the Polish Section (Department of Translation Studies, University of Vienna)

Und dann - schon später werden wir das alles (alles?) übersetzen in eine zweite Sprache und ein zweites Mal versuchen zu berühren das lange Beben der Eukalyptusrinde auf den Zweigen in australischer Luft und ein zweites Mal werden wir geduldig entdecken unter unseren Fingern auf den gleichen Felsvorsprüngen und -spalten neue Töne und Wörter in der Landschaft

Contents Tabula gratulatoria

ix

Publications by Dieter Kastovsky

xi

Introduction (by Arthur Mettinger)

xix

Anders Ahlqvist Old Irish duilem 'maker, creator'

1

John Anderson and Fran Colman The importance of being Leofwine Horn

7

Leiv Egil Breivik and Toril Swan The desemanticisation of existential there in a synchronic-diachronic perspective

19

Uwe Carls Anti- and pro- in Indian English

35

Hans-Jürgen Diller The meaning of naked and a theory of metaphor

45

Wolfgang U. Dressier and Μάτια Ladänyi On contrastive word-formation: German and Hungarian denominal adjective formation

59

Andreas Fischer The greening of greening

75

Jacek Fisiak Middle English beck in the Midlands: the place-name evidence

87

Manfred Görlach Conceptual and semantic change in the history of English

95

Klaus Hansen Genitive case or determinative phrase? The status of '5 forms in Modern English

Ill

Raymond Hickey Direction and location in Modern Irish

125

Werner Hüllen John Locke, semanticist

141

viii

Table of contents

Andreas Η. Jucker Thou in the history of English: a case for historical semantics or pragmatics?

153

Barbara Kryk-Kastovsky Diminutives: an interface of word-formation, semantics and pragmatics

165

Roger Lass Which end is your head on? Typological remarks on Modern Persian

175

Leonhard Lipka Word-formation and (proper) names: a neglected field

187

Magnus Ljung Text condensation in the press: the case of compound adjectival premodifiers

205

Stephen J. Nagle and Patsy L. Holmes On the semantics of English double modals

217

Terttu Nevalainen and Helena Raumolin-Brunberg The third-person singular -(E)S and -(E)TH revisited: the morphophonemic hypothesis

235

Matti Rissanen Paths of loan-word grammaticalisation: the case of according to

249

Herbert Schendl The third person present plural in Shakespeare's First Folio: A case of interaction of morphology and syntax?

263

Gabriele Stein The function of word-formation and the case of English -cum-

277

Robert Stockwell and Donka Minkova What happened to Old English clitic pronouns and why?

289

Gunnel Tottie Tall As He Was: On the meaning of complement-os-constructions

307

Werner Winter Eyes and ears

323

Subject index

329

Tabula Gratulatoria Adelaar Willem, Leiden Aijmer Karin, Göteborg Allwood Jens, Göteborg Alves Helio Osvaldo, Minho Andersen Henning, Los Angeles Ash Sharon, Philadelphia Askedal John Ole, Oslo Awedyk Wieslaw, Poznan Bailey Charles-James, Hawaii Bakken Bjorg, Bergen Bech Kristin, Bergen Beedham Christopher, St. Andrews Bergner Heinz, Glessen Bergs Alexander, Düsseldorf Beuketna Frits, Leiden Blaylock Curtis, Urbana Boas Hans Ulrich, Erfurt Boeder Wilfried, Oldenburg Bonheim Helmut, Köln Brekle Herbert, Regensburg Brinton Laurel, British Columbia Bühler Hildegund, Wien Byles J. Montgomery, Nicosia Cacoulhos Ann R., Athens Comrie Bernard, Leipzig Cowie Ciaire, Sheffield Cunchillos Carmelo, Rioja Dardano Maurizio, Roma Darski Jozef, Poznan Denison David, Manchester Di Meola Claudio, Roma Disterheft Dorothea, Columbia Doleschal Ursula, Wien Dotter Franz, Klagenfurt Duffley Patrick J., Laval Durell Martin, Manchester Eliasson Stig, Mainz Emons Rudolf, Zürich Fabiszak Malgorzata, Poznan Fanego Teresa, Santiago de Compostela Fenk-Oczlon Gertraud, Klagenfurt Fill Alwin, Graz Firbas Jan, Brno ( t ) Fischer Olga, Amsterdam Fitzmaurice Susan, Flagstaff Fries Udo, Zürich Geckeier Horst, Münster Gneuss Helmut, München

Gracia Lluisa, Girone Haberer Adolphe, Lyon Hajicova Eva, Prag Hamans Camiel, Breda Hartmann Reinhard, Exeter Hentschel Gerd, Oldenburg Herman Luc, Antwerp Hladky Josef, Brno Hogg Richard, Manchester Hoyt Alexander D., Zagreb Hudabiunigg Ingrid, Chemnitz Ikegami Yoshihiko, Tokyo Jackson Howard, Birmingham Jahr, Ernst Hakon, Tromsö James Allen, Klagenfurt Kalogjera Damir, Zagreb Kawashima Atsuo, Kyoto Kayman Martin Α., Coimbra Kettemann Bernhard, Graz Kiefer Ferenc, Budapest Klein Jared S., Athens, Georgia Kniezsa Veronika, Budapest König Ekkehard, Berlin Krohn Hele, Helsinki Kürschner Wilfried, Vechta Kuteva Tania, Köln Kytö Merja, Uppsala Lehman Winfred P., Austin, Texas Lenerz Jürgen, Köln Leong Alvin, Singapore Lewandowska-Tomaszczyk Barbara, Lodz Loebel Elisabeth, Köln Löpez-Couso Angel, Valencia Lucko Peter, Berlin Lutz Angelika, Erlangen Macek Dora, Zagreb Mackenzie Lachlan, Amsterdam, Marcos Marin Francisco, Madrid Marenco Franco, Turin Markus Manfred, Innsbruck Martinez del Castillo Jesus, Almeria Mautner Gerlinde, Wien Mazzon Gabriella, Palermo Mendez Naya Belen, Santiago de Compostela Milfull Inge, Eichstätt Mogensen Jens Erik, Copenhagen Murray Robert, Calgary Nathan Geoffrey S., Carbondale

χ

Tabula gratulatoria

Niemi Jussi, Joensuu Nikolic-Hoyt Anja, Zagreb Nurmi Arja, Helsinki Olsen Susan, Leipzig Onega Jaön Susana, Zaragoza Oresnik Janez, Ljubljana Owens Jonathan, Bayreuth Petric Teodor, Maribor Pfeiler Barbara, Mirida Piller Ingrid, Hamburg Pintzuk Susan, York Plag Ingo, Siegen Pollner Claus, Leipzig Popelak Johann, Wien Punter David, Stirling Ramat Paolo and Anna Giacalone, Pavia Rauh Gisela, Wuppertal Rissanen Eeva, Helsinki Rutkowska Hanna, Poznan Säbel Joachim, Frankfurt Salvador-Rabaza Ramos Asunc.,Valencia Saric Ljiljana, Oldenburg Scalise Sergio, Bologna Schläger Jürgen, Berlin Schulte Michael, Kristiansand Schulze Rainer, Hannover Seoane-Posse Elena, Santiago de Compostela Seuren Pieter, Nijmegen Sevic Radmila, Novi Sad Sikorska Liliana, Poznan

Skrzypczak Waldemar, Torun Smith John Charles, Oxford Stamenov Christo, Sofia Stein Dieter, Düsseldorf Stekauer Pavel, Prsov Svartvik Jan, Lund Szwedek Aleksander, Bydgoszcz Tabakowska Elzbieta, Krakow Tissari Heli, Helsinki Traugott Elisabeth, Stanford Van Gelderen Elly, Tempe Van Kemenade Ans, Nijmegen Van Schoonewald C.H., La Roche/Foron Vater Heinz, Köln Vennemann Theo, München Viereck Wolfgang, Bamberg Vogel Petra, Osnabrück Warvik Brita, Turku Wasik Zdzislaw, Poznan & Opole Wegener Heide, Potsdam Weigand Edda, Münster Weiss Corinna, Wien Weydt, Harald, Frankfurt Wildgen Wolfgang, Bremen Winters Margaret E., Carbondale Wotjak Gerd, Leipzig Wright Laura, Cambridge Wyler Siegfried, St. Gallen Zabotkina Vera, Kiew Zettersten Arne, Lund Zimmermann Rüdiger, Marburg

Institut Für Anglistik und Amerikanistik der Universität Wien Institut für Anglistik der Universität Graz Institut für Anglistik der Universität Innsbruck Institut für Anglistik und Amerikanistik der Universität Klagenfurt Institut für Englische Sprache der Wirtschaftsuniversität Wien Institut für Sprachwissenschaft der Universität Klagenfürt Institut für Sprachwissenschaft der Universität Wien Institut für Romanistik der Universität Wien Institut für Germanistik der Universität Wien Institut für Theoretische und Angewandte Translationswissenschaft der Universität Graz Institut für Übersetzen und Dolmetschen der Universität Wien Seminar für Englische Philologie der Eberhard-Karls-Universität Tübingen Institut für Anglistik der Bergischen Universität-Gesamthochschule Wuppertal School of English der Adam Mickiewicz-Universität Poznan Societas Linguistica Europaea European Society for the Study of English Austrian Association for University Teachers of English

Publications by Dieter Kastovsky 1. Monographs 1968 Old English deverbal substantives derived by means of a zero morpheme. [Tübingen University Ph. D. dissertation 1967]. Esslingen/N.: Langer. 1971 Studies in morphology. Aspects of English and German verb inflection. (Tübinger Beiträge zur Linguistik 18). Tübingen: Narr. 1982 Wortbildung und Semantik. (Studienreihe Englisch 14). Tübingen Düsseldorf: Francke/Bagel.

2. Articles 1969 Wortbildung und Nullmorphem. Linguistische Berichte 2: 1-13. [1974]Repr. in: Werner Bauer, Matthias Hartig, Herwig Krenn, Edith Mayer, Klaus Müllner and Hartmut Pott (eds.), Studien zur generativen Grammatik. (Athenaion Skripten zur Linguistik), 134-157. Frankfurt/M.: Athenaion. [1981]Repr. in: Leonhard Lipka and Harmut Günther (eds.), Wortbildung. (Wege der Forschung 564), 306-323. Darmstadt: Wissenschaftliche Buchgesellschaft. 1971a The Old English suffix -er(e). Anglia 89: 285-325. 1971b Assimilationserscheinungen in der Morphologie des englischen und deutschen Verbs. In: Karl-Richard Bausch and Hans Martin Gauger (eds.), Interlinguistica. Sprachvergleich und Übersetzung. Festschrift zum 60. Geburtstag von Mario Wandruszka, 149-173. Tübingen: Niemeyer. 1971c Strukturelle Morphologie I, II. Radioscript/Study material. Funkkolleg Sprache. Süddeutscher Rundfunk Stuttgart. 197ld Generative Morphologie I, II. Radioscript/Study material. Funkkolleg Sprache. Süddeutscher Rundfunk Stuttgart. 1972 Probleme der historischen Morphologie aus generativer Sicht: das englische konsonantische Präteritum und 2. Partizip. In: Karl HyldgaardJensen (ed.), Linguistik 1971. Referate des 6. Linguistischen Kolloquiums 11. - 14. August 1971 in Kopenhagen. (Athenäum-Skripten Linguistik 1), 252-265. Frankfurt: Athenäum. 1973a Strukturelle Morphologie I, II. In: Klaus Baumgärtner and Hugo Steger (eds.), Funkkolleg Sprache. Eine Einführung in die moderne Linguistik. Volume 1, 152-172. Frankfurt: Fischer. 1973b Generative Morphologie I, II. In: Klaus Baumgärtner and Hugo Steger (eds.), Funkkolleg Sprache. Eine Einführung in die moderne Linguistik. Volume 1, 365-394. Frankfurt: Fischer. 1973c Causatives, Foundations of Language 10: 255-315.

xii

Publications by Dieter Kastovsky

1974a Strukturelle Morphologie I, II (Studienmaterial). In: Klaus Baumgärtner Hugo Steger (eds.), Lehrgang Sprache. Eine Einführung in die moderne Linguistik, 242-276. Weinheim/Tübingen: Beiz. 1974b Generative Morphologie I, II (Studienmaterial). In: Klaus Baumgärtner Hugo Steger (eds.), Lehrgang Sprache. Eine Einführung in die moderne Linguistik, 565-603. Weinheim/Tübingen: Beiz. 1974c Word-formation, case grammar and denominal adjectives. Anglia 92: 1 54. 1974d Introduction. In: Hans Marchand, Studies in syntax and word-formation. Selected articles. On the occasion of his 65th birthday on October 1st, 1972, edited by Dieter Kastovsky. (Internationale Bibliothek für allgemeine Linguistik 18), 10-22. München: Fink. 1976a Zur Analyse von Nomina Actionis. In: Kurt Braunmüller and Wilfried Kürschner (eds.), Akten des 10. Linguistischen Kolloquiums, Tübingen 1975. Volume 2. Grammatik, 77-90. Tübingen: Niemeyer. [1981] Repr. in: Leonhard Lipka and Hartmut Günther (eds.), Wortbildung. (Wege der Forschung 564) , 377-390. Darmstadt: Wissenschaftliche Buchgesellschaft. 1976b Intensification and semantic analysis. Some notes on Bolinger's Degree words [Review article]. Foundations of Language 14: 377-398. 1976c Word-formation, or: At the crossroads of morphology, syntax, semantics, and the lexicon. L.A. U. T. 1977a Introduction (with Herbert E. Brekle). In: Herbert E. Brekle and Dieter Kastovsky (eds.), Perspektiven der Wortbildungsforschung. Beiträge zum Wuppertaler Wortbildungskolloquium vom 9.-10. Juli 1975. Anläßlich des 70. Geburtstages von Hans Marchand am 1. Oktober 1977. (Wuppertaler Schriftenreihe Linguistik 1), 7-19. Bonn: Bouvier. 1977b Sentence adverbs in nominalisations. In: Herbert E. Brekle and Dieter Kastovsky (eds.), Perspektiven der Wortbildungsforschung. Beiträge zum Wuppertaler Wortbildungskolloquium vom 9.-10. Juli 1975. Anläßlich des 70. Geburtstages von Hans Marchand am 1. Oktober 1977. (Wuppertaler Schriftenreihe Linguistik 1), 116-128. Bonn: Bouvier. 1977c Problems of word-formation. In: Christian Gutknecht (ed.), Grundbegriffe und Hauptströmungen der Linguistik, 301-335. Frankfurt: Hoffmann & Campe. 1977d Word-formation, or: At the crossroads of morphology, syntax, semantics, and the lexicon. Folia Linguistica 10: 1-33. 1978 Zum gegenwärtigen Stand der Wortbildungslehre des Englischen. Linguistik und Didaktik 36: 351-366. 1979 (with Herbert E. Brekle, Leonhard Lipka, Gabriele Stein). Nachruf auf Hans Marchand. Anglia 97: 287-289. 1980a Zur Situation der lexikalischen Semantik. In: Dieter Kastovsky (ed.), Perspektiven der lexikalischen Semantik. (Wuppertaler Schriftenreihe Linguistik 2), 1-13. Bonn: Bouvier.

Publications by Dieter Kastovsky

xiii

1980b Selectional restrictions and lexical solidarities. In: Dieter Kastovsky (ed.), Perspektiven der lexikalischen Semantik. (Wuppertaler Schriftenreihe Linguistik 2), 70-92. Bonn: Bouvier. 1980c Perspektiven der lexikalischen Semantik - Perspectives of lexical semantics. In: Werner Habicht, (ed.), English and American studies in German. Summaries of theses and monographs, 19-21. Tübingen: Niemeyer. 1980d Zero in morphology. A means of making up for phonological losses? In: Jacek Fisiak (ed.), Historical morphology. (Trends in Linguistics. Studies and Monographs 17), 213-250. The Hague: Mouton. 1981a Wortbildung bei der Wortschatzarbeit. Der fremdsprachliche Unterricht 59: 169-176. 1981b Generative Grammatik und Fremdsprachenunterricht. Infinitiv und Gerundium. In: Peter Kunsmann and Ortwin Kuhn (eds.), Weltsprache Englisch in Forschung und Lehre. Festschrift für Kurt Wächtler, 151-174. Berlin: Schmidt. 1981c Intersection of syntax and the lexicon: lexical converses. In: Jürgen Esser and Axel Hübler (eds.), Forms and functions. Papers in general English and applied linguistics presented to Vilem Fried on the occasion of his sixty-fifth birthday, 123-136. Tübingen: Narr. 198Id Lexical fields and word-formation. In: Horst Geckeier, Brigitte Schlieben-Lange, Jürgen Trabant, Harald Weydt (eds.), Logos semantikos. Studio linguistica in honorem Eugenio Coseriu 1921 - 1981. Volume 3. Semantik, 429-445. Berlin, New York, Madrid: de Gruyter/Gredos. 1981e Implizite Kasus. In: Günther Radden and Renö Dirven (eds.), Kasusgrammatik und Fremdsprachendidaktik, (anglistik und englischunterricht 14), 189-200. Trier: Wissenschaftlicher Verlag Trier. 1982a Word-formation. A functional view. Folia Linguistica 16: 181-198. 1982b "Privative opposition" and lexical semantics. Studio Anglica Posnaniensia 14: 29-45. 1985a Typological changes in the nominal inflectional system of English and German. Studia gramatyczne 7: 97-117. 1985b Deverbal nouns in Old English and Modern English: from stem-formation to word-formation. In: Jacek Fisiak (ed.), Historical semantics. Historical word-formation. (Trends in Linguistics. Studies and Monographs 29), 221-262. Berlin: Mouton de Gruyter. 1986a Diachronic word-formation in a functional perspective. In: Dieter Kastovsky and Aleksander Szwedek (eds.), Linguistics across historical and geographical boundaries. In honour of Jacek Fisiak on the occasion of his fiftieth birthday. (Trends in Liguistics. Studies and Monographs 32), 409-421. Berlin: Mouton de Gruyter. 1986b The problem of productivity in word-formation. Linguistics 24: 585-600. 1986c Word-formation and pragmatics. In: Otto Rauchbauer (ed.), Festschrift für Siegfried Korninger. (Wiener Beiträge zur Englischen Philologie 80), 63-78. Wien: Braumüller.

xiv

Publications by Dieter Kastovsky

1986d Problems in the morphological analysis of complex lexical items. Acta Linguistica Academiae Scientiarum Hungaricae 36: 93-107. 1987a Word-formation, boundaries and the concept of motivation. In: Brigitte Asbach-Schnitker - Johannes Roggenhofer (eds.), Neuere Forschungen zur Wortbildung und Historiographie der Linguistik. Festgabe für Herbert E. Brekle zum 50. Geburtstag, 399-405. Tübingen: Narr. 1987b Boundaries in English and German morphology. In: Wolfgang Lörscher and Rainer Schulze (eds.), Perspectives on language in performance. Studies in linguistics, literary criticism, language teaching and learning. To honour Werner Hüllen on the occasion of his sixtieth birthday. (Tübinger Beiträge zur Linguistik 317), 159-170. Tübingen: Narr. 1988a Structural semantics or prototype semantics? The evidence of wordformation. In: Werner Hüllen and Rainer Schulze (eds.), Understanding the lexicon. Meaning, sense and world-knowledge in lexical semantics. (Linguistische Arbeiten 210), 190-203. Tübingen: Niemeyer. 1988b Typological changes in the history of English morphology. In: Udo Fries and Martin Heusser (eds.), Meaning and beyond. Ernst Leist zum 70. Geburtstag, 159-178. Tübingen: Narr. 1988c (with Gero Bauer) Introduction. In: Dieter Kastovsky and Gero Bauer (eds.), Luick revisited. Papers read at the Luick Symposium at Schloß Liechtenstein, 15. - 18. 9. 1985. Tübingen: Narr. 1989a Morphophonemic alternations and the history of English: examples from Old English. In: Manfred Markus (ed.), Historical English. On the occasion of Karl Brunner's 100th birthday. (Innsbrucker Beiträge zur Kulturwissenschaft. Anglistische Reihe 1), 112-123. Innsbruck: University of Innsbruck. 1989b Typological changes in the history of English word-formation. In: Heinz Joachim Müllenbrock and Renate Noll-Wiemann (eds.), Anglistentag 1988 Göttingen. Vorträge, 281-293. Tübingen: Niemeyer. 1989c Word-formation. In: Ren£ Dirven, Wolfgang Zydatiß and Willis J. Edmondson (eds.), A user's grammar of English: word, sentence, text, interaction, 171-214. New York, a.e.: Lang. 1989d01d English morphonological processes and morphology. In: Wolfgang Grosser, Karl Hubmayer, Franz Wagner and Wilfried Wieden (eds.), Phonophilia. Untersuchungen zur Phonetik und Phonologie. Festschrift für Franz Zaic, 83-94. Salzburg: Abakus. 1990a The typological status of Old English word-formation. In: Sylvia Adamson, Vivien Law, Nigel Vincent and Susan Wright (eds.), Papers from the 5th International Conference on English Historical Linguistics. (Current Issues in Linguistic Theory 65), 205-223. Amsterdam: Benjamins. 1990b Whatever happened to the ablaut nouns in English - and why did it not happen in German?. In: Henning Andersen and Konrad Koerner (eds.), Historical linguistics 1987. (Current Issues in Linguistic Theory 66), 253-264. Amsterdam: Benjamins.

Publications by Dieter Kastovsky

xv

1990c The interaction of semantic and formal structures in the lexicon. In: Jerzy Tomaszczyk and Barbara Lewandowska-Tomaszczyk (eds.), Meaning and lexicography. (Linguistic and Literary Studies in Eastern Europe 28), 75-91. Amsterdam: Benjamins. 1990d Word-formation and translation. Meta 35:45-49. 1991 Introduction. In: Dieter Kastovsky (ed.), Historical English syntax. (Topics in English Linguistics 2), 1-10. Berlin: Mouton de Gruyter. 1992a Semantics and vocabulary. In: Richard M. Hogg (ed.), The Cambridge history of the English language. Volume 1. The beginnings to 1066, 290407. Cambridge: Cambridge University Press. 1992b Typological reorientation as a result of level interaction: the case of English morphology. In: Günter Kellermann and Michael D. Morrissey (eds.), Diachrony within synchrony: Language history and cognition. (Duisburger Arbeiten zur Sprach- und Kulturwissenschaft 14), 411-428. Frankfurt/M.: Lang. 1992c The formats change - the problems remain: word-formation theory between 1960 and 1990. In: Martin Pütz (ed.), Thirty years of linguistic evolution, 285-310. Amsterdam: Benjamins. 1993a Inflection, derivation and zero - or: what makes OE and German derived denominal verbs verbs? Views 2: 71-81. 1993b The structure of the lexicon and language teaching. In: James E. Alatis (ed.), Georgetown University Round Table on Languages and Linguistics 1993, 521-531. Washington, D.C.: Georgetown University Press. 1994a Typological differences between English and German morphology and their causes. In: Toril Swan, Endre Merk, Olaf Jansen Westvik (eds.), Language change and language structure: older Germanic languages in a comparative perspective, 135-157. Berlin: Mouton de Gruyter. 1994b Introduction. In: Dieter Kastovsky (ed.), Studies in Early Modern English. (Topics in English Linguistics 13), 1-10. Berlin: Mouton de Gruyter. 1994c Historical English word-formation. From a monostratal to a polystratal system. In: Rolando Bacchielli (ed.), Historical English word-formation. Papers read at the Sixth National Conference of the History of English, 17-31. Urbino: QuattroVenti. 1995a Morphological reanalysis and typology: the case of the German r-plural and why English didn't develop it. In: Henning Andersen (ed.), Historical linguistics 1993. (Current Issues in Linguistic Theoiy 124), 227-237. Amsterdam: Benjamins. 1995b The syntactic aspects of word-formation: where are we today?. In: Gunnel Melchers and Beatrice Warren (eds.), Studies in Anglistics. (Stockholm Studies in English 85), 157-169. Stockholm: Almqvist & Wiksell. 1995c Allgemeine Prinzipien der Wortbildung. In: Rüdiger Ahrens, WolfDietrich Bald, Werner Hüllen (eds.), Handbuch Englisch als Fremdsprache, 104-109. Berlin: Schmidt.

xvi

Publications by Dieter Kastovsky

1995d Wortbildungssemantik, ein historischer Lagebericht. In: Ulrich Hoinkes (ed.), Panorama der Lexikalischen Semantik. Thematische Festschrift aus Anlaß des 60. Geburtstags von Horst Geckeier, 385-398. Tübingen: Narr. 1996a Verbal derivation in English: A historical survey. Or: Much ado about nothing. In: Derek Britton (ed.), English Historical Linguistics 1994. (Current Issues in Linguistic Theory 135), 93-117. Amsterdam/Philadelphia: Benjamins. 1996b Morphological reclassification: The morphological and morphophonemic restructuring of the weak verbs in Old and Middle English. In: Jürgen Klein and Dirk Vanderbeke (eds.), Anglistentag 1995 Greifswald. Proceedings. Volume 17,273-284. Tübingen: Niemeyer. 1996c The place of word-formation in grammar: A historical survey. In: Kurt R. Jankowsky (ed.), Multiple perspectives on the historical dimensions of language, 227-243. Münster: Nodus. 1996d The representation of space in English derivational morphology. In: Ren£ Dirven and Martin Pütz (eds.), The construal of space in language and thought. (Cognitive Linguistic Research 8), 197-208. Berlin: Mouton de Gruyter. 1996e Categorial restructuring of the weak verbs in Late Old English Middle English. In: Jacek Fisiak (ed.), Middle English miscellany. From vocabulary to linguistic variation, 29-45. Poznan: Motivex. 1997a Einige Überlegungen zur Wortbildungstheorie. In: Ulrich Hoinkes and Wolf Dietrich (eds.), Kaleidoskop der Lexikalischen Semantik, 305-313. Tübingen, Narr. 1997b Morphological classification in English historical linguistics: The interplay of diachrony, synchrony and morphological theory. In: Terttu Nevalainen and Leena Kahlas-Tarkka (eds.), To explain the present. Studies in the changing English language in honour of Matti Rissanen. (Mimoires de la Sociötä Ndophilologique de Helsinki 52), 63-75. Helsinki: Sociöti N6ophilologique. 1997c Sign-oriented vs. form-oriented linguistics and word-formation. Studia Anglica Posnaniensia 31: 79-90. 1997d (with Barbara Kryk-Kastovsky) Morphological and pragmatic factors in text cohesion. In: Heinrich Ramisch and Kenneth Wynne (eds.), Language in Time and Space. Studies in Honour of Wolfgang Viereck on the occasion of his 60th birthday. (ZDL-Beiheft 97), 462^75. Stuttgart: Steiner. 1997e (with Barbara Kryk-Kastovsky) How do you know what I'm talking about? On the semantics and pragmatics of referring. In: Raymond Hickey and Stanislaw Puppel (eds.), Language History and Linguistic Modelling. A Festschrift for Jacek Fisiak on his 60th Birtday. (Trends in Linguistics. Studies and Monographs 101), 1665-1679. Berlin: Mouton de Gruyter.

Publications by Dieter Kastovsky

xvii

1998a Morphological restructuring. The case of Old English and Middle English verbs. In: Richard Hogg and Linda van Bergen (eds.), Historical Linguistics 1995. Volume 2: Germanic Linguistics, 131—147. Amsterdam/ Philadelphia: Benjamins. 1998b Fremdsprachenausbildung für Europa. Universitäre Austausch- und Entwicklungsprogramme innerhalb der EU und Mittel-/Osteuropa. Erziehung und Unterricht. Österreichische Pädagogische Zeitschrift 9/10: 769-774. 1999a Hans Marchand's theory of word-formation: Genesis and development. In: Uwe Carls and Peter Lucko (eds.), Form, Function and Variation in English. Studies in Honour of Klaus Hansen, 19-39. Frankfurt am Main, etc.: Peter Lang. 1999 Inflectional classes, morphological restructuring, and the dissolution of Old English grammatical gender. In: Barbara Unterbeck and Matti Rissanen (eds), Gender in grammar and cognition. Volume 2. Manifestations of gender. Berlin: Mouton de Gruyter, 1999. 1999 English and German morphology: a typological comparison. In: Wolfgang Falkner and Hans-Jörg Schmid (eds.), Words, Lexemes, Concepts Approaches to the Lexicon, 39-51. Tübingen: Narr. 1999 On writing a history of English: The 'local' and the 'global'. The European English Messenger 8/1: 13-15. 2000 Words and word-formation: Morphology in the OED. In: Lynda Mugglestone (ed.), Lexicography and the OED. Pioneers in the untrodden forest, 110-125. Oxford: Oxford University Press.

3. Reviews 1970 Wienold, Götz, 1967. Genus und Semantik. Meisenheim an der Glan: Hain. Indogermanische Forschungen 75: 256-265. 1971a Fleischer, Wolfgang, 1969. Wortbildung der deutschen Gegenwartssprache. Leipzig: VEB Bibliographisches Institut. Indogermanische Forschungen 76: 328-338. 1971b Wagner, Karl Heinz, 1969. Generative grammatical studies in the Old English language. Heidelberg: Groos. Anglia 89: 482-492. 1972a Agricola, Erhard, 1968. Syntaktische Mehrdeutigkeit (Polysyntaktizität) bei der Analyse des Deutschen und des Englischen. (Schriften zur Phonetik, Sprachwissenschaft und Kommunikationsforschung 12). Berlin: Akademie-Verlag. Anglia 90: 180-188. 1972b Bechert, Johannes, Daniele Clöment and Wolfgang Thümmel, 1970. Einführung in die generative Transformationsgrammatik München: Hueber. Anglia 90: 363-366. 1972c Cattell, N. R., 1968. The new English grammar: a descriptive introduction. Cambridge, Mass.: M.I.T.-Press. Anglia 90:494-497.

xviii

Publications by Dieter Kastovsky

1973a Hammarström, Göran, 1966. Linguistische Einheiten im Rahmen der modernen Sprachwissenschaft. (Kommunikation und Kybernetik in Einzeldarstellungen 5). Berlin: Springer. Anglia 91: 366-370. 1973b Ljung, Magnus, 1970. English denominal adjectives.A generative study of the semantics of a group of high-frequency denominal adjectives in English (Acta Universitatis Gothoburgensis 21). Lund: Almqvist & Wiksell. Anglia 91: 505-514. 1974 Heger, Klaus, 1971. Monem, Wort, Satz. (Konzepte der Sprach- und Literaturwissenschaft 8). Tübingen: Niemeyer. Anglia 92: 185-194. 1978 Liefrink, Frans, 1973. Semantico-syntax. London: Longman. Anglia 96: 164-174. 1982 Gutch, Donald, 1975. Einführung in die anglistische Sprachwisssenschaft. (Studienreihe Englisch 2). Bern, München; Francke. / Mindt, Dieter, 1975. Moderne Linguistik. (Studienreihe Englisch 13). Bern, München: Francke. Anglia 100:442-450. 1986 Bauer, Laurie, 1982. English word-formation. London - Cambridge: Cambridge University Press. System 4: 349-352.

4. Editions 1967 Eugenio Coseriu, Probleme der romanischen Semantik. Vorlesung gehalten im WS 1965/1966 an der Universität Tübingen. Autorisierte Nachschrift besorgt von Dieter Kastovsky und Wolfgang Müller. Tübingen: Selbstverlag. 1973 Eugenio Coseriu, Probleme der strukturellen Semantik. (Tübinger Beiträge zur Linguistik 40). Tübingen: Narr. [2nd edition 1976]. 1974 Hans Marchand, Studies in syntax and word-formation. Selected articles by Hans Marchand. Published in honour of his 65th birthday. München: Fink. 1977 (with Herbert E. Brekle) Perspektiven der Wortbildungsforschung. Beiträge zum Wuppertaler Wortbildungskolloquium vom 9. - 10. Juli 1975. Anläßlich des 70. Geburtstages von Hans Marchand am 1. Oktober 1977. (Wuppertaler Schriftenreihe Linguistik 1). Bonn: Bouvier. 1981 Probleme der lexikalischen Semantik. Beiträge zum Wuppertaler Kolloquium über lexikalische Semantik vom 2 . - 3 . Dezember 1977. (Wuppertaler Schriftenreihen Linguistik 2). Bonn: Bouvier. 1986 (with Alexander Szwedek) Linguistics across historical and geographical boundaries. In honour of Jacek Fisiak on the occasion of his fiftieth birthday. 2 volumes. Berlin: Mouton de Gruyter. 1988 (with Gero Bauer) Luick revisited. Papers read at the Luick Symposium at Schloß Liechtenstein, 15. - 18. 9. 1985. Tübingen: Narr. 1991 Historical English syntax. (Topics in English Linguistics 2). Berlin: Mouton de Gruyter. 1994 Studies in Early Modern English. (Topics in English Linguistics 13). Berlin: Mouton.

Introduction Arthur Mettinger

Anybody who has met Dieter Kastovsky recently will find it hard to think of him as rapidly approaching his sixtieth birthday. This human dynamo seems to simultaneously give papers in all four corners of the world, to produce an ever growing number of books and articles (as is shown by his impressive bibliography), to deftly steer the courses of more than one department at the University of Vienna, or to be organising one of his firstat-Schloss-Liechtenstein-now-in-Tulln-conferences on various aspects of the history of the English language. Moreover, while other people after a day's or a week's work enjoy the simple pleasures of home life, Dieter and his charming wife can be seen waltzing in Vienna's most beautiful ballrooms or listening to Wagnerian tunes in Vienna's State Opera. And yet, there was a strong feeling among Dieter's friends that compiling a festschrift on the occasion of his sixtieth birthday would be a nice thing to do. It soon turned out, in fact, that the great number of eminent linguists among Dieter's friends who wanted to contribute would have made this volume about three times its size - so the editors decided to strictly limit the topics of contributions to those areas of general and English linguistics that have formed the core of Dieter Kastovsky's scholarly work for the past thirty years. The contributions in this volume thus mirror the wide scope of Dieter's interests: His work on the history of English (mainly Old English) includes also more general issues in historical linguistics such as linguistic typology, internal mechanisms of linguistic change, and historical morphology. Inflectional morphology, and particularly morphonological processes, have been treated by Dieter Kastovsky synchronically, diachronically, and from a contrastive point of view. The area that he has been strongly associated with for a long time is derivational morphology, i.e. word-formation: in a great number of descriptive and theoretical works he has dealt with various aspects of, mainly English, word-formation (formal and functional aspects, problems of productivity, links between word-formation and semantics). Moreover, in the course of time Dieter Kastovsky has become one of the most important historiographers of central European wordformation in general and Hans Marchand's work in particular. Semantics

XX

Introduction

(mainly structural but recently also conceptual) has been another recurring theme in Dieter's work and has for the past few years been complemented by a growing interest in pragmatics. At the hub of this web of work, then, there is an interest in the formal aspects of the word. From there, thematic connections branch out into the neighbouring areas of phonology, syntax and semantics. The study of word-meaning itself carries over into the lexicon and into pragmatics. Accordingly, all the contributions in this volume take up the topics mentioned above or combinations of them and elaborate them in accordance with the respective contributors' interests in one or more areas of linguistic research. In a book dedicated to an Anglicist, it is of course not surprising that the great majority of papers deal with the English language. Among these papers, the ones with a diachronic perspective take a clear preponderance - something which is appropriately iconic of Dieter Kastovsky's own bibliography. The papers that do not use English data, operate with theoretical concepts which have been an ongoing concern of the festschriftee over the years: semantics (Ahlqvist), linguistic typology (Dressier and Ladänyi, Lass), analogy (Winter), linguistic historiography (Hüllen), morphological transparency (Hickey). Almost half the papers deal with questions of semantics in one way or another: some in connection with grammaticalisation (Breivik and Swan, Rissanen) and/or language change (Ahlqvist, Görlach), some from a descriptive point of view, which may be lexical (Diller, Fischer) or syntactic (Nagle and Holmes, Tottie). Jucker establishes a link with pragmatics while Hüllen takes a historiographic interest in the beginnings of conceptual semantics. Two contributions raise questions with regard to the meaning and morphological status of names (Anderson and Colman, Lipka), relatively new territory in the study of morphology. (By virtue of the examples used, the latter author also takes the reader on a journey through the festschriftee's biography.) Related issues in the description of derivational processes are also dealt with in several papers focussing on individual word-formation elements or -patterns in present-day English (Carls, Fischer, Ljung, Stein) - these naturally touch upon the issue of morphological productivity, a concept which is central to Dressier and Ladany's paper. This paper, on the other hand, is also an exercise in contrastive word-formation, a line of interest which is also pursued by KrykKastovsky, whose own interest in pragmatics is shared by Jucker. The morphological status of items is also at issue in the contribution of Stockwell and Minkova, which straddles the borderline of syntax and morphology. Another borderline which has been a major interest of Dieter Kastov-

Introduction xxi sky's is that between phonology and morphology. Morphonological considerations, then, play a major role in Nevalainen and Raumolin-Brunberg' contribution and also in the one by Hickey. Inflection and the change of inflectional sub-systems are at issue in the papers by Hansen and Schendl. Finally, Fisiak's contribution takes us back to a central interest of an Anglicist: the interconnection between external and internal aspects in the history of the English language. Mirroring the scope and richness of Dieter Kastovsky's work, the contributions in this volume show what a central and integrating role 'the word' plays in almost all matters linguistic. Moreover, however, they want to be interpreted as a personal gift to a famous scholar, a wonderful entertainer, and a great friend. Arthur Mettinger

Old Irish düilem 'maker, creator9 Anders Ahlqvist

This little case study in word-formation and historical semantics describes the development of an interesting Irish lexical item. Its main agenda is to show that the original meaning of duilem was probably more concrete than usually suggested and that its development seems to follow a general tendency, whereby lexical items with more concrete meanings are to be viewed as primary when compared with other lexical items that have similar, but more abstract meanings. According to the standard dictionary of Old Irish (Quin 1983: 254), the normal meaning of duilem is 'Creator', with a capital C, thus obviously referring to a divine creator of some sort. Likewise, a current dictionary of Modern Irish ( 0 Donaill 1977: 461) lists it as follows: "duileamh [...] {of God) Creator". According to Thurneysen (1946: 172), duilem consists of a noun duil 'creature' and the obsolescent agency suffix -em, which is an nstem; he adds that the original 'suffix seems to have been -iamon- . The meaning of duil is described in some detail in the Old Irish dictionary (Quin 1983: 254), beginning with the observation that "in general sense" duil means "element, being, creature, thing". The (eighth- and ninth-century) Old Irish glosses (for which see further Thurneysen 1946: 4-6) on Latin texts, have duil to gloss Latin elementum, res and creatura. "Hence", the dictionary continues, it is found to refer to "created things in general, including not only material objects and living beings, but invisible and spiritual essences"; it also means "element, esp. one of the four elements common in later and scientific texts", as well as "thing in the widest sense". The etymology of duil has not been entirely clarified. According to Lambert (1996: 215), we are here dealing with an "etymologie incertaine". McCone (1994: 218), on the other hand, analyses it as deriving from IndoEuropean *ctuh- 'breath' + -//, thus cognate with Greek θύμος and Sanskrit ctümas 'smoke'. This seems to fit with duil as a translation of Latin creatura. He also maintains that duilem is a direct caique on Latin creator, in its Christian meaning. This may be correct, but must, for the time being, be regarded as not proven. The main difficulty is that, as mentioned above (Thurneysen 1946: 172, see further Kelly 1976: 43), the agent suffix -em was already obsolescent in the Old Irish period. Of the other words in -em

2

Anders Ahlqvist

mentioned by Thumeysen, none look like implying Christian origin. They are: brithem 'judge', airem 'ploughman', ddlem 'distributor', mraithem 'traitor, flaithem 'lord', fechem 'debtor', orbam 'heir' and luam 'helmsman'. In words with Christian connotations, one might rather have expected a suffix borrowed from Latin -drius (see Thumeysen 1946: 172 for its various manifestations in Old Irish). The probably earliest attestations of duil and duilem known thus far are probably those found in Morand's Testament. This text deals with kingship in early Irish society: it describes advice, given by the old judge Morann, to a young king, about how to look after his kingdom, in accordance with principles of natural justice and truth. It has, by its three editors so far, been dated as follows. Thumeysen (1917: 77) would put the date of composition around 800. Kelly (1976: xxix) has come to the conclusion that the linguistic evidence "gives us a compilation date of c. A.D. 700, though much of the text must have had a previous oral or possibly manuscript evidence". More recently, I (Ahlqvist 1984: 152) have stated that it dates at the latest from the seventh century. The passage containing duil and duilem has been edited as follows: Apair iris, ad-mestar duile duilemon doda-rösat amal do-rösata; nach r£t nad asa maTnib mTastar, nicope län-torath do bera. (Thumeysen 1917, 94) Apair fris, ad-mestar duili diiilemon tod[a]-rösat amal to-rrösata; nach r6t nad asa moinib miastar, nicope läntoruth toda-b6ra(?) [Tell him, let him estimate the creations of the creator who made as they were made; anything which he will not judge according to its profits will not give them with full increase] (Kelly 1976, 10-13) Apair fris: ad-mestar duil e duilimain todrösat amail to-rrösata; na-ch rdt nad· asa muinib miastar, nicope läntorath dö toböra. [Tell him: he must estimate an article by the artisan who has made it as they have been made; also, anything he will not estimate by goods it may be exchanged, that will not give him a complete result.] (Ahlqvist 1984, 158) Thumeysen gives no translation for the passage, but his reading is obviously based on the same assumption as Kelly's, namely that the original had had something rather like duili duilemon, in which duili is an accusative plural and duilemon a genitive singular. This is not the place for a full discussion of the numerous and complex philological issues: for that all three editions should be consulted.1 The immediately most relevant of these issues is, to my mind, at any rate, that

Old Irish düilem 'maker, creator'

3

duil (acc. sg.) e (prep, 'out of cp. Lat. ex) duilimain (dat. sg.) is based veiy closely on the actual reading (duileduilimaiw) in what nowadays is generally considered to be much the best of the manuscripts available (see further Pokorny 1919, Thurneysen 1921, Kelly 1976, and Ahlqvist 1984: 152). Turning to the actual content of the passage concerned, it seems, first of all, an unlikely notion that a divine Creator, whether pagan, as Kelly (1976: 43) has suggested, or Christian, could be regarded as a - by definition variable - measure for anything. Or, to put the same argument in slightly different terms, such a fundamental difference is implicit between a divine Creator and His creations that a direct comparison between the two would appear to serve no purpose with any conceivable practical application in real life. Furthermore, the alternative that might see a divine Creator's creations as the basis for measuring other objects would seem logically rather futile, since divine Creators by definition are viewed as having created everything that exists: the possibility of such comparisons is thereby excluded. On the other hand, it seems natural that information about the quality of manufactured items might be sought from a comparison with other items emanating from an identical source. Furthermore, this fits nicely with the further notion that a proper result is conditional on the assessment being performed in a manner appropriate to commerce in a barter economy. In support of the notion of regarding duil and duilem in a concrete and practical fashion, it may be useful to consider the fact that the passage mentioned above is followed by two series of sentences, all of which begin with admestar, and to take note of the further important fact that most of these sentences deal with concrete and practical matters relevant to everyday life, as for instance: (1)

Ad mestar asa inricib ibar [He must estimate yew by its sellable products];

and (2)

Ad mestar umae asa daingnib denib dlumaicdib [He must estimate copper by its solid, strong and hard products]. (Ahlqvist 1984: 158-159, see also Thurneysen 1917: 94 and Kelly 1976:12-13)

If duil, and the agent noun duilem formed from it, originally referred to something fairly concrete, i.e. 'material objects' and their 'makers', it then follows that the subsequent development of these two words - to mean not

4

Anders Ahlqvist

only 'thing' and 'maker', but also 'creature' and 'creator' - adheres to a well-known pattern of semantic change. Bloomfield has described it in these terms: "The surface study of semantic change indicates that refined and abstract meanings largely grow out of more concrete meanings" (1933: 429). It is entirely obvious to me that a 'maker' who has an observable physical shape must be regarded as more concrete than, and therefore also semantically primary to, a divine Creator that has no observable physical shape or concrete manifestation. This holds true regardless of the various possible stands one may adopt regarding the fundamentally theological question of whether He actually exists or not. Additionally, it seems relevant to point out that there are other signs of abstract thinking developing out of more concrete modes, in Ireland and elsewhere, during the mediaeval period. Thurneysen (1928: 279), for instance, in discussing the early Irish grammatical text Auraicept na nEces (i.e. Scholars' Primer), has provided some very interesting comments about grammatical schemes being seen concretely, "als Realitäten" [as realities], and the spread of "Abstraktionsfähigkeit" [the faculty of abstraction] into early Ireland. Finally and to sum up, this paper has tried to show that the original basic meaning of duil and dwlem is likely to have been more concrete than usually suggested in our sources. In support of this proposition, contextual and other evidence from an interesting early text has been adduced, which demonstrates the survival of the proposed original meaning in attested material. Additionally, it has been shown that this proposed original meaning suggests a semantic development of the word which follows the normal direction taken by semantic change from generally more concrete to generally more abstract senses. Note 1. Nevertheless, it may be found helpful by some readers to have the three versions analysed and glossed word by word. Thus, Thurneysen's version may be understood as follows: Apair [2 sg. ipv. 'tell!']yra [prep. pron. 3 masc. sg. 'to him'], admestar [3 sg. pres. subj. 'may he estimate'] duile [acc. pi. 'creations'] duilemon [gen. sg. 'creator's] dodarösat [3 sg. rel. perf. + 3 pi. inf. pron. 'who has made them'] amal [conj. 'as'] do rösata [3 pi. perf. pass, 'they have been made']; nach [nom. sg. 'any'] ref [nom. sg. 'thing'] nad [neg. rel. 'which not'] asa [prep. + poss. pron. 3 masc. 'of its'] mamib [dat. pi. 'riches'] mJastar [fut. 3 sg. pass, 'will be estimated'], nicope [neg. + cop. 3 sg. fut. 'will not be'] län-torath [nom. sg. 'full result'] dobera [3 sg. fut. rel. 'which it will give']. Similarly, Kelly's text may be understood (I leave unglossed those words which agree with Thurneysen's edition) thus: Apairfris,ad-mestar duili

Old Irish düilem 'maker, creator'

5

duilemon todfaj-rosat amal torrosata; nach ret nad asa moinib miastar, nicope läntoruth [dat. sg. 'with full increase] toda-bera [3 sg. fut. + 3 pi. inf. pron. 'will not give them']. Finally, mine is to be interpreted (again, I only gloss words with substantial differences) as follows: Apair fris: admestar duil (acc. sg. 'article') e [prep, 'out o f , by'] duilimain [dat. sg. 'maker, artisan'] todrosat [3 sg. rel. perf. + 3 fern. sg. inf. pron. 'who has made it'] amail to rrosata; na-ch [nom. neut. sg. 'any' + conj. 'and'] ret nad· asa muinib [dat. pi. 'exchangeable goods'] miastar, nicope läntorath do [prep. pron. 3 masc. sg. 'to him'] tobera.

References Ahlqvist, Anders 1984 Le Testament de Morand. Etudes celtiques 21: 151-170. Bloomfield, Leonard 1933 Language. New York: Holt. Kelly, Fergus 1976 Audacht Morainn. Dublin: Institute for Advanced Studies. Lambert, Pierre-Yves 1996 Lexique etymologique de l'irlandais ancien de Joseph Vendryes. Paris: Centre National de la Recherche Scientifique and Dublin: Institute for Advanced Studies. McCone, Kim 1994 An tSean-Ghaeilge agus a Riamhstair ['Old Irish and its Prehistory']. In: Kim McCone et al. (eds.), Stair na Gaeilge [History of Irish], 61219. Maynooth: Department of Old Irish. 0 Dönaill, Niall 1977 Focloir Gaeilge-Bearla [Irish-English Dictionary]. Dublin: Stationery Office. Pokomy, Julius 1919 Zu Morands Fürstenspiegel. Zeitschrift für celtische Philologie 13: 43-46. Quin, Ernest Gordon 1983 Dictionary of the Irish Language. Compact Edition. Dublin: Royal Irish Academy. Thuraeysen, Rudolf 1917 Morands Fürstenspiegel. Zeitschrift für celtische Philologie 11: 56106.

1921 1928 1946

Zu Morands Fürstenspiegel. Zeitschrift für celtische Philologie 13: 298-299. Auraicept na n-Iices. Zeitschrift für celtische Philologie 17: 277-303. A Grammar of Old Irish. Dublin: Institute for Advanced Studies.

The importance of being Leofwine Horn

£

John Anderson and Fran Colman

1. Names as a category The schema of word classes proposed in Anderson (1992, 1997) rejects the idea that names are a subclass of noun, distinguished, say, as 'proper' vs. 'common'. Rather, it is recognised there, as it has been elsewhere, that the distribution of names warrants their being seen, along with pronouns, as syntactically equivalent to (traditional) noun phrases or (more recently) determiner phrases. Indeed, determiners themselves can be seen as a device for making nouns (and their subordinates) capable of a name-like distribution: John = he = that boy. Names and pronouns and determiners allow the arguments that satisfy the valency of a predicator to fulfil a referentiality requirement. Determiners differ from the others, and specifically from pronouns, in being 'transitive', requiring complementation by a 'noun phrase', on which they confer a name-like notional character (reflected in the name-like distribution). The complement 'noun phrase' introduces elements with lexical semantic content, whereas transitive and intransitive pronouns (i.e. determiners and pronouns) are differentiated among themselves only in relation to notional categories that are typically expressed grammatically, such as definiteness, deixis, gender, number, case, person. Pronouns are 'shifters'; their reference is associated with the immediate context of discourse. Names, on the other hand, though they may not each be unique to a particular referent, are typically used as if they were. Names are more numerous than pronouns, but they are typically classifiable only along an even more reduced set of grammatical dimensions: case; gender, involving the sub-dimensions inanimate, specifically place (Torquay) vs. animate, animal (Fido) vs. human, female (Elizabeth) vs. male (Charles). And even those are scarcely so salient with respect to name-giving that a speaker of English can rely on being able to attribute the appropriate category term to an English name with an unknown referent. In this sense, names lack lexical semantic content and show minimal grammatically relevant notional categorisation. Anderson (1992, 1997) distinguishes word classes with reference to combinations of the simplex notional features Ν (referentiality) and Ρ

8 John Anderson and Fran Colman

(predicability), such that names and pronouns are unique in being characterised by sole presence of the Ν feature, as shown in the following table: Table 1. Word classes name noun {Ν} {N;P}

adjective {P:N}

verb {P;N}

finite {P}

functor {}

{N} items have full referentiality; with the others, Ν is 'diluted' by presence of the other feature or is absent; Ρ is associated with those items that may be predicators. The representations for noun, adjective and verb indicate a decreasing proportion of N, and thus of denotational discreteness and capacity to be made referential as a complement of a transitive {N}, a determiner. Representations involving both Ν and Ρ define lexical categories, and further such categories can be allowed for by second-order combinations of Ν and P; the other categories in the table are functional. Of these, {P} is realised e.g. by 'periphrastic do' in English and also morphologically; functors specify semantic/grammatical roles and may be realised adpositionally (at, by), or morphologically (case). Among {N} items, non-names are realised by pronouns or determiners, or morphologically. Names warrant their non-lexical status by virtue of the absence with them of any lexical structure, and of any associated denotational function. This status is highlighted by a study of both the sources of names and the role of names in the derivation of members of other word classes. We cannot attempt a full survey of such material here. We consider the derivational processes associated with names briefly in section 2 before turning in the following sections to the system of Old English personal names, which exhibits name-sources of different kinds which nevertheless share the 'de-lexicalisation' that we would expect to be associated with the assumption of name status, on the basis of the account of their categorisation we have just outlined. After considering the sources for the 'core' system and external constraints on their implementation, we turn to the character of by-names. In both areas name-formation brings along loss of denotational function. 2. Names, derivation and the sources of names As indicated, here we look at the role of names in lexical processes. We contend that the character of this role reflects the status of names as lacking lexical denotation. Firstly we take up derivations based on names.

The importance of being Leo/wine Horn

2.1.

9

Name-based derivation

Names participate in derivational processes creating members of other categories, notably nouns. An obvious example of such is Present-day English -wm-formation, many exemplars of which are name-based, examples such as Platonism, Thatcherism, Maoism. Here the lexical semantics of the derived noun is based on speakers' encyclopedic knowledge of the person associated with the name; the name itself has no lexical semantics to transmit. This is also true of the productive adjective formations exemplified by Kafkaesque and Wagnerian. So too with lexical items based on place names, as illustrated by Balkanize, Balkanization, alongside the personal-name-based macadamize, where again the lexical semantics of the derivative is encyclopaedia-based. The bases in both cases have only grammatical meanings to transmit, concerning whether the source of the derivative is human or a place. Some name-based derivatives are conversions: these include the class traditionally labelled as eponyms, such as sandwich, hoover, based on personal names, directly or indirectly (via corporate names). Again, the lexical semantics of the derived item is constructed from encyclopedic knowledge. And often the encyclopedic basis for the derived semantics is lost, as far as ordinary speakers are concerned, as with, say, arras - or sandwich. A name may combine with an item of a different class to form a compound which is not a name, as with peach-melba, which is a lexical noun. Some confusion has been occasioned by the further, or alternative development that some such formations can undergo, eventuating in a compound name. For instance, Carroll (1985: ch.8, §1.1) mounts an assault on Mill's (1843) position that names only refer, on the basis of the example the Willis Avenue Bridge, whose "referent must be a bridge" (p. 167). But this is a compound based on some elements of a lexical sort, whose content carries over into the compound item: as Carroll himself says (p. 168), the item is "structurally complex". And even if it is a name (as is in many such cases confirmed by dropping of the the, as in Sydney Harbour Bridge or Cumberland Street), this is problematical only if we insist that complex items must be homogeneous, so that a name may not incorporate material which enables lexical generalisations to be made. This capacity is not a property of the name as a whole. All we are saying is that such items may remain compositional. But they may not, for many speakers, or they may be recomposed, as illustrated by those denizens of Edinburgh who allegedly believe, despite the spelling, that Canongate is a kind of door for admitting large weapons or involves an allusion to a clerical way of walking.

10 John Anderson and Fran Colman

Anderson (1997: 19) suggests that examples like the younger Bill or the Bill we used to know also involve conversion to noun status; so too, presumably, in origin, the two Ronnies. But the only increase in content involves the addition of countability to the grammatical notions that can be associated with names, in recognition of loss of singleness of reference. So too, more commonly, given the conventions for name-assignment in the relevant culture, with plurals like the Smiths. Perhaps one should think rather of such examples as a 'blend' of determiner phrase and name, as a strategy for coping with non-singularity, not a typical property of names. 2.2.

Derivation of names

We have alluded to one variety of name-based name already: that resulting from the derivation of corporate names from personal names. These may be singular (Ford) or plural (Boots), but even the singular can take plural concord, reflecting that collective has been added to the grammatical notions attributable to such names. Here, then, we have a minor conversion within the class of names. The other major simple name-to-name process involves compounding: MaryBeth etc. Further 'compounding' is allowed by the possibility of combining different traditions of name-giving, such as those combined in assigning both a personal (or 'first' - or 'second', if you're Hungarian) and a family name to a single person. In origin, one element in such compounds may be an additional name, or byname, typically derived from a descriptive lexical item, such as constitutes a nickname or a label of status or occupation (see below). Other name-based names involve elements in origin lexical, such as diminutive affixes: Billy etc. Unlike with grammatical affixes, such as gender suffixes, e.g. as in Davidina, the content of lexical affixes is eminently bleachable and may be pragmatised to signal, in the case of diminutives, intimacy. This is observable even when diminutives are attached to lexical nouns: thus, a Greek child may be addressed by a parent as a meyalo peöaki 'big chx\d-diminutive\ But the marked tendency to delexification found with names accords with their lack of a denotational function. And we find a similar bleaching tendency with patronymic formations (Johnson etc.), as well as with family names based on occupation (Baker etc.). This typically occurs even with those additional names, or nicknames, which also are entirely lexically based, and chosen on the basis of denotation, as directly or ironically characterizing (Slim), or for perversity (Blue applied to a redhead). Slim may have spread a bit but still be socalled, and Young Bill may be a grandfather (with no Old Bill to partner). This reflects the dominance of reference over denotation involved in

The importance of being Leo/wine Horn 11 naming; denotation is a repair strategy for referential ambivalence, or a conceit, which will be dispensed with whenever possible. Bynames constitute a type of personal-name formation particularly pertinent to this discussion, and to the history of English onomastics: in particular, 'second names' appended to some names of people living in Anglo-Saxon England. Such names are discussed at §4 below, but may be exemplified here by Leofred Brun, ^Elfwine Mus, Leofwine Horn. The by-name in each instance is derived from a common-word cognate (respectively: brun 'brown', mus 'mouse', and horn 'horn'; for the last, Tengvik (1938: 337) suggests a nickname 'nose', or, euphemistically, 'possibly some other part of the body markedly projecting': see §4 below). Such names may appear to have lexical semantic denotation, to imply, for instance, that a person X has the attribute Y; and may seem therefore to contradict the theoretical claim at § 1 above that names are purely referential. On the contrary, as argued at §4 below, an analysis of the function of such names offers support for the theory of names promoted here. 2.3.

Systems of name-giving

We can talk of onomastic systems of varying rigidity. One kind of minimal system is where there exists a stock of traditional names which may be assigned to particular referents, usually differentiated by gender, as with many systems of personal names. Another kind is where the assignment is pre-determined by birth or marriage contract, as with many systems of family names. The fact that different systems of names have different typical sources (the name borne by a saint or god or hero, the first name of one's paternal grandfather, the family name of one's father or mother, a Latin ordinal (Septimus), an item that alliterates with one's father or mother's name, etc.) enables various inferences to be drawn, so that, for instance, a group referred to as the Smiths are likely to be related by birth or marriage (and may even include related people not called Smith). But this constitutes encyclopedic knowledge, not lexical - metalinguistic at best, rather than linguistic. Often a naming system is based on some segment or segments of the name or non-name lexicon, as with, in the first case, family names based on a place name, and, in the second, occupational names (another possibility for English family names, for instance). The Old English system of personal names shows lexical cognates restricted to certain lexical items, items which fall into certain lexical classes. As is typical, the names are de-lexified, such that jokes involving the names (such as Wulfstan's Sermo Lupi...) specifically depend on the re-lexification of an item that has no

12 John Anderson and Fran Colman denotational function. The same seems to be true of the bynames which also occur on the coins which are a major source of evidence concerning Old English personal names and their evolution. We turn now to look at these systems in more detail, focusing on name-data from Anglo-Saxon coins. These allow us to make use of evidence afforded by numismatic reconstruction. 3. Old English personal names 3.1.

Old English systems of name giving

With respect to Old English personal-name formation (see §3.2 below), a case may be made for an onomastic system that draws on a pool of elements identifiable as sourced from common words (typically declensionclass words) denoting a limited set of properties (Colman 1992: 21-25). But whether or not the rim of such a pool may be delimited precisely is dependant on both the limitations of surviving Old English source material, and the theoretical basis for classification of the items selected for conversion to personal-name elements. (For a critique of the concept of an onomastic system as presented in Colman 1992, see Coates 1993: 11831184; and further Colman 1996: 13-16.) External constraints on the naming of persons in Old English seem to have applied in various ways. At least in the earlier Anglo-Saxon period, a principle of alliteration with the names of parents and other family members seemed to have operated (Ström 1939: xxxiv ff.; Stenton 1924 [1970: 87]): "[s]o, King Alfred's name accords to some extent, with a line of brothers whose names all had the first element v®thel-" (Colman 1992: 26). Even such innovation as might be attributed to alliterative principles of name-giving appears to be On the wane by the late period, as evidenced, for instance, by the ubiquitous Godwine (Smart 1981: xv, and see §4 below), and by the name Deorman, shown by Nightingale (1982) to identify members of the same family operating as moneyers (see §3.3 below) before and after the Norman Conquest. The evidence from the latter would be in favour of a 'constraint' whereby the same name was passed from father to son. 3.2.

Old English name-formation

As intimated in what precedes, interpretation of the core system of Old English name-formation provides evidence to support the theoiy outlined

The importance of being Leofwine Horn

13

at §§1 and 2 above. Details of Old English name-formation to be found in, e.g., Clarke 1992, Colman 1984, 1992 and references therein, may be summarised here as follows. Old English personal names, formed from common words, were either monothematic (e.g. Brid, brid 'bird'; Cild, cild 'child') or dithematic (e.g. Godwine, god 'good' + wine 'friend'; Deorman, deor 'dear' + man 'man'). While such names as these dithematic ones may suggest semantic content, 'oxymoronic' names such as Daegnieht, dceg 'day' + nieht 'night', Wigfrith, wig 'war' + frith 'peace' testify to the lexemic emptiness of such formations (see further, e.g., Colman 1992: 67-69). 3.3.

Sources of Old English names

Old English personal names are recorded in a variety of sources, including manuscripts containing histories (notably that of Bede), and legal documents, runic inscriptions, and, of particular significance here, coins. The coins, made of silver, were stamped from dies engraved with the name of the king (on the obverse, or 'heads' side), and that of a moneyer, and an abbreviation of a place name (on the reverse, or 'tails' side). The moneyer was not the die-engraver, but an administrative officer responsible for the weight and purity of silver (see, in particular, Smart 1990: 445). The placename identifies the mint-town at which the coin was struck and issued; but the dies were engraved centrally at London (and possibly at York), and distributed to the forty-plus mints as required (for details of reconstructions of methods of coin-manufacture, chronological ordering of different coin-types, etc. see Sellwood 1962; Petersson 1969; Colman 1984, 1992: 16-19, 159-161). 4. Old English bynames Fellows-Jensen (1968: xxix) defines a byname as "a characterising name which was given to a man by his contemporaries and borne by him in addition to his pers[onal] n[ame]". A byname may be classified as typically a nickname (e.g. Slim, Blue, cited at §2.2. above), but might also be derived from a noun denoting status or occupation (e.g. Thegn, thegn 'servant': Colman 1992: 105). Alternatively, the critical discussion in Smart (1990: 444-445), based on an interpretation of thegn as referring to a higher class of officer, minister, a person of rank (see Bosworth-Toller 1043-1044), raises the possibility of Thegn as a patronymic ('son of a thegn'), or as a

14

John Anderson and Fran Colman

nickname bestowed jocularly, belonging in the latter case to "the class of true bynames". Two questions may be raised. Are the bynames evidence of the evolution of surnames (for which otherwise the existing material comes from Post-Conquest sources: Smart 1990: 437)? And how can the formation of bynames, as described at §2.2 above, be reconciled with a theory invoking lack of lexical semantic content for names? First of all, however, it must be established that sequences of two personal names (such as exemplified at §2.2 above) indeed represent a name plus a byname, rather than two separate names representing two different people (as suggested by the treatment in the British Museum Catalogue, where two personal-name forms on the reverse of a coin, e.g., (LIFWINE HORN) are treated as Lifwine (a form of Leofwine) and Horn: Smart 1990: 437). The earlier views of Hildebrand (1881: vii), which observe that in instances of double moneyers' names the first name is a very commonly used name (such as Godwine: §3.1 above), and are translated and summarised in Smart (1990: 437), conclude that "certain moneyers may have found it necessary to assume a byname in order not to be confused with another colleague of the same name", and refer to the second name as "an adopted surname [Italics FC]". This may appear to be an answer to the first question posed above. But, as observed by Smart (1990: 443), the chronological distribution of double names shows that the practice, at its height during the reigns of Cnut and his two sons (Harold and Harthacnut), subsequently declined during the reign of Edward the Confessor (104266). This pattern, regarded by Smart as "perplexing", suggests that the second names are not to be regarded as surnames: surnames do not 'die outSmart's arguments do, however, confirm Hildebrand's view that the second names were bynames (rather than evidence of two moneyers). A brutally brief illustration of Smart's detailed arguments might include the following. At least two coins from the late Anglo-Saxon period have forms of moneyers names where the syntactic evidence of a determiner confirms the byname status of the second name (Smart 1990: 443-444). Of significance incidental to the present discussion is the form (EDPALD DEALDA) (Hildebrand 1881: Harold I, London 597) Eadwald the Alda, of which Smart (1990: 443) observes: "the derivation is ... the OE definite article followed by eald 'old'". In the absence of any plausible etymology for a name Thealda, Smart's is the most convincing interpretation; and in this case, the coin-form supplies early eleventh-century evidence for the use of 'the', as opposed to the OE nominative masculine singular determiner form (se). This would support Traugott's (1972: 19) suggestion, that while φβ> is first attested in manuscripts of the twelfth century, "presuma-

The importance of being Leoßvine Horn

15

bly pe was used in the spoken language considerably earlier". Other arguments adduced by Smart in favour of classification of second names as bynames include the observations that, although dithematic name formations were the most common types, the second name is almost invariably a monothematic name, and that the first name is invariably more common than the second. This last observation accords with that of Hildebrand (see above). But Smart is not convinced by Hildebrand's idea that the byname served to distinguish two people with the same name, citing "[o]bverse die-links between surnamed moneyers and their namesakes" (1990: 451): that is, coins struck from the same obverse die, but different reverse dies. For example, (vELFWINE MVS) and (^LFWINE) alone, occur on coins from the same obverse die, suggesting that the same man was identified by either name plus byname, or name alone. The bynames are not, therefore, to be interpreted as evidence that, for instance "there was another quiet, sober or chaste Lifwine minting at Rochester or an unmouselike ^ilfwine at IIchester... The use of bynames seems to have become fairly widespread in late Anglo-Saxon times, due no doubt to the increasing monotony of OE name-giving. Equally ... the use of bynames at this period was essentially optional and unstable" (Smart 1990: 451). The numismatic evidence would suggest that the bynames lacked referential discrimination based on denotational function (and see again the discussion of names such as Slim at §2.2 above). Note *

Conventions adopted are as follows: - ( ) enclose spelling forms. -Coin-spellings are given in capitals, the closest representation of the coin epigraphy -Personal name forms are given in bold; cognate lexical items are given in italics -Glosses to lexical items are enclosed in single inverted commas

References Anderson, John 1992 Linguistic representation: structural analogy and stratification. Berlin - New York: Mouton de Gruyter. 1997 A notional theory of syntactic categories. Cambridge: Cambridge University Press.

16

John Anderson and Fran Colman

Bosworth, Joseph and T. Northcote Toller 1898 An Anglo-Saxon dictionary. Oxford: Oxford University Press. [Supplement by T. Northcote Toller. Oxford. 1921; Addenda and Corrigenda by Alistair Campbell, Oxford. 1972.] Carroll, John M. 1985 What's in a name? An essay in the psychology of reference. New York: W.H. Freedman. Coates, Richard 1993 Review of Fran Colman 1992: Money talks: reconstructing Old English. Linguistics 31:1183-91. Clarke, Cecily 1992 Onomastics. In: Richard M. Hogg (ed.), The Cambridge History of the English language, Volume 1, 1452-1489. Cambridge: Cambridge University Press. Colman, Fran 1984 Anglo-Saxon pennies and Old English phonology. Folia Linguistica Historica 5: 91-143. 1992 Money talks: reconstructing Old English. Berlin - New York: Mouton de Gruyter. 1996 Names will never hurt me. In: M.J. Toswell and E.M. Tyler (eds.), Studies in English language and literature: 'Doubt wisely': Papers in Honour of E.G. Stanley, 13-28. London: Routledge. Fellows-Jensen, Gillian 1968 Scandinavian personal names in Lincolnshire and Yorkshire. Copenhagen: Akademisk Forlag. Hildebrand, Bror Emil 1881 Anglosachsiska mynt funna i Sveriges jord, ny tillökt upplaga. Stockholm. Mill, John Stuart [1956] A System of Logic. London: Longman. Nightingale, Pamela 1982 Some London moneyers and reflections on the organisation of English mints in the eleventh and twelfth centuries. Numismatic Chronicle 142: 34-50. Petersson, H. Bertil 1969 Anglo-Saxon currency: King Edgar's reform to the Norman Conquest. Lund: C.W.K Gleerup. Poole, Reginald Stuart (ed.) 1887 British Museum catalogue. Anglo-Saxon coins. Volume 1 edited by Herbert Grueber and Charles Keary, Volume 2 edited by Charles Keary. London: Longmans. Sellwood, David 1962 Medieval minting techniques. British Numismatic Journal 31: 57-65

The importance of being Leo/wine Horn

17

Smart, Veronica 1981 Sylloge of coins of the British Isles: Cumulative index of volumes 1-20. London: Oxford University Press and Spink & Son Ltd. 1990 Osulf Thein and other double moneyers names on the late Anglo-Saxon coinage. In: Kenneth Jonsson (ed.), Studies in Late Anglo-Saxon coinage, 435-453. Stockholm: Swedish Numismatic Society. Stenton, Frank Merry 1924 Personal names in place names. English Place-Name Society 1, Part 1, 165-189 (reprinted in: Dorothy M. Stenton (ed.), 1970, Preparatory to Anglo-Saxon England, 84-105. Oxford: Oxford University Press.) Ström, Hilmer 1939 Old English personal names in Bede's History. (Lund Studies in English 8.) Lund: C.W.K. Gleerup. Tengvik, Gösta 1938 Old English bynames. (Nomina Germanica 4.) Uppsala: Almqvist and Wiksell. Traugott, Elizabeth Closs 1972 A history of English syntax. New York: Holt, Rinehart and Winston.

The desemanticisation of existential there in a synchronic-diachronic perspective Leiv Egil Breivik and Toril Swan

1. Introduction It is well understood that there are two kinds of there in present-day English, existential there {there\) and locative there (there-i). The two kinds are exemplified in the following sentences: (1)

There] are linguists in the English Department who are worryingly ignorant of historical syntax.

(2)

The linguist is over there^ behind the lamppost.

It is a characteristic feature of there\ that it behaves like a subject noun phrase with respect to a number of syntactic operations, e.g. inversion and subject raising. As far as semantics is concerned, native speakers of English do not feel that there\ can refer to a concrete place. That is, they do not feel that it carries the meaning, 'in the English Department' in (1). There2, on the other hand, functions as a fully fledged locative adverb, being semantically unmarked in relation to here in the deictic distinction of proximity. It is often assumed that the distinction between the two there's was less clear in the earlier stages of the language. For example, Jespersen (1969: 129) claims that "it is evident that [thereoriginated as the ordinary there, a 'pronominal adverb' meaning 'at that particular place', but in the course of time it has diverged very considerably from its origin, not only in pronunciation ..., but in other respects as well". A similar development is hypothesised by Bolinger (1977: 120), Jacobsson (1951: 211), KruisingaErades (1953: 69) and Lyons (1967: 390) inter alia. However, scholars do not agree as to when there \ separated from there2 and thus began to function as an expression without (concrete) locative reference. In some of our previous publications (e.g. Breivik 1990 and Breivik-Swan 1994), we have assumed that there 1 already functioned as an empty slot-filler in Old English. In this connection, it is interesting to note that Quirk (1951) argues that there 1 occurs in the earliest charters and poetry. According to

20 Leiv Egil Breivik and Toril Swan him, a number of unambiguous instances are to be found in Beowulf. One of his examples is: (3 )

I>aer wees sang ond sweg samod cetgcedere fore Healfdenes hildewisan ... 'There was singing and music joined together in the presence of Healfdene's warlike chieftain' (Beowulf 1063-1064)

A similar view is held by Traugott, who claims that peer in sentences like (3) functions as "an empty subject-marker rather than a true adverb" (1992: 218). Mitchell, on the other hand, finds examples like (3) unconvincing, in that "peer could anticipate line 1064" (1985: 625). He does, however, admit that Old English peer had developed some sort of introductory function. In his pioneering study of the diachrony of there i, Nagashima (1972) argues that it was well established in Old English but that it carried a much stronger demonstrative-evocative force than its present-day English descendant, a hypothesis which Mitchell (1985: 626) finds "very plausible". In the present paper, we shall re-examine the semantic relation between the two there*s in a synchronic-diachronic perspective. In doing so, it will also be necessary to discuss the syntactic role of there \ at various stages in the history of the English language.1 Our contention is that this morpheme has functioned as a non-referring expression without any demonstrativeevocative force since the earliest records. In section 2, data will be cited which support our central claim about the semantics of there \ in earlier English. We shall mainly be concerned with Old English since, as we have seen, scholars disagree as to what relation there \ bore to the fully locative pro-form there2 in this period. The meaning of there\ from Middle English onwards is less controversial; most scholars addressing this issue claim or assume that Middle English there\ functioned as an empty subject-position holder (cf. Breivik 1990: chapters 1 and 4, Fischer 1992: 235, Kohonen 1978: 98-99, 180-181). Section 3 examines our findings in the light of the concepts of grammaticalisation and subjectification, which we believe are essential to understanding the historical development of there \ as well as its role in present-day English.2 The main results of our investigation are summed up in section 4.

The desemanticisation of existential there

21

2. The meaning oithere\ in earlier English Consider first the Old English example in (4), which contains an unambiguous instance of there2. If we follow Mitchell's (1985: 625) line of argumentation (see above), it is also possible to interpret the there's in (5)-{9) as instances of there2 (since they could conceivably be co-referential with a locative adverbial in the same sentence) or, alternatively, as ambiguous between there\ and there2. Another possibility is to adopt the analysis proposed by Quirk and Wrenn, who claim that Old English peer in such sentences is a particle which "is used expletively and not with its full locative meaning ..." (1957: 93). Quirk and Wrenn's analysis is consistent with Nagashima's proposal referred to above that "the introductory peer is not so 'empty' as it is in [present-day English]; it seems to have a much stronger demonstrative force than [present-day English] there" (1972: 143). (4)

On he sippan hwearfhamweard to Babylonia. jjaer wceron cerendracan on anbide eallre worolde... 'He afterwards went home to Babylon. There ambassadors from all over the world were waiting' (Orosius 136: 3-4)

(5)

... Jjaer is sum baladung onpcere scegne. 'there is some excuse in your assertion' (vElfric 1 III: 185)

(6)

Gifö aer beob fiftig wera wunigende on pam ear de ... 'If there are fifty men living in the place' (^lfric 1 XIII: 196)

(7)

Pa geseah se halga wer pcet Jjaer scet an deofol on pcere cu hryege 'then the holy man saw that there sat a devil on the cow's back' OElfric 2 XXXI: 1046-1047)

(8)

And f>aer wceron on pam carcerne twa hund and eahta and feowertig wera, and nigon andfeowertig wifa, da se haliga Andreas panon onsende. 'And there were in the prison two hundred and forty-eight men and forty-nine women, whom the holy Andrew sent from that place'(Bückling 293: 13-15)

(9)

J)aer bid swyde mycel gewinn betweonan him. 'There is very much strife among them' (Orosius 20: 17-18)

However, the hypothesis that Old English there\ has retained some of its concrete locative meaning and demonstrative force becomes less con-

22 Leiv Egil Breivik and Toril Swan vincing if we extend the database to other relevant sentences. Note, first of all, that there\ does not occur in the following examples: (10)

On dcere byrig wees sum pegn bonifacius gehaten ... 'In that city there was a certain noble, named Bonifacius' (vElfric 1 XXII: 72-73)

(11)

is ponne on western medmycel duru pcet mannes heafod ge pa sculdro magan in... 'There is on the western side a moderate-sized door, through which a man's head and shoulders may enter' (Bückling 127: 8 10)

To judge from our material, it is problematic to interpret there \ in sentences like (5)-(9) as a meaningful element which contrasts with its absence in sentences like (10)—(11). Just like (5)-{9), (10)—(11) are existential sentences, that is, they express propositions concerning existence. We propose that the rationale for the use and non-use of there \ in these sentences is to be sought in syntax rather than in semantics. A few general remarks on Old English syntax are therefore in order. It is widely recognised that Old English had some sort of verb-second constraint (cf. Pintzuk 1995, Stockwell 1977, Stockwell and Minkova 1991, Vennemann 1974 inter alia). It should be pointed out, however, that Old English was far from consistent in this respect: declarative main clauses could also exhibit verb-first and verb-third order. As shown by Breivik (1990: chapter 4), the verb-second pattern exhibited by (10) is the prototypical constituent order of existential sentences in Old English texts; only 15 per cent of such sentences contain there(Verb-initial sentences like (11) are extremely rare.) Our claim is that when there \ does occur in Old English existential sentences, it merely functions as a slot-filler. In sentences like (5), there\ is inserted as an empty topic to move the finite verb into second position, just like its Modern German equivalent es as in Es sind viele hier gewesen (cf. Hier sind viele gewesen).3 We would claim that the Old English empty topic there \ was syntactically reanalysed as an empty subject when English changed typologically from verb-second (TVX) to verb-medial (SVX), and that we see the germ of this process in Old English. Thus we occasionally find Old English there ι in contexts where the verb-second constraint does not obtain, e.g. in example (7) where it appears in a subordinate clause. The use of there ι became more and more frequent towards the end of the Middle English period, as a result of the increasing fixation of the SVX syntax (cf. Breivik 1990: chapter 4).4

The desemanticisation of existential there

23

It is interesting to note that there \ and zero are often used interchangeably in the same existential sentences in Old and Middle English manuscripts of the identical text. Again, this suggests that there\ was not felt to be a meaning-bearing element. The extracts from the Lauderdale MS and the Cotton MS of Orosius in (12) provide an example of the alternation of there\ and zero. Sweet's (1883) edition (referred to as Orosius in the present paper) is based on the Lauderdale MS, while Bosworth's (1859) edition is based on the Cotton MS. (12)

a. Ac sippan Scipia geascade pat pa foreweardas wceron feor dam fastenne gesette, 7 eac pat J)aer none odre near naran, he pa diegellice geladde his firdbetuhpam weardum... 'But when Scipio learned that the scouts were set far from the fastness, and also that there were no others nearer, he then secretly led his army between the warders' (Orosius 200: 11-14) b. ... pat none odre near naran... (Bosworth 95: 13)

The existential sentence in (12a) contains par (=there\) while this morpheme is absent in the corresponding construction in (12b), which apparently expresses the same proposition as its counterpart in (12a). This brings us to our next point. The hypothesis that there \ functioned as a non-referring expression in earlier English receives further support from the fact that it was interchangeable with the classic dummy form it. The examples in (13>—(16) illustrate the use of existential it in Old English and Early Middle English. Existential it, which was quite common in both Old and Middle English, lingered on into Modern English in archaic ballad style (cf. Visser 1970: 42-43). (13)

... forpon hit was an geleafa & an hiht on pa halgan prynesse ar Cristes tocyme... 'because there was one belief and one hope in the Holy Trinity before Christ's advent' (Bückling 81: 26-27)

(14)

Pa weard hit swa mycel age fram pam here, pet man ne mihte gepeoncean ne asmagian hu man of earde hi gebringon sceolde 'Then there arose so much terror from the force that men could not consider nor conceive how they might be driven from the land' (Chronicles 1006: 32-34)

(15)

be liht and hit wes liht. 'be light and there was light' (Homilies 139: 15)

24 Leiv Egil Breivik and Toril Swan (16)

Ac hit bied sume de to michel par of penced ... 'But there are some who think too much of them' (Vices 137: 20-21)

Before proceeding to a discussion of other data, we shall briefly substantiate the claim made above that it could be used as a semantically empty subject in earlier English. There can be no doubt that Old English hit 'it' is merely a slot-filler in meteorological statements like hit rind 'it is raining'. In such cases hit has no identifiable meaning of its own and it is not in paradigmatic contrast with other noun phrases. Note also that dummy it was by no means obligatory in Old and Middle English (cf. Mustanoja 1960: 131, 433; Visser 1970: 4; Wahlen 1925: 10-11). Wahlen (1925: 9) claims that hit never had any semantic function whatever in constructions of the types: hit rind, hit is stille, and that its secondary occurrence here was due to a wish for conformity with the enormous number of sentences containingfromthe outset a morphologically detached subject, in this case with a real semantic function, and a predicate. Wahlen's claim is supported by the fact that the use of a formal subject in meteorological statements is a comparatively recent innovation in IndoEuropean languages. Brugmann (1917: 1) writes: "Nur germanische und romanische Sprachen kennen dieses pronominale Element in dieser Verwendung als echt einheimische Erscheinung. Wo auf slavischen Gebiet ono 'es' so gebraucht auftritt..., beruht das nach aller Wahrscheinlichkeit auf Nachahmung der deutschen Ausdrucksweise". As stated above (note 4), many Romance dialects do not have a formal subject corresponding to English it (cf. Spanish llueve 'it is raining'). It is also worth observing that Old English hit rind is synonymous with Old Norse rignir (cf. Modern Norwegian det 'it' regnar), Gothic rigneip and Latin pluit (cf. Modern French il pleut 'it is raining'). Consider in this connection the following extracts from the Lauderdale MS and the Cotton MS where existential it is commutable with zero: (17)

a. ALfter pcem for Hannibal ofer Bardan pone beorg, peh pe ymb pone tieman wceren swa micel snawgebland swa pcette cegper ge para horsa fela forwurdon ge pa elpendas ealle buton anum, ge pa men selfe uneade pone ciele genceson. 'Afterwards Hannibal went over the mountain Barda, although there was about that time so great a snowstorm that many of the

The desemanticisation of existential there 25 horses persished, and all the elephants but one; and the men themselves could hardly bear the cold' (Orosius 186:33-34, 188:1-2) b. ... pehpe hitymbepone timan wceron swa micel snaw-gebland ... (Bosworth 90:5-6) The relevant construction in (17a) has no existential particle, while hit (= existential it) has been inserted in (17b). In view of the above, it is perhaps not surprising that there\ could be used instead of it in non-existential sentences. Visser (1970: 52) points out that there\ "often takes the place of older English it in such sentences as 'it behoueth but one stroke' [> 'ther behoueth but one stroke wel sette' (Caxton, Jason 22)]...". Finally, consider the following examples which both refer to the spatial position of an object: (18)

Also ther was ther an jnbacetour fro the Kynge of Scottys who is now put in grete trobyll be hys son and other of the lordys ofhys londe. (Paston Letters 411:27-29)

(19)

For there is here no woman that I wyll abyde with ... (Caxton 42:34)

In (18) the post-verbal ther (= there2) denotes distal space in relation to the writer, while in (19) here denotes proximal space. The data in (18)(19) are consistent with our general hypothesis about the function of there \ in earlier English. The fact that there γ co-occurs with the fully fledged locative adverbs therei and here makes it unlikely that it functions as a term in the deictic distinction of proximity. Or, to be more specific, it is difficult to interpret there\ in (18)—(19) as a referring expression which implies contrast with the proximal term here. The evidence, then, is overwhelming that there \ has functioned as a non-referring expression since Early Old English. Nonetheless, as was stated in the introduction, it is evident that there\ originated as /Aere2· Danish, Norwegian, and Swedish are, genetically, closely related to English. It is no doubt significant that these languages also have an existential particle that is formally identical with a locative adverb; cf. Danish and Norwegian der\!deri and (Southern) Swedish där\/där2. A possible explanation for how there \ split off from there2 is to be found in Western (1921). In his discussion of what he calls the skinnsubjekt 'sham subject' in Norwegian, Western claims (1921: 61) that der\ (= there 1) ultimately

26 Leiv Egil Breivik and Toril Swan derives from the fully stressed locative adverb deri, as in (20). In order to make the referent of deri explicit, a co-referential locative adverbial was often added, as in (21). According to Western, the addition of a second locative led to a semantic reanalysis of der2: der became redundant as an indication of place; it lost its stress as well as its locative meaning. In Western's view, it is only a short step from the use of der as a deictic adverb in (21) to its use as a dummy subject in (22). (20)

Der2 bor en gammel mann. 'There2 lives an old man'

(21)

Der2 bor en gammel mann, i det huset. 'There2 lives an old man, in that house'

(22)

Derj bor en gammel mann i det huset. ' There ι lives an old man in that house'

Western does not deal with the equivalents of der\ and deri in other languages. However, mutatis mutandis, his hypothesis about Norwegian der\ offers a plausible explanation for how there\ came to be employed as an empty topic/subject in English. Needless to say, the separation must have occurred before the Old English period.

3. Therei as an example of subjectification in grammaticalisation The diachronic patterns dealt with in the present paper illustrate the processes of grammaticalisation and subjectification as described by Traugott, who has discussed the relation between syntactic and pragmatic-semantic change in a number of publications (Traugott 1989, 1995, 1996; Traugott and König 1991; Hopper and Traugott 1993; inter alia). She defines (1995: 32) grammaticalisation as "the process whereby lexical items or phrases come through frequent use in certain highly constrained local contexts to be reanalysed as having syntactic and morphological functions, and once grammaticalised, continue to develop new grammatical functions ...". An example of grammaticalisation is the development of be going to/gonna as a marker of prospective temporality in English. In certain contexts, the progressive form of the main verb go plus a subordinator introducing a clause of purpose came to function as an auxiliary. Once the reanalysis had occurred, be going to could undergo changes typical of auxiliaries such as phonological reduction. In gonna three morphemes have been reduced into one.

The desemanticisation of existential there

27

Subjectification has been defined in various ways. Traugott regards subjectification as a pragmatic-semantic process whereby meanings become increasingly based in the speaker's belief or attitude towards the proposition. Subjectification in grammaticalisation, then, "is, broadly speaking, the development of a grammatically identifiable expression of speaker belief or speaker attitude to what is said" (1995: 32). In this connection, it should be mentioned that there is widespread agreement among linguists who study grammaticalisation from a pragmatic-semantic perspective that it tends to be accompanied by an increase in pragmatic significance and subjective expressiveness (cf. Burridge 1995, Conradie 1995, Heine et al. 1991, Pagliuca 1994, Stein 1995 inter alia). If our hypothesis concerning the origin of there \ is correct, the (pre-Old English) separation of there \ and there2 is an example of grammaticalisation. The change hypothesised in section 2 illustrates many of the factors typically involved in grammaticalisation. (i) The change occurred in a very local context, as for instance that of (21). (ii) The shift from a locative adverb to an empty topic involved reanalysis. As pointed out by Langacker (1977) and Traugott (1995), reanalysis is the most widespread mechanism leading to grammaticalisation. (iii) Although the process of grammaticalisation must have started very early, traces of the first stages are present in Modern English, as there\ and there2 are still identical in their written form, (iv) The original semantics has been lost and there\ can no longer refer to a concrete location. Semantic loss or bleaching is a characteristic feature of grammaticalisation. The subsequent history of there 1 also illustrates a typical case of grammaticalisation. In Old English, there \ was very constrained with respect to the syntactic operations it could undergo, usually occurring as an empty slot-filler in the TVX structure (cf. section 2). It would appear that the subject function of there\ developed from its topic function, by extension of TVX —• SVX. Once there\ had been reanalysed as a subject, it was generalised to contexts which were unavailable before. For example, our material from Late Middle English contains sentences where there \ has undergone subject raising. It is also worth mentioning that present-day English there\ resembles Traugott's prototypical example of grammaticalisation, be going to, in that it has undergone phonological reduction; while there2 is usually pronounced /öea[r]/, the normal pronunciation for there 1 is /öa[r]/. Equally important, our data provide evidence in support of Traugott's hypothesis of unidirectional increase in subjectification over time. Through repeated use in local syntactic contexts (preceding a subject noun phrase conveying new information), there\ has increasingly acquired the

28 Leiv Egil Breivik and Toril Swan status of a preservative signal. Today the speaker uses it as a signal to tell the addressee that she must be prepared to direct her attention towards an item of new information (cf. Breivik 1981 and 1990: chapter 3). Thus there \ has increasingly come to serve pragmatic, interpersonal, speakerbased functions, whereas in the earliest stages of English it had a purely syntactic function, as a topic-/subject-position holder. The semantic development of there \ illustrates yet another factor which is often involved in grammaticalisation: "One kind of semantic-pragmatic unidirectionality often associated with grammaticalization is increase in abstractness" (Traugott 1996: 183). An example of this is gonna, which is less concrete than go. Although the original semantics of there\ has been lost, this loss has been balanced out by the development of a more abstract meaning. Our data suggest that this abstract meaning has developed during the Modern English period. There is no indication in our material that therej had a pragmatic, speaker-based function in Old and Middle English. As stated in the introduction, there \ and there-i have different categorial status in present-day English: the former functions as a subject noun phrase, while the latter is a fully fledged adverb. There are also important semantic and pragmatic differences between the two. Nonetheless, we would claim that the bond between them has not been completely severed. Although there\ has undergone desemanticisation and can no longer refer to a concrete location, it still carries some locative meaning. However, the location in question must be regarded as an abstraction and metaphorical extension. There\ designates what Fauconnier (1985) calls a mental space, i.e. a space where conceptual entities are located. This is precisely why there ι can be used as a presentative signal. Its function is to bring something into the addressee's awareness. By contrast, there2 in example (2) designates a concrete location in the presence of the speaker. Our account of the meaning of present-day English there\ is very much in the spirit of Bolinger's (1977) and Lakoffs (1987) proposals. Bolinger (1977: 93-94) maintains that there\ designates the 'awareness' of 'bring into awareness' and that 'awareness' is an 'abstract location'. A similar proposal is made by Lakoff, who substitutes the concept of mental space for awareness: "A mental space is a medium in which thoughts occur and in which conceptual entities are located" (1987: 542). It should be pointed out that neither Bolinger nor Lakoff examines existential constructions in a diachronic perspective.

The desemanticisation of existential there

29

4. Conclusion In the preceding sections, we have examined the relationship between there \ and therei, with particular reference to the semantics of there j. Our account, which makes crucial reference to the concepts of grammaticalisation and subjectification, offers a principled explanation for the historical development of therei, as well as for its function in present-day English. Our data show that there \ and therei were already distinguished in Old English in that the former could be used without being in paradigmatic contrast with the proximal adverb here or indeed without having any concrete referential meaning at all. We have hypothesised that the separation of the two there's represents the first stage in a process of grammaticalisation. As a result of semantic and syntactic reanalysis, there \ came to be employed as an empty topic in Old English. In the second stage (Old and Middle English), there \ underwent further grammaticalisation in that it was reanalysed as a subject (empty subject-position holder) by extension of TVX —> SVX and then generalised to contexts which were unavailable before. Finally, the syntactic development has been accompanied by an increase in subjectification; there\ has come to serve pragmatic, interpersonal, speaker-based functions, as an item signalling new information (subjectification in grammaticalisation). Thus, in present-day English, there\ and thereι play distinct roles, as a dummy subject/pragmatic marker and as a non-proximal locative adverb respectively. Nonetheless, it would appear that there\ has not undergone complete desemanticisation; it retains a vestige of spatiality. However, the location which there \ designates must be regarded as an abstraction and metaphorical extension. Or, to put it in cognitive terms, it designates a mental space where conceptual entities are located. Increase in abstractness is one kind of semantic-pragmatic unidirectionality which is often associated with grammaticalisation. Notes 1. For a full discussion of the syntactic development of therei, see Breivik (1990). 2. Section 3 is a revision of Breivik (1997: 40-43). The concepts of grammaticalisation and subjectification, which play a prominent role in this section, are also discussed in Swan (1997, 1999). 3. For a discussion of the syntactic similarities between Old English and Modern German, see Vennemann (1974). Vennemann does not address the use of empty topics in the two languages. 4. Our historical data are compatible with Haiman's (1974) hypothesis about the development of dummy subjects. He argues that obligatory subject pronouns

30

Leiv Egil Breivik and Toril Swan (including dummy subjects) are the property of languages which either have or have had the verb-second constraint (type A languages). Type A languages comprise all the Germanic languages plus French and Romansh. While the Germanic languages are type A languages by heredity, French and Romansh are type A languages by borrowing from Germanic dialects. Haiman's hypothesis is that we begin with a language having underlying verb-initial (VSX) order. In order to satisfy the verb-second constraint, various constituents of the sentences can be fronted. If nothing can be fronted, as in impersonal passives, a dummy pronoun is inserted to move the verb into second position (cf. Es wird gesungen). From this stage the language eventually develops SVX order, and dummy subjects begin to appear in positions where the verb-second constraint does not require them. Thus, in Modern English, hardly any traces remain of the original connection between there \ and the verb-second constraint. Although the sequence of events Haiman hypothesises certainly seems plausible, we would like to emphasise that we do not necessarily subscribe to the theoretical framework he uses to explain this sequence. We shall return to the development of dummy subjects in type A languages in connection with our discussion of the of use existential it in sentences like (13)-(16).

Primary texts iElfric 1 = Mlfric's Lives of saints, Vol. I.i, edited by Walter W. Skeat (London: EETS, 1881); Vol. I.ii, edited by Walter W. Skeat (London: EETS, 1885). Reprinted as one volume 1966. iElfric 2 = /Elfric's Lives of saints, Vol. Il.i, edited by Walter W. Skeat (London: EETS, 1890); Vol. Il.ii, edited by Walter W. Skeat (London: EETS, 1900). Reprinted as one volume 1966. Bückling = The Blickling homilies, Vol. I, edited by Richard Morris (London: EETS, 1874); Vol. II, edited by Richard Morris (London: EETS, 1876); Vol. Ill, edited by Richard Morris (London: EETS, 1880). Reprinted as one volume 1967. Bosworth = King Alfred's Anglo-Saxon version of the compendious history of the world by Orosius, edited by Joseph Bosworth (London: Longman, Brown, Green, and Longmans, 1859). Caxton = The book of the knight of the tower (translated by William Caxton), edited by M.Y. Offord (London: EETS, 1971). Chronicles = Two of the Saxon chronicles parallel, Vol. I, edited by Charles Plummer - John Earle (Oxford: Oxford University Press, 1892). Homilies = Old English homilies and homiletic treatises of the twelfth and thirteenth centuries, Vol. I.ii, edited by Richard Morris (London: EETS, 1868).

The desemanticisation of existential there

31

Orosius = King Alfred's Orosius, Vol. I, edited by Henry Sweet (London: EETS, 1883). Paston = Paston letters and papers of the fifteenth century, Vol. I, edited by Norman Davis (London: Oxford University Press, 1971). Vices = Vices and virtues, Vol. I, edited by Ferdinand Holthausen (London: EETS, 1888).

References Andersen, Henning (ed.) 1995 Historical linguistics 1993. Amsterdam: Benjamins. Bolinger, Dwight 1977 Meaning andform. London - New York: Longman. Breivik, Leiv Egil 1981 On the interpretation of existential there. Language 57: 1-25. 1990 Existential there. A synchronic and diachronic study. 2nd ed. Oslo: Novus. 1997 There in space and time. In: Heinrich Ramisch and Kenneth Wynne (eds.), Language in time and space: Studies in honour of Wolfgang Viereck on the occasion of his 60th birthday, 32-45. Stuttgart: Franz Steiner Verlag. Breivik, Leiv Egil and Toril Swan 1994 Initial adverbials and word order in English with special reference to the Early Modern English period. In: Dieter Kastovsky (ed.), Studies in Early Modern English, 11-43. Berlin - New York: Mouton de Gruyter. Brugmann, Karl 1917 Der Ursprung des Scheinsubjekts 'es' in den germanischen und den romanischen Sprachen. Leipzig: B. G. Teubner. Burridge, Kate 1995 Evidence of grammaticalization in Pennsylvania German. In: Andersen (ed.), 59-75. Conradie, J. 1995 On subjedification in modal adverbs. In: Henning Andersen (ed.), 93-103. Amsterdam: Benjamins. Fauconnier, Gilles 1985 Mental spaces. Cambridge, Mass.: MIT Press. Fischer, Olga 1992 Syntax. In: Norman Blake (ed.), The Cambridge history of the English language. Volume 2, 207-408. Cambridge: Cambridge University Press.

32

Leiv Egil Breivik and Toril Swan

Haiman, John 1974 Targets and syntactic change. The Hague: Mouton. Heine, Bernd, Ulrike Claudi and Friedrike Htlnnemeyer 1991 Grammaticalization. A conceptual framework. Chicago: University of Chicago Press. Hopper, Paul J. and Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Jacobsson, Bengt 1951 Inversion in English with special reference to the Early Modern English period. Uppsala: Almqvist & Wiksell. Jespersen, Otto 1969 Analytic syntax. New York: Holt, Rinehart and Winston. [First published 1937.] Kohonen, Viljo 1978 On the development of English word order in religious prose around 1000 and 1200 A.D.: A quantitative study of word order in context. Abo: The Research Institute of the Abo Akademi Foundation. Kruisinga, E. and P. A. Erades 1953 An English grammar. 8th ed., Volume 1. Groningen: Noordhoff. Lakoff, George 1987 Women, fire, and dangerous things. What categories reveal about the mind. Chicago: University of Chicago Press. Langacker, Ronald W. 1977 Syntactic reanalysis. In: Li (ed.) 1977, 57-139. Li, Charles N. (ed.) 1977 Mechanisms of syntactic change. Austin - London: University of Texas Press. Lyons, John 1967 A note on possessive, existential and locative sentences. Foundations of Language 3: 390-396. Mitchell, Bruce 1985 Old English syntax. Volume 1. Oxford: Clarendon. Mustanoja, Tauno F. 1960 A Middle English syntax. Part 1. Helsinki: Soci&0 N^ophilologique. Nagashima, Daisuke 1972 A historical study of the introductory there, Part I. Studies in Foreign Languages and Literatures 8: 135-168. Pagliuca, William (ed.) 1994 Perspectives on grammaticalization. Amsterdam: Benjamins. Pintzuk, Susan 1995 Verb-seconding in Old English. In: Andersen (ed.), 387-399. Quirk, Randolph 1951 Expletive or existential there. London Mediaeval Studies 2:32.

The desemanticisation of existential there

33

Quirk, Randolph and Charles L. Wrenn 1957 Art Old English grammar, 2nd edition. London - New York: Methuen. Stein, Dieter 1995 Subjective meanings and the history of inversions in English. In: Dieter Stein and Susan Wright (eds.), Subjectivity and subjectivisation. Linguistic perspectives, 129-150. Cambridge: Cambridge University Press. Stockwell, Robert P. 1977 Motivations for exbraciation in Old English. In: Li (ed.) 1977, 291314. Stockwell, Robert P. and Donka Minkova 1991 Subordination and word order change in the history of English. In: Dieter Kastovsky (ed.), Historical English syntax, 367-408. Berlin New York: Mouton de Gruyter. Swan, Toril 1997 From manner to subject-modification. Journal of Nordic Linguistics 20: 179-195. 1999 A feast of senses: Rhetorical devices in the prose of Salman Rushdie. Nordlit 6. Traugott, Elizabeth Closs 1989 On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65: 31-55. 1992 Syntax. In: Richard M. Hogg (ed.), The Cambridge history of the English language, Volume 2, 168-289. Cambridge: Cambridge University Press. 1995 Subjectification in grammaticalisation. In: Dieter Stein and Susan Wright (eds.), Subjectivity and subjectivisation. Linguistic perspectives, 31-54. Cambridge: Cambridge University Press. 1996 Grammaticalization and lexicalization. In: Keith Brown and Jim Miller (eds.), Concise encyclopedia of syntactic theories, 181-187. Kidlington, Oxford: Pergamon. Traugott, Elizabeth Closs and Ekkehard König 1991 The semantics-pragmatics of grammaticalization revisited. In: Elizabeth Closs Traugott and Bemd Heine (eds.), Approaches to grammaticalization, 189-218. Amsterdam: Benjamins. Vennemann, Theo 1974 Topics, subjects, and word order: From SXV to SVX via TVX. In: John M. Anderson and Charles Jones (eds.), Historical linguistics. Proceedings of the First International Conference on Historical Linguistics, Volume 1, 339-376. Amsterdam - Oxford: North-Holland. Visser, Fredericus Th. 1970 An historical syntax of the English language. Volume 1. Leiden: E.J. Brill.

34

Leiv Egil Breivik and Toril Swan

Wahl6n, Nils 1925 The Old English Impersonalia. Gothenburg: Elander. Western, August 1921 Norsk riksmäls-grammatikk. Kristiania [Oslo]: Aschehoug.

Anti- and pro- in Indian English Uwe Carls

Indian English as one of the oldest among the new national varieties of English is clearly marked by specific features in vocabulaiy, pronunciation and grammar. In the course of the adaptation process of English to the new natural, cultural and social environment, designations had to be found for naming new or at least prototypically different phenomena, properties or relations. In addition to this, many lexical Indianisms have come into being whose coinage was motivated by a number of other factors. Among them there is a marked tendency for a deliberate nationalisation of Indian English by using loans from indigenous languages (mainly from Hindi) for denotata of pan-Indian dimensions (cf. Doordarshan 'the official Indian TV station', Bharat Bandh 'a nationwide strike'), a trend towards a higher degree of regularity and explicitness in word-formation, as in Indian English upliftment (with the word-class marker -ment) for English English uplift, car-lchild-/cycle-!sheep-lifter (on the analogy of English English shop-lifter), or maid-runner for English English procurer and a certain preference for 'stilted' expressions as in Indian English ward for child (cf. Carls 1994: 205-207). Apart from borrowing and the transfer of already existing English designations to new denotata, word-formation is by far the most frequently used method for the creation of new lexical items. Indian English makes full use of all kinds of word-formation available in the system of English: There are determinative compound nouns (beer-bottle 'a bottle of beer', booth-capturing 'the stealing of ballot-boxes'), prefixations {anti-Dalit 'a member of the former untouchable class in India'1, delink 'to remove links'), suffixations (rewardee, debtful 'indebted', history-sheeter 'a criminal; i.e. someone who has a histoiy sheet, i.e. a police record'), zeroderivations {four-twenty 'a scoundrel, swindler, cheat, i.e. someone to whom article twenty-four of the penal code applies'), namaskar, v. 'to perform a namaskar, i.e. a greeting by bowing the head and putting palms of hands together'), backderivations (force-land, v. < force-landing), blends (trishaw < tricycle + rickshaw), clippings (Pak < Pakistani, tat < tatty·, o f f , n. < day o f f ) and acronyms (AIR = All India Radio). Theoretically, the peculiarities of a national variety of English in the domain of word-formation may be of three different types:

36

Uwe Carls

1. They may consist in new models. That is new formal patterns and/or new types, i.e. semantic structures assigned to these new patterns or to already existing ones. New models are relatively rare in the national varieties. In Indian English a case in point is the suffix-like word-formation element -walla(h)/-wala(h), which is of Indian origin (Hindi suffix wala 'doer' New Oxford) and which has found its way (via Anglo-Indian) into Indian English. It is marked as a "combining form" (cf. Chambers, New Oxford, Collins) with the types 'someone employed in, or concerned with, a specific type of work; someone who occupies an eminent position in an organisation, etc.' {Chambers) or, rarely, 'inhabitant of...' cf.: boxwalla 'tradesman' (Little Oxford) competition wallah 'a member of the Indian Civil Service who obtained appointment by the competitive system instituted in 1856' (Chambers) Congresswallah (56/9/91/27)3 Bombay-wallah (New Oxford) A new word-formation type is represented by a determinative compound like beer-bottle 'a bottle of beer'. The meaning of this type ('N2 filled with N l ' ) constitutes a new type as compared to English English because this semantic structure is usually not assigned to a compound of the model "Nl + N2" but to an «^construction. 2. A national variety may form new individual words on the basis of already existing models and types, which is obviously what most frequently happens, and 3. It may give preference to certain models or types, though the individual coinage itself is not necessarily variety-specific (cf. Carls 1999: 149ff.; Hansen 1991b: 35ff.). In the present paper, an attempt is made to examine the prefixes antiand pro- with regard to these theoretical possibilities. The words formed by means of these prefixes are antonymous and represent parallel models and types. The material stems from my own corpus of Indian English lexical items, which is mainly based on newspapers and magazines (cf. list of sources). It also includes Hawkins's glossaiy (cf. Hawkins 1990) and the lexical items listed in LABE as well as in other dictionaries and reference books. As most words with prefixes of Latin origin, words with the prefixes anti- and pro- should be considered pseudo-prefixations on the basis of their semantic structure (cf. Marchand 21969: 134 ff.; Hansen et al. 1990: 68ff.). In a word like pro-national the prefix pro- has the function of a

anti- and pro- in Indian English

37

preposition (= forixn favour of) which only refers to the underlying noun (N) or proper noun (Nprop), i.e. to nation in pro-national, and not to the whole stem -national. The suffix {-at) is a mere formal device functioning as a marker of the word-class of adjective. Thus, the morphological structure that corresponds to these semantic relations should be "(pro- + N) + adjectival suffix". However, most of these adjectives seem to be formed on the analogy of Anglicised forms of corresponding words in Latin or Greek (cf. Hansen et al. 1990: 69). For this reason and because they have a nonprefixed counterpart in most cases (national in our example), they shall be treated like prefix derivatives. These Latinate prepositional prefixes are a convenient means of integrating prepositional phrases as prenominal modifiers into a noun phrase (cf. anti-Congress line (Congress = Congress Party) (56/9/91/42), pro-ban lobby (527/12/92/35)), which might be one reason for their frequent use in journalistic style. With regard to the three types of peculiarities in a national variety that are theoretically possible, the analysis of the anti- and pro- combinations showed the following results: 1. There was no specifically Indian English model or type of anti- or procombinations. The lexemes anti-worker/-working class, pro-democracy (camp) (/El9/9/95/11) and pro-national liberation movement (stand) (NA1/10/89/12), for instance, are completely on a par with the system of English word-formation, though English English usage seems to prefer anti-labour, pro-democratic, and pro-nationalist(ic)(Collins). In a few instances, pro- seems to be unnecessary from a strictly semantic angle; cf. the bold words in the following examples: a This is true despite the presence in their ranks of certain individuals with rightist pro-feudal and even anti-secular sympathies. (AH23/10/88/15) b. ... that the Planning Commission is now divided between propublic sector and pro-private sector protagonists: proliberalisationists and anti-liberalisationists: pro-Centre and pro-States factions. (7^8/1/88/8) In these examples, the pro- formations could be replaced by feudal sympathies, public sector and private sector protagonists, liberalisationists and, in view of the omnipresent antagonisms between the Centre ('the central government of the Union') and the States ('the individual states of the Union'), Centre and States factions would have done, too. In these sentences, pro- and anti- seem to have a stylistic function. On grounds of their antonymous character they are apt to function as the rhetorical figure

38

Uwe Carls

of antithesis, highlighting the contrasted items. In the remaining examples (pro-public sector versus pro-private sector and pro-Centre versus proStates), pro- has essentially the same focussing function. This time, however, the antonymous character of the words rests on the word-formation stems. 2. There are, however, quite a number of new individual coinages in Indian English. Their specifically Indian English character is evident when antiand pro- are prefixed to stems that themselves represent Indian English lexemes or sememes; cf.: anti-Naxalism (operation) ('Maoist agrarian terrorism') (7/723/3 /96 /10), anti-Congress (line) (£6/9/91/42,); pro-zamindari (elements) ('the system of land-tenure and taxation involving zamindars, i.e. landowners who lease their land to tenant farmers') (7710/2/89/Sect2/3), pro-Congress (organisations) (513/12/91 /73f.) The specifically Indian character of a number of anti- and proformations with a variety-neutral stem like anti-land reforms, anti-worker or pro-negotiation (cf. below) is still open to further research. They were represented in the material underlying this paper, and their potential "Indianness" is based on the fact that they do not occur in the London-Lund, Brown or London-Oslo-Bergen text corpora or in the dictionaries used as references. In the following I will describe the models and semantic types of the adjectives and nouns represented in the data. The description follows Hansen et al. (1990: 73f., 82), however, the semantic classes of the underlying nouns have been integrated into the meanings of the types. The description of the pattern is followed by the meaning of the type. Thus, anti-communal represents the model or pattern "anti- + N/Nprop + adjectival suffix", with "N/Nprop" referring to the noun/proper noun (commune) underlying the adjective (communal), and the semantic type 'against/opposed to the institutions, principles, ideologies etc. connected with N/(the views etc. of) Nprop'. The following models and types were represented in the material: (A) Adjectives 1. "anti- IIpro- + N/Nprop + adjectival suffix" ('against/opposed to//in favour of the institutions, principles, ideologies etc. connected with N/(the views etc. of) Nprop'):

anti- and pro- in Indian English

39

anti-communal (force) (/El 5/5/91/3), anti-Nehruvian (politicians) (5^2/12/95/12), anti-secular (sympathies) (Λ^23/10/88/15), proKhalistani (head priests) ('the name to be given to a separatist State of Punjab') (NA10/4/88/2) 2. "anti- Hpro- + N/Nprop + 0Adj" ('against/opposed to//in favour of the institutions, principles, ideologies etc. connected with N/(the views etc. of) Nprop'):4 anti-adivasi ('member of the Aboriginal tribal peoples of India') anti-Dalit (... that the cutting edge of the Shiv Sena thrust is antiMuslim, anti-Dalit, and even anti-adivasi. 77*25/1/87/27) anti-bandh (campaign) ('strike') (NA 1/10/89/7) anti-BJP (= Bharatya Janata Party, a fundamentalist Hindu party in India; "Now Arjun Singh is seen to be anti-BJP while you are considered close to the BJP." 527/12/92/90) anti-cow slaughter (agitation) (527/12/92/33) anti-harijan (pogroms) {NA 18/10/87/2) anti-hoarding (cell) ('the hoarding of essential food supplies') (/E8/1/88/3) anti-land reforms (attitudes) (NA 10/4/88/2) anti-naxalite (squad) (NA 18/10/87/14) anti-peasant (measures) (NA 18/10/87/4) anti-people (government) (#723/3/96/3) anti-poor (To the ordinary Congress party worker, however, reforms are anti-poor\ BI\ 1/3/96/63) anti-Rajiv (campaign) (IW\ 5/6/91/12) anti-Ram (utterances) (IEl9/9/95/8) anti-repression (day) (Mil 0/4/88/8) anti-reservation (stir) ('a policy that reserves certain posts in the public services or other institutions for, or certain grants to underprivileged sections of the population') (7V7/18/4/91/5) anti-sati (ordinance) ('the Hindu custom of a widow immolating herself on her husband's funeral pyre') (/£6/10/87/5) anti-stability (... both are in fact essentially anti-stability because... 1/4/91/4) anti-upper caste (rhetoric) (/£ 19/9/95/8) anti-women (... refuted the Congress (I) charge that the BJP was "anti-women" [!]... ZE16/5/91/6) anti-worker (government policies) (^424/9/89/6) anti-working class (labour bill) (7VX20/4/88/3)

40

Uwe Carls

pro-BJP (wave) (7^15/5/91/7) pro-Congress (organisations) (SI3/12/92/7If.) pro-coimtry (... that a majority of the Christian community is more pro-country than pro any particular party. 7^17/5/91/3) pro-Khalistan (elements) (7^17/5/91/1) pro-monopolist (measures) (7^424/9/89/9) pro-movement (leaders) (7£19/9/95/9) pro-negotiation (cadres) (SI 3/12/92/11) pro-peace (stand) {NA1/10/89/12) pro-people (image) (5/28/10/95/7) pro-poor (programme) (5713/3/96/71) pro-public!private sector (protagonists) (7£8/l/88/8) pro-Ram (wave) {IE\ 5/5/91/4) pro-rich (The legislation was termed "pro-rich" by the CPI(M), ...Λ11/11/95/7) pro-rupee (statements) (St\ 1/11/95/15) Prefix-derivations with both prefixes were originally confined to an attributive function (cf. Marchand 21969: 143) and even today this use seems to dominate clearly. As the examples underlined in the preceding paragraph indicate, Indian English seems to be less hesitant to use them as predicatives, too. (B) Nouns 1. "anti- + N" ('against/opposed to the institutions, principles, ideologies, etc. of N'): anti-communalism (7^23/10/88/10), anti-congressism (ΜΏ1/4/91/4) anti-Naxalism (7/723/3/96/10) 2. "anti- H pro- + N/Nprop + -istl-ite" ('person who is against/opposed to Ν // in favour of N'): anti-liberalisationist(s) (7^8/1/88/8), anti-reservationist {NH11/4/91/1); cmti-naxalite {NA 18/10/87/14) // proliberalisationist(s) (7^8/1/88/8), pro-zamindar(s) (7710/2/89/Sect2/3) 3. Concerning the third type of peculiarity mentioned above, it can be concluded that the adjectival model 2 (anti-adivasi; pro-Khalistan) is by far the most frequent one in the material. The adjectival model 1, which is

anti- and pro- in Indian English 41 well represented in the respective lists of cmti- and pro- formations presented in (British) dictionaries, seems to be less frequent. In summary, it can be asserted that Indian English has not developed new types of anti- and pro- combinations. Instead, it makes frequent and creative use of existing types with a clearly marked preference for the type 'against/opposed to//in favour of the institutions, principles, ideologies etc. connected with N/(the views etc. of Nprop)' of the model "anti- II pro- + N/Nprop + 0Adj". Moreover, there seems to be less hesitation than in English English to use this type in predicative position.

Notes 1. As a rule, definitions and equivalents refer to the unknown elements only. 2. Force-landing itself is also Indian English, which prefers compounds of this model to English English phraseologisms of the model forced landing (Collins) ; cf. also IndE colour pencil vs. EngE coloured pencil. 3. Newspaper sources are quoted according to the following convention: abbreviation of name day/month/year/page. For full names of newspapers see "Sources" section of the bibliography. 4. For those anti- and pro- combinations that are used predicatively, the respective sample sentence is quoted.

References Cambridge International = Procter, Paul (ed.) 1995 Cambridge International Dictionary of English. Cambridge: Cambridge University Press. Carls, Uwe 1982 The status of English in India. Linguistische Studien, Reihe A, 100 Berlin: 80-86. 1991 Notes on the History of English in India. Zeitschrift für Anglistik und Amerikanistik 39: 51-57. 1993 Types of Lexical Peculiarities in Indian English. In: Günther Blaicher and Brigitte Glaser (eds.), Anglistentag 1993 Eichstätt. Proceedings vol XV, 205-216. Tübingen. 1999 Compounding in Indian English. In: Uwe Carls and Peter Lucko (eds.), Form, Function and Variation in English. Studies in Honour of Klaus Hansen, 141-154. Frankfurt am Main - Berlin - Bern - New York - Paris - Wien: Peter Lang. Collins = Hanks, Patrick (ed.) 1986 The Collins Dictionary of the English Language. London: Collins.

42

Uwe Carls

Chambers = 1998 The Chambers Dictionary. New Edition. Edinburgh: Chambers. Cobuild = Sinclair, John (ed.) Collins Cobuild English Dictionary. 1995. 2nd edition. London: Collins. Concise Oxford = Thompson, Delia (ed.) 1995 The Concise Oxford Dictionary. 9th edition. Oxford: Clarendon. Hansen, Klaus, Uwe Carls and Peter Lucko 1996 Die Differenzierung des Englischen in nationale Varianten. Eine Einführung. Berlin: Schmidt. Hansen, Barbara, Klaus Hansen, Albrecht Neubert and Manfred Schentke 1990 Englische Lexikologie. Einfuhrung in Wortbildung und lexikalische Semantik. 3rd edition. Leipzig: Enzyklopädie Verlag. Hawkins, R. E. 1982 Α Supplement of Words from India, Pakistan, Bangladesh and Sri Lanka, in: Swannell, J., (ed.), The Little Oxford Dictionary of Current English. 3rd edition. Oxford - Calcutta - Bombay - Madras: Clarendon. IABE = Nihalani, Paroo, R. K. Tongue and Priya Hosali 1979 Indian and British English. A Handbook of usage and pronunciation. Delhi - Oxford: Oxford University Press. Kachru, Braj B. 1972 Lexical innovations in South Asian English. International Journal of the Sociology Language 4: 55-73. 1994 English in South Asia. In: Robert Burchfield (ed.), The Cambridge History of the English Language. Volume 5. English in Britain and Overseas: Origins and Development, 497-553. Cambridge: Cambridge University Press. Little Oxford = Swannell, Julia (ed.) 1985 The Little Oxford Dictionary of Current English. 5th edition. Oxford Calcutta - Bombay - Madras: Clarendon. Marchand, Hans 1969 The categories and types of Present-Day English word-formation. 2nd edition. München: Beck. New Oxford = Pearsall, Judy 1998 The New Oxford Dictionary of English. Oxford: Oxford University Press. Oxford Advanced - Crowther, Jonathan 1995 Oxford Advanced Learner's Dictionary. 5th edition. Oxford: Oxford University Press.

anti- and pro- in Indian English

43

The following corpora were used: Brown, LOB, London-Lund

Sources The sources are quoted according to the following convention: day/month/year/page. BI HT IE IT IW NA NH S St TI

Business India. The magazine of the corporate world. Bombay. The Hindustan Times. New Delhi. Indian Express. New Delhi. India Today. Bombay. The Illustrated Weekly of India. The Feature Magazine. Bombay. New Age. New Delhi. National Herald. New Delhi. Sunday. Calcutta. The Statesman. Weekly. News and Comments from The Statesman of Calcutta & New Delhi. The Times of India. New Delhi.

The meaning of naked and a theory of metaphor Hans-Jürgen Diller

0. Apology The topic of this contribution, by a 65-year-old to a Festschrift celebrating a 60-year-old, may strike readers as somewhat bizarre. I will be drawing on corpus evidence and, in the language of commerce, some of the wares on display are liable to be offensive to some customers. A short explanation of my interest in the subject is therefore in order. In a seminar on metaphor I came across an analysis of two lines in Matthew Arnold's 'Dover Beach' (1867): (1)

[... ] down the vast edges drear And naked shingles of the world. (Goatly 1997: 114)

'Naked' is regarded as metaphorical, its "actual unconventional referent" being "the concept 'uncovered'". It is also claimed that the "conventional colligate" of the lexeme naked is the noun body {shingles is not discussed). I was unhappy with this analysis, first because of the rendering of 'naked' as "uncovered", secondly because I felt that the "conventional colligate" of naked was not body so much as any expression specifically referring to one or several human beings. I also felt that a decision between these two views had rather important implications for the interpretation of metaphors containing naked, but the way to decide the issue was not so clear. The aim of this paper, which over long stretches is a study of the colligations of naked, is an attempt to reach a reasoned decision based on as much evidence as possible, both empirical and intuitive. I will first examine the notion of "conventional colligate" and then study colligations in which naked typically occurs. On the basis of these colligations I will first identify classes of objects that are typically called naked, i.e. "carriers" of the property "naked",1 and then distinguish different meanings of naked. The paper will end with a brief discussion of the lexeme's inferential features.

46 Hans-Jürgen Diller 1. Terminology and methodology: colligations and conventional colligates Colligation is defined as a syntactic relationship between two words, whereas collocation is any kind of co-occurrence (Goatly 1997: 8). An important point which Goatly does not discuss is the fact that two lexemes may enter into more than one syntactic relationship. Adjectives and nouns, for instance, may co-occur as modifier and noun in a nominal group, or as subject and predicate in a clause. This has consequences for the notion of "conventional colligate", the search for which is an important step in the interpretation of metaphor. Simplifying somewhat, a metaphorical expression is an "unconventional colligation", whose interpretation involves replacing one of the colligates by an alternative expression which would make the colligation conventional. As 'naked shingles' is, in Goatly's opinion, an unconventional colligation, he looks for a conventional colligate for naked and finds it in body. Apparently he does not interpret 'shingles' as denoting "herpes zoster", which would make the colligation perfectly conventional; but then of course the task would be to find some substitute for herpes zoster that would colligate conventionally with the world. This point, however, affects an individual interpretation rather than the principle, which seems sound. The problem is merely that there may be too many possible - and quite conventional - colligates and no principled way to choose between them. Moreover, Goatly seems to silently assume that the conventional colligate has to fill exactly the same position as the unconventional one by which it is evoked. In the special case of (1) - and presumably many others - this requirement clashes with his other aim of making the conventional colligate as specific as possible (1997: 90), because in 'Dover Beach' naked is used attributively and expressions colligating with attributive naked are far less restricted than expressions colligating with predicative naked. The latter are thus likely to yield more specific colligates. While it is easy to see that the search for conventional colligates may result in putting the linguistic cart before the psychological horse, it is less easy to propose an alternative way of identifying the tenor of figurative expressions. I will try three ways supplementing each other. The first is an attempt at constructing an ostensive definition, the second I will call "inferring the carrier", the third, more empirical and most time-consuming, will look at the evidence supplied by the British National Corpus.

The meaning of naked 47 2. Ostensive definition and inferring the carrier The procedure of ostensive definition is usually illustrated with nouns (cf. Leisi ([1952/] 1971: 33-40), but the ostensive definition of a noun like tower does not differ crucially from that of an adjective like naked {ibid., 48). Let us assume that we have to answer question (2a) and there is a naked person or a nude painting in sight, and let that (real or painted) person's name be Qwertyu.2 We will utter something to the effect of (2b): (2a)

What is [viz. the meaning of] naked!

(2b)

Qwertyu is naked.

(2c)

Now my eye is naked.

(2d)

Qwertyu is naked.

If we have only a naked wall to point at, we may of course use it for our ostensive purposes, but our chances of success will be slender. Taking off my glasses, pointing at one of my eyes and saying (2c) may be slightly better, but still not very good. In a terminology that is popular today, we could explain these differences by saying that there is a cline in prototypicality from naked people to eyes and then to walls. The eye is more prototypical than the wall because by taking off my glasses I can illustrate the fact that "naked" is typically a temporary state which is the result of "volitional, controllable behaviour" (Hamann 1991: 658). But what is more important is that ostensive definition, pointing at something present in the physical world, does not help us to resolve our "body or person" problem. Let us now assume that we know the meaning of naked but not the referent of Qwertyu and that we hear (2d) without a pointing gesture to accompany it and without a naked person in sight. In that case most of us would infer that Qwertyu is a (momentarily invisible) person not wearing any clothes at the moment. This is strong evidence that the prototypical carrier of the property "naked" is a human person rather than merely a body, for knowledge of a property-denoting lexeme often permits us to infer the class of objects to which the carrier belongs. Such knowledge is of course not conveyed by ostensive definition, it is acquired by frequent exposure to the use of a word in "real language", as the corpus linguists like to say. To this evidence we will turn now.3

48

Hans-Jürgen Diller

3. Corpus.evidence 3.1.

Tense

Unfortunately but not surprisingly, we don't find utterances like those listed under (2) even in a 100-milIion-word corpus like the BNC with its 2045 occurrences of naked. By "like" I mean statements where naked is used predicatively and the subject refers to a specific, living person that is actually naked at the time of utterance. What we do find is one description of naked figures (G30 188) and a narrative employing the historical present (FAJ 2253). It is not difficult to see why this should be so: people don't appear naked, giving occasion for remarks like (2d), when a recording corpus linguist is around. It is thus no surprise that naked colligates with the past tense far more frequently than with the present: against 21 instances of am, is and are naked the BNC contains 116 of was and were naked. It is a cultural fact that corpus data on the use of naked are likely to remain less plentiful and representative than we would like. Most of our examples will thus come from narrative texts. 3.2.

Carriers

The words denoting carrier and property can colligate in various syntactic constructions (all examples from the BNC): (3a)

Kiku was naked.: subject and complement ('predicate') in clause

(3 b)

naked abuse: head and attribute in nominal group

(3 c)

stiletto-bearing heels naked in their needfor power, head and postpositive in nominal group

(3d)

I stalked the room naked', subject and supplementive in clause

(3e)

But they do like them naked: object and object complement in clause.

For our purposes it will be enough to distinguish between attributive and non-attributive uses. The advantage of this simple dichotomy is that it corresponds well with the colligating nouns: in a random sample of 500, non-attributive naked colligates almost exclusively with expressions referring to human beings, whereas colligations with attributive naked are divided much more evenly between different classes of carriers:

The meaning of naked 49 Table 1. Carriers of 'naked'

Heads or carriers human beings parts of the body abstract nouns others Total:

with attributive naked

79 88 42 76 285

with non-attr. naked

201 3 1 10 215

The few "other" carriers in the third column are of some interest. Among them are animals, trees, a king in chess, buildings, a sword and the bud of some plant. These non-human carriers are responsible for specialized meanings, many of which are so conventionalized that they are listed in the dictionaries. Very often the reference is generic rather than specific. It goes without saying that in these contexts nakedness does not call forth the reactions associated with human nakedness in our culture. But there is always something that has metaphorically taken the place of the clothes and whose absence is indicated by the adjective: plumage, fur, foliage, the other chess pieces, the scabbard. Like literal clothes they form a barrier between carrier and outside world, which is absent in the case of nakedness. In the terminology of Goatly and many others, these uses of naked are dead metaphors. More importantly for our purposes, the nouns colligating with attributive naked denote two important groups of carriers apart from human beings: parts of the (human) body and abstractions. The noun body itself behaves like the expressions referring to body parts rather than like those referring to persons: 77 times in the entire corpus it is the head of attributive naked, only 5 times does it function as the subject of predicative naked. The abstract nouns refer usually to human emotions such as ambition or hate or to human conditions such as distress. The number of nouns referring to body parts is even slightly higher than the number of nouns referring to persons. That, however, is solely due to the frequency of naked eye, which occurs no less than 40 times in the random sample (and 151 times in the entire corpus). It is always a generalized naked eye, never that of any individual.4 But even if we subtract the forty instances of naked eye, the question remains: why do nouns referring to the body and its parts or abstractions colligate with naked so much more frequently as heads of a nominal group than as subjects of a clause? Two different questions, two different answers. The frequency of body parts with attributive naked and their rarity in other positions seems to be due to the fact that we perceive our fellow

50 Hans-Jürgen Diller humans as wholes rather than as ensembles of parts. It has to be remembered that most examples of naked come from narrative texts. Formally, sentences like (4a-b) describe states that may last for a long time. But in the chain of narrated events they represent first perceptions by a character in the narrative: (4a) (4b)

Qwertyu was naked. ?

Qwertyu's arms were naked.

(4c)

Qwertyu entered the room. Her arms were naked.

(4d)

Qwertyu entered the room. She was wearing an attractive blouse which left her slim, graceful arms free! naked.

Unless there is a taboo on naked arms in Qwertyu's culture, (4b and c) would be a strangely sensational way of describing her appearance. (5)

He hadjust time to notice that the arm round his neck was naked. (Collins [1868/] 1968: 239)

(5) is a highly unusual way of becoming aware of another person.5 Naked body parts are not mentioned when they are perceived but when they come into action, e.g. when naked shoulders are kissed, naked backs lashed, naked feet have to walk on hot ground, or naked arms encircle your neck. We know that normal human beings are equipped with all these parts, and when their nakedness becomes relevant for the story, there is usually enough time to mention it. In (5) the two moments of perception and of entering the action coincide. With naked abstractions matters are not much different. As I said, they tend to be human emotions or conditions, that is they are nominalized properties or states in which people find themselves; when an abstraction is characterized as naked, this means that the property in question is particularly obvious, that the carrier - of hate, ambition, self-interest, distress and the like - is either unwilling or unable to hide it. In such cases naked is something of an intensifier. But to explicitly predicate nakedness of such a state would be saying that a state is in a (temporary) state or a property has a (temporary) property. It's not very often that we can usefully say such things. That is why we find sentences like (6a) but hardly (6b): (6a)

Isembard's deep eyes flashed to Harry's face and lingered there, untroubled by the naked hate and defiance that stared back at him.

The meaning of naked

(6b)

9

51

The hate in Harry 's face was naked.

In the random sample the only instance of non-attributive naked after an abstract noun is (7)

[...] and the distress of the children, although more naked than that of their mothers, was less terrible.

Although one should be cautious with justifications of single examples, the sentence seems to appear natural because of the comparison and the second adjective, terrible. But this is a minor question. What emerges clearly from the data presented so far is that the usual carrier of the property "naked" is either a human being, or (parts of) the human body, or a human quality or condition, and that syntactically emotions and conditions are treated very much like the body. And there is a good reason for that: they are both "part" of a person. Since the person is, so to speak, naked "via" his/her body, body parts, emotions or condition, it is sometimes useful to refer to him or her as the "ultimate" carrier of the property of nakedness. However, there is also an interesting difference between abstracts and the body: while the naked body is often perceived as attractive, naked abstractions are almost always unattractive. We find naked aggression (10 times), power (9),/ear (7), terror (4), hate, hostility (3 each), vulnerability (2), and when we read of naked anti-unionism, capitalism, cynicism, imperialism, self-interest or scepticism (1 each), we feel that these things are (morally!) ugly and ought to be hidden from the general view. Even the naked truth or Truth, which appears 18 times (including 5 times in a pop song), does not as a rule appear to be agreeable. Apart from its temporariness, there is yet another respect in which the nakedness of abstractions resembles that of people and body parts and does not resemble that of non-human objects such as rocks, trees, buildings, or buds: it is the product of some action for which we make someone responsible. "Naked anti-unionism", for instance, seems more objectionable than the hidden variety, because the anti-unionist isn't bothering to cloak it. When someone other than the ultimate carrier is responsible, as in the case of fear, we feel pity for him (or, perhaps more frequently, her). Interestingly, there are some semantic tests which suggest that our three classes of carriers entail three different meanings of naked (Kastovsky 1982: 113f.): (8a)

*He and his feet were naked.

52

Hans-Jürgen Diller

(8b)

*He was naked, but his feet were not.

(8c)

*His naked lust andfeet

but: (8d)

His feet and arms were naked.

(8e)

His naked lust and desire

3.3.

Modifications of naked

3.3.1. Location Modifications of naked are far more frequent in non-attributive than in attributive use. This is probably true of other adjectives as well, but here it has the additional consequence that modified nakedness is almost exclusively predicated of human beings, either directly or indirectly via their physical or psychological "parts". As we saw in Section 2 from (2d), the default meaning of naked is "entirely naked". But we can also express the fact that someone is naked in parts, although only in a limited number of ways: (9a)

Qwertyu is nakedfrom the waist up / above the waist.

(9b)

Qwertyu is nakedfromthe waist down / below the waist.

(9c)

Qwertyu is naked to the waist.

Significantly, (9c) is synonymous with (9a). The waist seems to be almost the only part of Qwertyu's body that can be used to qualify the statement in (2d). The only exception I have come across is (10) in the "Introduction" to Browning's drama Pippa Passes:6 (10)

All but naked to the knee! (1.155)

A sentence like (11) sounds very odd, although we can of course utter it if our feelings about shoulder-free ball gowns are correspondingly strong: (11)

Qwertyu is nakedfrom the bust up / above the bust.

The meaning of naked

53

Once the sexually distinctive parts of the body are covered, it is no longer considered naked, or so it seems at the moment. It does happen, of course, that Qwertyu wears some pieces of clothing, though not in the socially required places. The BNC is quite generous with examples, of which I give a small selection: (12a)

He was naked, apart from his polished calf-length military boots.

(12b)

Without ceremony she dropped her silk gown, standing naked but for her briefs and her tattoos, which were all black

(12c)

John Lennon dropped in on it, and got embarrassed when the poet, naked exceptfor a pair of underpants draped around his head, kissed him on the cheek

(12d)

I had read with interest of the lady golfer who, when confronted with a naked man wearing only a bowler hat, asked him whether he was a member, [...].

What is expressed here is the item that covers a part of the body rather than the clothed or covered part itself. We don't seem to say (13)

He was naked exceptfor his feet.

And, as we saw in the previous section, we do not say (8b): *He was naked, but his feet were not. Note also that the clothing items are in most cases carefully specified, so as to make the appearance of the entire person particularly grotesque. So far, the "limitations" on nakedness have implied that the private parts are not covered, the only (partial) exception being (12b). But the BNC contains some interesting evidence that a body can be called naked even when it is covered and perhaps not even directly visible: (14a)

Nina danced naked under a black veil.

(14b)

Duncan knew she was naked underneath.

(14c)

Clutching a blanket partly around her, she was obviously naked.

This suggests that the denotatum of naked is a state in which the body is not so much without any cover as without clothes and that a full description of the meaning of naked would have to include the meaning of clothes. But even that is not the whole truth. Sometimes, though rarely, a carrier is described as naked even when (s)he is clearly clothed:

54 Hans-Jürgen Diller (15a)

[...] nervously caressing the nakedflesh beneath her skirt.

(15b)

She generally wasn 't aware of her naked skin underneath the cover of her clothes!

In (15a) most hearers would probably assume that "she" was not wearing any underwear, but such an assumption does not suggest itself for (15b), and even for (15a) it is by no means necessary. It is of course dangerous to generalize from these few examples, but it does seem significant that (15a&b) differ from (14) not only in respect of the covering items (clothes vs. others) but also in respect of the carrier of the nakedness: under veils and blankets it is the entire person that is naked, under pieces of clothing it is only part of the body: the flesh or the skin. Of course we can say (16), but corpus evidence suggests that opportunities to do so are rare: (16)

McQwertyu felt an uncomfortable draught, because he was naked under his kilt.

If the entities in (14-16) are clearly covered or even clothed, what justifies us calling them naked? My tentative answer is that a barrier between the carrier and the outside world has been removed with the effect that the human body (or part of it) is more accessible than in the "proper" order of things. That would suggest that impropriety (Leisi's "slight offensiveness") is indeed an important component of the meaning of naked (cf. Leisi [1952/] 1971: 45). 3.3.2.

Degree

Apart from identifying the location of non-nakedness, we can, of course, specify the extent or degree of nakedness. The majority of pertinent modifiers are maximizers, which "denote the upper extreme of the scale" (Quirk et al. 1985: 590): bollock (1), butt (1), completely (19), entirely (3), quite (4), stark (36), stark bollock (1), totally (12), utterly (1); total: 81. A sizable minority are approximators, which denote a point near the maximum (Quirk et al. 1985: 597): all but (1), almost (16), nearly (6), virtually (3); total: 26. An important group of modifiers, not classified by Quirk et al. (1985), is formed by half (with and without hyphen: 78) and the much rarer part (1), partially (1) and partly (2). The last two groups make it particularly clear that naked denotes not a scale but the maximum end of a scale. It is the modifiers, not the adjective, which place the carrier on the scale. The maximizers, by contrast, merely reinforce the position which is already expressed by the adjective. The singularity of naked shows in

The meaning of naked

55

maximizers like stark (with 36 the most frequent one), butt, bollock and stark bollock, the latter three having only naked as a possible head. The frequency of these modifiers suggests that naked is not normally a gradable adjective. In this respect it behaves very much like other privatives, such as bald, blind or deaf. But graded naked does occur. The BNC contains 6 comparatives, 3 superlatives, and one case of naked modified by the intensifier very. There are also 3 diminishers, which indicate a point near the minimum end and "roughly mean 'to a small extent'" (Quirk et al. 1985: 597): a bit, slightly, somewhat. With regard to gradability naked seems rather like empty, of which there are 13 comparatives and 7 superlatives. Rather like it, but not quite. For we do find (17a) and similar sentences in the BNC, but nothing resembling (17b): (17a) (17b)

The flask was a deal emptier than it had been before the dinner break, (in BNC) 7 At/Near/Toward the end of her act the stripper was more naked than at the beginning.

There are reasons to suspect that the gradable and the non-gradable naked are not quite the same. Such suspicion is supported above all by the diminishers and the superlatives, but also by the comparatives and the one case of naked modified by very. What is called 'most naked' is the 'forms' of 'the immense, amorphous city', the 'exercising of user power', and (generically) 'man', the 'most naked of all the apes': the carrier is never a person. What is called 'a bit naked' is a development plan, and no one is said to 'be' slightly or somewhat naked: the persons concerned only 'feel' that way, e.g.: (18)

Yet, she felt slightly naked in having to get into the ambulance without a coat or even something on her head.

It seems that all instances of graded naked are in a sense metaphorical, those with more and very being the most interesting ones: (19a)

She was more naked than if she were wearing nothing.

(19b)

He turned up at the dance studio in a pair of very skimpy tight shorts and observing that everyone else was very white and wore towelling track suits, and seeing that he was very tall and very naked, got nervous so went outfor a walk and smoked a joint.

56

Hans-Jürgen Diller

From (19a) we must conclude, paradoxically, that you can be more naked than 'completely naked' (='wearing nothing'). And from (19b) it is clear that a person called very naked wears more clothes than one described merely as naked. 4. Inferential features and the interpretation of metaphor To explain these paradoxes, it is convenient to invoke Lipka's distinction between distinctive (especially denotative) and inferential features (1990: 114f.). Among the denotative features of naked are [+OBSERVABLE] and [-GRADABLE]. But the state of nakedness is regularly associated with psychological reactions, both in the carrier and in the beholder. We may thus posit an inferential feature of naked which could be paraphrased as "causing some of the reactions associated with nakedness". The reactions may be of the most various kinds: embarrassment or shame, shamelessness or defiance, invitation to or availability for intimacy - the list is open, the appropriate reaction will depend in part on whether the carrier is perpetrator or victim of the nakedness. But as a reaction it can be weaker or stronger. If the inferential feature, i.e. the reaction, is important enough, the denotative feature [-GRADABLE] of naked can be reversed by the presence of an intensifier, and the extent of physical nakedness may even become irrelevant (cf. (18) and (19)).7 It seems doubtful that such inferential features can be formulated as terms of a semantic metalanguage clearly distinct from their objectlanguage homonyms. Such distinctness may not even be desirable. In order to make sense of a word like naked when the carrier is clearly not naked, we have to draw on world knowledge, which tells us that human nakedness is the exception rather than the rule and thus the result of "volitional, controllable behaviour" (cf. Section 2). It prompts us to try to identify the person responsible for that behaviour. In the case of non-human nakedness it prompts us to identify the objects) that could be regarded as taking the place of the clothes. But it is our knowledge of linguistic usage which tells us that it is people rather than bodies that nakedness is predicated of and which makes us suspect that, while we 'cover' our bodies, we 'clothe' ourselves. Another corpus study might tell us if that suspicion is correct.

The meaning of naked

57

Notes 1. The term carrier is borrowed from Halliday (1994). 2. Qwertyu is better than, for instance, X, because X is usually the name of a variable. What we need is a value of that variable. Qwertyu, being just a series of typewriter keys, seems ideally neutral with respect to gender, culture, etc. 3. For help with the BNC I have to thank Patricia Sift and Tobias Rademann. 4. Dickens uses the colligation to satirize American prudery in Martin Chuzzlewit, ch. 22, where the American authoress, Mrs. Hominy, takes offense at Martin's "talk about a naked eye" in the presence of a lady (Dickens [1844/] 1951:369). 5. The sentence describes an attack from behind narrated by a very verbose Victorian lady (Wilkie Collins, The moonstone, Second Period, First Narrative, Chapter 1). 6. Browning [1905/] 1940: 167. I owe literary examples to Ewald Mengel and Fritz-Wilhelm Neumann. 7. I use the term "intensifier" like Quirk et al. (1972: 439; 1985: 589) and above all Bolinger (1972: 17) to refer to: "any device that scales a quality, whether up or down or somewhere between the two".

References Bolinger, Dwight 1972 Degree words. The Hague: Mouton. Browning, Robert 1940 The poetical works of Robert Browning. London: Oxford University Press. First published [1905]. Collins, Wilkie 1966 The moonstone. Edited with an introduction by J.I.M. Stewart Harmondsworth: Penguin. First published [1868]. Dickens, Charles [1951] The life and adventures of Martin Chuzzlewit. Edited with an introduction by Geoffrey Russell, London: Oxford University Press. First published [1844], Goatly, Andrew 1997 The language of metaphors. London: Routledge. Halliday, Michael A.K. 1994 An introduction to functional grammar. Second edition. London: Arnold.

58

Hans-Jürgen Diller

Hamann, Cornelia 1991 Adjectives. In: Arnim v. Stechow and Dieter Wunderlich (eds.), Semantik - Semantics. Ein internationales Handbuch der zeitgenössischen Forschung. An International Handbook of Contemporary Research. Berlin - New York: Mouton de Gruyter. Kastovsky, Dieter 1982 Wortbildung und Semantik. Düsseldorf - Basel: Bagel - Francke. Leisi, Ernst 1971 Der Wortinhalt. Seine Struktur im Deutschen und Englischen. Heidelberg: Quelle & Meyer. First published [1952]. Lipka, Leonhard 1990 An outline of English lexicology. Tübingen: Niemeyer. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik 1972 A grammar of contemporary English. London: Longman. 1985 A comprehensive grammar of the English language. London: Longman.

On contrastive word formation: German and Hungarian denominal adjective formation Wolfgang U. Dressier and Maria Ladanyi

1. Introduction Continuing earlier work on the theory of word formation with particular emphasis on productivity (Dressler/Ladanyi 1998, 1999) we try to develop a functionalist approach to contrastive word formation and apply it to two typologically different languages, one predominantly inflecting-fusional and the other predominantly agglutinating. Our theoretical stance is Natural Morphology1, but in contrast to previous work on word-formation rules within this framework, we will also discuss paradigmatic relations (cf. van Marie 1985, Pounder 1987, Bauer 1997, Dressler/Ladanyi 1998). The theory of Natural Morphology consists of three subtheories: The first is the subtheory of universal preferences. It explains why the semiotically based preferences for iconicity, morphotactic and morphosemantic transparency, biuniqueness hold in predictably similar or dissimilar ways for the different parts of morphology (cf. Dressler in Dressler et al. 1987). These parts are ordered on a continuum of prototypical inflection non-prototypical inflection - non-prototypical derivation2 - prototypical derivation - non-prototypical compounding/derivation3 - prototypical compounding. Denominal adjective formation belongs to prototypical derivation, because it serves lexical enrichment (see (b) on the next page), is non-obligatory, full of rule competition, changes word class,4 and is reappliable.5 The second subtheory is that of typological adequacy. It explains why agglutinating morphology is richer and thus also more productive than inflecting-fusional morphology - especially in inflection but at least partially also in derivation (cf. Pöchtrager et al. 1998 and the data in Ülkü 1980). In Dressier and Ladänyi (1998, 1999) we have found such difference in morphological richness in deverbal verb formation between agglutinating Hungarian and inflecting-fusional German. This paper deals with the question whether a similar difference in morphological richness (as defined by productive morphological rules (= MRs)) can also be found with denominal adjectives (and if not, why).

60

Wolfgang U. Dressler and Märia Ladänyi

The third subtheory is that of language-specific system adequacy. It is of special importance for this contribution. In our view, it has productive MRs at its core, as only productive MRs are fully system-adequate. Therefore, morphological descriptions of individual languages should start with, and concentrate on, the fully productive rules. This is what this contribution will do. For a functional approach to word-formation both the communicative and the cognitive functions of language are important, whereas for inflection the cognitive function is less important. To both of the main functions, the following lower-order specific functions of morphology are subordinated: (a) The function of (a) word-internal motivation (already identified by Saussure) holds for all areas of morphology, both in terms of morphosemantic and morphotactic motivation. (b) In addition, word-formation (but not inflection) has the lexical function of lexical enrichment (already identified by Bally), i.e. of providing new labels for cognitive and/or communicative needs.6 These new words then have the same lexical referential function as old words, be they simplex or complex words. (c) Finally, the textual function of morphology, i.e. of serving cohesion and coherence (cf.. Dressler 1985)7 is much more rarely used and thus of less pertinence. Within word-formation, especially compounding and nominalization of verbs serves to recall and combine concepts in the earlier co-text or discourse. Denominal adjective formation rarely serves this function. According to Peirce (1965: 11.332), "the essential function of a sign is to render inefficient relations efficient". Since efficiency is a matter of degree, more efficient semiosis is clearly preferred, and therefore, within naturalness theory, 'degree of efficiency' has been used as an equivalent of 'degree of naturalness' of an operation on a given universal parameter (such as iconicity, etc., see above). Thus, the function of motivation is better served by morphosemantically and morphotactically transparent word-formation techniques than by more opaque ones. And all three functions of (a) to (c) are better served by productive than by unproductive rules, i.e. by the core of grammatical morphology. Similarly, the notions of typological adequacy and system adequacy are concepts of relative efficiency. Biuniqueness is a semiotic ideal both for the lexicon and for morphology, i.e. it may hold both for whole words (as stored in the lexicon) and for

Ort contrastive word formation

61

their parts (as combined by morphological rules in case of morphological concatenation). Now, in semiotic terms, words are primary signs, morphemes only secondary signs. Thus the biuniqueness of a morphological part of a word is subordinated to the biuniqueness of the whole word. This is the main functional reason for idiosyncratic lexicalization of complex words and for the relative unimportance of reference of word parts. 2. Contrastive word formation Contrastive linguistics is a branch of comparative typological linguistics and deals, as any comparative endeavour (cf. Jucquois 1993), with both (a) 'similitudo in dissimilitudine' and (b) 'dissimilitudo in similitudine' of (typically) two different languages.8 Clearly, within the first subtheory of Natural Morphology (universal preferences) we focus on similarities, within the third (language-specific system adequacy) on dissimilarities, and within the second (typological adequacy) in our case as well. Next we have to establish the range of what is to be compared. Within the class of denominal adjectives (all productive ones formed by suffixation) we will concentrate on the productive core. Any comparison assigns and presupposes structure within the objects compared. Here, beyond what is given by the framework of Natural Morphology, we want to attend particularly the hierarchical structure of the word-formation classes compared. This will allow us to establish correspondences between both structures and items within structures. Only with the above presuppositions can we hope to approach the explanation of similarities and dissimilarities and to find implications between them. 3. The gradient of productivity We start from the assumption (cf. Dressier and Ladänyi 1998, 1999) that word-formation productivity, as grammatical, rule-governed creativity, is a primitive and prototypical property of word-formation rules within the potential system of grammar, in analogy to grammatical rules of inflection, syntax and phonology. This implies that unproductive rules are marginal within morphological grammar (including word-formation). This is clearly a qualitative, not a quantitative conception of productivity. Therefore, if we assume that, as many other concepts of naturalness theory, also productivity is gradual, we require a hierarchy of grammatical productivity

62

Wolfgang U. Dressler and Märia Ladänyi

criteria of word-formation rules. The one suggested below is based on an analogical hierarchy established for inflectional morphology (in Dressier 1997, Dressier and Thornton 1996). 3.1.

First-grade productivity

The highest degree of derivational productivity holds for word-formation rules which (in addition to applying to indigenous bases) can integrate foreign words, since it is more difficult to integrate words coming from a foreign language [marked alternative] than indigenous words [unmarked alternative].9 This high degree of productivity holds for German relational adjective word-formation rules with the suffixes -ig (cf. E. -y), -isch (cf. E. -ish) and -lich (cf. E. -ly), illustrated with bases loaned from French or English: (1)

gelee-ig, rass-ig, trend-ig, zwei-etag-ig·, rass-isch, film-isch; sportlich, weekend-lich

Hungarian denominal adjective formation is highly productive with the relational suffixes -s (-a/o/e/ös), -i, -beli, -φύ/φΰ (which applies only to compounds or syntactic phrases), and non-relational -nyi, as in (2)

multimidid-s, marketing-es, graffiti-s, internet-es; ombudsman-i, logisztika-i, operator-i = kameraman-i; (multi)media-beli, kommuna-beli, kibuc-beli; gyors-modem-u, nagy-kapacitäs-u, nagymemoria-ju (referring to a PC with a quick modem, big capacity,

3.2.

Second-grade productivity

A hierarchically lower, but still important productivity criterion is derivation from abbreviations. Abbreviations are not formed via grammatical word-formation rules but by extragrammatical means (cf. Dressier and Merlini Barbaresi 1994: 36-40) and are therefore both marginal and marked within the lexical stock of a language. There are no German relational adjectives formed from abbreviations 0naz-istisch is no counter-example, because it is not derived from Nazi but a parallel formation to Naz-ismus). Among the Hungarian sufFixations enumerated in 3.1., only -i suffixation cannot apply to abbreviations. Examples are:

On contrastive wordformation 63 (3)

MSzP-s [emespe:s], Fidesz-es, MDF-es [emDe:efe§] 'being a party member of the parties MSZP, Fidesz, MDF'; PM-beli, kolesz-beli Penzügy-miniszterium 'finance ministry', /collegium 'youth hostel'; magas-IQ-jit, alacsony-GDP-ju 'with high IQ, with low gross domestic product'

Examples with -nyi are questionable: (4)

?

USA-nyi, ?NA TO-nyi 'USA/NATO-sized', potential but not accepted variants of: USA-nagysäg-ύ, NATO-meret-ü 'having the size of USA/NATO'

3.3.

Third-grade productivity

Hierarchically still lower as productivity criterion is (the rather infrequent) shift of a derivative from one (prototypically) recessive to another (prototypically more stable and thus more productive) word-formation rule. A German example of such suffix replacement is (see Pounder 1987, Osman 1994): (5)

G. -ig > -isch, when pejorative: argwöhn-ig > argwöhn-isch 'distrustful' ( neid-isch 'envious' ( szemerm-es, szorgalm-as 'modest; diligent'. 3.4.

Fourth-grade productivity

The last and hierarchically lowest criterion is the word-formation productivity of affixations as directly observable in fully grammatical indigenous neologisms formed from indigenous bases. This is productive derivation from unmarked bases, sc. from fully integrated indigenous bases. (For further subhierarchies cf. Dressier and Ladanyi 1999). All previously mentioned word-formation rules apply to such bases, plus word-formation rules that only apply to such bases and are thus less productive than the other ones. These are German suflfixation with -e(r)n, as in the derivations from old loan words:

64

Wolfgang U. Dressler and Märia Ladänyi

(6)

basalt-en, granit-en, schokolad-en 'made of basalt, granit, chocolate' and Hungarian suffixation with privative -talanltelenl(a)tlanl(e)tlen (cf. Kiefer 1998: 5.3)

3.5.

Unproductive word-formation rules

Unproductive denominal adjectival suffixes are G. -bar (e.g. frucht-bar 'fertile' Frucht 'fruit'), -sam (e.g. betrieb-sam 'busy' Betrieb), Η. -atosietes (e.g.felelm-etes 'horrible' adv. ur-as-an 'very well, in the way of a lord'

Also, word-formation synonymy is often only partial anyway,11 as in (24)

H. kossuth-i magatartds 'behaviour characteristic of Kossuth' = 1 •) G. ein Kossuth 'sches Benehmen Η. kossuth-os magatartds 'behaviour similar to that of K.' = G. ein Kossuth-haftes Benehmen

However, even in such cases of lexical oppositions, Hungarian has more suffix competition than German, as in (25)

H. hegy-es tdj 'a mountainous landscape' = G. eine berg-ige (gebirg-ige) Landschaft Η. hegy-i tdj 'a landscape in the mountains' = G. eine Landschaft in den Bergen

Summing up, our hypothesis of greater morphological richness leading to more word-formation rule competition and thus more word-formation synonymy seems to be confirmed.

68

Wolfgang U. Dressler and Μάπα Ladänyi

5. The hierarchical structure of the class of denominal adjectives The basic unit of the hierarchy is the derivational paradigm, defined as a set consisting of a base and of its parallel derivations which belong to the same category and share the same or nearly the same word-formation meaning. Parallel sets form a derivational microclass, e.g. the microclass of adjectival derivations from nouns, as in G. Kind 'child' -» adj. kind-lich and pejorative kind-isch, Weib 'wife' —> adj. weib-lich and pej. weib-isch, etc. (Pounder 1987). Also here we have to distinguish (cf. Zemskaja 1992: 17-19) actual, concrete word-formation paradigms such as the above from the corresponding abstract word-formation paradigm within language as a potential system, which is the expression of the word-formation meanings of the respective word-formation rules as applied to a specific base (such as G. Kind). The derivational microclass is then established by the systematically recurring complementary distribution of the word-formation meanings of the word-formation rules taking part in the abstract paradigms. In our above example, there is a systematic difference in connotational meaning between the two word-formation rules involved. Another type of microclass is established by word-formation rules with synonymous wordformation meanings, as in Hungarian felelem 'horror' -> adj. felelm-es = felelm-etes with productive -es and unproductive -etes suffixation (cf. 3.2.1, 3.2.5). If we also consider sufFixoids, then -fele, -szeru and -fajta have to be included into this microclass. The concept of a derivational microclass thus represents a meaningful generalisation over (sometimes only partial) systematic, grammaticalised word-formation synonymies within the lexicon of a language (cf. 4). A derivational paradigm is the systematic synchronic core of a word family (in the traditional sense, not in the definition of Corbin 1997, see below under word-formation nest). Thus a derivational microclass can be formed from derivational paradigms, whereas no meaningful classes can be formed from word families. Hierarchically in between the concepts of word family and word-formation paradigm, we can situate the concept of word-formation nest (G. Wortbildimgsnest < R. slovoobrazovatel'noe gnezdo), which consists of all synchronic and morphosemantically identifiable derivations from the same lexical base.13 In contrast, a (traditional) word family contains also derived words which are not connected synchronically (but only diachronically) and have no identifiable morphosemantic relation which conforms with the word-formation meaning of the respective word-formation rule.

On contrastive wordformation 69 Sometimes both word-formation paradigms and the microclasses being constituted by them consist just of one word-formation rule, i.e. the sets (of sets) of a basis with one derivative, as is the case with the privative suffix -tlan and its allomorphs: -talanltelenl(a)tlcml(e)tlen (the details about the distribution of the four allomorphs see in Kiefer 1998), e.g. fiil-etlen 'without a handle'. A derivational class consists of microclasses with the same base category and output category and the same word-formation or basic/hierarchical superordinate word-formation meaning.14 Thus the German derivational class of denominal relational adjectives (cf. Pounder 1987) includes the microclass (with relational meaning) of the abovementioned word-formation paradigms of Kind ->· adj. kind-lich; kind-isch, etc., and the microclass of word-formation paradigms, such as: (26)

Stein 'stone' -> stein-ig 'ston-y' and stein-ern 'made of stone' Glas 'glass' -> glas-ig, gläs-ern

Another microclass (albeit marginal because of the meaning differences) is established by paradigms such as (27)

Sinn 'sense' -» sinn-ig 'thoughtful'; sinn-lich 'sensuous' Erde 'earth' -» erd-ig 'earthy'; erd-lich 'earthly' Draht 'wire' ->· draht-ig 'wiry' (of persons); draht-lich 'by wire' Zeit 'time' -» zeit-ig 'timely'; zeit-lich 'temporal'

Since by now we have, within the same derivational class, two microclasses which contain -ig derivations, and two which contain -lich derivations, these microclasses overlap (cf. also 4.3.3). The hierarchically next higher concept which represents a valid generalization over word-formation potentials of word subclasses is a formal set of microclasses. This concept comprises microclasses with the same sets of lexical bases and the same output categories, e.g. the set of all denominal adjectives. This set differs, in German, from the respective derivational class only if we include the privative suffixoid -los (e.g. sinn-los 'senseless'), because these formations belong to the formal set of microclasses of denominal adjective formation, but the word-formation meaning of -los clearly differs from that of relational adjective formation. As can be shown, the corresponding Hungarian hierarchy of denominal adjective formation is more elaborate than the German one. First, Hungarian has not just more productive word-formation rules but as a consequence also more microclass overlap (cf. 3.6 and 4.2 above). In

70

Wolfgang U. Dressler and Märia Ladänyi

(17) and (18) we already illustrated the two overlaps between the microclasses of -i and -beli suffixation, of -/ and -(e)s suffixation and of -(e)s and ~(j)u suffixation. The difference between Hungarian and German does not lie in the number of potential overlaps (two for both languages) but in the number of actual overlaps. It is much higher in Hungarian and reflects the correspondingly higher number of synonymous Hungarian wordformations, which is in turn a consequence of greater word-formation rule productivity, of course. Furthermore, Hungarian has more microclasses which consist just of one word-formation rule as, for example, in geographical adjective formation, the bases which take the locative case -ben!ban belong to the microclass of adjective formation with -i and -beli (cf. (17)), but those that take the locative case -e/o/ön belong to the microclass of adjective formation with only -i suffixation, as in (28)

Budapest, Loc. Budapest-en, adj. budapest-i

As a consequence, Hungarian has more microclasses than German within the derivational class of denominal relational adjective formation. This distributional difference holds also if we include Latinate suffixation (cf. Dressier and Ladänyi 1998: 45-46). Finally, the formal set of microclasses with denominal adjectives adds only a single word-formation rule (and microclass) to the derivational class of relational adjectives in German, namely -los formation (cf. note 3), which is a suffixoid, not a prototypical suffix, Hungarian adds two sufflxations: -tlan and -nyi (cf. (2)).

6. Conclusion Typologists generally acknowledge that inflectional morphology plays a greater role in agglutinating languages than in inflecting-fusional languages and is therefore also considerably richer (cf. Pöchtrager et al. 1998). We have postulated the same for derivational morphology (Dressier and Ladänyi 1999). We have defined morphological richness in terms of productive morphological patterns, i.e. in terms of productive morphological categories, rules and paradigms. The greater word-formation richness of Hungarian (vs. German) is characterized by a greater number of productive word-formation rules, by greater productivity, and (as a consequence) by more rule competition and rule synonymy. In other words, Hungarian uses prototypical (e.g. suf-

On contrastive wordformation

71

fixing) word-formation of adjective formation both more intensely and more systematically, for expressing relations between nouns. This holds for other, non-relational adjective formations as well. As a result, the hierarchical paradigmatic organization of denominal adjective word-formation is more elaborate. This greater vertical depth of such derivational class hierarchies and a greater number of competing word-formation rules on each class level seems to be typical of agglutinating languages and distinguishes them from many inflecting-fusional languages. It is even more evident with deverbal verb formation, as we have shown in Dressier and Ladänyi (1999). This is probably due to the greater syntactic relevance of verb formation. Morphological richness as we define it (Dressier and Ladänyi 1999), is a hyponym of morphological complexity. Morphological complexity reflects the total number of morphological patterns in a language, both productive and unproductive. It has to include unproductive patterns because it is calculated in terms of the total learning effort devoted to morphology. Morphological richness, on the other hand, refers exclusively to the active mechanism of morphology. It thus excludes unproductive patterns because - according to realistic models of the mental lexicon - they are lexically stored. The ideal type of an inflecting-fusional language has an inflection which is both rich and very complex. The ideal type of an agglutinating language, on the other hand, has a very rich morphology, which corresponds perfectly to the much larger role that morphology plays in respect to syntax in this language type, when compared with inflecting-fusional languages. But there are, ideally, no unproductive morphological patterns in agglutinating inflection which would extend complexity beyond richness. However, in word-formation, also agglutinating languages, such as Hungarian, have many unproductive rules. This reflects that productivity universally plays a greater role in inflection than in word-formation. Still Hungarian, as an agglutinating language, has relatively fewer unproductive word-formation rules than inflecting-fusional languages, and this is also true for denominal adjective formation (3.2.5).

Notes 1. Cf. Dressler et al. 1987; Kilani-Schoch 1988; Dressier 1997; the special issue of Sprachtypologie und Universalien-Forschung 49,3 (1996). 2. For studies on diminutives and other alteratives cf. Dressier and Merlini Barbaresi (1994).

72

Wolfgang U. Dressler and Märia Ladänyi

3. For reasons of space and homogeneity we will not deal with suffixoids such as G. -haft, -mäßig, -weise, -los, Η. -file, -fajta, -szerü. We will also exclude clearly "foreign" suffixes. 4. Kastovsky (1986: 594) elevates this prototypical property of derivational morphology to one of his two main functions of word-formation, viz. the function of "syntactic recategorization". 5. Cf. Dressier (1989). Thus we reject Spencer's (1998) syntactic analysis of relational adjectives as inappropriate to the morphological character of adjective formation. 6. Cf. Dressier (1977: 14; 1978: 426). On labelling see especially Kastovsky (1986: 595), Seiler (1975). 7. Kastovsky (1986: 595) attaches this function to his "syntactic recategorization" function, whereas Kastovsky and Kryk-Kastovsky (1998) treat the cohesive function of word-formation on its own. 8. Sometimes a) is called confrontative linguistics and only b) contrastive linguistics. Since these two perspectives are like the two sides of a coin we will consider both perspectives under the label of contrastive linguistics. 9. In the area of German and Hungarian denominal adjective formation we do not find the further differentiation between the case of foreign words that have unfitting properties, which must be fitted, i.e. accommodated to the properties of the respective language-specific system adequacy of the target language, and the case when there are no such unfitting properties. The latter is the case for our adjective formations, because foreign nouns must not be morphologically fitted into German or Hungarian. 10. More details in Kiefer (1998) and Ladänyi (1998). 11. More details in Ladänyi (1998) and Kiefer and Ladänyi (2000). 12. -sch is an allomorph of -isch when attached to proper names (Kühnhold et al. 1978). 13. Cf. Uluxanov (1977: 63); Barz (1988: 96). Corbin's (1997: 9) definition of a "morphological family" is very similar. 14. Exceptionally a derivational class may consist of just one microclass: then the assumption of a derivational class must be justified by symmetry to a sisterclass which consists of more than one microclass.

References Barz, Irmhild 1988 Nomination durch Wortbildung. Leipzig: Enzyklopädie. Bauer, Laurie 1997 Derivational paradigms. Yearbook of Morphology 1996:243-256.

On contrastive wordformation

73

Corbin, Danielle 1997 La representation d'une "famille" de mots dans le Dictionnaire derivationnel du frangais et ses corrölats th^oriques, möthodologiques et descriptifs. Recherches Linguistiques de Vincennes 26: 5-37. Dressier, Wolfgang U. 1977 Elements of a polycentristic theory of word-formation. Wiener Linguistische Gazette 15: 13-32. 1978 On a polycentristic theory of word-formation. PICL 12: 426-429. 1985 Morphology. In: Teun van Dijk (ed.), Handbook of discourse analysis. Volume 2,77-86. London: Academic Press. 1989 Prototypical differences between inflection and derivation. ZPSK 42: 3-10. 1997 On productivity and potentiality in inflectional morphology. (CLASNET Working Paper 7.) Montreal: Univ. de Montreal. Dressier, Wolfgang U., Willi Mayerthaler, Wolfgang U. Wurzel and Oswald Panagl 1987 Leitmotifs in Natural Morphology. Amsterdam: Benjamins. Dressier, Wolfgang U. and Lavinia Merlini Barbaresi 1994 Morphopragmatics. Berlin: Mouton de Gruyter. Dressier, Wolfgang U. and Märia Ladänyi 1998 On grammatical productivity of word formation rules. Wiener Linguistische Gazette 62/63: 29-55. 1999 Productivity in word formation (WF): a morphological approach. Ms. for Acta Linginguistica Hungarica. Dressier, Wolfgang U. and Anna M. Thornton 1996 Italian nominal inflection. Wiener Linguistische Gazette 55-57: 1-24. Jucquois, Guy 1993 Le comparatisme, une mise en situation. Univ. de Louvain. Kastovsky, Dieter 1986 The problem of productivity in word formation. Linguistics 24: 585600. Kastovsky, Dieter and Barbara Kryk-Kastovsky 1998 Morpholexical and pragmatic factors in text cohesion. In: Heinrich Ramisch and Kenneth Wynne (eds.), Language in Time and Space, Festschrift for W. Viereck, 462-475. Stuttgart: Steiner. Kiefer, Ferenc 1998 Alaktan. In: Katalin έ. Kiss, Ferenc Kiefer and P6ter Siptär: llj magyar nyelvtan, Budapest: Osiris. Kiefer, Ferenc and Ladänyi, Märia 2000 Morfoszintaktikailag semleges sz0k£pz£s. In: Ferenc Kiefer (ed.), Strukturalis magyar nyelvtan 3. Alaktan. Budapest: Akadömiai Kiadö.

74

Wolfgang U. Dressler and Maria Laddnyi

Kilani-Schoch, Marianne 1988 Introduction a la morphologie naturelle. Bern: Lang. Kühnhold, Ingeburg, Oskar Putzer and Hans Wellmann 1978 Deutsche Wortbildung 3: Das Adjektiv. Düsseldorf: Schwann. Ladänyi, Märia 1998 Szinonimia 0s/vagy funkciömegoszläs a magyar mell6kn6vk0pz0sben. In: T.Gecsö and Marcellin Spannraft (eds.), Α szinonimitäsrol, 97-104. Budapest: Tinta Könyvkiadö. Osman, Nabil 1994 Kleines Lexikon untergegangener Wörter. München: Beck. Peirce, Charles Sanders 1965 Collected papers. Cambridge, Mass.: Harvard Univ. Press. Pounder, Amanda 1987 Systemangemessenheit in der Wortbildung am Beispiel desubstantivischer Adjektivableitung im Deutschen. PhD dissertation, University of Vienna. Pöchtrager, Markus, Csanäd Bodö, Wolfgang U. Dressler and Teresa Schweiger 1998 On some inflectional properties of the agglutinating type, illustrated from Finnish, Hungarian and Turkish inflection. Wiener Linguistische Gazette 62/63: 57-92. Seiler, Hansjakob 1975 Die Prinzipien der deskriptiven und der etikettierenden Benennung. In: Hansjakob Seiler (ed.), Linguistic Workshop 3, 2-57. München: Fink. Spencer, Andrew 1998 Relational adjectives and the redundancy of lexical categories. In: Geert Booij et al. (eds.), Proceedings of the First Mediterranean Conference of Morphology, 19-28. Patras: University of Patras. Ülkü, Vural 1980 Affixale Wortbildung im Deutschen und Türkischen. Ankara: Üniversitesi Baslmevi. Uluxanov, Igor' S. 1977 Slovoobrazovatel 'naja semantika ν russkom jazyke. Moskva: Nauka. van Marie, Jaap 1985 On the paradigmatic dimension of morphological creativity. Dordrecht: Foris. Zemskaja, Elena 1992 Slovoobrazovanije kak dejatel 'nost'. Moskva: Nauka.

The greening of greening Andreas Fischer

1. Introduction Words have lives. They are born through coinage, processes of word formation, or borrowing; during their lifetime they may develop different senses; eventually they may become obsolete and die. The life span of words, however, is potentially unlimited and, moreover, no death is absolutely final. Seemingly dead, that is, archaic or obsolete words may come back to life at any time (as long as they are recorded somewhere), and it is more appropriate, therefore, to use another metaphor and to speak of the hibernation rather than the death of words (Schäfer 1983). Information technology, for example, has breathed new life into seemingly dead words such as computer and interface} The term computer was used from 1646 to 1855 with the sense 'one who computes', while the current sense 'machine for computing' has been attested since 1897. The term interface is a 19th-century word meaning 'a surface lying between two portions of matter or space' (attestations 1882, 1883). Marshall McLuhan revived it in 1962 in the sense 'a means or place of interaction between two systems, organisations, etc.', and the new concrete sense '(an) apparatus designed to connect two scientific instruments, devices, etc., so that they can be operated jointly' is attested from soon after that date (1964). In the following study I am concerned not with Lazarus-like revivals, but with a group of words that have had a marginal existence throughout history: the colour verbs, as I will call them in this paper. More specifically I will look at the recent flowering of one of them, the verb to green and its deverbal noun greening. The so-called colour terms have been the subject of intense scrutiny both in individual languages and cross-linguistically, but attention has been focused primarily on the adjectives, while the far less frequent colour verbs have remained largely unstudied.2 I use the term colour verbs for formations of the type (in English) to black or to blacken with transitive and/or intransitive meaning ('to make sth. black' or 'to turn black'). The study is based on information found in dictionaries (notably the OED) and, for present-day usage, in the British National Corpus.

76 Andreas Fischer 2. The morphology and semantics of colour verbs For the brief discussion of colour verbs below I will restrict myself to Berlin and Kay's (1969: 4) eleven basic colour terms: white, black, red, green, yellow, blue, brown, purple, pink, orange and grey. In the tables they are listed according to the developmental sequence suggested by Berlin and Kay (1969: §2). I have done this for purely practical purposes, however, and do not suggest that this sequence is in any way important for the colour verbs. According to the OED the following colour verbs are attested in the history of English:3 Table 1. English colour verbs transitive white (OE) black (OE) red (OE) green (OE) yellow (OE) blue (a1300) brown (OE/al300) purple (OE) pink (1828/46) orange (1542) grey (OE)

intrans.

OE OE whiten cl315 al225 blacken fal225-1736 fOE-1422 redden 1570 OE alOSO/1821 1558 1606 fl709 1570 c 1300/1859

transit.

intrans.

al300 1552 1611

al633 al300 1711

1968

1890

—

—

—

—

1432-50 1898/1927 —

al618

1646 1854 —

13..

—

pinken —

—

The colour verbs in English are formed according to two distinct patterns: conversion and derivation by means of the suffix -en. The first of these patterns, conversion, is both older and more common than the second. According to the OED it is evidenced from Old English onwards, although one can only speak of conversion in the true sense of the word after the loss of endings caused the adjectives and the verbs to be formally indistinguishable.4 Still according to the OED, all eleven basic colours except orange are attested,5 with first attestations ranging from the Old English period {white, red, green) to the 19th and 20th centuries (pink). No particular chronological sequence and no preference for either the transitive or the intransitive meaning is evident. It would appear that the possibility to form colour verbs by conversion was always (and is still) latent in English, and that such verbs are actually coined and used when there is a need for them.

The greening of greening 77 The situation is different with the second, derived pattern, since only four such verbs are attested {blacken, whiten, redden, pinkeri). The apparently limited productivity of the pattern is due to a phonological restriction: "In all stages of productivity [of the suffix -en], only stems ending in a stop or fricative have lent themselves to this type of derivation [...]" (Marchand 1969: 272). The four colour adjectives in question {black, white, red, pink) all end in a plosive while the remaining seven do not. The 'division of labour' between the two patterns is unclear: black and blacken, white and whiten as well as the recent pink and pinken exist side by side, while redden has replaced the older red. A brief look at German for purposes of comparison shows a very different situation:6 Table 2. German colour verbs. transitive weiss schwarz rot grün gelb blau braun violett rosa orange srau

weisse(l)n schwärzen röten begrünen vergilben bläuen bräunen

intransitive —

—

erröten (er)grünen vergilben blauen —

reflexive —

sich schwärzen sich röten —

—

—

sich bräunen

—

—

—

—

—

—

—

—

—

—

ergrauen

—

Morphologically, German colour verbs are marked mostly by /-mutation and/or by preflxation. /-mutation and prefixation together tend to distinguish transitive verbs from intransitives quite clearly, but the marking does not seem to follow any completely logical principles, /-mutation, for example, is found everywhere except in the two intransitive verbs blauen and ergrauen, while preflxation is found above all, but not exclusively, with the intransitives (the only prefixed transitives being vergilben and begrünen). From the point of view of morphology, then, the English and the German colour verbs present a very different picture. They have one thing in common, however: while the available morphological means would make it possible in each case to establish a logical and coherent system, in practice this has not happened, and the colour verbs appear to be a weakly structured part of the lexicon. The reason for this is not easy to find, but it

78 Andreas Fischer may have to do with the semantics of these verbs, an aspect I will turn to in the next section. The discussion so far may have created the impression that to redden, for example, is the only possible expression for the notion 'to make/turn red', but verb-adjective combinations like to make!colour!paint sth. red (transitive) or to become/turn red (intransitive) may, of course, also be used. In fact, it would seem that in actual use these verb-adjective combinations are much more frequent than the colour verbs themselves. Generally speaking the colour verbs are employed above all for what may be called natural processes, that is for changes of hue/colour due to some natural cause. Among these natural causes we find the sun or blood for reddening (sky, clouds; skin, face, ?water); spring for greening (nature); light together with old age, for yellowing (paper, leaves, teeth, skin); etc. In addition to their use in connection with such natural processes, colour verbs may denote very specific, technical ones. In this latter case they form part of specialised vocabularies.7 The close connection between colour verbs and certain veiy specific processes (be they general or technical) means that colour verbs are frequently found in collocations.8 The relative scarcity of the colour verbs is thus directly related to a rather narrow semantic range and a high degree of collocability.

3. To green and greening The above general points may be illustrated with the verb to green and its uses. In its intransitive use to green is found from Old English to the present, always in the 'natural sense' of'to become green, as growing herbage; occas. to appear or look green; to become covered with verdure, to be 'clothed' with green. (Also with over.)' (OED). The earliest and the latest quotations in the OED are the following: (1)

α1000 Boeth. Metr. xi. 57 (Sedgefield), Hcefl> se celmihtija...Pcet jewrixle jeset... wyrta growan, leaf grenian. 1899 Daily News 15 Apr. 8/1, The wild-rose briars will be shooting strongly, the elder greening over.

The transitive sense, documented from 1570 onwards, is glossed as 'to colour or dye green; to soil or stain with green; to impart a green colour to; to cover with verdure or vegetation (also with over); to 'clothe' with

The greening of greening 79 green'. The earliest and the latest examples may again serve as illustrations: (2)

1570 B. Googe Pop. Kingd. 10, The Rest with silver garnisht is, and plaitedfine and neat Least it shoulde greene his holy hands. 1891 T. Hardy Tess I. iii, The..white frock..which she had so carelessly greened..on the damping grass.

The OED, finally, lists two technical senses of the transitive verb, namely (2.b) 'Oyster-culture. To turn (oysters) green in the gills by putting them in pits.' and (2.c) ''Plumbing. To rub (new sheet-lead) with some green vegetable' (obsolete). The OED also lists greening as a verbal noun and a participial adjective, respectively. These are the two latest examples: (3)

1883 Stallybrass tr. Grimm's Teut. Myth. III. 959 note, The withering or greening of a tree is bound up with the fate of a country. 1858 Bailey Age 52, Slowly greening woods Make dim the distant view.

The above dates suggest that to green and greening did not develop any new senses in the first half of the 20th century, but this was to change soon. In the second volume of the Oxford Dictionary Additions Series (Simpson and Weiner 1993) the verb to green is listed with the new, additional senses (2.d) 'to render (an urban area) more green or rural in appearance, esp. by planting trees, etc. and developing parkland; also, to reclaim (a desert area)' (attested since 1979) and (2.e) 'to render (a person, etc.) sensitive to ecological issues; hence, to make (something) less harmful to the environment, to adapt along environmentally friendly lines' (attested since 1985). The deverbal noun greening, likewise, has the new senses (2.d) 'The process or result of planting or cultivating greenery, esp. as part of a programme of urban revitalization' (1955) and (4) 'The process or result of making or becoming aware of ecological issues, or of espousing environmentalism. Freq. in political context' (1986). A slightly earlier, figurative sense (3) is glossed as 'Rejuvenation, renewal, esp. after a period of stagnation or decay; an instance of this, orig. and chiefly U. S.' It owes its existence to a book by C. A. Reich, first published in 1971, and entitled The greening of America. These new 'ecological' senses have developed organically from the original transitive sense of the verb and the deverbal noun: greening 'making green/covering with green' leading to 'making urban or infertile land green' leading to both 'rejuvenation' and 'developing ecological principles and practices'. Insofar as greening is

80 Andreas Fischer associated with the healthy growth of vegetation and, generally, with spring and summer as against autumn and winter, it has always had inherent positive connotations. Deliberately or inadvertently, the ecological movement has made use of the term and the positive values it transmits and it has made green, once (almost) a colour like any other, a synonym of 'ecological'. As a result one would expect to green and greening to be used more frequently than other colour verbs in present-day English and one might want to learn something about the relative frequency of its older and its newer, 'ecological' senses. Colour verbs are rare, however, and one needs a very large body of texts to find a sufficient number of examples. In order to investigate the uses of to green and greening in present day English I therefore turned to the largest corpus presently available, the British National Corpus (BNC), which contains over 100 million words of present-day (post 1960s) British English.9 It is not easy to keep the colour verbs apart from the more frequent homophonous adjectives, and even a tagged corpus like the BNC is likely to contain many wrongly assigned tags. Since the present study is no more than an exploratory survey I decided to avoid this problem for the time being by disregarding the ambiguous word forms green and greens and to concentrate on the more easily identifiable greened and greening™ A computerised search for these forms in the complete corpus yielded eight instances of greened and 64 instances of greening (actually 67, but three of them turned out to be names). Contrary to my expectations these numbers establish to green as a colour verb with no more than average frequency of use. Some colour verbs are attested more frequently (yellowed/yellowing: 169 examples, greyed/greying: 137, redden/reddening: 71),11 while others are decidedly rare (whited: 10; blued!blueing: 9; pur10 pled: 4; oranged: 1 ). The corpus of 72 examples is given at the end of the paper, and my final remarks will be concerned with the senses of green attested in it. Six examples13 illustrate the basic intransitive and transitive senses 'to become green, as growing herbage; to become covered with verdure' and 'to cover with verdure or vegetation' respectively, and in three further examples14 greening/greened means 'turning/having turned green (through some cause)': "a suspicious woman in greening black clothes" (21), "copper strips and inlays, greened with age and atmosphere" (40), "the teeth it exposed were greened with mould" (56). Two special senses are also on record: in example (69) "greening a sumptuous 4-iron" the verb must be a golfing term meaning 'to drive an iron into the ground/green',15 while in examples (71) and (72)16 "greened up" probably has the slang sense 'made to appear 'green', simple, or gullible' (OED green v., sense 3).

The greening of greening

81

The deverbal noun greening is also attested with two specialised 17 senses: it may refer to a potato disease and 'a vegetable colouring matter' (OED greening vbl. n., sense 2.b).18 The examples discussed so far number 24, which leaves 48, all of which show the new 'ecological' senses in one form or another. Some illustrate the sense 'planting or cultivating greeneiy, esp. as part of a programme of urban revitalisation' ("existing derelict land was needed for greening the cities" (1)), while others exemplify the more abstract 'The process or result of making or becoming aware of ecological issues, or of espousing environmentalism' ("the sudden greening of British politics" (3)). Yet others could have either of these senses or might even mean 'rejuvenation, renewal' ("The greening of Europe" (20) or "the greening of Sellafield" (55)), but cannot be classified without taking into account more context. Two facts are very obvious, however: the new 'ecological' senses dominate all others, and they have led to a revitalisation, a veritable prospering (or greening?) of green and greening. Noun phrases of the type greening N> N's greening and, above all, the greening of Ν are especially popular: in the first 20 examples of the corpus alone we find "greening the cities" (1), "greening inner cities" (5), "Mrs Thatcher's greening" (11) and no fewer than eleven examples of the type the greening of N: "the greening of the countryside / all Whitehall departments / the market / the greys / the electorate / the Kremlin / the Archbishop /commerce / the Cornish Alps / Europe". Is there any person, body or entity, one might ask, that is not subject to greening? "The greening of greening'': in this brief paper I have suggested ways in which the colour verbs of English and other languages may be studied and, focusing on to green and greening, I have shown how the ecological movement has led to the rejuvenation or renewal, in short the greening, of a once rather marginal colour term. Appendix: Greened and greening in the British National Corpus19 (instances of greened are marked *) (1) (2) (3) (4)

Yet only weeks ago he was saying that existing derelict land was needed for greening the cities [...]. (A2P:408) It is the blueing, not the greening, of the countryside. (A3T: 145) Appropriate too, because the sudden greening of British politics [...] introduces a catalytic ingredient into the brew known as Thatcherism. (A4U:092) The Secretary of State paved the way for the greening of all Whitehall departments through a White Paper [...]. (A50:600)

82

Andreas Fischer

(5)

The activities include [...] "greening" inner cities by developing disused allotments. (A7H:1301) Below him in the valley, the river slid by greening banks [...]. (A7J:2235) Are we witnessing the next step in the greening of the market, or is it just the latest move in the marketing of the greens? (AAG:224) [...] We're au fait with entropy, Gaia, Ecosphere, synergy Words for greening. But their meaning's A linguistic mystery [...]. (AAP:364)20 Eco-politics: The greening of the greys. (AB6:0189) While the greening of the electorate was rapidly taken aboard [...]. (AB6:0346) [...] the day Mrs Thatcher's greening was declared in a speech [...]. (AB6:1154) One British reaction at the time to the greening of the Kremlin [...]. (AB6:1155) *The electorates are being greened. (AB6:1442) The recent greening of the world means that [...]. (ABS.0260) [...] The Times ran an editorial entitled "The Greening of the Archbishop". (ABV:0368) To a greening valley white with sheep. (AC6:1412) MORE on the greening of commerce. (ANX:2789) *[...] the grey rocks [...] still more beautifully enriched by many greened mosses [...]. (B3H:1079) The Greening of the Cornish Alps. (B7H:0456) The Greening of Europe. (B7L:2360) [...] a suspicious woman in greening black clothes [...]. (C85:3643) *[...] in Freiburg narrow, calmed streets with insufficient space for treeplanting have been greened by the simple expedient of growing climbing plants [...]. (C8F:0914) [...] dramatic specific results, such as Buxtehude's growth of cycling or Moabit's greening of the environment. (C8F:1141) Such benefits of environmental traffic management might include the greening of the streetscape [...]. (C8F:1159) There is, surprisingly, no specific encouragement for greening project areas in the Feet First proposals [...]. (C8F: 1406) The published Rob Gray's booklet The Greening of Accountancy [...]. (CBX:2739) The TUC has issued a policy paper, Greening the Workplace [...]. (EC0:081) Affected by GREENING. (EEA:039) Greening of potatoes. (EEA:085) How far does the greening penetrate? (EEA:219) [...] free from cuts, blemishes, greening and rots. (EEA:352)

(6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22)

(23) (24) (25) (26) (27) (28) (29) (30) (31)

The greening of greening (32) (33) (34) (35) (36) (37) (38) (39) (40)

(41) (42) (43) (44) (45)

(46) (47) (48) (49) (50) (51) (52) (53) (54) (55) (56) (57)

83

[...] so that greening is reduced to a minimum. (EEA:423) Freedom from internal bruising, disease, cracking and greening. (EEA:474) Freedom from internal bruising, disease and greening. (EEA:505) [...] evenly shaped tubers over 25mm with no greening. (EEA:531) The tubers should ideally contain little disease, greening or severe mechanical damage. (EEA:546) [...] maraschino; vegetable greening; a little water [...]. (EFU.1376) [...] use a very little vegetable greening [...]. (EFU:1378) [...] the redundant vegetable greening [...]. (EFU:1394) *[...] a vista of church spires, each one different in design yet each unmistakably decorated by copper strips and inlays, greened with age and atmosphere. (HA5:0029) [...] Higher environmental standards are now a reality [...] in this ever greening world. (HAU:243) [...] vital in ensuring the greening of the whole Company. (HBC:591) [...] vital in ensuring the greening of the whole Company. (HBC:600) The "greening of work environments" is something that is being taken seriously both internally and externally. (HBG:200) [...] interior and exterior appearance of premises, recycling, "greening" buildings with plants and flowers, providing no-smoking areas [...]. (HC3:418) The "cost of greening" must be fairly shared [...]. (HH3:02939) Greening. (HH3:04988) Now there is a danger that this "poor world" will fall victim of the greening process. (HH3:08273) And this poverty will need correcting or greening will translate into more domination. (HH3.08278) Your understanding of the earth's plight is greening everything and everyone around you. (HH3:11048) [...] energy-saving, waste recycling and the greening of derelict land. (HJ 1:21507) Many people believe that the increased coverage is contributing to the "greening of public opinion" [...]. (HJ1:23640) Greening the office. (HJM:071) Calling for a "greening of politics" Gorbachev said [...]. (HKP:1689) [...] former Science Editor of the Financial Times David Fishlock wrote about the Greening of Sellafield. (HPB:076) *The teeth it exposed were greened with mould [...]. (HTM: 1230) Then spring again, and with the greening of the land the creature gave birth to birds [...]. (HTM: 1356)

84

Andreas Fischer

(58) *[...] the trees [...] are still half-wintry at the top when already fully greened lower down. (FA2:0676) (59) The magnificent Tijou gates of the main entrance were no longer opened and the drive to the house was greening with disuse. (GWG:2309) (60) Training Course: Greening the Content of Local Plans. (GXG:0845) (61) The 1990 White Paper requires the government to publish an annual progress report showing how far it has got with the "greening" of its policies and actions. (GXG:3963) (62) [...] the interests of pedestrians, riders and cyclists can be greatly enhanced as part of a wider "greening" of local transport policies. (GXG.5336) (63) Greening the Machinery of Government suggests that [...]. (J2R:405) (64) A multi-national project, "Greening the Danube" has been launched [...]. (J3E:075) (65) WWF calls for greening of world trade body. (J3H:210) (66) Smith received the award for his commitment to "greening" the Labour Party [...]. (J3E:351) (67) [...] Phil Sedwill is to take over the "Greening The Tees Corridor" scheme as habitat creation project officer. (K4W:05405) (68) THE GREENING of Edinburgh may take a step forward [...]. (K56:123) (69) [...] he followed that perfectly shaped tee shot by greening a sumptuous 4iron. (K5A:4172) (70) [...] a major study by the Robert Gordon University, aimed at improving the greening of Scottish industry and commerce. (K5D:05954) (71) *Bit all all mix - all me greened up. (KDA:4120)21 (72) *God knows what this survey's gonna make of this all me greened up and right shitted up [...]. (KDA:4121).

Notes 1. All historical information on English words in this paper is taken from the second edition of the OED (Simpson and Weiner 1989) on CD-ROM (1992). 2. The classic study is Berlin and Kay (1969). The 1991 reprint of this book contains an additional "Bibliography of color categorization research, 19701990". For English see Wyler (1992). Berlin and Kay (1969) do not mention colour verbs at all; Wyler deals with them on a single page (1992:137). 3. The date of first attestation in the left hand column (in brackets) is that of the colour adjectives. 4. Thus Old English grene (adjective) and grenian (verb) became homophones only in Middle English. 5. A verb to orange is not listed in the OED, but my search of the 100-million word British National Corpus (see below) has brought to light one example:

The greening of greening

6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19.

20.

21.

85

As they quitened [sic, quietened?], the sun-colours on the stone yellowed, oranged, reddened. (AON: 2421) A cross-linguistic study of colour verbs might yield interesting results. Examples of technical senses of green are given in the next section. For colour collocations see Bennett (1988). See Aston and Burnard (1998) and Burnard (1995) for details. My thanks go to lie. phil. Hans-Martin Lehmann for helping me extract the relevant data. I realize that the omission of the uninflected form green and of (he/she) greens may skew my results. In a full study all forms of the verb will have to be included. Black/blacken and brown present special problems and I have not yet counted them. See note 5. Note also Max Apple's The oranging of America (1976), a title clearly modelled on Reich's The greening of America (1971). Examples 6, 16, 18, 57, 58, 59. Examples 21,40, 56. This sense is not attested in the OED. See note 21. 9 examples (28 to 36), all from the same text (EEA = Potato marketing materials, 1991). This sense is not recorded in the OED. 3 examples (37-39), all from the same text (EFU = An omelette and a glass of wine by Elizabeth David, 1987). The passages are identified by a three character code for each text excerpted (for example A2P or AAG), followed by a number for the segment. For details see Burnard (1995). Omissions marked [...] are mine. This text looks like some sort of verse: We 're au fait with entropy / Gaia, Ecosphere, synergy / Words for greening. / But their meaning's / A linguistic mystery. (AAP - Guardian, electronic edition of 19891222,1989, world affairs material). Examples (71) and (72) are spoken English (KDA = 132 conversations recorded by one person between November and December 1991 with eleven interlocutors).

References Aston, Guy and Lou Burnard 1998 The BNC handbook: Exploring the British National Corpus with SARA. (Edinburgh Textbooks in Empirical Linguistics.) Edinburgh: Edinburgh University Press. Bennet, Thomas J.A. 1988 Aspects of English colour collocations and idioms. (Anglistische Forschungen 197.) Heidelberg: Carl Winter.

86

Andreas Fischer

Berlin, Brent and Paul Kay 1969 Basic color terms: Their universality and evolution, [repr. 1991] Berkeley: University of California Press. Burnard, Lou (ed.) 1995 Users' reference guide for the British National Corpus. Oxford: Oxford University Computing Services. Marchand, Hans 1969 The categories and types of Present-Day English word-formation: A synchronic-diachronic approach. 2nd, completely revised and enlarged edition. [1st edition Wiesbaden: Otto Harassowitz, I960]. München: C.H. Beck. OED = Simpson, John A. and Edmund S.C. Weiner (eds.) 1989 The Oxford English dictionary. 20 volumes. 2nd edition. Oxford: Clarendon Press. Schäfer, Jürgen 1983 Tod und Winterschlaf in der lexikographischen Tradition: Ein Problem moderner Lexikographie. In: Jörg Hasler (ed.), Anglistentag 1981 [Trier]: Vorträge, 35-43. (Trierer Studien zur Literatur 7.) Frankfurt/M. - Bern: Peter Lang. Simpson, John and Edmund Weiner (eds.) 1993 Oxford English Dictionary Additions Series. Volume 2. Oxford: Clarendon Press. Wyler, Siegfried 1992 Colour and language: Colour terms in English. (Tübinger Beiträge zur Linguistik 364.) Tübingen: Gunter Narr.

Middle English beck in the Midlands: the place-name evidence Jacek Fisiak

Historical word-geography of English, particularly Old and Middle English, is still a very much underdeveloped area of research although its importance has been recognized for a long time (Mcintosh 1973, 1978). The situation is even worse when it comes to the use of the onomastic evidence which often can provide the density of coverage better than texts, sparse and unevenly distributed in space in the earlier history of English. The pioneering efforts going back to the beginning of the twentieth century (Ekwall 1913; 1963; see also Kristensson 1989) have not been followed on a massive scale although some important works have appeared, more recently those by Kristensson (e.g., 1991), who has demonstrated beyond any doubt the relevance of the onomastic word-geography for the history of English. The present paper is intended as a modest contribution to the field and will deal with the distribution of ME beck, a word of Scandinavian origin (OSc bekkr), in the Midlands. ME bee has survived in ModE dialects as beck in the area between Lines A and Β as indicated on Map 1 (see below). According to Smith (1956:26): bekkr (bekkjar gen.sg.) ON, 'a stream, a beck', is extensively used throughout the Danelaw and the NCy (except for Nb), taking the place of broc and burna. It frequently has a village or other place-name prefixed, as in Dalby, etc. and occasionally an older r.n., as in Dover Beck Nt. It replaces OE broc in Holbeck Nt, bece in Skirbeck L as well as burna in some Du names. In minor names and f.ns. it is common in Cu and Y, but rare in some parts of L....

In what follows we shall look at the occurrence of beck in place-names and occasionally by-names in the southern part of the Danelaw, i.e., some of the Midland counties, from the time of Domesday Book to 1500. The material comes from relevant volumes published by the English PlaceName Society (EPNS) as well as from Ekwall's Concise dictionary of English place-names (1960), Mill's Dictionary of English place-names (1991) and Cameron's Dictionary of Lincolnshire place-names (1998).

88 Jacek Fisiak Each place-name will be given in the form of its first occurrence. If it appeared in Domesday Book (DB), the DB form will also be provided. LEGEND •

SCOTLAND

ME place andfieldnames with beck

A Northern boundary of ModE beck Β Southern boundary of ModE beck [Orton-Wright 1974: 84] ····

The Great Scandinavian Belt [Samuels 1985: 270]

A

Β

»•••J WALES

Mapl. Middle English place and field names including the element beck

Middle English beck in the Midlands: the place-name evidence

89

Material LINCOLNSHIRE: Fulebec 1986DB, 1130 (FULBECK, DLPN), Handebec 1210 (HANBY, DLPN), Hamundebek 1315 (HAMMOND BECK, DLPN), Holebec 1198 (HOLBECK, DLPN1), Pincebec 1086DB, 1188 (PINCHBECK, DLPN), Scirebec 1086DB, 1125 (SKIRBECK, DLPN), le Bek fii.2 1457 (Beck Field, EPNS LXXIII, 33), bucbek' fii 1298 (Buck Beck, EPNS LXIII, 117), Beckfurlonges fii 1457 (in SCARTHO, EPNS LVIV-LXV, 239), beckedaile fii Hy3 (beck Planatation, EPNS LXIV-LXV, 239), atte Bekke de Nirth'kelsey 1329 (North Kelsey Beck, EPNS LXIV-LXV, 183), Atte becke 1327 (BECKS GATE, EPNS LXIV-LXV,60), Cuntebecsic fii Hy3 (in CAISTOR, EPNS LXIX-LXV, 93), Atte Bek' de Ellesham 1308 (The Beacon in ELSHAM, EPNS LXIV-LXV, 107), ultra cokebec fii ml3 (in ESHAM, EPNS LXIV-LXV, 107), bee 12c fii (in HARBROUGH, EPNS LXIV-LXV, 143), atte Bek 1332 (in HORKSTOW, EPNS LXIV-LXV, 163), le braythebecke fn cl295 (in IMMINGHAM, EPNS XIV-LXV, 169), Staynwellebekfurlanges fii 1245 (in IMMINGHAM, EPNS LXIV-LXV, 170), le wellebeckesse fii 1334 (in IMMINGHAM, EPNS LXIV-LXV, 171), Bek fii 1281 (in KIRMINGTON, EPNS LXIV-LXV, 218), Beckedaile fii Hy3 (in NETTLETON, EPNS LXIV-LXV, 245), Helingbeck' fii Hy3 (in STALLINGBOROUGH, EPNS LXIV-LXV, 276), Bechmar fii cl240 (in ULCEBY, EPNS LXIV-LXV, 296), Seggebecke fii 1210 (in FOTHERBY, EPNS LXXI, 21), the bek' fii 1497 (in LITTLE GRIMSBY, EPNS LXXI, 23), Cringelbec fn 1147 (in NORTH ORMSBY, EPNS LXXI, 34), Thorne ad Bee Hy3 (in Cabourne, EPNS LXXI, 70), Bekfeld' 1202 (BECKFIELD, EPNS LXVI, 3), Bekfurlonges 1327 (OWERSBY BECK, lost, EPNS LXVI, 83), Ouerbeck' fh 13c (in BINBROOK, EPNS LXVI, 10), Wellebeck fii eHy3 (in CLAXBY, EPNS LXVI, 26), Vuerhoubeck fii Hy3 (in NORMANBY LE WOLD, EPNS LXVI, 77), Wrangbek fh 1393 (in OWERSBY, EPNS LXVI, 94), Northebek fh cl376 (in MIDDLE RASEN, EPNS LXVI, 113), Becklangfurlonges fh cl244 (in WEST RASEN, EPNS LXVI, 118), ultra bee fii cl 115 (in USSELBY, EPNS LXVI, 170). NOTTINGHAMSHIRE: Forewidebec fii 13c (in RUFFORD, EPNS XVII, 298), Ballandebec fii 1226 (in LAXTON, EPNS XVII, 299), Binetheglasbek fh cl220 (Glassbyke Close, EPNS XVII, 314), Holebeck fii 13th (Holebeck Close, EPNS XVII, 317), le Beck fii 1257 (Beck Close, EPNS XVII, 319), Hollebec 1227 (HOLBECK, EPNS XVII, 83), Holebeck cl 180 (HOLEBECK, EPNS XVII, 189), Maplebec 1086DB, 1165 (MAPLEBECK, EPNS XVII, 189) Oswardesbec 1130 (OSWALDBECK, lost, EPNS XVII, 43), Welbech cll61 (WELBECK, EPNS XVII, 103), Wodebec cl300 (WOODBECK, EPNS XVII, 58), Holbeck Woodhouse 1330 (in HOLBECK, EPNS XVII, 84), Wottonbeke Hy6 (WOODEN BECK HILL, EPNS XVII, 41), Mortonbek fii, 1330 (Beck Meadow and Nook,

90

Jacek Fisiak

EPNS XVII, 311), Crossek fo 1333 (Crees Beck Furlong, EPNS XVII, 320), Brillebecdyk 1335 (Brimblebeck lane in SUTTON ON TRENT, EPNS XVII, 196), DERBYSHIRE: Hegelbec 10865DB, Heselbek Hy3 (HAZLEBADAGE, EPNS XXVII, 118), leHoldebec fii 1297 (in BAKEWELL, EPNS XXVII, 38), Staynbekford fii 1330 (in DOVERIDGE EPNS XXIX, 552), Karebek fo 1226 (in NORTH WINGFIELD, EPNS XXVIII 335), Bectune 1086DB, Becton 1236 (BEIGHTON, DEPN), Beksiche 1389 (Becksitch Lane in BELPER, EPNS XXIX, 526), Thirbeck 1325 (FIRBECK HO, EPNS XXVIII, 328), LEICESTERSHIRE: Cotesbece 1986 DB (COTESBACH, DEPN). RUTLAND: Beclane fo 1495 (in HAMBLETON, EPNS LXVII - LXIX, 184), le beck fh 1366 (in NORTH LUFFENHAM, EPNS LXVII-LXIX, 260). NORTHAMPTONSHIRE: Walbek fh 1301 (Walbeck, EPNS x, 284, Brakebec fo Hy3 (in DRAUGHTON, EPNS X, 260), Holebeke diche fo 1216 (in HANGING HOUGHTON AND LAMPORT, EPNS X, 284), Fulbek fh 1400 (in IRTHLINGBOROUGH, EPNS X, 260), Holbek fo 1435 (in DRAUGHTON, EPNS X, 260). HUNTINGDONSGIRE: Holebek fo 1279 (in EYNESBURY, EPNS III, 292). BEDFORDSHIRE: Walebek fo 1273 (in CLOPHILL, EPNS III, 292). BUCKINGHAMSHIRE: Becentone 1986DB, Bechentone 1160 (BEACHAMPTON, EPNS II, 59). CAMBRIDGESHIRE: le Bekke fo 15c (Beck Close, EPNS XIX, 368), Richard de Bolebek 1251 (BULLBECK MILL, EPNS XIX, 72), le Bek fo 1342 (in WICKEN, EPNS XIX, 312). NORFOLK: Crowbeck fo 1232 (in WORSTEAD, EPNS LXXII, 207), Bec 1086 DB (BECK, DEPN), BEKKES fo 1368 (in ASHBY, EPNS LXXII, 44), debecke 1466 (DEWBECK, EPNS LXXII, 71), Bec' fo cl230 (DU BECK, EPNS LXXII, 43). SUFFOLK: Becles 1086DB, Beclis 1157 (BECCLES, DEPN), Gosebech 1179 (GOSEBECK, DEPN). WARWICKSHIRE: Withebec fo Ed III (n. Rugby, EPNS XIII, 321). The largest number of occurrences of bec can be found, as one might expect, in Lincolnshire, Nottinghamshire, Derbyshire, and Norfolk. In Northamptonshire their number is relatively small despite strong Scandinavian influences. A fairly large number of bec forms remains outside the Great Scandinavian Belt (Samuels 1985), as can be seen in Map 1. It

Middle English beck in the Midlands: the place-name evidence

91

seems that if the concept is to remain valid, the Belt should include all the areas where Scandinavian influence is visible not only in lexical borrowings but also in place-names (see the distribution of place-names in -by, -thorp, -thwaite, etc.). Comparing the distribution of Middle English bee in place-names with the southern boundary of the occurrence of the Modern English dialectal word beck (Line B) we can see the receding movement of the isogloss in the northerly direction as in the case of other Scandinavian borrowings, for instance kirk.

Notes 1. DLPN = Cameron, Kenneth. 1998. A dictionary of Lincolnshire place-names, DEPN = Eckwall, Ellert. 1960. The concise Oxford dictionary of English place-names, EPNS = English Place-Name Society. 2. fn = field name 3. Hy3 = the reign of Henry III (1216-1272)

References Anderson, Thorsten and Karl Inge Sandred (eds.) 1978 The Vikings. Proceedings of the Symposium of the Faculty of Arts of Uppsala University, June 6-9, 1977. Uppsala: Uppsala University Press. Breivik, Leiv E., Arnoldus Hille and Stig Johansson 1989 Essays on English language in honour of Bertil Sundby. Oslo: Novus Forlag. Cameron, Kenneth 1959 The place-names of Derbyshire. Parts 1-3. (The English Place-Name Society XXVII-XXIX). Cambridge: Cambridge University Press. 1985-97 The place-names of Lincolnshire. Parts 1-5. (The Place-Name Society LVIII, LXIV, LXV, LXVI, LXXI, LXXIII). Nottingham: English Place-Names Society. 1998 A dictionary of Lincolnshire place-names. Nottingham: English Place-Name Society. Cox, Barrie 1989-92 The place-names of Rutland (English Place-Name Society LXVIILXIX). Nottingham: English Place-Name Society. 1998 The place-names of Leicestershire. Part one. (The Survey of English Place-Names LXXV). Nottingham: English Place-Name Society.

92

Jacek Fisiak

Ekwall, Ellert 1913 Die Ortsnamenforschung ein Hilfsmittel für das Studium der englischen Sprachgeschichte. Germanisch-Romanische Monatsschrift 5: 592-608. 1960 The concise Oxford dictionary of English place-names. (4th edition) Oxford: Oxford University Press. 1963 A problem of Old Mercian phonology in the light of West Midland place-names. Namn och Bygd 51:16-48. Gover, J.E.B., Mawer, Allen and Frank Merry Stenton 1936 The place-names of Warwickshire. (The English Place-Name Society XIII). Cambridge: Cambridge University Press. 1975 The place-names of Northamptonshire. (The English Place-Name Society X). Cambridge: Cambridge University Press. 1979 The place-names of Nottinghamshire. (The English Place-Name Society XVII). Nottingham: English Place-Name Society. Kristensson, Gillis 1989 Place-names and linguistic geography. In: Lei ν Ε. Breivik, Arnoldus Hille and Stig Johansson (eds.), 155-164. Oslo: Novus Forlag. 1990 Middle English kirk and cognates in the Midlands. In: Lars-Erik Edlund and Gunnar Persson (eds.), Language - The time-machine, 5161. Stockholm: Almqvist & Wiksell. Mcintosh, Angus 1973 Word geography in the lexicography of Medieval English. Annals of the New York Academy of Sciences 211: 55-66. 1974 Middle English word-geography : Its potential role in the study of the long-term impact of the Scandinavian settlements upon English. In: Thorsten Anderson and Karl Inge Sandred (eds.), 124-130. Uppsala: Uppsala University Press. Mawer, Allen and Frank Merry Stenton 1925 The place-names of Buckinghamshire. (The English Place-Name Society II). Cambridge: Cambridge University Press. 1926 The place-names of Bedfordshire and Huntingdonshire. (The English Place-Name Society III). Cambridge: Cambridge University Press. Mills, Anthony D. 1991 A dictionary of English place-names. Oxford: Oxford University Press. Orton, Harold and Nathalia Wright 1975 A word geography of England. London - New York - San Francisco: Seminar Press. Reaney, Peter H. 1943 The place-names of Cambridgeshire and the Isle of Ely. (The English Place-Name Society XIX). Cambridge: Cambridge University Press.

Middle English beck in the Midlands: the place-name evidence

93

Samuels, Michael L. 1985 The Great Scandinavian Belt. In: Roger Eaton, Olga Fischer, Willem Koopman, Frederike van der Leek (eds.), Papers from the 4th International Conference on English Historical Linguistics (Amsterdam, 10-13 April 1985), 269-281. Amsterdam: Benjamins. Sandred, Karl Inge 1996 The place-names of Norfolk. Part two. (The English Place-Name Society LXXII). Nottingham: English Place-Name Society. Sandred, Karl Inge and Bengt Lindstrom 1989 The place-names of Norfolk. Part one. (The English Place-Name Society LXI). Nottingham: The Place-Name Society. Smith, Albert Hugh 1956 The place-name elements. Part one. (The English Place-Name Society XXV). Cambridge: Cambridge University Press.

Conceptual and semantic change in the history of English1 Manfred Görlach

1. Introduction Authors dealing with semantic change in various European languages including English have tended to describe the phenomenon with ill-defined and sometimes idiosyncratic categories. Others have relied on cultural history (that is, changes in objects or concepts) or on psychology rather than on a more narrowly defined linguistic categorisation. Allowing for the specific problems of reconstruction of linguistic items in early periods, I still believe that a method based on functional structuralist principles yields the best insights into the synchronic meaning of, say, lexical items in Old English, and for a quantificational diachronic description of changes in content through various periods of English. In a second step of analysis we may then try to account for such changes by identifying intralinguistic conditions (such as differentiations of former synonyms) and extralinguistic causes (such as cultural change or contact-induced developments).

2. Basic assumptions Various methods have been proposed in the course of the past three thousand years to account for the classification of meaning, for determining the relationship between concepts and things, and for explaining semantic change. It will therefore be good to start with some basic assumptions on how I wish to define the topic and how I describe semantic relations. This will be followed by a classification of reasons adduced to explain change of meaning and end with a few problem cases that might better be excluded from my discussion. A topic like this is difficult to treat on a purely theoretical level; I will therefore illustrate my arguments with many exampies. Meaning, thus the general consent, is to be defined as the solidary and symmetric relationship of form and content, which together constitute the linguistic sign visualised in the universally used model, the semiotic triangle (form-content-referent). The relationship is arbitrary, that is, we cannot

96

Manfred Görlach

make assumptions about the meaning on the basis of the form, and vice versa. Even onomatopoeic signs are not exempt from this statement: the relationship in these is also conventional, that is, it is by agreement among the members of the same speech community. A very important conclusion from all these insights is that the arbitrariness makes it possible for both the form and the content to change {sound change and semantic change), as long as new conventions are agreed on by the speech community. It is also assumed that meanings of individual words are largely independent of context (so that lexical semantics and dictionary entries become possible) and that meanings (or sememes) are discrete on the level of langue (as phonemes are on the phonological level). This discreteness is established by oppositions, a fact which makes meaning language-specific (and period-specific), one of the major problems encountered in translation. The semantic model here chosen has recently been criticised as not adequate; adherents of prototype semantics have argued that meanings are not entirely discrete even on the level of langue and that we have to base our descriptions on central meanings (cf. the summary in Taylor (1989), and the test we can do with 'seats', cf. Gipper (1973); note that classification will not only depend on national differences but also differ from one informant to another in the same speech community). It is obvious that quantification in the sense as defined here becomes impossible when this model is applied, and its uses for historical linguists have indeed not been totally satisfactory so far (cf. Geeraerts 1997).3 3. Methods of historical semantic analysis Words (and contents) stand in paradigmatic and syntagmatic relationships, much like elements in phonological analysis. Contrasting words of similar content yields distinctive features (semantic components, semes) whereas the combinability of words intimates in what semantic fields they operate. The principle can be easily illustrated with dictionary entries of bank, boy, and bull. However precisely we can define lexical meaning synchronically in a homogeneous system, semantic identity is impossible to establish (or to be expected) across time and space. One of the best sources of information is to compare translations. For instance, where translations of the same text yield identical results for different stages of the same language, or for genetically related languages, the easiest assumption is that at least the

Conceptual and semantic change in the history of English

97

meaning required for the specific context has remained stable (possibly together with designation, i.e., identity of referent), compare the following specimens (from Görlach 1991): (1)

a. se dysiga man getimbrode hys hüs ofer sawi/ceosel (Mt 7.26) a foolish man built his house upon the sand ein Mann baute sein Haus auf Sand b. foxas habbaö holu and heofenan fuglas nest (Mt 8.20) the foxes have holes and the bryddes of the ayer have nestes die Füchse haben Gruben und die Vögel unter dem Himmel haben Nester

Such data are, however, exceptional for at least three reasons: 1) The words man, house, sand, fox, nest have well-defined referents, that is, their meanings can be expected to be unambiguous (with the exception of secondary metaphorical extensions). 2) These referents and the linguistic signs designating them have remained remarkably stable over time, to the extent that a total equation is found in the case of nests, and only a few sound changes have affected the forms of house, sand, and fox. 3) The fact that we have to do with translation equivalents identifies the meanings as renderings of words in the common source language, here Latin - on which much more detailed semantic information is available than on early vernaculars. In the histoiy of linguistics such bilingual texts have of course been of greatest importance - from the Rosetta Stone onwards. We need to look at words whose meanings, though they are translation equivalents, are vague or ambiguous. We would need to identify and then exclude emergency equations made to cover up lexical gaps. For instance, the Christmas story (Luke 2.1-21) has the praeses Cyrinius, a title which has given translators great concern - think of Luther's awkward coinage Landpfleger. Now the West Saxon translation has the term dema whose well-established meaning is 'the man who gives a sentence (döm), that is, a judge.' Should we assume a semantic caique in the West Saxon translation, then? I think not: however well the Old English word serves for the context (the dema's functions provide at least a considerable metonymic overlap), we cannot claim that there was a permanent, systematic extension of the meaning of dema in Old English. Another problem has to do with the inadequacy of paradigmatic relations that can be set up on the basis of translations. If the same Latin term

98

Manfred Görlach

is rendered by different Old English or Middle English words can these be considered synonyms? In order to form a hypothesis about possible semantic distinctions we face two further problems: 1) No prediction is possible about which features of the infinite number of possible logical distinctions are semantically relevant in an individual language. 2) It is not immediately apparent which notional or referential distinctions have remained stable over time. Let me illustrate the problem with an example. When I analysed the semantic contrast between members of the semantic field 'dwelling' (the words taken from Roget no. 192) in Görlach (1997: 125) I was able to use my near-native competence in English for the semantic classification. If I had decided to provide a similar stemma for 17th-century English, I could have used the respective chapter in Wilkins for a carefully analysed field of logical (if not semantic) distinctions - compare the field of 'vehicles' which I did take from Wilkins (1668: 257, in 1991: 188). However, Weimann's taking over of my graph based on Modern English structures to apply to Old English lexis (in 31995: 166) was not considerate and theoretically unreflected: his transfer suggests that notional and semantic distinctions can be assumed to have remained stable over time. This, however, is certainly not the case with respect to: (1) the objects - because living conditions have drastically changed since Anglo-Saxon times; (2) the function of generic as against specific concepts {bar may have covered a referential space that is impossible to reconstruct); (3) the entire semantic field which linguistic and cultural contact may have completely changed (one of the consequences being that the meaning of many individual words has drastically shifted - the congeners found in English bower and German Bauer certainly do not mean 'room in a house' as the word used to designate in Germanic; contrast the stability of nest above). Let me illustrate how semantic components can change (and sometimes quite unexpected ones become distinctive, only to be lost again in later development) by an example, which has to do with the integration of a loanword: One of the most famous test cases for structural semantics is the field of seats. Pottier's analysis was based on his native French; it established that apparent factual distinctions such as the number of legs, etc., were semantically relevant. When Köhler translated the paper (for Geckeler 1978) she was not aware of the fact that the distinctions in German were not identical, not even in Standard German. (In Austrian there is

Conceptual and semantic change in the history of English

99

no Stuhl except in medical contexts, so individuals always have to sit on a Sessel.) Later on, Gipper (1973) showed that there is also a great deal of vagueness involved - informants did not fully agree in their subclassification of drawings of various types of chairs, even though they were all speakers of Standard German. In fact, the examples have been used to illustrate the advantages of prototype semantics. In Old English, the field of 'seats' appears to have been beautifully simple. A generic term setl 'seat' could be used for compounds if an additional distinction was intended (as in cynesetl = 'throne'). The only lexical distinction was according to number of persons, a stöl being for one person, a bene for many. Accordingly, a cynestöl was also possible as a synonym for cynesetl (which could also mean 'capital city'). In addition, there was sceamol which cut right across all these, meaning 'stool', 'footstool', 'bench', and 'table of money-changers'. (The German set Sessel, Stuhl, Bank, Schemel retains many of the old distinctions, but was of course subject to notional and semantic changes, too.) The English system changed with the advent of chair which, as a consequence of the socio-linguistic prestige of French, introduced a new and quite unexpected seme, namely 'for noble persons' (which was, of course, accompanied by a change in comfort and design in the thing). Since the new component 'for noble persons' was very salient, it could even happen that the most important old (and modern) distinction as to 'number of persons' was neutralised: in Middle English king and queen could sit together on one chair. Note that the seme 'for noble persons' got lost again with the 'democratisation' of seating distinctions. My next problem involves the use of both paradigmatic and syntagmatic features. The well-defined field of terms for 'human settlement' was more or less distinguished in linear order according to size. The most frequent Old English words were tun and ceastre (cf. the tables 10.3-5 in Görlach 1997: 135-136). However, the West Saxon Gospels have regular castel for 'village'. Are we to posit a feature 'foreign' or 'biblical' when describing its meaning? A similar problem relates to Middle English city, which came to complement a field which was complicated by various factors, such as: 1)The vagueness of the general language (in touri) contrasting with administrative distinctions (by which toun came to be fixed as a translation equivalent of French ville - which initially suffered from the same problem of vagueness as toun did). 2) Regional differences according to which burgh was more frequent for 'town' in the north.

100

Manfred Görlach

3) A social/stylistic distinction according to which city at least initially had higher prestige, and could again be preferred with reference to foreign towns. Under these conditions the limitations of paradigmatic-structural analysis become all too apparent. It is therefore good to explore whether syntagmatic analysis can help - since for older stages we have to determine meanings from texts, this is the most usual form of semantic reconstruction anyway. We will have to face the fact that most textual occurrences of words are vague and few collocations semantically restrictive enough to gather precise information on a word's textual meaning. This statement is even true for literary writers who can be expected to choose their words carefully, but whose choice is partly determined by stylistics and poetic concerns as is illustrated by Chaucer and Puttenham's use of torn. When Chaucer talks about the village priest in the General Prologue he introduces him as a povre persoim of a toun (line 478). The reader/listener is unable to gather from the context whether this toun is (in modern terms) a farm, a hamlet, a village or a town. The uncertainty is dissolved only a few lines later when Chaucer describes the parson's parish as Wyd was his parisshe, and houses fer asonder (line 491) which excludes both 'farm' and 'town'. Puttenham, two hundred years later, talks about "the best towns and cities" which are to serve as a pattern for correct speech (see the quotation in Görlach 1991: 237). We might wish to assume that a semantic contrast existed between the two words (which could of course be intended), but we are warned by Puttenham's practice in the same passage to link together two lexical items as synonyms in order to illustrate the stylistic grace of the English language that lies in copiousness of speech, an ideal widespread in 16th-century authors. (The same passage has sounds and voyces, poet or maker, marches and frontiers, in the same stylistic function). 4. Reasons for change This section will correlate diachronic changes with their possible causes. As stated above, there are words which exhibit no change over time, as to meaning, concept, referent, or even form. Examples are nest and fox in my first excerpt; by contrast hüs > house can be said to show a change in the thing, but not in meaning since it meant 'human habitation' in Old English, and it still does. Before the reasons are classified that can be adduced to account for semantic changes, it may be useful to summarise the types of

Conceptual and semantic change in the history of English

101

stability (S) and change (C) encountered in the history of individual words): Table 1. Stability (S) and change (C) in selected English items meaning

S c C(?) s c

fields: C

C

c s c

s c c

examples (with German and Dutch in comparison) nest, fox = Ge, Du fish = Ge, Du earth = Ge, Du house = Ge, Du town Φ Ge, Du

s

C(?) c

cook (vb) vehicle

concept S C

c

referent S S

Where semantic change has occurred, it can be quantified by counting the senses (sememes) and the semes in them. Although every classification is open to objections, I will here try to account for change under twelve headings; note that several reasons can concur to produce the change: a) Intralinguistic reasons 1) The desire for precise, unambiguous reference has led to more or less monosemous words in the case of ones which were formerly polysemous or vague, that is, by a reduction of senses, or specification: döm, mete, toun. 2) This is the case in particular where synonyms were available which could be differentiated, affecting the range of meaning, often in complex relationships {mercyI pity, cf. Görlach 1997: 150), or the syntactic behaviour (corn/grain) etc. 3) Metaphors, by contrast, extend the semantic range: a glittering personality is a film star (cf. 'metonymy', if we wish to make the distinction, which leads to similar extensions, but is not intralinguistic). 4) The fossilisation of syntagmatic relations leads to the pre-dominance of certain connotations or restrictions of combinability: evil/ill (and later on bad) started off as being largely synonymous (cf. Görlach 1991: 202). b) Language contact phenomena 1) The identification of translation equivalents leads to semantic extension when unshared senses are transferred to the target language. This happened to knight (which became the equivalent of chevalier) and at present is hap-

102

Manfred Görlach

pening to the term for the (computer) mouse in many languages. In Old English many central terms of Christianity were equated with heathen concepts, such as sin and lord. 2) Loanwords affect the meaning of native items where they serve to complete sets, especially in closed inventories. Thus the meanings of yellow and red were necessarily changed by the adoption of orange. The modern borrowing ofjob into various European languages changed the meaning of the older words for 'work, profession' etc. c) Revival of senses Connotations or entire senses can be added to existing ones when the range of meaning is complemented or corrected by senses drawn from older texts. d) Terminological specification Terminologies are often created by the adoption of loanwords or the coinage of derivations or compounds. In many cases, however, old words are made to serve by a redefinition of their meaning. This can lead to an extension (the old and new meaning coexisting, though in different registers) and, in due course, to secondary reduction (when the old sense is given up): acre, yard ('3 feet'), borough, lent (formerly, the common word for 'spring', replaced by ver in the 14th century and by spring in the 15th centuiy; cf. German Lenz). Equally, words which come to be restricted to terminologies may undergo specification. Whereas a feet was a general vessel in Old English (cf. German Faß, Gefäß), it is now used only in the jargon of various trades such as brewers or tanners. A conceit is now either a type of 'vanity' or as a specialised literary term, 'an elaborate image or far-fetched comparison'. The contrary development, the adoption of a word into general usage, leads to generalisation (a thing, or German Ding and Sache, used to be legal terms and are now as generic or vague as possible), or can lead to metonymic reinterpretations: terms relating to the medieval lore of how body fluids led to specific types of character (choler, gall, phlegm and blood) were popularly understood by the effects the mixtures had, and gave us the modern meanings of humour (orig. 'fluid') and temperament and complexion, both orig. 'mixture'; cf. Schäfer (1966) and my summary (Görlach 1991: 206).

Conceptual and semantic change in the history of English

103

e) Metonymy If two referents stand in a causal, temporal, instrumental, part-for-whole or similar encyclopaedic/situational relation, meanings can be transferred from one to the other, bead, formerly 'prayer' (cf. German Gebet, beten) was transferred to the glass pearls, individually moved on after each prayer. In "I'll have/drink another glass" metonymy may be involved, as may be clipping, the expression alternatively arising from the phrase 'a glass [of beer]'. The two prevailing senses of game arose by semantic specification ('entertainment' developing to 'contest played according to rules') and by metonymy ('sport' > 'the sport: the chase'> 'object being chased' > 'venison'). f ) Transfer to similar objects This category may include metaphor and metonymy. Note that new oppositions must arise to justify the assumption that semantic change has happened. beam was one of the Old English words for 'tree' (cf. German Baum), but is now applied only to a product, and to a 'ray' sharing the notional similarity of being straight. A coach is either horse-drawn (the older meaning now specified by differentiation from other horse-drawn vehicles) or a bus or railway wagon, obviously because of the similarity of form and function (and the desire of owners to make the new type of transport attractive). g) Change of conceptual structure and h) Change of material culture This leads on to the question of how notional changes affect meaning. For instance, interpretations of diseases and the relationship of body and character have undergone revolutionary developments since the Middle Ages. Compare, for a similar consequence, the modern distinction of living beings as being either animals or plants. Both concepts were impossible to think in the medieval system, something that moved was a beast or deer (if big), a fish (if in the water), a fowl/bird (if in the air), a worm (if small and crawling), leaving many types of animals unclassified. Dragons were either worms or birds. Something green might be a tree, a shrub, a herb, a flower, but certainly not a plant, a word that was used only to designate a 'sprout', until Albertus Magnus (patron so to speak of Cologne University) came to redefine the system of natural philosophy. We still live with me-

104 Manfred Görlach dieval categories, however, when talking about crayfish (above) and German Walfisch for 'whale' and Lindwurm for 'dragon', although bats, bumblebees and (flying) dragons have long ceased to be 'birds'. Note that even Linne's classification of 1735 which distinguished between six classes (mammals, birds, amphibians, fishes, insects, worms) does not satisfy a modern biologist, although it did away with the above misclassifications (cf. the example of seats, discussed above). On a different level, and leading on to attitude, is the question of whether the meaning of virtue has changed, from 'manliness, valour' to 'excellence' in general and on to the Victorian exclusive concern for 'sexual cleanness'. The description of Chaucer's Knight in the General Prologue lists the categories he believed and excelled in, all of which now need semantic interpretation since the value system of feudal society is gone, and with it, the meanings of chivalry, truth, honour, freedom and courtesy.4 (Note that one meaning of the latter survives as curtsey in the metonymic extension to the gesture, the obeisance). i) Change of attitudes and cultural values The speaker and the addressee have an important role to play in the interpretation of the meaning of an utterance. Their attitudes influence meaning in many ways. Euphemisms serve to disguise a tabooed content, most often relating to death, sex and excrements. One of the solutions is to adopt a foreign word; other forms are the choice of a generic term {undertaker, originally: 'entrepreneur', German Unternehmer, hussy orig. 'housewife' or siege 'toilet' orig. 'seat', possibly reinforced by ellipsis from seat {of ease)). In German Stuhl came to be used for 'toilet' and then, by metonymy, to 'excrement' in medical contexts. j) Clipping The reduction of a complex expression to one of its constituents, being a purely formal process, leaves strangely mutilated items even when morpheme boundaries are observed (unlike pram and bumf). The modern meaning offlour shows no relation to flower, only the full phrase flour of wheat 'the best of wheat' explains the origin. Similar cases are fall {of leaves) and train {of wagons), compare again German Zug. We might wish to include here lexicalisations in which the syntactical context is deleted (but could classify these problems as derivational rather than as involving a change of meaning); what springs in SPRING designating either the first season, the source of a watercourse, or the flexible metal contrivance?

Conceptual and semantic change in the history of English

105

What draws or is drawn in the various senses of draft, draught, drawer and drawing? Note the three successive processes of ellipsis, which led to the homonymous senses of capital {letter, Middle English; stock, 16C; city, 17C). It is even open to debate whether these processes should be classified under semantic change and not rather under word formation, as the subtractive equivalent of compounding. k) Contamination andfolk etymology Historically incorrect associations and interpretations can also lead to semantic change. Numbles means 'loins, fillet', and this is the origin of humble pie which was reinterpreted and made into the phrase to eat humble pie 'to be forced to do something humiliating' under the influence of humble, sandblind is 'half blind', and might we not suspect that stirrup for many speakers is now possibly connected with 'stir up' a horse rather than with the word's etymology of stig + räp (German Stegreif, Steigbügel)? Compare German Schwager 'coach-driver' which is not 'brother-in-law' but comes from chevalier, and many other instances. In all these changes it would be good to know whether the speed of change has anything to do with transparency: does the meaning develop less quickly as long as the word or its components are properly recognised, a hypothesis not only applicable to folk etymology but also to cases like hussy vs. housewife? I) Syntagmatic changes There are many cases where the change can only, or best, be explained by syntagmatic relations. This is particularly the case where lexical 'items' have become semantically bleached. Consider the present-day compatibility of verbs of'becoming', as analysed in Leisi (1973: 185). A solution for a proper description is here not to consider the verb by itself but to see the verbal phrase as a universal semantic unit.

5. Disguised diachronic relations Various forms of metaphoric but especially metonymic relations depended on very specific historical situations, the knowledge of which has long been lost. The word cockpit can serve as an example: originally used for a 'pit for cockfighting' and as such still transparent and used today, the modern dominant sense arose as follows (Barnhart 1988: 185-186):

106

Manfred Görlach

a) the name of a London theatre name The Cockpit, built on a site where a cockpit once stood (1587), b) the buildings housing the Treasury and Privy Council built on this site, c) the part of a warship where junior officers were quartered (1706) (the connection with the previous senses is opaque), d) the place where the pilot sits in an aeroplane. In a similar way, the history of seemingly transparent words like cocktail is totally opaque. The great range of senses now exhibited by stock falls into a number of disconnected semantic fields; it is entirely uncertain how many of these came about, and whether we have, in the OED entry, modern descendants of the same item. 6. Changes in other languages and repeated borrowings Specimens quoted above from other languages are not meant to imply that we can describe change of meaning across languages. Various misleading examples are found in the literature, as when arrive is said to have become generalised from 'arrive at the shore'; it has, but this is not a question of English, which borrowed the generic meaning from French. Repeated borrowing of different senses of the 'same' word likewise is not semantic change, even though the range of meaning is of course extended in the process. Compare various senses of English case, or loanwords in European languages which have consecutively been borrowed from the English word twist the notions relating to 'drink', 'yarn' and 'dance'. A different selection of meanings with subsequent development explains the non-equivalence of Modern English pit and German Pfütze 'puddle', both fromputeus 'well'. 7. Conclusion Meaning and its change are especially stimulating fields of linguistic investigation. The description remains problematic for various reasons which we may wish to summarise again in order not to forget how difficult the entire undertaking is: (a) The discreteness of meanings cannot be established for all semantic fields; it is very doubtful for earlier stages of languages where gaps in the documentation grievously restrict our data. (b) The changes that happened on the semantic, notional and encyclopedic levels cannot in all cases be strictly kept apart. This is one of the elemen-

Conceptual and semantic change in the history of English

107

tary difficulties of historical thesauruses which have to take into account changeability on all three levels, which leaves no stable framework for the classification of individual items. (c) The reasons for changes are very often entirely opaque, the situations in which they occurred not being sufficiently documented. It is easy to see that a fuller account of semantic change, if it were available, might again be used for better insights into cultural history, the two being closely interrelated. (d) The conditions and causes for semantic change found on various levels (formal, social, attitudinal etc.), mean that the interplay of these is not easily reconstructable; this is even more so when we look at the point of divergence, that is, when changes were initiated, first as ad-hoc usages, only later to become accepted into the common system. In spite of all these reservations, (historical) semantics remains one of the most fruitful fields of linguistic investigation - for some linguists, including the honoree, these problems may in fact constitute a special attraction.

Notes 1. It would seem to be equivalent to carrying Sachertorte to Vienna (coals to Newcastle) to write again on semantic change and dedicate the draft to a specialist in the field, whose chapter on "Semantics and vocabulary" in CHEL, 1 has set standards (Kastovsky 1992), but I will do so all the same hoping for a particularly critical reader in DK. 2. Most of the examples are taken from the relevant chapters in my Linguistic history (Görlach 1997) and my Early Modern English (Görlach 1991); for methods of structuralist semantic analysis also compare Baldinger (1980). 3. Admittedly, my analysis may occasionally suggest discreteness where it does not exist, or did not exist for the contemporary users; a case in question is the Venn diagram for 'mercy' (Görlach 1997: 150) in which senses of the Middle English items are rendered by Latin equivalents, but the vagueness and semantic overlap of items in the field are suggestive rather than strictly provable. 4. Semantic analysis of curteis/curteisye is complicated by the fact that Chaucer slyly allows the pilgrims who claim to be curteis their own definitions and then accepts these as a participant observer, cf. the resulting range of meanings attested by the knight, the prioress, and the friar (lines 46, 132 and 250 of the General Prologue).

108

Manfred Görlach

References Baldinger, Kurt 1980 Semantic theory: towards a modern semantics. [Span, original Teoria semäntica 1970.] Oxford: Blackwell. Barnhart, Robert Κ. 1988 The Barnhart dictionary of etymology. New York: H. W. Wilson. Dornseiff, Franz 1933/34 Der deutsche Wortschatz nach Sachgruppen. Berlin: de Gruyter. Geckeier, Horst 1978 Strukturelle Bedeutungslehre. (Wege der Forschung 426). Darmstadt: WBG. Geeraerts, Dirk 1997 Diachronic prototype semantics. A contribution to historical lexicology. Oxford: Oxford University Press. Gipper, Helmut 1973 Sessel oder Stuhl. In: Lothar Schmidt (ed.), Wortfeldforschung, 371398. (Wege der Forschung 250.) Darmstadt: WBG. Görlach, Manfred 1994a Einführung in die englische Sprachgeschichte. Third edition. Heidelberg: Quelle & Meyer. 1994b Einföhrung ins Frühneuenglische. Second edition. Heidelberg: Winter. [1978] English version: Introduction to Early Modern English. Cambridge: University Press, 1991. 1997 The linguistic history of English. London: Macmillan. 1999 English in 19th-century England. Cambridge: Cambridge University Press. Leisi, Ernst 1973 Praxis der englischen Semantik. Heidelberg: Winter. Macleod, Iseabail 1990 The Scots thesaurus. Aberdeen: Aberdeen University Press. Menner, R.J. 1945 Multiple meaning and change of meaning in English. Language 21: 59-76. Pottier, Bernard 1964 Vers une sömantique moderne. Travaux de Literature et Linguistique 2, 107-137; German translation by G. Köhler in Geckeler 1978, 4589. Roberts, Jane and Christian Kay 1995 A thesaurus of Old English. 2 volumes. London: King's College.

Conceptual and semantic change in the history of English

109

Roget, Mark Peter 1852 Thesaurus of English words and phrases, classified and arranged so as to facilitate the expression of ideas and assist in literary composition. London: Longman [facsimile reprint London: Bloomsbury, 1987]. 1964 Roget's international thesaurus. London & Glasgow: Collins. Samuels, Michael 1972 Linguistic evolution. Cambridge: Cambridge University Press. Samuels, Michael, Christian Kay, et al. forthc. A historical thesaurus of the English language. Oxford: Oxford University Press. Schäfer, Jürgen 1966 Wort und Begriff 'humour' in der elisabethanischen Komödie. Münster: Aschendorff. Waldron, Ronald A. 1979 Sense and sense development. 2nd edition. London: Deutsch. Wartburg, Walter von and Rudolf Hallig 1952 Begriffssystem als Grundlage der Lexikographie. Berlin: Sitzungsberichte der Berliner Akademie. Wehrle - Eggers 1967 Deutscher Wortschatz. Ein Wegweiser zum treffenden Ausdruck. 13th edition. Stuttgart: Klett. [1961], Weimann, Klaus 1982 Einführung ins Altenglische. Heidelberg: Quelle & Meyer. [21990, 3 1995]. Wilkins, John 1668 An essay towards a real character, and a philosophical language. London. [Facsimile edition EL 119].

Genitive case or determinative phrase? The status of's forms in Modern English Klaus Hansen

English has undergone a deep change in its grammatical structure since it has developed from a language expressing syntactic relations still largely synthetically, by means of inflectional endings, into a language expressing them mainly analytically, by means of function words and/or fixed word order. As for the nouns, this has also had an influence on the development of the morphological category of case. Old English still showed a morphological system of four cases (nominative, genitive, dative, accusative) and in old use even traces of a fifth case, an instrumental, formally identical with the dative (cf. ALlfred kyning häteö gretan Wcerferd biscep his wordum lufllce ond freondlice [King Alfred's Preface to the translation of Pope Gregory's Cura Pastoralis]). Besides, Old English still had several classes of declension with different sets of inflections and also full vowels in the endings; cf.:

sing.

Pi.

α-stems: masc.

ö-stems: fem.

nom. gen. dat. acc.

(se) QHZS) (pcem) (pone)

cyning

(seo) (poire) (pcere) Φα)

tain

nom. gen. dat. acc.

ΦΦ (pära) φαιή) Φα)

cyningas

φα) φαra) (pcem) φ-α)

tall·

-es -e cyning

-a -um -as

-e -e -e -a -um -a

Towards the end of the Old English period, however, most inflections were reduced to -e [-a], which resulted in a coincidence of the respective case forms. The only distinctive case marker that was lastly left was the phonetically stable ending -es of the singular genitive of the former α-stems. It was therefore extended as a general genitive marker to nearly all other nouns and also to the plural. Here, however, it coincided with the reduced form of Old English -as so that the plural genitive was now marked only by its prenominal position and distinguished from the singular genitive at best by the context in which it occurred (cf. ME sing, pe king :: pes hinges

112

Klaus Hansen

cnihtes / pis kinges time - pl. alle langes :: allre kynges kyng [cf. Berndt 19893: III]). 1 By the end of the Middle English period, even -e was deleted and the category of case thus reduced to a binary opposition between the genitive as its marked member and a non-genitive (often called common or general case) as its unmarked member. This reduction of the morphological category of case was closely interrelated with an increasing use of prepositions (mainly of and to) to express the case functions, and later also by the emergence of fixed patterns of word order. Besides, the genitive was replaced by possessive forms of the type the man of lawe his tale (Chaucer), i.e. by constructions with a possessive pronoun, mostly his. They had their origin in sentences like Old English Her Romane Leone Jjäem päpan his tungan forcwfon and hiseagan ästungon (Anglo-Saxon Chronicle; cf. Brunner 1951: II, 27), where the dative followed by a noun phrase with his could be interpreted as a possessive dative. As such, it occurred then in examples like Old English pä Gode his name ... clgdon (cf. ibid.: II, 28) whose use, however, seems to have been largely restricted at first to the informal spoken language. But from the 15th century onwards such possessive forms became quite common even in literary and more formal written texts, probably due to the fact that the loss of the dative inflection had resulted in a merger of the Aw-constructions with genitive phrases whose -(e)s ending coincided with the reduced forms of his ([iz] or even [z/s]);2 cf.: (1) the

kynges kyngys sonne (cf. also Wyld 19563: 3 1 4f.) kyng (h)ys

The merger strengthened the use of the Aw-construction but also made the [iz] forms ambiguous and thus open to reanalysis. As for the genitive, this supported a reinterpretation of the case ending -(e)s as a phrase marker spelt 's (N-j > NP'j; cf. the king of Englands daughter). Besides, it contributed to restrictions in the use of the former genitive, which as NP'S now mainly occurred with nouns marked as (HUMAN) or at least (ANIMATE) and with a possessive function as its central or prototypical meaning. As for the Aw-construction, the merger with the genitive suggested a reinterpretation of his as a postposed clitic attached to the preceding noun phrase (NP + [his + N] > NP'J + N; cf. the spelling) and contributed to the predominant use of his in this construction, occasionally even with an antecedent marked as (FEMALE) (e.g. Mrs. Francis her mariage, but also: my

The status of 's forms in Modern English

113

moderys sake, the queneys modyr, Margaretys doughter [cf. Wyld 19563: 315f.]). All this changed the status of the s-forms and resulted in a loss of the morphological category of case since they had constituted its only marked member. Nonetheless many grammars still use the term genitive to refer to the noun phrases with 's and postulate a category of case for Modern English as well; cf. e.g. Quirk et al. (1985: 318): "We shall distinguish between two cases of nouns: the unmarked COMMON CASE (e.g.: boy in the singular, boys in the plural) and the marked GENITIVE CASE (e.g. : boy's in the singular, boys' in the plural)." In other grammars, the forms with 's are only described as such but also still referred to with the term genitive or possessive case (cf. e.g. Thomson/Martinet 19864: 29f.; Alexander 1988: 51). Occasionally, however, and especially with respect to the so-called "group genitives", one also finds hints in the description of these forms that an interpretation o f ' s as a morphological case marker is no longer appropriate. Thus Palmer (1924: 36f.) already stated that "the genitive [z], [s] or [iz] may be considered either as an inflexion or as an independent word": "On logical grounds it is an independent word, for it may be separated from the noun and attached to the last member of the noun-group: ÖQ maen ai so: jestodi ζ fa:ös ... ÖQ kir) ov irjgtend ζ kraun ... Traditionally, however, it is an inflected form of the noun to which it is nominally attached." "If an inflexion", Palmer concluded, "it should come under the heading of the accidence of the noun; if an independent word, it should come under the heading of postpositions or of syntax." He thought, however, that all this was "more of academic than of practical interest" and that for the purpose of his description it was "more convenient ... to treat the form in question as an inflexional modification of the noun". Objections to an interpretation of the's forms as genitives were also put forward by Quirk et al. (1985). They still postulate a morphological categoiy of case for the noun but admit that "it is arguable ... that... the common/genitive distinction in present-day English is not really a case distinction" (318) and that with respect to noun phrases with 's, especially the so-called "group genitives", "it is necessary to revise the idea ... that the genitive is a noun inflection" (328). But what then are the peculiarities in the use of the "genitive" that make it doubtful whether 's can still be regarded as an inflectional ending that constitutes a morphological category of case?

114

Klaus Hansen

1. The use of 's as a phrase marker In Old English, not only the head-noun of a genitive phrase but also its determiner or an adjective attributed to it were case-marked (cf. e.g. ... tö gepeodnisse Jjaes heofonlTcan lifes [King Alfred's translation of Bede's Historia ecclesiastica, Lib. IV, Cap. XXIV]). In the course of Middle English, the inflection of the determiner and the adjective was given up and case-marking came to be restricted to the noun. As a result of this, the nominal genitive ending took on the character of a phrase marker. It was no longer attached to the head-noun as such but to the respective phrase as a whole; cf.: (2)

[the old king]'5 daughter. NP's + N

For genitive phrases with a non-final head-noun this meant that the smarker was shifted from their head to their end and thus attached to their last word. Examples to illustrate this are the phrases traditionally described as "group genitives". Originally, they had the genitive ending added to their head;3 cf.: (3)

the kynges doughter ofEnglande (1520; cf. Wyld 19563: 3 1 8)

But then the s-marker was shifted to the end of the phrase and became a kind of postposed clitic attached to it as a whole; cf.: (4)

[the king of England] 's daughter. NP's + N

The change of the s-marker from a case-ending of the noun (N-s) to a phrase marker (NP'J) is also evidenced by phrases with nouns like wife, which as genitives (i.e. with -s as an inflectional ending) should show an alternation of their stem (cf. the old wives tale [Chaucer] and also the plural form wives) but now occur unaltered with 's attached to them as a phrase marker: [my w i f e ] f a m i l y . The development of the ^-marker from a case-ending added to the noun (N-5) to a postposed clitic attached to the noun phrase as a whole (NP's) may well have been influenced or at least strengthened by the hisconstruction described above, since the "group genitive" became fully established in the second half of the 16th century when the use of phrases with his was quite common; cf.:

The status of 's forms in Modern English

(5)

115

[the king of England]'« / (h)is daughter

As to their syntactic status, 's phrases take the position of the article or a demonstrative in the superordinate noun phrase and may therefore be regarded as complex (phrasal) determiners or "determiner phrases" (Taylor 1996: 116). This implies that adjectives referring to the head of the superordinate phrase have to follow them; cf.: (6)

NP -> Det (+ Adj) + Ν

Det

art the dem this poss my NP'« the young

(new) car

The '« phrase usually occurs as a definite determiner with specific reference and serves to determine or specify an instance (or instances) of what is denoted by the head of the superordinate noun phrase (cf. e.g. [the student]'« dictionary, [my children]'« books, [the veteran]'« eldest son). This is why the examples discussed so far have been described as "determinative genitives" (Quirk et al. 1985: 326f.) or "specifying genitives" (Kruisinga 19325: 55). But there are also examples like a [student's dictionary], (my) [children's books] or a white [sailor's uniform] which still show traces of the old genitive and have '« added only to the preceding noun (N'«) and not to a noun phrase (NP'«).4 Such forms with '« have been characterised as "modifying genitives" (Quirk et al. 1985: 327f.) or "classifying genitives" (Kruisinga 19325: 55). They are used with generic reference to modify, i.e. restrict, the type or class referred to by the noun with which they are combined so that the combination as a whole denotes a subclass of the class denoted by its head; cf.: a student's dictionary 'a dictionary for students', children's books 'books for children'. Combinations with a "classifying genitive" show the character of compounds or compoundlike complex lexemes. They are therefore lexical units rather than syntactic units and have constituted a pattern of wordformation in which «/'«/«' no longer occurs as a case marker but as a kind of derivative or linking element (historically also based on the plural ending as, e.g., in salesman)·, cf.: craftsman, widow's pension - widows'/widows benefit, girl'si girls'!girls school, salesman (see Marchand 19692: 65-69 and cf. German Ankunftszeit, Hilfskraft). As lexicalised

116

Klaus Hansen

units, combinations with a "classifying genitive" tend to take on additional semantic components and occur with a more or less idiomatic meaning (cf., e.g., childsplay 'something that is very easy to do'). A determiner or an adjective preceding such combinations refers to their head and not to their constituent with 's; cf.: the new [student's dictionary] :: [the new student] 's dictionary·, his extravagant [tailor's bills] :: [his extravagant tailor]'5 bills. In the spoken language, combinations with a "classifying genitive" may be marked as lexical units by having their primary stress on the first constituent; cf.: her ['child's face] :: [her ,child]'j xface\ 1birds-nest, 1childs play, 1girls' school, lwomen's ^movement. Besides, the use of the indefinite or the zero article and the meaning of the preceding adjective may point to the classifying function of the noun with 's; cf.: a children's home, dressed in labourers' clothes', a grantmaintained boys' school. To avoid ambiguity, 's forms may be replaced by prepositional phrases; cf.: the children's home -» the home of the children :: the home for children.

2. Restrictions of the use of the "genitive" The loss of the morphological category of case is also reflected by restrictions of the use of the "genitive" since, by the beginning of early Modern English, it came to be largely limited to phrases with heads marked as (HUMAN) or at least (ANIMATE), i.e. with personal name (cf. Mary's friend, Shakespeare's plays) and nouns denoting persons (cf. my father's car) or animals, especially higher animals (cf. the cat's tail). In addition, the heads now were mainly - names of organisations, corporations or institutions; cf.: Labour's new leadership, BBC's night programme, the Church of England's comment;5 - collective nouns "which emphasize the aspect of 'organized individuals'" (Quirk et al. 1985: 324); cf.: the cabinet's first step, the club's annual meeting, the government's proposals·, - geographical names and nouns denoting places or regions, especially if used with "human associations" (Zandvoort 19612: 104); cf.: Cairo's antipathy to Khartoum, England's positive thinking, Russia's biggest trading partner, but also: London's boroughs, Europe's largest coastal resort, beyond Spain's borders; the town's Methodist Church, the country's leaders, the islands wealth;

The status of 's forms in Modern English

-

117

other nouns with "human associations" or "of special relevance to human activity" (Quirk et al. 1985: 324); cf.: history's big guns, life's batterings, time's arrow, - temporal nouns; cf.: yesterday's press, next year's elections, this weekends issue of Parade magazine. Inanimate, especially concrete, nouns now usually require an o/-phrase (cf. the door of the house), although there seems to be a tendency in PresentDay English to also extend the use o f ' s to such nouns, at least in certain text types, e.g. newspaper articles; cf.: the buildings thousand rooms, the car's automatic gearbox, the drug's chemical effect, the record's grooves, the fund's advantage (see also Raab-Fischer 1995). The lexical constraints described are closely interrelated with functional or semantic restrictions concerning the use of the 's phrase. The genitive had expressed a wide range of semantic relations depending on the meanings of the nouns from which it was formed and with which it was combined. But with the emergence of the ^construction and under the influence of the Ä/s-construction its use came to be more or less limited and, with respect to some of the semantic relations, a certain complementary distribution developed between the's phrase and the o/*-phrase. As a result, the 's phrase is now primarily used to express a possessive relation (in the widest sense of the word) so that this relation may be regarded as its central or prototypical meaning (cf. my father's car, Mary's friend). The second important function assigned to the 's phrase is its role in expressing a subjective relation, which occurs mainly if the head of the superordinate phrase is a noun related to a verb (esp. an intransitive one) (cf. the girl's arrival, the French government's protest, Mr Gore's comments). Other semantic roles assigned to the 's phrase, e.g. those described by Quirk et al. (1985: 322) as the "genitive of attribute" (cf. the man's kindness, Rupert Murdoch's ambition) or the "genitive of origin" (cf. the student's paper, Shakespeare's plays), may be subsumed under the two main functions, whereas the "descriptive genitive" (cf. a children's home) and to some extent also the "genitive of measure" (cf. at a mile's distance, a two hours' walk) are restricted to complex lexical units. As for the "objective genitive", a n ' s phrase can only be used to express it if the meaning of its head and/or of the head of the superordinate phrase makes the semantic relation unambiguous or if the context helps to identify it; cf.: the patient's examination :: the doctor's examination·, her husbands re-election, Mr Yeltsin's hospitalisation). Otherwise an o/-phrase will be preferred to express this semantic relation; cf.: the story of Mike 'the story about Mike' :: Mike's story 'the story told by Mike'; the loss of his wife, the prosecution of Mr Shayler). Consequently, "there is a tendency for genitives to be taken as

118

Klaus Hansen

subjective, and for o/-constructions to be taken as objective" (Quirk et al. 1985: 322, Note [a]).6 Thus club in the club's support will be interpreted as the subject rather than as the object of the verb underlying support unless there is a context to point to the contrary, as in the following case: The imminent takeover of [Manchester] United, by BSkyB, has angered a section of the club's support... The supporters may not be able to avert the inevitable, but... (T 19/11/98: 56). The o/-phrase came to be used exclusively for the "partitive genitive" (cf. a glass of red wine, the rest of the family) and for the "appositive genitive" (cf. the art of printing, a brute of a husband). In all other cases, it occurs as a potential variant of the 's phrase. Where both a n ' s phrase and an o/-phrase are acceptable, their use may be conditioned by the following factors: a) The length and complexity of the phrase Longer and more complex phrases tend to be postposed with of rather than to be preposed with 's; cf.: He is the son of the man I met yesterday rather than He is the man I met yesterday's son, which may, however, occur in informal spoken usage. b) The principle of end-focus The choice between a n ' s phrase and an o/-phrase may also be determined by the "principle of end-focus" (Quirk et al. 1985: 323) according to which the communicatively more important constituent of the superordinate noun phrase is shifted to its end; cf. my brother's daughter (:: son) - the daughter of my brother (:: my sister) and I learned that as well as being the patients' physician he was their advocate for dignity, privacy and individuality. (IR 20/11/98: 6) The price for refusing to come up with the goods was apparent in the next morning's papers. Readers of the Sun didn't even need to turn to the back page... (120/11/98: 27) The principle of end-focus may also have contributed to the use of 's phrases with inanimate nouns; cf.: When the French national assembly passed the bill's first reading ... (DT 20/11/98: 42) - ... the film's defining characteristic ... (IR

The status of 's forms in Modern English

119

20/11/98: 17) - ... almost unanimous hostility to the ban's lifting. (GE 24/11/98: 4) In the spoken language, this principle may be overridden by the use of contrastive stress; cf.: This, from an Avmerican's perspective, is unfortunate and unwise instead of...,from the perspective of an American, ... c) The use of spoken or written language Whether the head-noun of an 's phrase is used in the singular or in the plural is explicitly expressed only in the written language; cf.: my client's interests - my clients' interests. In the spoken language, one has to use an o/-phrase to make the form unambiguous: the interests of my client - the interests of my clients', my son's/sons' children —> the children of my son/sons. The spoken form is only unambiguous in the case of nouns with an irregular plural (cf. the child sichildren's parents) or in certain contexts (cf. the teachers' common room, the agricultural ministers' meeting, in many critics' end-of-year lists). d) The text type 's phrases seem to be used more widely in newspaper texts than in other types of written language and occur there not least in headlines, "for which the compactness of the case-form is very convenient, so that it is often used where it would ordinarily be inappropriate" (Strang 19682: 110; see also Raab-Fischer 1995); cf. e.g.: Shops' charity hopes dashed: Hopes that struggling village shops could get charitable status have been dashed ... (GE 20/11/98: 7).

3. The use of the spelling ('s) The use of an apostrophe in spelling also points to an interpretation o f ' s as a phrase marker rather than as an inflectional ending marking a case form of the noun since it has not become common for the phonetically identical j-ending of the plural forms (cf. the king's / the king of Englands daughter - the kings of England). The apostrophe was introduced into the spelling of English to mark the elision of letters and thus came to be used in the written forms of words that had been reduced in their spoken form. As to the spelling ('s), it still represents the reduced unstressed variants of word-forms like us, is or has

120

Klaus Hansen

(cf. let's go, it's true, he's gone), but in early Modern English it also occurred for a corresponding form of his, as can be seen from the following examples in Shakespeare's plays: He that cannot make a leg, put o f f s cap, kiss his hand and say nothing... (All's Well 11,2: 9f.) - May they perceive's intent! (Coriolanus 11,2: 160) - ... wounds received for's country, (ibid. 11,3: 171) - There's no man in the world more bound to's mother ... (ibid. V,3: 159) (cf. also Jespersen 1961: VI, 301). The spelling of the phrase marker may therefore well go back to the possessive pronoun his. Its interpretation as a reduced form of this word is supported by the fact that ('s) came to be established at the end of the 17th century when his was still widely used as a marker of possessive phrases. At least, the hisconstruction may have strengthened a use of the apostrophe to signal that s was no longer an inflectional ending of the preceding noun but a marker of the whole phrase in which it occurred as its head. In view of all this, it seems to be inappropriate to still interpret 's as a case marker and to postulate also for Modern English a morphological category of case with the genitive as its marked member. It is inappropriate since 1) the "genitive" can no longer be formed from all nouns but has been limited largely to those marked as (HUMAN) or used with "human associations"; 2) 's is not added to the noun only (N's) but attached to the phrase headed by it (NP's), as indicated by the use of an apostrophe in the spelling, which would be uncommon for an inflectional ending. But opinions are divided as to the status o f ' s resulting from this. Some grammarians have pointed out that 's now "ranks rather with derivative affixes than with inflexional endings" (Vachek 1961: 25). Thus Kruisinga (19325: 64) already stated that "the facts as to the forms in -s used as a genitive compel us to consider the genitive as a derivative rather than an inflectional form"; and Vachek compared the use o f ' s with that of suffixes in derivatives from word-groups like John Bullish or dog-in-themangerism, coming to the conclusion "that the ModE Possessive Form ... appears to be gradually acquiring adjectival character, and has already covered a great part of the road ... leading towards full adjectivization" (Vachek 1961: 26). He had to admit, though, that there are some "obstacles" standing in the way of such an interpretation, especially the use o f ' s in examples like a man's blood (as against a man's hat), where "the indefinite article clearly belongs to the Possessive Form, not to the uncountable noun" and where "one might perhaps argue that the suffix 's is added not to the bare substantival form man alone, but to the word-group a-man con-

The status of's forms in Modern English

121

sidered as a whole" (ibid.: 27). This, however, means that an interpretation of the 's forms as adjectival derivatives (suggested by the parallelism between instances like a man's hat and a new hat) is at best true for the socalled "classifying genitives", i.e. for the forms used as constituents of lexical rather than of grammatical units. It does not apply to the more common 's phrases described as "determinative" or "specifying genitives". It seems therefore more to the point to treat 's as a kind of "grammatical particle" (Culpeper 1997: 52) or postposition parallel to the preposition of (cf. Palmer 1924: 36f., quoted above). Unlike the preposition, however, the postposition never occurs as an independent word but only as an enclitic 's attached to the final word of the phrase and determined by it also in its pronunciation (cf. the oldjudge's [iz] ~ doctor's/man's [ζ] ~ poet's [s] home). This makes it appear more appropriate to characterise 's as a postposed clitic (or: enclitic [cf. Quirk et al. 1985: 328]). With respect to their central or prototypical meaning, the terms possessive case (cf. e.g. Thomson/Martinet 19864: 30) or possessive form (cf. e.g. Vachek 1961: 25; EG 19844: 63; Collins 1990: 102) have been used to refer to t h e ' s forms. Summing up our discussion of these forms, it seems to be most appropriate to characterise them with regard to their predominant grammatical use as determinative phrases marked by enclitic 's*

Notes 1. This is still true for the spoken form of Modern English (cf. my [smo] friends) so that Quirk et al. (1985: 319) characterise the "plural genitive" as a "zero genitive". In the written language, it is now marked explicitly by means of an apostrophe following the plural -s (cf. my sons' friends). Such forms are likewise referred to in our description by the term's form/phrase. 2. Up to the 16th century, the genitive ending was often still used in its syllabic form [-iz] even after non-sibilants, not least for rhythmical reasons (cf. To shew his teeth as white as whales bone [Shakespeare: Love's Labour's Lost V,2: 332]). Besides, his had also been reduced to 's (cf. below: 3. The use of the spelling /g'niKs/ anuas 'from above'

The above form means 'downwards from a position of being up', that is in a direction away from being up, though in a direction to the speaker or other referent. As opposed to the base morphemes, all of the specifiers in (4a - c) above have varying phonetic realisations depending on whether the initial vowel of the base morpheme they are combined with has the phonetic feature [+palatal] or [-palatal] (the graphemic representation, however, is unaffected by this conditioning). This latter feature is an essential element of Irish phonology and is used in the grammar and lexicon to indicate categories (like case and number) and distinguish words respectively. In traditional distinctive-feature descriptions, palatality is indicated by a combination of [+high] and [-back]. However, for Irish there are many compelling reasons for postulating a single feature [+/-palatal] (indicated by a subscript /j/ in transcriptions in this paper) as this feature can behave differently from all other distinctive features that may be postulated, above all because of its high functionality on the morphological and lexical levels. a.

{S-}

s

b.

{AN-},

an /on/ /anj/

c.

{H-}

th

/s/ v

/h/ /h/

/ V [-palatal] / _ V [+palatal] / V [-palatal] / _ V [+palatal] / V [+palatal] / _ V [-palatal]

The semantic features as given earlier for both the base and specifying morphemes appear in pairs for each of the actual forms given to the right of the arrows in aA - cB (see (8) below). An example of such a combination would be the following form. (7)

sios /sji:s/ [+above, +direction from]

The six possible combinations can be formulated as follows.

128 Raymond Hickey

aA bA cA aB bB cB

{S-}+{-UAS} {AN-}i+{-UAS} {H-}+{-UAS} {S-}+{-IS} {AN-h+{-IS} {H-}+{-IS}

-> ->

-»

— »

— »

/su:s/ suas /a'nu:s/ anuas /hu:s/ thuas /sji:s/ sios /ö'nji'.s/ anios /hi:s/ thios

'to above' 'from above' 'above' 'to below' 'from below' 'below'

Typical occurrences of the six final forms are to be found in the following sentences. (9)

a.

Chuaigh Diarmuid suas an staighre. 'Dermot went up the stairs.'

aA

b.

Thdinig Diarmuid anuas an staighre 'Dermot came down the stairs.'

bA

c.

Τά Diarmuid thuas an staighre. 'Dermot is upstairs.'

cA

d.

Chuaigh Brid sios go dti an trd. 'Bridget went down to the strand.'

aB

e.

Thdinigh Brid anios on trd 'Bridget came up from the strand.' Τά Brid thios ar an trd. 'Bridget is down on the strand.'

bB

f.

cB

As in English, where one has both 'up' and 'over/above' as expressions of location, one has in Irish locational morphemes which need an additional feature to distinguish them from one another. This is the feature [+/relational]. Consider the following examples. (10)

a.

Τά Seamus thuas sa seomra. 'Seamus is up in the room.'

b.

Tdim chun dul thar muir. (am-I to go over/above sea) Ί am going overseas.'

In the last sentence it is necessary to specify that X is above Y whereas in (10a) this is not the case. Thus by the addition of a feature one can distinguish semantically the different forms thuas 'up'and thar 'over/above'.3

Direction and location in Modern Irish

(11) 2.1.

{H-}+{-UAS} {H-}+{-AR}

-> /huis/ thuas ->· /hair/ thar

129

[+above,-relational] [+above, +relational]

Nominalising locative forms

A further phonetic realisation of both {UAS-} and {IS-} occurs when they combine with the suffixal morpheme {-TAR} which nominalises the base morpheme which is modified. In synchronic phonological terms one can postulate that there is a shift of base-final /-s/ to /-x/ before a suffix beginning in /t/. (12)

{UAS-} {IS-} {-TAR}

/urx/ / # _ { - T A R } /i:x/ / # {-TAR} /tar/

{UAS-}+{-TAR} {IS-}+{-TAR}

/uixtor/ uachtar -> /iixtor/ iochtar

'the above' 'the below'

3. Expressing lateral proximity The group of lateral morphemes can, for the purposes of classification, be divided into two further subgroups each of which is differentiated from the vertical group by referring to lateral as opposed to vertical direction. The lateral group contains a pair of complementary morphemes and a number of specifying morphemes which, however, do not exhibit the regularity of the vertical group with {-UAS} or {-IS}. The forms of the base morphemes are as follows: (13)

LATERAL BASE MORPHEMES C: {-ALL} [-proximate] 4 D: {-ONN} [+proximate]

There is a degree of irregularity in the realisation of these morphemes in modern Irish: {-ONN} has the written form onn and, in western Irish, the pronunciation /u:n/ due to vowel raising before former long nasals. The feature [+/-proximate] is a postulate to account for the compound meaning of the base and specifying morphemes and can be understood as referring to whether the motion is from a point close to the speaker/referent [+proximate] or from some location at a distance towards the speaker/referent [-proximate].

130 Raymond Hickey The specifying morphemes in this case have the same forms as (6a - c) above. There are no variant realisations as both the base morphemes C and D begin with back vowels and hence do not trigger palatalisation of a preceding consonant. (14)

{S-} {AN-} ι {H-}

/s/ /an/ /h/

[+direction to] [+direction from] [-direction]

There are, however, only four possible combinations of C, D and a - c in present-day Irish. (15)

aC: bC: cC: bD:

{S-}+{-ALL} {AN-}i+{-ALL} {H-}+{-ALL} {AN-}i+{-ONN}

/sa:l/ sail /a'nail/ anall iha'XJ thall -> / a W n / anonn

'over across' 'from across' 'beyond, there' 'across to'

The following fifth combination can be regarded as obsolete.5 (16)

aD:

{S-}+{-ONN} [+direction to, +proximate]

-»

/su:n/ sonn

'hither, here'

Furthermore, it did not have the sole meaning which one would be led to expect from the semantic labelling offered here, that is, [+proximate] and [+direction to] but also the combination [+proximate] and [+stationary]. It is also interesting to note that the form which would be generated with the morphemes D and c in this context neither exists nor has it ever existed, its function being discharged by a formally dissimilar item, abhus. (17)

Dc:

{H-}+{-ONN}

/hu:n/*thann

[-direction, +proximate]

The word abhus is a petrified form which, unlike those in aC - bD, cannot be analysed morphologically.6 One can therefore attribute its semantic features to the item as a whole. (18)

/a'vAs/ abhus [+proximate, -direction] 'here, on this side'

Direction and location in Modern Irish

131

In addition, no abstraction is necessary as this is a unique lexical item which is morphologically opaque. It is seen with its semantic complement in the following expression. (19)

thall agus abhus 'here and there' (lit.: 'there and here')

The following are some representative sentences with examples of aC tobD. (20)

3.1.

a.

Cuirtear sail thar an abhainn iad. 'They were ferried across the river.'

aC

b.

Τά siad ag teacht anall anois. 'They are coming from over there now.'

bC

c.

Bhi sifein thall i Sasana uair amhäin. 'She herself was over in England once.'

cC

d.

Chuaigh si anonn uaim. 'He went away from me.'

bD

Degrees of lateral proximity

A further aspect of lateral proximity in Irish concerns a second group of morphemes which express degrees of proximity. There is a triad of items here, each of which consists of a base and specifying morpheme. (21)

E: F: G:

{-SU} {-SIN} {-SUD}

[+proximate] [-proximate] [-proximate]

!sy/J seo /sjinj/ sin /sju:d/ sitid

The difficulty which arises here is that {-SIN} and {-SUD} while both having the feature [-proximate] are distinguished by the extent to which that feature applies, much along the same lines as English made (and in some varieties still makes) a distinction between there and yonder. When dealing with three (or more) grades of a single feature the binary distinction of plus and minus, +/-, is insufficient. To overcome this I will use a Greek letter, starting with a, which then denotes decreasing grades of the feature it qualifies. By these means one can show formally that the degree of proximity with {SIN} is greater than with {SUD}. (22)

{SU} {SIN}

[aproximate] [ßproximate]

132

Raymond Hickey

{SUD}

[yproximate]

The forms Ε - G are all combinable with a single morpheme, but only with this one, as shown in the following. (23)

d.

{AN-}2

/an/ an

[+locative]

This converts the forms Ε - G, which unlike earlier base morphemes are free and are regularly used to form demonstrative pronouns, into forms which denote the location in question. (24)

3.2.

Ed: Fd: Gd:

{AN-h+{-SU} {AN-h+{-SIN} {AN-}2+{-SUD}

/an'sjAI cmseo -» /en'SjinjI ansin -> /on'sjUid/ ansiud

'here' 'there' 'yonder'

a.

Nil Pädraig anseo inniu. 'Padraig is not here today.'

[aproximate]

b.

Τά do cheannsa ansin ceart go leor. 'Your one is there all right.'

[ßproximate]

c.

Rachaimid ansiud tar eis an dinneir. 'We will go over there after dinner.'

[yproximate]

Analysis with the features [+ /-confined] and [+/•-stationary]

Closely related to the last instances of lateral morphemes is a further group of locational morphemes whose members share a feature which separates them from the foregoing, namely [+/-confined]. (25)

Bhi a athair amuigh sa ghort i rith an lae. 'His father was out in the field during the day.'

In the above sentence with the (complex) locational morpheme amuigh there is an implied contrast between the field [-confined] and a further element with the feature [+confined], for example, a house. This accounts then for the semantic distinction between (25) and (26). (26)

Bhi a athair sa ghort i rith an lae. 'His father was in the field during the day.'

Direction and location in Modern Irish

133

The two base morphemes which are used when the notion of confinement, or its absence, is to be expressed are the following. (27)

H: I:

{IST-}7 {AM-}

/isjtj/ ist/am/ am-

[+confined] [-confined]

For each of these there are two specifying morphemes. (28)

e: f:

{-ACH} {-1}

/ax/ ach /i,i:,A/ (u)igh

[-stationary] [+stationaiy]

This then results in the following combinations. (29)

He: Hf: Ie: If:

{IST-}+{-ACH} {IST-}+{-I} {AM-}+{-ACH} {AM-}+{-I}

-> /isytjaxl isteach /isj'tji/ istigh /a'max/ amach / s W amuigh

'into' 'inside' 'out to' 'outside'

Three particular points have to be discussed in connection with this set of morphemes. Firstly, the feature [+/-stationary] is necessary as there is a formal distinction which depends on it, see e and / in (28). Contrast the following two sets of sentences, the first of which does not distinguish formally between [+stationary], see (30a), and [-stationary], see (30b). (30)

a.

b.

c. d.

Τά an pictiuir thar barr an dorais. [is the picture over top door-GENITIVE] 'The picture is over the door.' D 'imigh si thar sdile le fiche blian. [went he over brine/sea with twenty years] 'He went abroad at twenty years of age.' Τά Seän istigh sa teach. 'Seän is inside in the house.' Tar isteach sa chistin mäs mian leat. 'Come into the kitchen if you want to.'

One can thus mark the forms He - If as follows. (31)

{IST-}+{-ACH} {IST-}+{-I} {AM-}+{-ACH} {AM-}+{-I}

[+confmed,-stationary] [+confined, +stationary] —> [-confined, -stationary] [-confined, +stationary]

134 Raymond Hickey The second point concerns alternative phonetic realisations for Hf and If. Basically, the front vowel /i/ is to be found after palatal consonants, i.e. with istigh, and ltd after non-palatal consonants, i.e. with amuigh. There are variants with both istigh and amuigh in which a long ivJ occurs (in both cases). However, there does not seem to be any semantic distinction between forms with a short and those with a long vowel. Rather the conditioning appears to be phonological. A third point concerns an additional form related to amuigh where an 8 /l-/ is prefixed to the adverb and an /s/ is infixed, the form then being followed by the preposition 'of (or a compound form of preposition+article). The semantic function of this prefixed /l-/ is to add the notion of relationality, in a similar fashion to that described above in (20a+b). The relational element may simply be implied in elliptical sentences such as (32b). Note that the prefix /l-/ is also found with the vertical morpheme {-UAS} [+above], see (32c). (32)

{L-}+{ASM-}+{-I}

[+relational,-confined,+stationaiy]

a.

Τά duine eigin lasmuigh den teach. 'Someone is outside the house.'

b.

Τά duine eigin lasmuigh. 'Someone is outside.'

c.

Lastuas den bhalla ud. 'Beyond (i.e. further up than) that wall over there.'

4. Cardinal point reference Finally it remains to deal with the morphemes which are used for the cardinal points in Irish. The four base morphemes can be formalised as follows. (33)

{UA} {AS} {ER} {IR}

-> -» -»

/us/ uaidh /asleas /ε r,/ oir /i:r/ iar

'North' 'South' 'East' 'West'

Each of these can be combined with one of three specifying morphemes which have the semantic features as specified in (4) above. However, only (34c) {H-} [-direction] can retain its graphemic form with all the cardinal

Direction and location in Modern Irish

135

points; (34a) {S-} [+direction to] and (34b) {AN-}l [+direction from] are variously modified, in some cases quite radically. a.

{AN-} ι [+direction from]

/adlad /any an fan! an /any cm

North South East West

b.

{S-} [+direction to]

/o: h l o t h /o:j/ odh Isis /Sj/ s

North South East West

c.

{H-} [+stationary]

Ihlth

North, South, East, West

If one divides the cardinal points into two planes, a longitudinal (vertical) and a latitudinal (horizontal) one, as was done with some of the other directional and locational morphemes, then one can formulate the conditions for the realisation of the specifying morphemes on the horizontal plane (East, West) and, in the case of the morpheme with the feature [+stationary], on both planes (North-South as well). (35)

[+direction from]

{AN-

/on// _ V [-palatal] /an//_ _V[+palatal]

East West

[+direction to]

{S-}

Isll

V[-palatal] / S j / / _ V[+palatal]

West East

[/t-stationary]

{H-}

Ihl 1

V[+/-palatal]

North, South, East, West

The vertical plane for {AN-}i and {S-} does not allow such a formulation. It shows synchronic irregularities in two instances: 1) in the form for North with the feature Indirection from], i.e. /gd/ where the occurrence of [d] cannot be accounted for in terms of palatalisation and 2) in the forms /o:h/ and /o:j/ which are not the expected forms but rather consist of a preceding directional preposition /ο:/ ό and a lenited initial consonant, i.e. Ν th and /j/ dh of the unlenited forms /tua/ tuaidh and /djas/ deas which do not occur as such, however.

136 Raymond Hickey 4.1.

Names of cardinal points

One can nonetheless postulate /tus/ and /djas/ as abstractions (along with forms for East and West) and thus by means of two suffixes derive the appellative nominal forms for the cardinal points. Note that the short vowel /a/ in the form for South is raised to /ε/ before the initial palatal consonant of the suffix /-Sjkjart/, i.e. one has /djesjkjart/. This raising of short vowels before palatals is the rule in Irish phonology so that there is external justification for the description being offered here. (37)

{-SC ART}

/sjkjart/ sceart

{-HAR}

/har/ th(e)ar

{TUA-, DAS-} + {-SCART} {TER-, TIR-} + {-HAR}

nominalising suffix (longitudinal) nominalising suffix (latitudinal)

/tuisjkjart/ tuaisceart /djESjkjart/ deisceart -> /terror/ (an) t-oirthear /tjiirhar/ (an) t-iarthar

'the North' 'the South' 'the East' 'the West'

The initial /-prefix (found before vowel-initial masculine nouns) is included here as it is present in the citation form of these nouns, i.e. in the nominative singular with the definite article. 4.2.

Direction, location and cardinal points

In conclusion one can combine the base morphemes in (33) with the specifying morphemes in (34) to attain the following actual forms. {AN-}i+{-UA} {S-}+{-UA} {H-}+{-UA} {AN-}i+{-AS} {S-}+{-AS} {H-}+{-AS} {AN-}i+{-ER} {S-}+{-ER} {H-}+{-ER} {AN}i+{IR}

-» - >

->

-» -» ->· — »

/d'dua/ aduaidh /o:'hu9/ 6 thuaidh /hus/ thuaidh /s'njas/ aneas /oi'jas/ ό dheas /has/ theas /s'nsrj/ anoir /serj/ soir /herj/ thoir /g'njiir/ aniar

'from the North' 'to the North' 'in the North' 'from the South' 'to the South' 'in the South' 'from the East' 'to the East' 'in the East' 'from the West'

Direction and location in Modern Irish 137 {S}+{IR} {H}+{IR}

/sji:r/ siar /hia-/ thiar

'to the West' 'in the West'

In Irish, as in English, one has the possibility of combining these forms with each other. Arranged according to specifying morphemes one can present three groups for the four cardinal points. These can be used, alone or in combination, as the sample sentences below illustrate. (39)

Stationary:

thuaidh, theas, thoir, 'in the north, south, thiar east, west'

Direction to:

ό thuaidh, ό dheas, soir, siar.

'to the north, south, east, west'

Direction from:

aduaidh, aneas, cmoir, aneas

'from the north, south, east, west'

a.

Bhi me thuaidh sa samhradh Ί was up north for the summer.'

b.

Rachainn ό dheas dä mbeadh me in ann. Ί would go south if I could.'

c.

Beidh siad ag teacht cmoir. 'They will be coming from the east.'

d.

An ghaoth aniar aneas. [the wind from-the-west from-the-south] "The South-West wind.'

5. Conclusion The consideration of the above adverbs in Modern Irish shows that the morphology of the language has a highly structured and largely regular system for the expression of the semantic notions of location and direction. Of course any sub-system in a language is the result of diachronic developments, but the considerable regularity of this section of Irish morphology offers firm evidence for the semantic status of the twin notions of location and direction, the expression of which has maintained its clarity, conferring on it a high degree of synchronic transparency which is unique in the language.

138

Raymond Hickey

Notes 1. The article by Stenson (1997) examines various mixed language idioms (IrishEnglish) of the type Nilair α turnäil si suas [when COMP turned she up] 'When she turned up' and confirms that these fit into the syntax of Irish, by corresponding structurally to constructions already available in the language, and are generally motivated by filling lexical gaps. 2. /h/, if it derives from a non-palatal segment, is always [h] irrespective of the value for /palatal/ of the following vowel. However, if its origin lies in the lenition of a palatal consonant then its realisation is [9] before a back vowel, compare a Shedin h ham/ [9 icwij], ' John' -VOCATIVE with a theach fa hax/ [a hae:x] 'his house'. 3. As with abhus, thar represents a single lexical item which could, however, be segmented morphologically as {H-}+{-AR} -> /ha:r/ thar. Little is gained by this segmentation, however, as {H-} would be the only specifying morpheme combinable with {-AR} and, more importantly, as the whole morphemic unit has only two features, [+stationary] and [+relational], the second of which would then apply to the base only. The feature [+relational] is insufficient, however, for a base morpheme as it gives no information as to location or direction. 4. For this morpheme the final consonant has been doubled which has the advantage of allowing a distinction between it and the specifying morpheme b. 5. This form is found in Dinneen (1927: 1087) where it is marked as 'early' (i.e. Middle or Old Irish). It is also listed in the Contributions to a dictionary of the Irish language, Vol S. (1953: 429). 0 Dönaill (1977: 1134) labels it as 'literary'. The developmental history of many such forms is treated in Ahlqvist (1976: 158-168). 6. Synchronically of course. Historically the form arose from /' bhfus which literally meant 'in repose' which then took on the meaning 'here' as seen in fan i bhfus 'stay here' (Dinneen 1927: 483). The modern spelling abhus reflects the indeterminate quality of the initial vowel which is unstressed. A parallel in English is afforded by such words as about, alive, asleep where a former preposition and noun have become a predicative adjective (Old English on släfye -> asleep). 7. In the treatment of this and the following morpheme the segmentation made is not correct when viewed diachronically. However, in a synchronic description which purports to account for native speaker's intuitions of morpheme structure this segmentation is appropriate. Note further that the palatal quality of the t in this morpheme can be given in the phonetic description of the base morpheme as it is always followed by a front vowel. 8. Diachronically this /I/ represents an elision of the word leath /lja/ with a following istigh or amuigh such that it was no longer recognized as such and came to be written as an initial / with these forms (cf. abhus / and its syncopated variant φ / (φ/ stands for the IPA symbol /Θ/ in Danielsson's notation). He restricts the syncopated form φ / to colloquial speech, and it appears as a minority variant in his transcriptions. Even in The opening of the unreasonable writing of our Inglish toung (1551), the full form Iφ! predominates. In all his works, Hart transcribes /ej>/ or spells -ΕΤΗ in the overwhelming majority of cases (Danielsson 1955, II, 175-176). Another piece of contemporary evidence can be adduced from the autobiography of the musician Thomas Whythorne (1961 [1576]). He was inspired by Hart's suggestions for phonemic spelling conventions, which he largely adopted in his manuscript. In Whythorne's first thirty instances of the third-person indicative present-tense suffix, only four appear with

238

Terttu Nevalainen and Helena Raumolin-Brunberg

-S, all of them syncopated (belongs, breedz, maks, remayns), while all the rest display the syllabic -ΕΤΗ (quotes from the Bible and verse passages were excluded from the sample). Writing in 1586, William Bullokar allows the use of -S as a poetic contraction of -ΕΤΗ. In the 1620s, Alexander Gill labels HAS as a Northern variant of HATH, regarding the forms with the dental fricative (HATH and DOTH) as normal, but adds that in other verbs -ΕΤΗ can shorten to -s or -z, or become -ez after a sibilant. Finally, in 1643 Richard Hodges notes that, although -(E)TH may appear in writing, it is commonly pronounced -s or -z in ordinary speech. Hodges' general usage is praised by Dobson (1957,1, 186) as "real and unaffected" and "in many ways somewhat advanced". It is, however, often forgotten (but observed by Danielsson 1955, II, 176), that Hodges, too, continues to use the unsyncopated suffix with the dental fricative in his transcriptions of liturgical speech in The English Primrose (1644). Variation is therefore clearly in evidence in the verbal suffix in the spoken registers of Londoners even in the middle of the seventeenth century. The past-tense and participial suffix -ED loses its vowel later than the nominal and third-person singular inflections. Hart provides evidence for both syncope and lack of it in his commentary and transciptions. His Orthography (1569) and Methode (1570), for instance, show syncope in 38 lexeme types ending in consonants, but full forms in 30 others (plus unsyncopated, full forms in another 30 verbs in -t and -d). Interestingly, the unsyncopated form also occurs after vowels and diphthongs in both what Danielsson (1955, II, 176-177) considers his ordinary speech and formal prose. Gill (1621) similarly uses both forms, though in his prose unsyncopated forms outweigh syncopated forms. By contrast, Hodges (1644) prefers syncope, except in his biblical passages, where he often retains the full form. In sum, it is suggested by the orthoepists examined by Danielsson (1955) and Dobson (1957) that the syncopated form had become the norm by the latter half of the seventeenth century. 2.2.

Orthographic evidence

Our orthographic evidence is drawn from the Corpus of Early English Correspondence (CEEC), a 2.7-million-word corpus of personal letters compiled at the University of Helsinki as part of the Sociolinguistics and Language History project (see, e.g., Nevalainen & Raumolin-Brunberg (eds.) 1996). The present data consist of samples ranging from 1,700 to 4,000 words by individual writers.2 Our aim was to include only letters edited from autograph originals, but some exceptions to this principle

The third person singular ~(E)S and -(E) TH revisited

239

could not be avoided. The letters by Godfrey Greene, Robenet Plumpton and Henry Savill go back to copies, since no authentic Northern material was available from their time period. We are naturally aware of the pitfalls of using ordinary spelling as evidence for morphophonemic variation. It is particularly the common sixteenth-centuiy use of the final (e) to signal the length of the preceding vowel that potentially conflicts with the syllabic interpretation of a suffix like -ΕΤΗ. Hart (1955[1551]: 122) explicitly condemned this "superfluous" vowel in monosyllables and final syllables. So did Thomas Whythorne (1961[1576]: 5), one of Hart's followers, who in his "new Orthografye" compares different ways of "to make the wurd long, & leaving owt the .e. at the end of those wurds".3 General orthographic developments certainly make the spelling-topronunciation mapping increasingly difficult, especially as regards later spellings, where regularisation had taken place in private texts, too, at least to some extent (Salmon 1999). On the other hand, our material contains a number of informants whose level of education was relatively low (merchant families like the Celys and the Johnsons, and the theatrical manager Philip Henslowe, among them). Because their letters frequently display 'phonetic' spellings, we are tempted to believe that many of their actual morphophonemic variants, too, may be represented in them. Table 1 gives the variant spellings of -(E)S and -(E)TH by a number of Londoners and residents and high administrative officers at the Royal Court in four periods. Some of their family have also been included, viz., Sabine Johnson, the wife of the London merchant John Johnson, and their servant Richard Preston, who mostly lived at the Johnson family estate in Northamptonshire. The informants include both native Londoners and immigrants, such as Edmund and John Paston and Philip Gawdy, who were born in Norfolk. The invariants eth, th, es, s in Table 1 stand for a whole range of spelling variants. It is interesting to find that, especially before 1500, the poststem vowel was most commonly (y) (see examples 1-2); it, too, is represented by e in the table. Examples (3-7) illustrate other variants. (1)

and also wat pontment makyt wyt Byfylde and all hoder men (CELY, Richard Cely sr, 11, 1477)4

(2)

as he that thenkys a long seson (STONOR, Richard Page,

141,

1482?) (3)

Pleasith it your Majestie to widerstände (GARDINER, Stephen Gardiner, 9, 1529)

240

Terttu Nevalainen and Helena Raumolin-Brunberg

(4)

all tynges that he deliverethe (JOHNSON, Anthony Cave, 8, 1542)

(5)

As the Lord knowth (JOHNSON, Sabine Johnson, 433,1545)

(6)

wcA as she sayes comes from her harte (HENSLOWE, Philip Henslowe, 277, 1593?)

(7)

and he desyers nothing more (LEYCESTER, Robert Dudley, 141, 1586)

Table 1 clearly shows that the unsyncopated forms were in the majority until 1600. Only two of our informants occasionally used -TH without a preceding vowel, Sir Thomas Wyatt and Sabine Johnson; in both cases it followed a stem-final vowel. Also, verbs with a syncopated -S occurred from Middle English onwards, but the first instance without a preceding vowel only turned up in the 1540s. The syncopated -S never acquired a significant position in this material. As expected, a stem-final sibilant is always followed by -ΕΤΗ in the material studied here. Table 1. Third-person suffix (sg and pi), London informants Preceding phoneme: Vowel Sibilant Before 1500 Betson Thomas Cely George Cely Richard sr Cely Richard jr Dalton John Maryon William Page Richard Paston Edmond2 Paston John2 1500-1540 Henry VIII Cromwell Thomas Gardiner Stephen More Thomas Wolsey Thomas Wyatt Thomas 1540-1560 Princess Elizabeth Cave Anthony Johnson John Johnson Otwell Johnson Richard

ΕΤΗ (ΕΤΗ) ΕΤΗ

(ΕΤΗ) ΕΤΗ

es, s eth, es, s

(ΕΤΗ)

—

—

ΕΤΗ

—

—

(ΕΤΗ)

ΕΤΗ, (j) ΕΤΗ —

ΕΤΗ ΕΤΗ ΕΤΗ (ΕΤΗ)

Other eth, es es, eth ΕΤΗ ES ES ΕΤΗ ES

—

es, eth

(ΕΤΗ)

eth, (es)

(ΕΤΗ) (ΕΤΗ) ΕΤΗ (ΕΤΗ)

ETH ETH ETH ETH ETH ETH

—

eth, th

(ΕΤΗ)

(ΕΤΗ) ΕΤΗ ΕΤΗ ΕΤΗ ΕΤΗ

(ΕΤΗ) — —

(ΕΤΗ) —

eth, es eth, es eth, (es) ETH, (es) ETH

The third person singular ~(E)S and -(E) TH revisited Johnson Sabine Preston Richard 1560-1600 Elizabeth I Cecil William Dudley Robert Henslowe Philip Gawdy Philip

eth, th, (es, 5) s, (es) eth, (es) ΕΤΗ eth, (es) es, eth ΕΤΗ

—

—

(ΕΤΗ) —

(ΕΤΗ) (ΕΤΗ) —

241

es, (s) ES eth, es, (s) ETH eth, es, s es, eth ETH, Os)

Note to Table 1. The suffixes in small capitals represent the main variants, i.e. those used in 80% of the cases or more. The endings in italics are rare variants, only up to 20% of the occurrences. The remaining endings represent frequencies between 21% and 79%. The forms in brackets are rare in absolute terms, only one or two occurrences.

Despite the small size of our Northern sample, the principal conclusion can be drawn from Table 2 that the syncopated form -S occurred in verbs with stems ending in sounds other than vowels or sibilants as early as the fifteenth century. This, of course, is well in line with the traditional view that the syncope of the vowel was first witnessed in the North. Our early sixteenth-century writers seem to have preferred suffixes with a vowel, indeed -ΕΊΓΗ in the majority of the cases. The syncopated -S occurs as a rare variant in the letters of our two informants writing after the 1540s. Table 2. Third-person suffix (sg and pi), Northern informants Preceding phoneme:

Before 1500 Greene Godfrey Plumpton Robenet 1500-1540 Dacre Thomas Wharton Thomas Wilson John 1540-1560 Savill Henry 1560-1600 Hutton Matthew

Vowel

Sibilant

—

—

(eth, th)

—

(eth, es) ETH ETH

—

ETH ETH

Other es, s (eth, es, s) eth, (es) ETH ETH, (es)

eth, (es)

—

es, (s)

(ETH)

(ETH)

ETH, (5)

Note to Table 2. For conventions, see Table 1.

The statements by the orthoepists surveyed above inspired a study of the nominal -(E)S endings (both the genitive and the plural) in our letter

242

Terttu Nevalainen and Helena

Raumolin-Brunberg

data. Appendix 1 indicates that the phonological conditions for the use of the syncopated forms differed from those of the verbal inflection. After a stem-final /n/ and Irl the syncopated -S was common from Middle English onwards, while in other nouns the suffix -ES was preferred up to 1600 (Aronoff 1989: 95; see further note 3). However, a comparison between Tables 1 and 3 suggests that the syncope of the vowel was earlier in the nominal suffix in London than in the verbal suffix. In this respect our material supports the evidence given by Hart's transcriptions of speech. There are, nevertheless, several reasons for interpreting our data with caution. The growing regularisation of spelling was already referred to above. Moreover, our material is based on published editions and not on manuscripts. This problem has been alleviated in a number of cases by collating edited letters against the originals (e.g. the Johnson, Henslowe and Gawdy collections, as well as Queen Elizabeth's early letters). Some editorial slips must nevertheless have remained in the material. Secondly, no attempt was made here to account for the phonotactic and prosodic conditioning factors that might have affected the realisation of the suffix. The length of a word, its stress pattern, and larger prosodic context were all important factors constraining the choice between the syncopated and unsyncopated alternatives, but they were not taken into account in our analysis (cf. Dobson 1957, II, 880-881; Strang 1970: 180). Thirdly, as has been shown by Ogura and Wang (1996), the spread of the suffix -(E)S appears to have followed a pattern of lexical diffusion. The choice of the third-person ending apparently varied from one lexeme to another, depending, among other things, on the overall frequency of the word. This is another conditioning factor that should be considered in a more comprehensive study. Despite all these reservations, we would, however, like to suggest that our material may be used as supportive evidence for the relatively late syncope of the vowel in the third-person verbal suffix -(E)s even in less carefully monitored registers in the South. Although the process had commenced in the North in the Middle English period, the syncopated suffix appears to have gained currency in London only in the second half of the sixteenth century (see Wakelyn's (1988: 112) similar conclusion on Queen Elizabeth's usage in the 1580s). This part of our study also speaks for the uncommonness of the syncopated form of the suffix -ΕΤΗ.

The third person singular ~(E)S and -(E)TH revisited 243

3. Two processes of change Now that we have established the approximate timing for the loss of the post-stem vowel in the third-person singular, it can be correlated with the diffusion of -S to the South of England. In our previous research we found that the replacement of the Southern suffix with the Northern throughout the country did not follow a neat S-curve of change. Figure 1 shows the variable usage of -(E)TH and -(E)S in three regions in the data contained in the Corpus of Early English Correspondence: London, East Anglia, and the North (from Nevalainen & Raumolin-Brunberg, forthcoming). These three regions are represented by writers whose domicile was 1) London, covering the City, Westminster (and the Royal Court), and Southwark (= Londoners); 2) Norfolk or Suffolk (= East Anglians); and 3) the counties north of Lincolnshire (= Northerners). The twenty-year sampling periods of the corpus are conflated into periods of forty years in Figure 1. Figure 1. Percentage of -S (vs. -TH) in the third-person singular indicative in three dialect areas; averages of individual means (1460-1680; excluding HAVE and DO; from Nevalainen & Raumolin-Brunberg, forthcoming). 1

% 50

14601499

15001539

15401579

15801619

16201659

16601680

At first, the results shown in Figure 1 appeared slightly anomalous to us. What seems to be a change in full swing in London at the end of the fifteenth century fails to be generalised, petering out at the beginning of the sixteenth. In this respect -(E)S resembles another Northern form that was similarly favoured in London in the Late Middle English period, the complex relative pronoun THE WHICH. It, too, failed to supralocalise in Early Modern English, although it had made its way into the documents

244

Terttu Nevalainen and Helena Raumolin-Brunberg

produced by the Chancery. This is, however, where similarities between the relative pronoun and the present-tense suffix end. The -(E)S form was not used in the early Chancery documents, but it was favoured by London merchants in the late fifteenth century. As we have demonstrated, however, it was the full form -ES that appeared in the letters of the Celys and their contemporaries, and not the contracted one. For reasons mostly unknown to us, this usage did not win enough support in the speech community, but it was rather the traditional Southern form, just like the relative WHICH, which started its crusade to the North in the next century as the dominant supraregional form. The second coming of -(E)S shows in the steep rise in the S-curve in Figure 1: the use of the form explodes in London in the period 1580-1619. This development coincides with the process of vowel loss in the thirdperson singular present-tense suffix. As we have seen, Hart only used the syncopated form of -S in his transcriptions, when he (rarely) resorted to it. He also testifies to the use of the contracted form of the native Southern suffix -TH, although only in about half of the cases of verbs ending in consonants other than sibilants. Where the two suffixes were rivals, then, was not so much the final consonant than the syncope of the preceding vowel. If the sibilant suffix and syncope went hand in hand in the North, it is no wonder that Southern orthoepists interpreted the syncopated -S as a colloquial contraction of -ΕΤΗ. This impression could certainly have been reinforced by the analogy of nominal suffixes, which had lost their vowel earlier in the South, too.

4. Conclusion As suggested by Kastovsky (1997: 133), the shift from vowel deletion to vowel insertion constitutes a case of rule inversion in the history of the preterite and past participle, and plural, genitive and third-person singular inflectional morphemes. Our late sixteenth-century London evidence can be taken to indicate that the third-person singulars underwent a period of transition that would have resulted in a suppletive paradigm, had it been generalised. The vowelless allomorph -S was coming to be accepted with all verbs except sibilant-final ones (and HAVE and DO), while the variant with the vowel, -ΕΊΉ, was particularly associated with sibilant-final verbs. It is true that the third-person singular present-tense suffix stands out as an anomaly in the inflectional morphology of Present-Day Standard English. Unlike the past-tense suffix -ED, which applies to all persons, the present-tense indicator is used solely in the third-person singular. Unlike

The third person singular ~(E)S and ~(E)TH revisited 245 the plural suffix -S in nouns, only the third-person singular is marked for number in verbs. Had the allomorphy that we have sketched for the late sixteenth-century London won the day, third-person singulars would now stand out as the exception to the morphophonemic rule as well. Appendix 1 Table 3. Noun suffixes (genitive and plural), London and Court informants. (For conventions, see Table 1.) Monosyllabic nouns with a All data long vowel or diphthong excluded η r other η r other Preceding phoneme Before 1500 s, (es) s, es ES Betson Thomas ES Cely George s, (es) s, es es, s es, s ES, (s) Cely Richard sr es, s S, (es) ES, (s) Cely Richard jr S es, s ES Dalton John ES ES Maryon William ES (es, s) s, (es) ES Page Richard ES s, es ES, (s) Paston Edmond2 ES ES, (s) ES, (s) Paston John2 1500-1540 s, es s, es ES, (s) Henry VIII s, (es) es, s ES, (5) Cromwell Thomas ES, (s) Gardiner Stephen s, es S, es s, (es) es, s ES, (s) More Thomas es, (s) es, s ES Wolsey Thomas es, s Wyatt Thomas s, (es) es, s 1540-1560 s S, (ei) ES, S s S, (es) Princess Elizabeth ES, s Cave Anthony S, (es) S, (es) ES, (s) s S, (es) ES, (s) s, es s, es ES s Johnson John s, es ES, ES, S S, (es) s, es s es, s Johnson Otwell ES, S s, es S, (es) ES, S s S, (es) ES, s Johnson Richard s, es s, es ES s, es s, es Johnson Sabine ES es, s es, s ES, (s) ES, (.s) ES, S ES, S Preston Richard 1560-1600 es, s Elizabeth I s, (es) s, es S, (es) S, es es, s s, es S, es ES, s S S, (es) Cecil William ES, S S, es s, es es, s S Dudley Robert s, es es, s s, es Henslowe Philip s, es ES, s s, es S, (es) ES, S s, es s, es ES, S Gawdy Philip S, (es) S, (es) ES, S

246

Terttu Nevalainen and Helena Raumolin-Brunberg

Notes 1. Our warm thanks go to Dieter Kastovsky, both for his mindfulness of the Eigenart of the Early Modem period in the history of the English language, and for his suggestion back in 1997 that we might benefit from a broader perspective on the issue of -S/-TH variation in Early Modern English in terms of both level interaction and its typological implications. 2. The random sampling was done as follows: after selecting appropriate informants, all the material in the corpus was analysed for writers with less than 4,000 running words. For people with more data, the material included in the analysis was limited to twenty WordCruncher screens, which roughly amounts to 4,000 words. The search for suffixes was carried out manually. 3. We tested the possible influence of the final (e) signalling the length of the preceding vowel on our material. Monosyllables with a long vowel or diphthong, such as write, feel, cause among the verbs, and need, queen, wife among the nouns were excluded from the analysis for 1540-1600. As regards the verb category "other", the results did not differ from Table 1 in any relevant manner, whereas in nouns it could be seen that the syncopated form was more common after /n/ and /r/ than in our previous analysis. It seems that the use of the 'superfluous' (e) condemned by Hart could be found with many nouns (see Table 3). The prefinal (e) in forms like fynes and year es, for instance, was probably not understood as part of the suffix but part of the stem, since with other words ending in /n/ and /r/ the syncopated form was clearly preferred. 4. The references in examples give the abbreviated name of the letter collection used, name of informant, page number and year of writing.

References Aronoff, Mark 1989 The orthographic system of an Early English printer: Wynkyn de Wörde. Folia Linguistica Historica 8: 65-97. Britton, Derek (ed.) 1994 English historical linguistics 1994. Amsterdam: Benjamins. Danielsson, Bror 1955 John Hart's works on English orthography and pronunciation, Part II, Phonology. Stockholm: Almqvist & Wiksell. Dobson, E.J. 1957 English pronunciation 1500-1700, Volume II, Pronunciation. Oxford: Clarendon Press. Haugland, Kari E. 1995 Is't allowe'd or ain't it? On contraction in early grammars and spelling books. Studia Neophilologica 67: 165-184.

The third person singular -(E) S and -(E) TH revisited

247

Hogg, Richard & Linda van Bergen (eds.) 1997 Historical linguistics 1995. Volume 2: Germanic linguistics (Current Issues in Linguistic Theory 162). Amsterdam & Philadelphia: John Benjamins. Holmqvist, Erik 1922 On the history of the English present inflections, particularly -th and -s. Heidelberg: Carl Winter's Universitätsbuchhandlung. Kastovsky, Dieter 1992 Typological reorientation as a result of level interaction: the case of English morphology. Kellermann & Morrissey (eds.), 411-428. 1997 Morphological restructuring: the case of Old English and Middle English verbs. Hogg & van Bergen (eds.), 131-147. Kellermann, Günter and Michael Morrissey (eds.) 1992 Diachrony within synchrony: language history and cognition (Duisburger Arbeiten zur Sprach- und Kulturwissenschaft 14). Frankfurt/M.: Peter Lang. Luick, Karl 1921 Historische Grammatik der englischen Sprache. Leipzig: Tauchnitz. Lutz, Angelika 1991 Phonotaktisch gesteuerte Konsonantenveränderungen in der Geschichte des Englischen. Tübingen: Max Niemeyer Verlag. Nevalainen, Terttu and Helena Raumolin-Brunberg (eds. ) 1996 Sociolinguistics and language history; Studies based on the Corpus of Early English Correspondence. Amsterdam & Atlanta GA: Rodopi. Nevalainen, Terttu & Helena Raumolin-Brunberg. forthc. The changing role of London on the linguistic map of Tudor and Stuart England. In: Dieter Kastovsky and Arthur Mettinger (eds.), The history of English in its social context. (Topics in Linguistics.) Berlin - New York: Mouton de Gruyter. Ogura, Mieko and William S.-Y. Wang 1996 Snowball effect in lexical diffusion: the development of -s in the third person singular present indicative in English. In: Britton (ed.), 119— 141. Salmon, Vivian 1999 Orthography. In: Roger Lass (ed.), The Cambridge history of the English language, Volume 3, Early Modern English. Cambridge: Cambridge University Press. Stein, Dieter 1987 At the crossroads of philology, linguistics and semiotics: notes on the replacement of th by s in the third person singular in English. English Studies 68: 406-431.

248

Terttu Nevalainen and Helena Raumolin-Brunberg

Strang, Barbara 1970 A history of English. London & New York: Methuen. Wakelyn, Martyn 1988 The archaeology of English. London: Batsford. Whythorne, Thomas 1961 The autobiography of Thomas Whythorne, Edited by James M. Osborn. Oxford: Clarendon Press.

Paths of loan-word grammaticalisation: the case of according to Matti Rissanen

1. Introduction Recent studies of grammaticalisation have mainly concentrated on native vocabulary.1 This is only natural as grammaticalisation, which represents one of the central language-internal processes of change, affects even the core vocabulary of the language, and tracing the changes in this part of the lexis is essential for understanding the nature of linguistic development. It seems, however, that a survey of the semantic developments of loanwords may add some details to our attempt to understand the process of grammaticalisation. It is in most cases easy to determine the approximate time of the introduction of a loan-word; this enables us to trace the developments gradually leading to grammaticalisation from the veiy beginning. It is also interesting to study the emergence of grammaticalised uses in relation to the establishment of the loan in the language. Finally, it seems that the establishment and spread of borrowed vocabulary, typically representing a 'change from above',2 is particularly sensitive to genre- or text-based constraints, which may give us some insight into the introduction of grammaticalised forms in the developing standard(s). Three kinds of grammaticalisation are typical of Middle English French loans: conjunctivisation, adverbialisation and prepositionalisation. There is plenty of evidence in present-day English lexis of each of these three types, and many loan-words show grammaticalized forms belonging to more than one of these groups (cf. Kortmann 1996: 24-28). The normal pattern and chronology of semantic development seems to be as follows: the word, either verb or noun, is borrowed into early Middle English with a fairly specific meaning. In later Middle English, its use and context of occurrence are generalized and its original prototypical meaning may give way to new prototypical meanings.3 This stage paves the way for grammaticalised uses, which may emerge in the fourteenth century and are gradually established in the fifteenth century and in Early Modern English. This pattern can be traced in the development of a number of French loans: according to, excepting), provided/providing, because (of), etc.4

250

Matti Rissanen

According, originally of course the present participle form of the verb accord, offers a nice example of the generalisation and grammaticalisation of an early French loan. The verb is borrowed as early as the twelfth century, its sphere of meaning expands rapidly, and the grammaticalisation of the -ing form seems to begin in the fourteenth century. Its grammaticalised forms comprehend all the three categories mentioned above: complex preposition according to, complex subordinator according as, and derived adverb accordingly. In the present paper I will concentrate on the development of the preposition.

2. Accord in Middle English The earliest extant occurrences of the verb accord, five in all, come from the Anglo-Saxon Chronicle, from annals dating from the mid-twelfth century:5 (1)

& syddan pees cynges mcen Heanriges manega him to gebugen. & wid hine acordedan pe ceror mid heora castelan him togeanes wceron. (,Anglo-Saxon Chronicle [CHRONE2] 248, annal for 1119)

(2)

On pis hinges time wes al unfrid & yfel & rceflac, for agenes him risen sona pa rice men pe wceron swikes, aire fyrst Balduin de Reduers; & held Execestre agenes him & te king it bescet, & siddan Balduin acordede. (Peterborough Chronicle [PETERB] 55, annal for 1135)

(3)

An se arcebiscop Turstein of Eoferwic weard purh pone papan wid pone cyng acordad. & hider to lande com. & his biscoprices onfeng. (Anglo-Saxon Chronicle [CHRONE2] 249, annal for 1120)

Also the Anglo-Saxon Chronicle 249 (1120); 250 (1121). As can be seen in these examples, the word was used in military/political contexts with reference to alliance both intransitively (1)(2), and transitively (3). The OED definition of the meaning of the intransitive use is "to come to an agreement or to terms; to be at one, to agree" (s.v. accord II 5), and of the transitive use, "to bring (persons) into agreement or harmony, to reconcile one with another" (I l).6 The action clearly takes place between human referents. I have not found any instances of the verb or noun accord in other extant 12th century texts and even in the 13 th century the word seems to

Paths of loan-word grammaticalisation: the case of according to

251

be rare. The only instance found in dictionaries or corpora is in The Owl and the Nightingale, in the Cotton MS from c. 1250. This is also the earliest instance of accord as a noun:7 (4)

pejwenebo atoneacorde, we m[a]je bet midfayre worde, witute eheste, & bute fi$te, plaidi midfoje... (iOwl and Nightingale MS C 181-184)

Both MED and OED quote a number of instances from c. 1300 (Robert of Gloucester's Chronicle·, South English Legendary). The increasing frequency of accord at the turn of the century is confirmed by the Helsinki Corpus material and by the Michigan Middle English Database. There are nine instances altogether in the Helsinki Corpus samples, recorded from the Early South English Legendary (2), Robert of Gloucester's Chronicle (5), the Earliest English Prose Psalter (1) and the Ayenbite of Inwyt (1), all dating from the period between c.1300 and 1350. In seven instances accord, verb or noun, refers to agreements or union between people, which is the original meaning of the loan. This is also true of the Owl and Nightingale instance quoted in (4) above. In some early HC instances, however, the referents are inanimate:8 (5)

Pulke ymage he weddede with a ring: ase a man dotp is wif, Clanliche to holden in spoushod: to hire al is lijf; "Aue Maria, graciaplena:" peos four wordes weren i-do And i-graued in pis ring: for huy a-cordeden wel par-to. (Early South English Legendary [SLEG] 434)

(6)

me ssel weje pet word er hit by yzed. And pou sselt ywyte / pet zopnesse halt pise rijtuolle waye. Vor zopnesse acordej) pe onderstondinge ofpe herte: and pet word of pe moupe / as hit is ine pe herte. (Ayenbite of Inwyt [AYENB] I 256)

In (5) accord is used intransitively, in the sense 'to be compatible or harmonious', 'to correspond', etc. (MED s.v. accorden 5a-b). In (6) the verb is used transitively, 'to reconcile or harmonize (things)' (MED loc. cit. 6a). Another important development in the generalisation of the meaning of accord is its use in impersonal constructions. The following instances date from the end of the 14th century:

252

Matti Rissanen

(7)

Manie such aduerbis, coniuncciouns, and preposiciouns ben set ofte oon for another, and at fre chois of autouris sumtyme; and now tho shulen be taken as it acordith best to the sentence. (Purvey Prologue to the Bible [WYCPROL] 1,60)

(8)

"But, " quod sehe, "I am certein by many resouns that schrewes ben unsely." "It accordeth, " quod I. (Chaucer Boethius [BOETHCH] 448.C2)

Examples (5)-(8) represent an important extension in the meaning of accord in regard to its grammaticalisation. It is of course impossible to say whether we are here witnessing a natural native development or whether the extended meaning is another borrowing from French. But this question is of less importance than the fact that the first step towards the grammaticalisation of the verb has now been taken (cf. Hopper and Traugott 1993: 96-97; Kortmann and König 1992: 680-681). The concrete 'agreeing' or 'allying', with reference to persons, has become 'corresponding' or 'harmonizing' between abstract concepts. These new meanings also support the transfer of accord from mainly historical/narrative genres such as chronicles and legends into religious writing of a more instructive/meditative type, as shown by examples (6)-(8). 3. Grammaticalisation of according (to) The earliest instances of the present participle forms {according, accordand, accordend, accordind) occur in the South English Legendary:9 (9)

It was a-cordinde to pe bodi in lengpe and in brede. (South English Legendary 270/329)

(10)

Pulke schulten so to purgatorie: for-to huy habben to ende ibroujhtpenaunce a-cordinde to heore sunne. (ibid. 421/38)

In (9) the present participle occurs with be; in (10) it is the verb of a non-finite clause. The construction is syntactically identical with the prepositional use, but a-cordinde is obviously still used as a verb, with penaunce as its subject. It is noteworthy that, as in (5) and (6), both the subject and the complement of the preposition are inanimate in these two passages. It seems that the South English Legendary is a particularly early text using the present participle of accord. The other instances quoted by MED

Paths of loan-word grammaticalisation: the case of according to

253

are from the end of the 14th century (Trevisa's translation of the Polychronicon) or later.10 Judging by the number of occurrences in corpora, according (to) remains relatively uncommon throughout the fourteenth century. In the Helsinki Corpus, the earliest instances of the present participle can be found in fifteenth-century texts.11 The prepositionalisation of according to takes place in the fifteenth century. This development can be illustrated by the following, highly simplified sequence of syntactic constructions: (a) (bi) (t>2) (ci) (C2)

χ accords to y and is/does z. χ is according to y and is/does z. χ according to y is/does z. χ is/does ζ according to y. according to y χ is/does z.

It seems that the non-finite use of type (b2) would offer the main source of the preposition as both (ci) and (C2) can be derived from it by simple change of word order. It is possible, however, that the be + -ing construction (bi), which is quite frequent in later Middle English, supported prepositionalisation. In this case, we would be dealing with reanalysis of the type [[x is according] to y] > [x is [according to y] (cf. examples 13 and 14 below).12 The following examples illustrate the path of grammaticalisation given in the schematic form above: Type (a): (11)

Lat its first see what preier is propirly in it-self, and per-after we mowe cleerlier knowe what worde wil best acorde to pe propirte of preier. {Cloud of Unknowing [CLOUD] 77)

Type (bi): (12)

said that this book was not accordyng in many places vnto the book that Gefferey chaucer had made (Caxton Prologue to Canterbury Tales [CAXTON] 91)

(13)

It hath be ordeyned that oon mesure and one Weight shuld be throughoute all this Realme of Englond, which Weight and Mesure shuld be according to the Standard of yor Escheker, And that every Man shuld by and seile by the same and with none oder {Statutes [STAT2] 551)

254 (14)

Matti Rissanen In pis tyme, aftir many tretis betwix pis lond and Fraims, and noping pat pei profered was acording to reson, pe kyng mad redy his schippis at Southampton to spede him to his conquest. (Capgrave [CAPCHR] 243]

Type (b2): (15)

haue others Chirche And iche of them gete other A presentacion from his patron by vertu of wiche Accorde / the seyd bysecher accordyng to this couenant resigned & gate A presentacion of the seyd abbot to the same william (Anthology of Chancery English [Michigan ME Database] 158)

Type (ci): (16)

Justice shuld be had and exercised amonge all yor Subgettes w'in this yor seid Realme according to thejfect of the seid Statutes and ordenances, {Statutes [STAT2] 551)

Type (c2): (17)

accordyng vnto ye prophecye of Abdy. ye fyrste chapytre. rehercyd by the apostle ye fyrst epistle. & fyrst chapitre vnto the Corinthees. I shall sayth god by hys prophete. distroye & brynge to nought ye sapience of sapient men: (Fitzjames [FITZJ] B4R)

The role played by constructions like (15) in the formation of the preposition, and its relationship to (16) and (17) is obvious. It is obvious, too, that in (12) according is still a verb, but it might be asked whether (13) represents an expression where reanalysis is taking place. The meaning of the verb is weakened from 'agreeing on' to 'equivalent to' and the construction is probably to some extent prepositionalised: [[should be according] to the Standard] is being reanalysed to [should be [according to the Standard]]. This kind of development seems even more likely in (14), with the idiomatic sounding expression "x is according to reason". The decisive step from verb to preposition is taken when the link between the subject of the sentence and according is broken. In (16) and (17) above, the prepositional quality of according to is proved by the fact that no clearly definable subject for accord exists. This development is not

Paths of loan-word grammaticalisation: the case of according to

255

surprising in view of the easy formation of unattached non-finite clauses in English (cf. Quirk & al. 1985: 15.59). It is a natural result of the shift of focus from the subject of according to the source, authority, etc., expressed by the complement of the prepositional according to, as in (18), in which the reference is rather indefinitely expressed by "all manner of other challenges" and specified by the prepositional phrase "according to the law": (18)

frome hensforth the chalenge comenly callid Riens Deyns le garde be no chalenge but utterly voide and of none effecte: Savyng to every person all maner of other chalenges what so ever they be according to the lawe. 0Statutes [STAT2] II 552)

The intensity of prepositional developments in Late Middle English is perhaps evidenced by the occurrence of the construction without to, as in the following passages. (19)

the seid Rauf... ordeined by his testament that (the) seidfeoffes of the seid Manere ... shuld immediatly qftir his decesse enfeffe the seide Iohanne terme of hur lyjf accordyng whiche will [= according to which will] the seid Ioh(anne by the sujfraunce of the seid feoffes after the deth of hur seid late husbond occupied (iChancery English 277)

(20)

the mesures and weightes w'in the seid Shire may be correct reformed amended and made according and after the mesure of the seid Standard before the feste of Seynt Mighell tharchangell next commyng. (Statutes [STAT2] II 552)

These instances may suggest a competing trend in the prepositionalisation of according. The possibility of a simple scribal slip cannot be excluded, but the adverbial use of according, quoted by the OED (s.v. accordingppl., aand adv.) is worth quoting in this context: (21)

Send me word by wrytting how he wilbe demeaned, & therafter I shall entreat him according. (71495 Plumpton Correspondence 110)

In the course of the fifteenth century the prepositional use gains ground rapidly. There was obviously a gap in the prepositional and conjunctive system in expressing source or relation. A comparison between a passage

256

Matti Rissanen

of the Wycliffite Bible translation of the late fourteenth century and Tyndale's Bible of the early sixteenth is illustrative: (22)

therfor the strengthe of the Lord be magnified, as thou hast swore. And Moises seide, Lordpacient, and of myche mercy, doynge awei wickidnesse and trespassis, and leeuynge no man vngilti, which visitist the synnes of fadris in to sones in to the thridde and fourthe generacioun, Y biseche, forjyue thou the synne of this thi puple, aftir the greetnesse of thi merci, as thou were merciful to men goynge out of Egipt til to this place. And the Lord seide, Y haue forjouun to hemt bi thi word. (Wyclif [WYCOLD] Numeri 14.17-20)

(23)

So now lat the power of my Lord be greate, acordynge as thou hast spoken sayenge: the Lorde is longe yer he be angrye, and full of mercy, and suffereth synne and trespace, and leaueth no man innocent, and visiteth the vnryghtuousnesse of the fathers vppon ye childern, euen vpon ye thirde & fourth generacion. be mercyfull I beseche ye therfore, vnto y synne of this people acordinge vnto thi greate mercy, & acordinge as thou hast forgeuen this people from Egipte euen vnto this place. And the Lorde sayed: I haue forgeuen it, acordynge to thy request. (Tyndale Numeri [TYNDOLD] 14.17-20)

The simple prepositions after and by are used by Wyclif where Tyndale uses according to, and the simple as instead of the conjunctive according

4. Establishment of the preposition In the course of the fifteenth century, Chancery English, particularly the early statutes dating from the end of the century, probably played an important role in the establishment of the preposition according to. Of the c. two dozen14 instances of prepositional use in the Helsinki Corpus, no less than nine occur in the c. 11,000 word sample of the Statutes of the Realm·}5 (24)

dyverse Yomen of the Corone ... which doo nor endevor not them selfe in yevyng their Attendaunce accordyng to their Dutie; (,Statutes [STAT2] II 533)

Paths of loan-word grammaticalisation: the case of according to

(25)

257

the King ... hath determyned hym self to passe over the See in to his Realme of Fraunce and to reduce the possession therof by the seid gee to hym and his heires Kinges of Englond according to his rightfull title (Statutes [STAT2] II 549)

See also (16), (18) and (20) above. It is obvious that this re-emerging genre of English writing, which demanded great accuracy in indicating the relations between concepts, supported the grammaticalisation of Middle English loan word constructions and particularly their establishment in the developing Southern Standard (cf. Rissanen, 1999b and forthcoming). It is also worth noting that even in other than statutory and documentary texts, the context often relates to the language of law: paye ... aftyr his deceesse, acordyng to his wyll (Reynes, Commonplace Book [REYNES] 294); punysshid acordyng to hys trespaces (Reynard 54), etc. As shown above, however, the roots of the grammaticalisation of according to go back to a time preceding the emergence of Chancery English. Another typical context of the prepositional according to is in religious texts. The reference is mostly to sacred writings, prophecies, God's law, etc. It is perhaps not a mere coincidence that the earliest instance of the prepositional use of according to in the Helsinki Corpus occurs in the Northern Homily Cycle (c. 1400): (26)

Pan Saint Peter come him till And said, "Lord, if it war pi will, I waldpe ask a lityll thing Acordand to pis same spekeing. If my broper in worde or dede Ogains me breke his broperhede, If he sal forgifnes haue Seuyn sithes if he will craue? " CNorthern Homily Cycle [NHOM] II, 71)

In the Helsinki Corpus sample of the Cloud of Unknowing, which probably dates from the beginning of the 15th centuiy, according to occurs seven times (sample of c. 15,000 words). All instances combine according with be; it is obvious that the construction is not yet completely prepositionalised in instances of the following type: (27)

And 3ifpee pink pat pis maner of worching be not acordyng to pi disposicion in body & in soule, pou maist leue it, {Cloud of Unknowing [CLOUD] 129)

258

Matti Rissanen

In the late 15th-centuiy sample from the Sermo Die Lüne, by Richard Fitzjames, according to occurs five times (sample of c. 5,000 words). In these instances the construction is clearly prepositional; note, however, the form unto following accordyng in (28): (28)

It plesyd god sayth the apostle, by prechyng of his holy gospell & wysdom therw' in conteyned. as the worlde acountyd folyssnes. euerlastyngly to saue all tho. whiche therunto geue fast & stronge fayth. accordyng vnto the wordes of our sauyour. (Fitzjames [FITZJ] B4V)

See also (17) above. The Corpus of Early English Correspondence Sampler (CEECS) gives us evidence of the spread of the preposition from legal and religious contexts to more everyday and even speech-based usage.16 This use is common in letters dating from the second half of the fifteenth century, with no less than 47 occurrences (in a sample of c. 85,000 words). As is the case with most developments of grammaticalisation affecting borrowed elements, the final establishment takes place in Early Modern English. The number of the instances rises to 58 in the 1500-1570 period in the Helsinki Corpus (El) and to 77 in 1570-1640 (E2).17 It spreads to all genres of texts and to all kinds of context indicating agreement, suitability, consistency, or reference (cf OED s.v. according adv. 4-5). It is interesting to note how the simple verb accord becomes rarer at the same time. There are 46 instances18 of the use of the simple forms of the verb accord, in addition to more than twenty instances of according in non-prepositional uses, in the 1420-1500 sub-period of the Helsinki Corpus. In each of the two Early Modern sub-periods, there are only two instances of the simple forms and not a single one of the form according without a following to. CEECS gives evidence of this development in letters: in the period 1420-1450, there are 12 instances of the simple forms and only two of according to}9 In the period 1450-1500, the number of the instances of according to rises to 47, while there are only three of the simple accord.

5. Summary The prepositionalisation of according to illustrates the pattern of loanword grammaticalisation in Middle English. The word is borrowed with a fairly restricted meaning soon after the Conquest; the sense is rapidly gen-

Paths of loan-word grammaticalisation: the case of according to

259

eralised and becomes more abstract, with reference to both inanimate subjects and complements. The development of the preposition may have been supported by the need for more accurate and explicit expression of relations between various concepts, created by the genres of religious, scientific and statutory writing. Although we are dealing with a typical change from above, the evidence from private correspondence indicates that the new deverbal preposition was readily accepted in less formal and genrespecific contexts, until it becomes by far the most common, prototypical use of accord in the sixteenth century.

Notes 1. See, for instance, the now classic studies by Traugott on the development of while (1982) and the modal auxiliaries (1989), and many of the examples in Hopper and Traugott (1993). See also Rissanen (1999a) and the articles in Rissanen et al. (1997). 2. For a discussion of 'change from above' and 'change from below', see Traugott and Romaine (1985). 3. For the prototypes in historical lexicology, see, for instance, Geeraerts (1997). 4. See, e.g., Rissanen (1998, 1999b and forthcoming) for brief discussions of because, provided and except. 5. The abbreviated titles in square brackets refer to the Helsinki Corpus list of texts (Kytö 1996: 167-230). The list can also be found on the internet: http://khnt.hit.uib.no/icame/manuals/HC/ 6. The definitions in the Middle English Dictionary are on the same lines. 7. I have checked the entries in MED and OED, the Helsinki Corpus and the University of Michigan Corpus of Middle English Prose and Verse (http://www.hti.umich.edu/english/mideng/). 8. The MED (s.v. accorden 5-6) quotes a few other instances of the use of the verb with inanimate referents. 9. These examples are quoted by the MED, s.v. according 2(a). 10. MED quotes an example of accordaunt from Shoreham's Poems, dating from c. 1330: Acordaunt to thy trauayl, Lord, graunte me thy coroune (Shoreham 85/177) MED suggests that this word, which also occurs in Chaucer, is borrowed directly from Old French. It may be regarded as another forerunner of the prepositional use of according to. 11. There is one instance from Chaucer's Astrolabe, but the manuscript of this text, too, dates from the fifteenth century. There are, however, a number of instances of according in the Chaucer manuscripts dating from the end of the 14th century.

260

Matti Rissanen

12. For a discussion of various types of reanalysis, cf. e.g. Hopper and Traugott (1993: 40-41), and in reference to deverbal prepositions, Kortmann and König (1992: 685-687). 13. The simple prepositions after and be occur as equivalents of the later according to even in the West-Saxon Gospels dating from c. 1000.1 hope to return to the grammaticalisation of the conjunctive link according as in another context. This link seems to develop later than according to: there are two Helsinki Corpus instances in the Statutes dating from c. 1490 and one in Cely Letters, from 1482: telde me that he had brokyn to hys master acordyng as Afydwyntter desyryde hym, (Cely Letters [RCELY] 151) 14. In many cases, it is impossible, and unnecessary, to define whether the use is truly prepositional or whether traces of the verbal character of according still remain. 15. In the Anthology of Chancery English, included in the Michigan Middle English Corpus, the prepositional according to occurs 18 times in the text of c. 87,000 words. 16. The relationship between the earliest non-official letters and spoken expression, however, needs more research. - For an introduction to the Corpus of Early English Correspondence, see Nevalainen and Raumolin-Brunberg (1996) andNurmi (1998,1999). 17. The size of each of these sections of the Helsinki Corpus is c. 200,000 words. 18. It should be noted, however, that the use of the simple verb is both textually and contextually restricted even in this period. Of the 46 instances, 17 occur in a single deposition, and 13 in the copy of a document in the London Chronicle, in the phrase it is accorded. 19. Both instances occur in a letter of 1448 by John Shillingford.

References Geeraerts, Dirk 1997 Diachronic prototype semantics: a contribution to historical lexicology. Oxford: Clarendon Press. Hopper, Paul J. and Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Kortmann, Bemd 1996 Adverbial subordination. A typology and history of adverbial subordinators based on European languages. Berlin - New York: Mouton de Gruyter. Kortmann, Bernd and Ekkehard König 1992 Categorial reanalysis: the case of deverbal prepositions. Linguistics 30: 671-697.

Paths of loan-word grammaticalisation: the case of according to

261

Kytö, Meija (comp.) 1996 Manual to the diachronic part of the Helsinki Corpus of English Texts: Coding conventions and lists of source texts. 3rd edition. Helsinki: Department of English, University of Helsinki. Nevalainen, Terttu and Helena Raumolin-Brunberg (eds.) 1996 Sociolinguistics and language history: Studies based on the Corpus of Early English Correspondence. Amsterdam: Rodopi. Nurmi, Arja (ed.) 1998 Manual for the Corpus of Early English Correspondence Sampler. Helsinki: Department of English, University of Helsinki. Nurmi, Arja 1999 The Corpus of Early English Correspondence Sampler (CEECS). ICAME Journal 23. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman. Rissanen, Matti 1998 Towards an integrated view of the development of English: notes on causal linking. In: Jacek Fisiak and Marcin Krygier (eds), Advances in English historical linguistics (1996), 389-406. Berlin - New York: Mouton de Gruyter. 1999a On the adverbialization of RATHER: surfing for historical data. In: Hilde Hasselgärd and Signe Oksefjell (eds), Out of corpora: Studies in honour of Stig Johansson, 49-59. Amsterdam - Atlanta, Ga.: Rodopi. 1999b Language of law and the development of Standard English. In: Irma Taavitsainen, Gunnel Melchers and Päivi Pahta (eds.), Writing in Nonstandard English. 189-203. Amsterdam and Philadelphia: Benjamins. fthc. Standardisation and the language of early statutes. In: Laura Wright (ed.), The development of Standard English: theories, descriptions, conflicts. Cambridge: Cambridge University Press. Rissanen, Matti, Merja Kytö and Minna Palander-Collin (eds.) 1993 Early English in the computer age: Explorations through the Helsinki Corpus (Topics in English Linguistics 11). Berlin - New York: Mouton de Gruyter. Rissanen, Matti, Merja Kytö and Kirsi Heikkonen (eds.) 1997 Grammaticalization at work: Studies of long-term developments in English. Berlin - New York: Mouton de Gruyter.

262

Matti Rissanen

Traugott, Elizabeth Closs 1982 From propositional to textual and expressive meanings: Some semantic-pragmatic aspects of grammaticalization. In: Walter P. Lehmann and Yakof Malkiel (eds.), Perspectives on historical linguistics, 245-271. Amsterdam - Philadelphia: Benjamins. 1989 On the rise of epistemic meanings in English: an example of subjectification in semantic change. Language 65: 31-55. Traugott, Elizabeth Closs and Suzanne Romaine 1985 Some questions for the definition of 'style' in socio-historical linguistics. Folia Linguistica Historica 6: 7-39.

The third person present plural in Shakespeare's First Folio: A case of interaction of morphology and syntax? Herbert Schendl

1. Introduction The aim of the present paper is to investigate morphological variation in and syntactic constraints on the 3rd plural present indicative in Shakespeare's First Folio, in particular in regard to the distribution of the variant vs. the zero variant. As is generally known, the emerging Early Modern English standard variety still had a morphologically rather complex system of the present indicative with the following main features: while the 1st person singular is consistently unmarked (Type: I make), the second person is inflected with ~(e)st after the informal personal pronoun thou (Type: thou makest), but uninflected with the pronoun you/ye; in the third singular we find the well-described variation between ~(e)th and -(e)s (he makes vs. he maketh). In the present plural indicative, the 1st and 2nd persons are predominantly unmarked, though occasionally the same inflected forms as with the 3rd plural occur; the 3rd person plural, on the other hand, shows variation between up to four variants in regionally unmarked Early Modern English texts (for the distribution of these variants see Holmqvist 1922, Wyld 1936, Franz 1939, Schlauch 1959, Barber 1997, etc.): (i) - 0 (drink), (ii) ~(e)s (drinkes), (iii) -(e)th (drinketh), (iv) ~(e)n (drinken). By Shakespeare's time, the -(e)n variant had become a literary archaism, and the old southern form -(e)th was very rare (Barber 1997: 169ff). Of the remaining two variants, the zero variant clearly predominated, though the -(e)i variant was by no means rare. The three main hypotheses about the origin of the ~(e)s variant are that it is due (i) to Northern influence, (ii) to analogical extension from the 3rd singular, (iii) to phonological factors (for a brief discussion of these hypotheses see Schendl 1996: 146f.). In a previous paper (Schendl 1996), I have provided convincing arguments for the first hypothesis, namely that Northern influence must be a main source for the 3rd plural ~(e)s variant. Of particular importance for this view is the fact that Early Modern English standard texts show quite a number of instances where the distribution

264

Herbert Schendl

of the zero vs. the -(e)s variant (and in some cases even of plural ~{e)th) depends on the nature and position of the subject in a way which clearly mirrors the so-called 'personal pronoun rule' (Mcintosh 1989) or 'Northern Present Tense Rule' (NPTR, cf. Montgomery 1994). Since this is a highly complex syntactic rule (see section 2 below), it is extremely unlikely that it could have developed independently, i.e. without Northern influence, in the standard language. 2. The Northern Present Tense Rule The Northern Present Tense Rule has been described in detail for the Northern and some Midland varieties of Middle English as well as for Old and Middle Scots, though surprisingly it has been neglected in most general handbooks of Middle English as well as in the comprehensive treatments of Early Modern English (Franz 1939, Schlauch 1959, Görlach 1991, Barber 1997). The rule states that in the 3rd plural present indicative, the zero variant of the verb is only chosen if the subject is a personal pronoun in contact position with the finite verb, as in they eat; with all other types of subjects, however, the ~(e)s variant is found, i.e. we have constructions such as they regularly eats (no contact position between personal pronoun and verb); people (iregularly) eats (nominal subject, with and without contact position); the people who eats (relative pronoun subject), etc. In his detailed study of the diachronic decline of this rule from Old Scots to late Middle Scots, Montgomery (1994) subdivides the rule into two related sub-rules, namely (i) the 'Type of Subject Constraint', and (ii) the 'Proximity to Subject Constraint': the former states that the verb is marked with ~(e)s "if its subject is anything but an adjacent personal pronoun", the latter that the verb is also marked with ~(e)s if the "personal pronoun subject is not adjacent to the verb", as illustrated in they eat and drinks (Montgomery 1994: 86).1 The Early Modern English examples under (1) and (2) from Elizabeth I and Beaumont and Fletcher clearly reflect the working of the 'Type of Subject Constraint' within a single sentence: here the -(e)s variant occurs after the nominal subjects wicked men and Poets, while the verbs in contact position with the personal pronoun subject they have zero marking: (1)

for wicked men giues this good turn to dignitie, that they spot them with their own infection (Elizabeth, Boethius 50.23)

Third person present plural in Shakespeare's First Folio 265 (2)

Poets, when they rage, Turnes gods to men (Q 1619) (Beaumont & Fletcher, Maid's Tragedy Β. 1.10.1.9)

Without being aware of the Northern Present Tense Rule, some earlier researchers into Early Modern English had already noted that plural ~(e)s hardly ever occurred if the subject was a personal pronoun (cf. Knecht 1911, Holmqvist 1922), i.e. that it depended on the 'Type of Subject', though these observations remained widely unnoticed. The working of the second sub-rule, the 'Proximity to Subject Constraint', on the other hand, had never been observed in non-regional Early Modern English texts until recently. As late as 1993, we find the following statement: However widespread the STC [Subject Type Constraint] may have been in Early Modem British and Irish varieties [...] there is no evidence to date of the PSC [Proximity to Subject Constraint] in other than northern British English. (Montgomery et. al. 1993: 338)

In Schendl (1996) it could, however, be shown that there is at least a handful of unambiguous examples of the 'Proximity to Subject Constraint' from Early Modern English standard texts for which Northern influence can be definitely ruled out. Among these are the writings of Th. Deloney and Elizabeth I, from which the quotations under (3) and (4) are taken. In each of these examples the first verb, which is in contact position with the personal pronoun subject, has the zero variant, while the co-ordinated nonadjacent second verb has the -{e)s variant: (3)

whereby they make their pottage fat, and therewith driues out the rest with more content. (Deloney, Jack of Newbury 72)

(4)

For if neither they can doo that they promise & wantes greatest good (Elizabeth, Boethius 48.11)

An excellent example of the working of both sub-rules in a single passage is provided in the quotation under (5) from the anonymous play Cromwell. Here we find the zero variant after adjacent they {they owe, They know), while -(e)s occurs after non-adjacent they and nominal subjects {My debters, they...forsweares..., saies). (5)

All parts abroade where ever I haue beene Growes wearie of me, and denies me succour; My debters, they that should releeue my

266 Herbert Schendl want, Forsweares my monie, saies they owe me none: They know my state too meane to beare out law. (Anon., Cromwell IV. 1.5ff.) The Northern Present Tense Rule was evidently a variable rule with limited currency in Early Modern English (cf. Schendl 1996: 150) and was thus not followed consistently even by a single author within a single sentence. In example (6), also from Elizabeth's translation of Boethius, the first two verb forms follow the 'Type of Subject Constraint', i.e., they show ~{e)s after nominal subject and zero directly after personal pronoun subject, while the last verb violates the 'Proximity to Subject Constraint' rule, since it has the zero variant in spite of its distance from the personal pronoun subject. (6)

My maydes knowes their lady, with me they come, & whan I parte, giue [I]place. (Elizabeth, Boethius 23.18)

The above examples have provided clear evidence of the presence of both sub-rules of the Northern Present Tense Rule in the emerging Early Modern English standard language. The following section will investigate whether these rules also operate in the language of the First Folio. j

3. Shakespeare's use of the 3 person plural variants It has been known for a long time that Shakespeare's language, especially that of the First Folio, has a substantial number of instances of the ~(e)s variant (and a small number of ~(e)th forms) in cases where present-day English would have a plural present indicative form in zero (cf. Poliert 1881, Spekker 1881, Smith 1896, etc.). In many of these instances, the Quartos and the Second and Third Folios replaced the ~(e)s by the zero variants, and later editions frequently preferred the readings which corresponded to modern usage. In the early studies on the present indicative -(e)s forms in Shakespeare, authors such as Poliert (1881), Spekker (1881) and Smith (1896) predominantly interpreted these inflected forms as 3rd person singular present indicatives, which were explained by specific Early Modern English semantic or subject-verb agreement rules. Such an interpretation was particularly advanced in the following cases (cf. also Barber 1997: 170; however, we also find unambiguous instances of zero plurals in these types both in Shakespeare and his contemporaries):2

Third person present plural in Shakespeare's First Folio

267

(i) co-ordinated singular subjects referring to the same or a closely related referent, i.e. being notionally singular, cf. (7) - (9): (7)

Where death and danger dogges the heeles of worth. (AWWmAAl)

(8)

Whil 'st man and Master laughes my woes to scorne: (ERR II.2.208)

(9)

Both winde and tide stayes for this Gentleman, (ERR IV. 1.53)

An example of the zero variant with such co-ordinated subjects is given in

(10): (10)

her smoothnes; Her verie silence, and per [sic!] patience, Speake to the people, (AYL 1.3.81)

(ii) collective nouns, as illustrated under (11) and (12), though in (12) the co-ordinated second verb have is the unmarked plural form: (11)

The Senate hath sent about three severall Quests, To search you out. (OTHl.2.54)

(12)

The people knowes it, And have now receiv'd his accusations. (ΛΛΤΙΙΙ.6.24)

The same nouns are, however, also found with unambiguous plural forms, cf. (13) and (14): (13)

The senate, Coriolanus, are well pleased To make thee consul. (COR II.2.137)

(14)

The people love me, and the Sea is mine; (ANT II. 1.14)

(iii) nouns such as looks, pains, passions, etc., cf. (15) to (17): (15) (16)

Her Looks doth argue her replete with Modesty, (3H6 III.2.100) And all my paines is sorted to no proofe. (SHR IV.3.48)

(17)

Beshrew me, but his passions moves me so, (3H61.4.160)

Here again the same nouns occur with unambiguous verbal plural forms, cf. (18) to (20):

268

Herbert Schendl

(18)

Heere comes the Queene, Whose Lookes bewray her anger. (3H61.1.240)

(19)

'Tis time to speake, My paines are quite forgot. (R31.3.122)

(20)

How all the other passions fleet to ayre, As doubtfull thoughts, {MV 3.2.114)

(iv) numerals as in (21) and (22), as well as certain pronouns, such as many (cf. ex. (30)), some, etc.: (21)

Two of both kindes makes up foure. (MND III.2.466)

(22)

looke where three farthings goes, (JN 1.1.151)

(v) preposed verbs, i.e., the construction VS, as illustrated in (23) and (24), though here again the zero variant occurs as well, cf. (25): (23)

Here comes Brabantio, and the Valiant Moore. (OTH 1.3.57)

(24)

Here comes the Townes-men, on Procession, (2H6 II. 1.83)

(25)

heere come our Wives. (1H4 III. 1.189)

The factors and types illustrated above may justify the interpretation of ~(e)s as singular forms in some cases, but such an interpretation seems less convincing or even doubtful in numerous other instances. Even supporters of an interpretation of the ~{e)s forms as singulars like Poliert (1881) have to admit that there are numerous doubtful instances, where none of the above explanations does apply. Today there seems to be wide agreement that at least such 'doubtful' cases are clear examples of plural -(e)s, i.e. the existence of an inflected 3rd plural present variant in -(e)s has been widely accepted for the Early Modern English standard language in general and for Shakespeare's language in particular (cf. Barber 1997: 170), though it still has not been given the attention it would deserve. Taking the existence of plural -(e)s in Shakespeare's language as a given fact, the following section will try to provide a systematic quantification of both the -(e)s and the zero variant, as well as investigate the distribution of the inflected variant in relation to the form and position of the subject.3 Pollert's dissertation from 1881 already provides a complete list of -{e)s verb forms after 'plural' subjects (with the above discussed qualification), so that a new study of this phenomenon may at first glance seem superfluous. However, neither Poliert nor any of the other studies mentioned above

Third person present plural in Shakespeare's First Folio

269

provide any systematic information on the type of subject (noun, personal pronoun, etc.) or on the distance between subject and verb; they equally lack any information on the relative frequencies of the inflected forms in relation to the zero plural verb forms - information which is of paramount importance for a qualified analysis of Shakespeare's verb morphology. Any quantification of a potentially ambiguous linguistic item faces the problem of how to deal with ambiguous cases and there are clearly two possibilities in our case. The first is to consider only those ~{e)s forms which cannot be explained as singular forms on the basis of 'rules' such as those under (i) to (v) above, i.e., those cases for which no singular explanation can be found. These amount to about 100 instances according to Poliert (1881) and Smith (1896). However, we cannot hope to arrive at any agreement on where a singular interpretation of the subject might apply, especially so in those cases where the same lexeme in subject position shows variation of the verb form in the corpus - a fact which may be due to variation in semantic interpretation, in syntax (concord), or in morphology. Since such a procedure would yield completely arbitrary figures, which would differ from one scholar to another, it has been rejected for the present study. The second possibility, which is the one chosen here, is to count all those ~(e)s forms as plurals in which modern English would have a plural, i.e. a zero form after the respective subject.4 Though this does evidently not take certain Early Modern English agreement rules and specific semantic restrictions into account, it seems the only viable solution for a linguistic phenomenon in which variation of both verb morphology and noun semantics is involved. 3.1.

Quantitative analysis

The following statistical analysis is restricted to the 36 plays of the First Folio and does not include the high-frequency auxiliaries have and be, which have, however, been analysed in regard to the possible working of the Northern Present Tense Rule. With these restrictions, the analysis has yielded the following results: The total number of 3rd present plural indicative verb forms in the First Folio amounts to 4106, of which 303 instances, i.e. about 7%, are -(e)s forms. These figures include, however, 1437 instances in the stage directions, which will be excluded from further discussion for a number of reasons: firstly, in 93% of the stage directions the verb precedes the plural subject (s) as against less than 5% of preceding verbs in the texts themselves, a fact which would clearly distort the overall results; secondly, the stage directions predominantly represent a few stereotypical patterns with

270

Herbert Schendl

only 5 instances of inflected verbs; finally, the stage directions have to be seen as separate (sub-)texts, possibly by (a) different authors). This leaves 2669 instances of plural verbs in the texts themselves, of which 89% represent the zero variant, while a still surprisingly high percentage of about 11% shows the inflected variant (1,5% of these are in a construction with preposed verbs, while in 9,5% the verb follows the subject). If we subtract the 623 instances of plural verbs with a personal pronoun subject (see next paragraph), the percentage of inflected plural verbs rises even to almost 15%. Concerning the type of subject combined with these plural variants, not a single one of the 294 plural ~(e)s forms occurs directly after a personal pronoun subject, in other words, the ~(e)s variant is only used after nominal subjects and other pronominal subjects, a fact which clearly mirrors the 'Type of Subject Constraint' discussed above. The working of this constraint is particularly evident in a number of cases where, in one and the same sentence, the personal pronoun subject they is directly followed by the zero suffix verb, while a following or preceding co-referential nominal or other pronominal subject combines with the ~(e)s variant, cf. the examples under (26) to (30), as well as example (37) below : (26)

Oh Gertrude, Gertrude, When sorrowes comes, they come not single spies, But in Battaliaes. (Ä4A/IV.5.71)

(27)

and on the answer, he payes you as surely, as your feete hits the ground they step on. (77VTII.4.276)

(28)

they laugh, that winnes. (OTH IV.1.141)

(29)

Well know they what they speake, that speakes so wisely. (TRO III.2.152)

(30)

Thereby to see the Minutes how they runne: How many makes the Houre full compleate, How many Houres brings about the Day, (3H6 II.5.265)

However, there are also a number of examples where both the verb adjacent to they and the second verb with co-referential other subject show the zero suffix, cf. (31): (31)

For they breath truth, that breath their words in paine. (R2II. 1.10)

Third person present plural in Shakespeare 's First Folio

271

Among the other types of subject, relative pronoun subjects show about 15% of inflected variants and are thus slightly above the overall percentage of 11%. However, a very clear and significant increase of inflected verbs to 42% is found after those heavy noun-phrases which consist of coordinated subject NPs. Though part of these co-ordinated subject nouns are near-synonyms and/or refer to similar concepts as illustrated above under (i) and below in (32) to (34), this is by no means always the case, as the instances under (35) to (37) illustrate. With all due caution, it seems that the heaviness of the NP is a factor in the choice of the inflected plural verb form. (32)

Death and Destruction dogges thee at thy heeles, (R3 IV. 1.50)

(33)

Which heaven andfortune still rewards with plagues. (TGV IV.3.35)

(34)

Hostilitie, and civill tumult reignes, Betweene my conscience, and my Cosins death. (JNYM.llSl)

(35)

Oh Cymbeline, Heaven and my Conscience knowes Thou didd'st unjustly banish me: (CKMIII.3.108)

(36)

But see where Somerset and Clarence comes: (3H6 IV.2.3)

(37)

Menacrates and Menas, famous Pyrates Makes the Sea serve them, which they eare and wound With keeles of every kinde. {ANT 1.4.56)

In a number of cases, the inflected variant provides the rhyme, a fact which may have been at least partly responsible for its choice, cf. (38) to (39): (38)

I know a banke where the wilde time blowes, Where Oxslips and the nodding Violet growes, (MND II.2.259)

(39)

Whiles I threat, he lives: Words to the heat of deedes too cold breath gives. {MAC II. 1.73)

The second sub-rule of the Northern Present Tense Rule, the 'Proximity to Subject Constraint' in its narrow form, i.e. the construction 'they V and Vs', however, does not occur at all in the First Folio. All instances of coordinated verbs after personal pronoun subject are of the form 'they (Χ) V and V', cf. examples (40) to (44): (40)

As in a Theater, whence they gape and point At your industrious Scenes and acts of death (JN II. 1.399)

272 Herbert Schendl (41)

Marry sir, they praise me, and make an asse of me, now my foes tell me plainly, I am an Asse: (TN V. 1.19)

(42)

And in their rage,... They whurle a-sunder, and dismember mee. (JZV III. 1.342)

(43)

Menacrates and Menas, famous Pyrates Makes the Sea serve them, which they eare and wound With keeles of every kinde. (ANT 1.4.56)

(44)

That they devoure their reason, and scarce thinke Their eies doe offices of Truth: (TMP V. 1.180)

In a few cases, however, a nominal subject is followed by a pair of coordinated verbs whose first member has the zero variant, while the second takes the inflected -(e)s variant, cf. (45) and (46): (45)

And these dread curses like the Sunne 'gainst glasse, (...), recoile, And turnes the force of them upon thy seife. (2H6 III.2.356)

(46)

Our doubts are traitors And makes us loose the good we oft might win, (MMIAM)

This type is not restricted to Shakespeare, but is also found in other Early Modern English authors, as the following examples illustrate: (47)

And not as flowers by the brightest rain Which grow up quickly and as quickly fades? (G. Peele, David and Bethsabe, quoted in Knecht, 1911: 65)

(48)

the Tartars and the Eastern theeues...Presume a bickering with your Emperour, And thinks to rouse us. (Marlowe, Tamburlaine /, 920)

(49)

Hel and confusion light upon their heads, That with such treason seek our ouerthrow, and cares so little. (Marlowe, Tamburlaine II, 2890)

Such cases have been explained in Schendl (1996: 151) as "overgeneralisations due to a reanalysis of the NPTR [Northern Present Tense Rule] in a subset of syntactic environments", i.e., as the result of a new variable rule marking the second of two co-ordinated present plural verbs with (e)s. This explanation rests on the attested occurrence of the Northern Present Tense Rule in the emerging standard language, and seems not unlikely

Third person present plural in Shakespeare 's First Folio

273

in view of the fact that the original rule was never fully accepted into standard English.

4. Conclusion As the above discussion has shown, the choice of a particular variant in the 3rd plural present indicative in Shakespeare's First Folio is at least to some extent influenced by the nature of the subject: (i) The 'Type of Subject Constraint' has been found to be a categorical rule in the case of personal pronoun subjects, which exclusively combine with the zero variant; with the other types of subject, however, the 'Type of Subject Constraint' is clearly a variable rule - a phenomenon which Shakespeare's language shares with that of a number of his contemporaries. With 11% or 15% respectively, the -(e)s variant is relatively frequent in Shakespeare, even though some of these instances may be due to semantic considerations and specific subject-verb agreement rules. Among the nominal subjects, those heavy NPs which consist of co-ordinated nouns have a significantly higher frequency of the inflected variant. (ii) No instance of the 'Proximity to Subject Constraint' with personal pronoun subjects occurs in the corpus, though, as shown in Schendl (1996), this rule is - though rarely - also attested in the emerging standard language. Summing up we can say that Shakespeare's First Folio shows some interaction between morphology and syntax in the case of the 3rd plural present indicative, in a way which mirrors the 'Type of Subject Constraint'. However, the 'Proximity to Subject Constraint', which would be proof of direct northern influence, is not attested in our corpus.

Notes 1. In recent years the Northern Present Tense Rule has increasingly attracted the interest of sociolinguists working on Black English Vernacular (or AfroAmerican Vernacular English) or early and non-standard American English, especially as an argument in the still on-going dispute on the origin of Black English Vernacular (cf. Bailey et al. 1989, Montgomery et al. 1993). 2. These and additional types in Shakespeare's language are extensively illustrated in Poliert (1881: 13-30) and Spekker (1881: 36-55); cf. also Knecht (1911: 15-48), Barber (1997: 169f.). 3. The few cases with plural -(e)th will also be included here under the cover term 'plural -(e)s' since this type of variation does not affect the topic of the present study.

274 4.

Herbert Schendl I am much indebted to Aline Längle, Sandra Edlinger and Werner Trnka for their help with the collection of the data.

References Porter, Charlotte and Helen A. Clarke (eds.) 1906 The complete works of William Shakespeare. Reprinted from the First Folio. 12 volumes. London: Harrap. Bailey, Guy, Natalie Maynor and Patricia Cukor-Avila 1989 Variation in subject-verb concord in Early Modern English. Language Variation and Change 1: 285-300. Bambas, Rudolph C. 1947 Verb forms in -S and -TH in Early Modern English prose. Journal of English and Germanic Philology 46: 183-187. Barber, Charles 1997 Early Modern English. 2nd edition. Edinburgh: Edinburgh University Press. Benskin, Michael 1992 Some new perspectives on the origins of Standard written English. In: Jan A. van Leuvensteijn and Jan B. Berns (eds.), Dialect and standard language in the English, Dutch, German and Norwegian language areas, 71-105. Amsterdam: North-Holland. Blake, Norman (ed.) 1992 The Cambridge history of the English language, Volume 2: 10661476. Cambridge: Cambridge University Press. Franz, Wilhelm 1939 Die Sprache Shakespeares in Vers und Prosa. ShakespeareGrammatik 4th ed. Halle: Niemeyer. Görlach, Manfred 1991 Introduction to Early Modern English. Cambridge: Cambridge University Press. Holmqvist, Erik 1922 On the history of the English present inflections particularly -th and -s. Heidelberg: Winter. Knecht, Jacob 1911 Die Kongruenz zwischen Subjekt und Prädikat und die 3. Person Pluralis auf -s im Elisabethanischen Englisch. Heidelberg: Winter. Kytö, Merja 1993 Third-person present singular verb inflection in Early British and American English. Language Variation and Change 5: 113-139.

Third person present plural in Shakespeare's First Folio

275

Mcintosh, Angus 1989 (1983) Present indicative plural forms in the later Middle English of the North Midlands. In: Angus Mcintosh, Michael Samuels and Margaret Laing (eds.), Middle English dialectology. Essays on some principles and problems, 116-122. Aberdeen: Aberdeen University Press. Montgomery, Michael 1994 The evolution of verb concord in Scots. In: Alexander Fenton and Donald A. MacDonald (eds.) Studies in Scots and Gaelic. Proceedings of the third international conference on the languages of Scotland, 81-95. Edinburgh: Canongate Academic. Montgomery, Michael, Janet M. Fuller and Sharon DeMarse 1993 'The black men has wives and sweet harts [and third person plural -5] jest like the white men': Evidence for verbal -s from written documents on 19A-century African-American speech. Lang»age Variation and Change 5: 535-557. Poliert, Karl 1881 Die 3. Person Pluralis auf s bei Shakespeare. Ph.Diss. Marburg. Reid, Wallis 1991 Verb and noun number in English: A functional explanation. London: Longman. Rodeffer, John D. 1903 The inflection of the English present plural indicative with special reference to the northern dialect. Baltimore: Murphy. Schendl, Herbert 1996 The 3rd plural present indicative in Early Modern English - variation and linguistic contact. In: Derek Britton (ed.), English historical linguistics 1994. Papers from the eighth international conference on English Historical Linguistics, 143-160. (Current Issues in Linguistic Theory 135.) Amsterdam: Benjamins. Schlauch, Margaret 1959 The English language in modern times (since 1400). Warszawa: Panstwowe Wydawnictwo Naukowe. Schneider, Edgar W. 1983 The origin of the verbal -s in Black English. American Speech 58: 99-113. Smith, C. Alphonso 1896 Shakespeare's present indicative s-endings with plural subjects. Publications of the Modern Language Association 11: 362-376. Spekker, Statius 1881 Über die Kongruenz des Subjekts und des Prädikats in der Sprache Shakespeares. Ph.Diss. Bremen.

276

Herbert Schendl

von Staden, Wilhelm 1903 Entwicklung der Präsens-Indikativ-Endungen im Englischen. Ph.Diss. Rostock. Stein, Dieter 1987 On the crossroads of philology, linguistics, and semiotics: Notes on the replacement of th by s in the third person singular in English. English Studies 68: 406-432. Stoelke, Hans 1916 Die Inkongruenz zwischen Subjekt und Prädikat im Englischen und in den verwandten Sprachen. Heidelberg: Winter. Wyld, Henry C. 1936 [1953] A history of modern colloquial English. 3rd edition. Oxford: Blackwell.

The function of word-formation and the case of English -cumGabriele Stein

The present paper consists of two parts: Part I illustrates the use of the element -cum- in present-day English. A new interpretation of its grammatical and lexical status is attempted on the basis of its use and characterisation in leading English dictionaries. Part II then explores how the lexical status suggested for -cum- might be integrated into a theoretical framework for modern English word-formation which will call for a reconsideration of the function of word-formation. 1. -cum- in present-day English As a lexicologist interested in neologisms in English I have in recent years repeatedly come across a type of word-combination that had not been familiar to me: the use of the element -cum- linking two words. Once alerted to this use, I started collecting occurrences from my reading. Here are examples encountered in the novels by five different authors, Amanda Brookfield, David Lodge, Mordecai Richler, Jane Rogers, and Andrew Rosenheim, where the combination in question has been highlighted: Carmen, their recently acquired Hispanic maid-cum-child minder, had been treated to a frilly black and white outfit in honour of the occasion, together with several torturous sessions on how to receive coats and flowers. (Brookfield 1997: 15f.) While doing a lot in terms of acquainting their father with the delights of caring for small infants, it was not so helpful where the marriage itself was concerned. It often seemed to Anita that the two of them now sat on opposite ends of an unstoppable see-saw of resentment and envy, typically characterised by spiralling arguments-cum-discussions as to whose suffering was the greatest. (Brookfield 1997: 47)

278

Gabriele Stein

They arrived at the pirate ship just as a cluster of whining children were being herded back out into the street. Luke climbed to the top of the metal laddercum-mast and swung himself out to one side, holding on with only one arm. (Brookfield 1997: 64) Putting a new sit in the sitcom, you might say. "But Tubby doesn't seem to be able to concentrate his mind on the problem", said Hal. "So we think he needs a script editor", said Ollie. "A kind of minder-cum-dramaturge", said Hal. (Lodge 1995: 182) When I climbed into the foothills of the Pyrenees the weather turned cooler, and it was raining when I reached St Jean Pied-de-Port (St John at the Foot of the Pass). It's a pleasant little market-cum-resort town of red gabled roofs and rushing brooks, nestling in a lumpy patchwork quilt of fields in various shades of green. (Lodge 1995: 288) In the bar-cum-cafi next to the church, a homely place of bare boards and refectory tables, I got into conversation with an elderly Dutch cyclist who claimed he had met an English pilgrim called Maureen in a refugio near Leön a week before. (Lodge 1995: 293) A neighbourhood that had elected the only Communist (Fred Rose) ever to serve as a member of parliament, produced a couple of decent club fighters (Louis Alter, Maxie Berger), the obligatory number of doctors and dentists, a celebrated gambler-cum-casino owner, more cut-throat lawyers than needed, sundry school teachers and shmata millionaires, a few rabbis, and at least one suspected murderer. (Richler 1998: 4) En route, Mendel, his name laundered to Matthew/Morrow, acquired a fourteen-room apartment on Park Avenue, serviced by a maid, a cook, a butlercum-chauffeur, and an English governess off the Old Kent Road for his firstborn son, Boogie, [...] (Richler 1998: 9-10) I tracked down and wrote or phoned mutual friends in Paris and Chicago and Dublin and, you know, that artsy desert pueblo-cum-Hollywood-shtetl in Arizona, favoured by short producers in cowboy boots, [...] (Richler 1998: 41) John was a brilliant criminal lawyer until he was undone by two costly alimony settlements and a deadly mixture of booze and irreverence. Defending a notorious swindler-cum-lounge lizard some years back, a man charged with the sexual assault of a woman he had picked up in the Esquire Show Bar, John made the mistake / of going out to a long liquid lunch at Delmo's [...] (Richler 1998: 81-82)]

The function of word-formation and the case of English -cum-

279

That inconsiderate bastard has died on me. Heart attack. He was en route to a reading-cum-book-signing at McGill when he went into a spin on Sherbrooke Street and collapsed on the sidewalk, clutching his chest. (Richler 1998: 369) "I am determined to dedicate the rest of my life to saving this country, which has been so good to me", said the inside trader, currency speculator, propertyflipper, hostile takeover bidder-cum-asset-stripper, tax evader, [...] (Richler 1998: 370) The weekend continued as it had begun. Very late that night they booked themselves into an empty little pub-cum-hoteh in Chapelmoor, a village twenty miles east of town. (Rogers 1986: 242) He combined a high level of genius with extreme personal dishevelment, and this accorded nicely with Madison's conception of what an erudite grammarian-cum-lexicographer should be. (Rosenheim 1993: 7)

Since this use of -cum- is neither described in any of the standard reference works on English word-formation nor in the books on recent neologisms, I turned for help to modern English dictionaries to see whether or not the item is listed and if so, as what kind of linguistic element. The OED describes cum as a Latin preposition with the meaning 'with, together with', whose use in English originated in Latin documents specifying the local names of combined parishes and benefices, as in Chorltoncum-Hardy, Stow-cum-Quy. At the end of its chronological account the OED writes that cum is "Freq. used as a combining word to indicate a dual nature or function". If we ignore the quotations provided for place names, the use of cum as a 'combining word' linking other words dates from the beginning of the 20th century. Four examples are provided, beginning with a quotation from R. Kipling: (1)

a. b. c. d.

1913: Easy motor-bike-cum-side-car trips round London 1939: [...] aphilosophic-cum-economic tinge 1959: Three short... dinner-cum-cocktail dresses 1959: The atmosphere of laboratory-cum-work shop

One of the weaknesses of the historical treatment in the OED is that it leaves us at a loss with respect to the grammatical function of cum in present-day English. Its synchronic use makes it a linguistic element in English, and the description as a 'combining word' ignores the part of speech feature so characteristic of all words in English (and, as is well known, many English words have multiple grammatical determination). Is it a preposition? Example (lb), α philosophic-cum-economic tinge, where cum

280 Gabriele Stein joins two adjectives, rules out an interpretation as a preposition. A semantic paraphrase would involve the conjunction and: a philosophic and economic tinge. Another parallel combined expression in modern English would be an o-form as in political and economic -» politico-economic. In the first example adduced by the OED (la), motor-bike-cum-side-car trips, cum could be regarded as a preposition corresponding to with. Both examples (la) and (lc), motor-bike-cum-side-car trips and dinner-cumcocktail dresses, would also be paraphrased by and and the normal expression of (la) nowadays is motor-bike and side-car. In addition, ellipsis in hyponymic compounds (as in example (lc)) is only acceptable when both compounds have the same determinatum and are linked by the conjunction and or or: Examples are: (2)

a front and back entrance; the pre- and post-war housing situation; try to switch it on or off Compare German: Wein- und!oder Sektgläser; ein- oder aussteigen

The more recent Oxford University Press product, the New shorter Oxford English dictionary on historical principles (NSOED) of 1993, no longer describes cum as Latin, but still as a preposition and the sense in question as 'And also: denoting a dual or combined nature or function'. In addition to the Kipling quotation, there is one with a repetition of cum: "He was the poet cum priest cum doctor". Delia Thompson's 9th edition of the Concise Oxford dictionary of current English of 1995 goes one step further. Cum is described as a preposition "usu. in comb? and its meaning given as 'with, combined with, also used as'. The example is a bedroom-cum-study. So much for the dictionary family published by Oxford University Press. Let us turn to Webster's third new international dictionary (W3). Cum is also classified as a preposition with two senses. For the sense 'with, combined with' an example from Newsweek is quoted: the entertainment-cum-profit motive. The sense 'including, along with' is illustrated with an example from the New Yorker: the members of this orchestra-cumballet, besides the instance a house-cum-farm. Some of the quotations listed in the OED and Wj have newspapers and news magazines as their sources and thus document that the use of cum is not restricted to fiction, as might have been suggested by my examples. The only dictionary that lists cum in the form in which it actually occurs, -cum-, is Collins Cobuild English language dictionary of 1987. The entry reads as follows:

The fiinction of word-formation and the case of English -cum-

281

-cum- is placed between two nouns to form a compound noun describing something or someone as being partly one thing and partly another. EG The light in the barn-cum-garage was on sort of teacher-cum-lexicographer.

a dining-cum-living room

a

The Cobuild dictionary leaves the linguistic element grammatically undescribed. This also holds for the new 1995 edition. The lemma form suggests a bound-morpheme status. It does not function as a prefix, nor is it used like a suffix. It cannot be described as a combining element because the Cobuild dictionary defines its label COMB as follows: In this dictionary COMB is used in the grammar notes beside entries to mean a combining word. This is a use of a word which only occurs in combinations with other words, like -legged in bandy-legged and bow-legged. The Cobuild lexicographers thus assume that legged is a word which could not occur on its own, just as -bearing, -bedded, also described as combining forms because of their occurrence in such instances as oilbearing rocky oxygen-bearing water, a twin-bedded room. Within word-formation theory the term 'combining form' is the one which is least clear and most in need of a consistent definition which outlines its relation to other bound morphemes in English, prefixes and suffixes in particular. The issue deserves a full and in-depth study which we cannot provide here (but see already in this context Stein (1977), Bauer (1979), and Warren (1990)). For our purpose, the linguistic characterisation of -cum- in modern English, we shall assume that -cum- is a bound lexical morpheme which has a distinct form and meaning of its own. Since its use is restricted to linking two nouns or adjectives and since it cannot occur in this function on its own, typographically reflected in the two hyphens, -cum- cannot be interpreted as a free coordinating conjunction. From a formal and grammatical point of view it can be clearly distinguished from prefixes and suffixes. The latter typically have a word-class determination in English: an English speaker knows that -dom, for instance, derives nouns, -less adjectives, and -ize verbs. The element -cum- has no part-of-speech determination. Prefixes and suffixes, in addition, have specific grammatical selection features: -less, for instance, will attach to nouns and verbs, e.g. fenceless, quenchless, but not to adjectives, e.g. *niceless. The grammatical bases selected by the prefix re- in English are nouns and verbs, e.g. reanalysis, to reorganize. By contrast, -cum- has a selectional feature which is bi-directional: the two bases which it links belong to the same

282

Gabriele Stein

word class. The examples provided suggest a predominant pattern n-cum«, adj-cum-adj seems to be less common. There seem to be further constraints or semantic selection features triggered by the meaning of -cum-. The 25 instances compiled lend themselves to the following interpretation: nouns 1. 'place'-CM/M- 'place': e.g. bar-cum-cafe - 9 instances 2. 'person/group of people'-cw/w- 'person/group of people', e.g. maid-cum-childminder, orchestra-cum-ballet - 9 instances 3. 'concrete object'-cum- 'concrete object': e.g. ladder-cum-mast 3 instances 4. 'abstract noun/action'-cum- 'abstract noun/action': e.g. arguments-cum-discussions - 3 instances adjectives The only example from the OED combines two relational adjectives: philosophic-cum-economic. The meaning descriptions given by the dictionaries consulted are similar, but some seem to be more appropriate than others: For the OED ("to indicate a dual nature or function") and the Cobuild dictionary ("being partly one thing and partly another") the two members of the combination are apparently on the same level. This calls for a differentiation of similar combinations in English: nouns 1. person + person: girl friend, owner-occupier 2. object + object: fighter-bomber, oak tree adjectives 1. relational adj + relational adj: auditory-visual, moral-philosophical 2. -o-form + full form: logico-semantic, socio-political

The function of word-formation and the case of English -cum-

283

3. colour + colour: blue-black, grey-green, bluish-grey, whitish-yellow, red-green (colour blindness) 4. other: bitter-sweet, deaf-mute, icy-cold, red-hot nouns!adjectives 1. full form + full form: 2. -o-form + full form:

German-American, Italian-Swiss Italo-American

The semantic relationship between the two members A and Β of these combinations is either a copulative one, A and Β are genuinely coordinated as in fighter-bomber, auditory-visual, deaf-mute, or it is a subordinating one, as in girlfriend, blue-black, red-hot, Italian-Swiss, Italo-American where the B-element is further specified, classified by the Α-element, e.g. a friend who is a girl, not a boy, a black colour that tends towards blue, not brown, an American of Italian extraction, not Spanish origin, etc. In contrast, the relation between the Α-element and the B-element in the -cMw-combinations is one where Β is subordinated to A, a butler-cumchauffeur is a butler who also performs chauffeur duties rather than a chauffeur with butler functions, and a pub-cum-hotel is a pub which has also accommodation rather than a hotel which is also a pub. The B-element is subsidiary, has an additional, extra function or role with respect to the Α-element. The semantic description "and also" given by the NSOED captures this quite well (at least better than its full characterisation "denoting a dual or combined nature or function", which would seem to put A and Β on the same level. Wj gives two senses: 'including, along with' is an adequate paraphrase. The second sense 'with, combined with' relates to abstract nouns linked by means of -cum- which are not described in any of the other dictionaries. So far we have seen that -cum- compounds differ in form, grammar and meaning from similar 'juxtaposed' noun-noun-compounds and adjectiveadjective compounds. The last question we have to tackle is the function of this type of compound formation. Apparently, all the compounds under consideration were coined because, in a specific situation of communication, English speakers sensed that the combination corresponded most adequately with what they wanted to say. The linguistic choice they had was between an ordinary syntactic expression or a word-formational one. They opted for the latter, "elevating" the items linked together to a single unit of reference. The striking

284

Gabriele Stein

difference between the nouns and the adjectives joined together is that in juxtapositions like fighter-bomber or deaf-mute the entity referred to singles out an object, a relation, a quality, etc. which, in that particular combination is well attested in the extralinguistic world so that the object, the relation, the quality, etc. could be deemed worth naming. In the case of -cwm-compounds, however, there is no such familiar occurrence. Instead, there is an extralinguistic situation where the referents of A- and the Belements happen to be present, therefore the naming is only a temporary one. The temporary aspect seems also to be reflected in the very order of the elements joined: the grammatical and semantic determinatum of -cwm-compounds is the Α-element which is contrary to English wordformation (but cf. Ungerer and Schmid, 1996). The whole unit thus resembles more a syntactic combination with the normal progression of endfocus on the B-element. From this point of view one might argue that -cwm-compounds have a linguistic position between syntactic structures and "traditional" word-formations. Another indication of the temporariness of the naming is the fact that -cwm-compounds are not listed in English dictionaries. This characterisation of -cwm-compounds also accounts for their stylistic effects and exploitation. Many of them strike one as deliberate, often humorous or slightly ironic, creations or tours de force by their coiners. 2. Implications for the function of word-formation The description given for -cwm-compounds in present-day English challenges contemporary definitions of word-formation which in general maintain that it is the task of the linguistic discipline to study the patterns by means of which "new words" are coined. Compounds with -cum- are ad-hoc-formations, but a specific type: the combinations are of temporary relevance only, with no implications of institutionalisation. There are thus new coinages, unlikely from the outset to be candidates for institutionalisation or lexicalisation. If such temporary relevance is to be integrated into a consistent functional description, these "traditional" characteristics of the "new word" will have to be re-examined though not necessarily abandoned. The field of word-formation suffers from the same weaknesses that most linguistic disciplines have: there is no consistent definition and delimitation of the field itself, basic concepts are not clarified or are intrinsically difficult to describe.

The function of word-formation and the case ofEnglish -cum-

285

Scholars have suggested various functions of word-formation. These suggestions are not wholly satisfactory for a number of reasons. First of all, it is not always clear to which sense of the multiply ambiguous word 'word-formation' these functions are attributed. Secondly, they do not seem to derive from a coherent principle. That is, a specification of one single functional principle from which others logically follow is preferable to an "unprincipled", heterogeneous and unrelated assembly of functions. And thirdly, the lack of a unifying principle is also responsible for a mixing of various levels. Let us therefore first of all consider the English term 'word-formation'. In the literature, one has usually distinguished three senses: (I) wordformation is a linguistic field of study (cf. German Wortbildmgslehre), (II) word-formation denotes a process of coining new words, a new word, and (III) a word-formation is a newly coined word, thus the result of the coining process. I wonder whether it would not be more appropriate to distinguish four senses by eliminating the ambiguity in (II) between the general and the individual: word-formation as the linguistic field of study; word-formation as a general linguistic technique or device, just as sentence formation; word-formation as the individual process of coining a word; word-formation as the result of such an individual process. We are now in a position to specify exactly for which sense a description of the function of word-formation is attempted. A delimitation of sense (I) would provide us with a description of the task(s) of wordformation research. It is sense (II) which calls for a functional description. In an attempt to introduce more consistency into word-formation theory I suggested in 1977 that patterned metonymic shifts should become an integral part of the field of word-formation since many coinings of words operated on specific senses of words, not always on all the senses of a particular word. That is, the formative basis was the lexical unit in the sense of Cruse (1986: 76). To relegate patterned semantic shifts into the field of historical semantics was not sustainable. In consequence, I also provided a modified definition of the field of word-formation which would then study all the patterned changes of existing words. Thus all the rulegoverned changes in the lexicon of a language would be investigated by one linguistic discipline, not several. My proposal is that it is the function of the linguistic technique or device to enable speakers of the language to perform patterned lexical operations on existing lexical units just as the linguistic technique of sentence

286

Gabriele Stein

formation enables them to perform patterned syntactic operations on existing lexical items. A lexical operation consists in providing a name for a meaning unit. Nothing is said about the transient nature or the permanence of the conceived content itself. The important point is that by providing a name which carries with it the status of a grammatical and semantic unit the entity so invested assumes the ordinary functions of content words: it names and it is subject to hypostatisation. These are common characteristics of words and therefore not particular "functions of word-formation". Since all common words fill a naming gap, this latter characteristic can also not be singled out as a word-formation function. Similar considerations hold for the "classifying" function. We are left with three further functions of word-formation that have been suggested: information condensation, syntactic recategorisation and pronominalisation. Insofar as patterned lexical operations are based on existing lexical units, the encapsulation of the old existing entity plus the additional one in one single name constitutes an information condensation, above all if contrasted with full syntactic structures. This may be a matter of degree because one might argue that a compound like letterbox is more condensed than bar-cumcafe. Syntactic recategorisation refers to such cases where for textual reasons a previous occurrence of a verb form of to arrive, for instance, is reshaped as arrival, or the being close of something as closeness. Although the grammatical recategorisation dominates, I would be happier with 'grammatical-semantic recategorisation'. Grammar cannot be separated from meaning. And there are other cases which have not been discussed but would also qualify for a function: semantic recategorisation as evidenced by such examples as stream - streamlet (diminutives), yuppie yuppiedom (collective nouns), open - reopen (prefixations), or the metonyms house (a building) - house (the people living in it), cup (container) cup (amount held by it). All these cases could be regarded as 'grammatical-semantic recategorisations' where either the grammatical or the semantic aspect dominates. And finally there is pronominalisation which might be illustrated with derivatives in -er where -er is said to replace a specific person, e.g. Father drove the car and speeded too much. The police questioned him as the driver of the car. Other instances would be -^-derivatives denoting a place: a nunnery, a swannery. In ad-hocformations we find other general nouns when speakers cannot find the appropriate item, e.g. the computer stuff,\ the tax fellow, the paper wallah. Since -er, -ery, stuff, fellow, wallah here replace a full lexeme, I would prefer 'prolexematisation'.

The function of word-formation and the case of English -cum- 287 Information condensation, grammatical-semantic recategorisation and prolexematisation are the typical instantiations when patterned lexical operations are performed on specific existing lexical units. They therefore derive naturally and in a principled way from the general function of wordformation outlined above and are not mutually exclusive, but often overlap.

References 1. Novels Brookfield, Amanda 1997 The godmother. London: Hodder & Stoughton. (Sceptre Paperback). Lodge, David 1995 Therapy. London: Penguin Books. Richler, Mordecai 1998 Barney's version. A novel. London: Vintage. Rogers, Jane 1980 Living image. London: Fontana Paperbacks (Flamingo edition). Rosenheim, Andrew 1993

Hands on. London: Mandarin Paperbacks.

2. Linguistic Studies Bauer, Laurie 1979 Against word-based morphology. Linguistic Inquiry 10: 508-9. 1983 English word-formation. Cambridge: Cambridge University Press. Cruse, D. Alan 1986 Lexical semantics. Cambridge: Cambridge University Press. Downing, Pamela 1977 On the creation and use of English compound nouns. Language 53: 810-842. Hansen, Barbara et al. 1985 Englische Lexikologie. 2nd edition. Leipzig: VEB Verlag Enzyklopädie. Hohenhaus, Peter 1996 Ad-hoc-Wortbildung. Terminologie, Typologie und Theorie kreativer Wortbildung im Englischen. Frankfurt a. M.: Lang. Kastovsky, Dieter 1982 Wortbildung und Semantik. Düsseldorf: Schwann-Behagel. 1986 The problem of productivity in word-formation. Linguistics 24: 585600.

288

Gabriele Stein

Lipka, Leonhard 1992 An outline of English lexicology. 2nd edition. Tübingen: Niemeyer. Stein, Gabriele 1971 Primäre und sekundäre Adjektive im Französischen und Englischen. Tübingen: Narr. 1977 English combining-forms. Linguistica 9: 140-147 1977 The place of word-formation in linguistic description. In: Herbert Ernst Brekle and Dieter Kastovsky (eds.), Perspektiven der Wortbildungsforschung, 219-235. Bonn: Bouvier. Ungerer, Friedrich and Hans-Jörg Schmid 1996 An introduction to cognitive linguistics. London - New York: Longman. Warren, Beatrice 1990 The importance of combining forms. In: Wolfgang U. Dressler et al. (eds.), Contemporary morphology, 111-132. Berlin - New York: Mouton de Gruyter.

What happened to Old English clitic pronouns and why?1 Robert Stockwell and Donka Minkova

1. Personal pronouns in Old English It is well known that special rules have to be formulated to account for the fact that in Old English texts, personal pronouns regularly appear in positions different from the normal full nominal positions. It is certainly possible to describe the pronominal positions without defining the pronouns themselves as having special morphosyntactic properties. However, the preferred analysis of generativists (e.g. Kemenade 1987) has been to assign to these pronouns a special morphosyntactic status, that of "special clitics".2 What is gained from this manoeuvre is syntactic invisibility, in effect treating their peculiar properties as morphological rather than syntactic. Given this status, the pronouns can be counted "outside" the normal constituent-order rules which in main clauses do not generally allow nominal objects to appear on the left of finite verbs. It is only by virtue of declaring pronominal objects to be in some way syntactically invisible that one can claim that like all the other Germanic languages Old English had a verb-second (V-2) constraint. Without some stipulation concerning the exceptionality of pronouns, the V-2 constraint would be massively violated by examples like (1): (1)

Chron 893

he him hasfde geseald gislas ond adas 'he to-him had given hostages and oaths' Chron 876 7 him pa gislas sealdon 'and to-him the hostages they-gave'

But not all pronominal objects are cliticised even in Old English. Koopman (1992: 61-63) has described five positions in which verbalobject pronouns cannot reasonably be said to be cliticised in view of the fact that they are in no way distinguishable from nominals in those positions. These are the following: (i) postmodified pronouns; (ii) coordinated pronouns (with another pronoun, or with a full NP); (iii) pronouns as part of a larger phrase;3 (iv) pronoun objects which are not in the clitic position;

290

Robert Stockwell and Donka Minkova

and (v) pronominal objects which are extraposed (to the right of the nonfinite main verb, or to the right of the main verb in a subordinate clause).4 Koopman's description of the non-clitic behaviour of the Old English pronouns is an important distributional amendment to the clitic analysis. In addition to raising serious issues about the surface viability of the "special clitic" reconstruction, which we will not address here, his data survey leads him to a conclusion which is especially relevant to our study, namely: "There is very little evidence that later texts show a greater degree of decliticisation than earlier texts" (1992: 84). This conclusion supports the view that whatever happened to pronominal clitics such that they disappeared, it happened in Middle English, not in Old English. The narrow focus of our investigation is to establish when and why OpronV is replaced by V ΟρΓΟη· The basis on which we define the crucial clitic positions is this: the pronouns in such positions are not "free", i.e., they appear in positions which are specific to them only and are not shared by nominals.5 Though this is not unrelated to V-2, we will ignore the potential connection; our topic is decliticisation.6 2. Personal pronouns in Middle English What follows here is an overview of the positioning of pronouns in some major Middle English texts. For obvious metrical reasons, the verse texts are almost a waste of time to study in respect to pronoun placement, but the prose texts show decliticisation in high frequencies starting with the 12th century Chronicle; and as we approach 1400 with Chaucer, the prose is almost fully decliticised while the verse contains a substantial mixture of archaic and modern pronoun placement. Between 1150 and 1400 object pronouns increasingly move to the right of their governing verb. An exception is The Ayenbite of Inwyt, a Kentish text of 1340 (holograph manuscript, precisely dated, all prose), which is almost a throw-back to Old English with respect to pronoun placement. For purposes of this argument, we provide only a few typical examples, with rough statistics over a large sample. (A full statistical sampling is in progress; there appears to be no doubt about the directionality and approximate dating of the change we are discussing.) 2.1.

Peterborough Chronicle 1122-1154

For the most part, the entries in the last 32 years of the Peterborough Chronicle - written down soon after the events it reports7 - are simply a

What happened to Old English clitic pronouns and why?

291

continuation of Old English in every respect except for the word order of subordinate clauses, which have changed from an earlier 60% SOV to essentially 100% SVO, and except for the large number of pronominal objects which are no longer preverbal. The significant numbers of pronominal objects that are postverbal really suggest that the clitic system has broken down. Kemenade has asserted that "we find a remarkable number of personal pronouns in postverbal position, i.e. to the right of the basegenerated position for the verb. This is a position where they rarely occurred in Old English, and that indicates a step towards the loss of clitic status" (1987: 191). Our counts of postverbal objects in this Chronicle fully support Kemenade's summary, except that we would want to put it more strongly: we claim it is very hard to defend the notion "clitic pronoun" in the Peterborough Chronicle continuations from 1122 to the end. Koopman has made careful statistical assessments of the same text (Koopman 1992: 70) - but he counted the entire text (i.e., all of the Laud ms, referred to as ChronE, with entries, as in the Parker ms, from the biblical beginnings). We do not agree that pronominal tokens in the entire text should be counted, if one is interested in empirical data relevant to the approximate dating of decliticisation, because the entries before 1122 are known to be copied from earlier chronicles, and those from 1122 onward are known to be roughly contemporaneous with the reported events. Koopman is aware, of course, of the fact that postverbal object pronouns are much more frequent in ChronE. He acknowledges that "Of all the texts ChronE shows the greatest number of constructions [with postverbal object pronouns] with relatively few [preverbal object pronouns]" (Koopman 1992: 76), but he does not draw what seems to us the obvious conclusion, namely Kemenade's conclusion cited above. In fact our own conclusion from these data goes much further than Kemenade's tentative recognition of incipient decliticisation. By the time of the Peterborough final continuation, written down in 1154 or slightly thereafter, the evidence does not justify the claim that the personal pronouns were special clitics at all. Koopman attributes the increase to a specific context, namely "in coordinated clauses without subject", as in: a

(2)

7 brohte him pider mid micelferd (ChronE 1140.14) 'and brought him thither with a large army'

Indeed, this is one of the environments in which one must assume decliticisation or non-clitic status at any date. Examples of this type are all after 1122 and most frequent in the final continuation from 1132. If co-ordinate clauses without subjects frequently contain postverbal pronominal objects

292

Robert Stockwell and Donka Minkova

("There is a strong tendency in ChronE to put the object after the verb in co-ordinate clauses" [Koopman 1992: 76]), this is the very strongest possible indication that the pronouns have decliticised because these clauses strongly favour verb-finality in earlier Old English. Contrary to Koopman's assertion that "Van Kemenade's claims about ChronE are not supported by my evidence" (ibid), we will now show that ChronE 1122-1154 overwhelmingly supports a much stronger claim than hers, namely the one we are making. This text contains 124 P-3 direct and indirect object pronouns. We have not counted "quirky subjects" in this set (him semed etc.) because though they have the same form, their function is utterly different. We have not counted prepositional objects (the objects in postpositions, the vast majority, are non-clitics by definition anyway). We have not counted benefactive datives (hi scoldon cesen hem cefcebiscop 'they should choose for themselves as archbishop...'). We have not counted P-l and P-2 pronouns at all, though the numbers are so small that they would not carry any weight one way or another. We have discriminated three clause types: main, formally subordinate, and those introduced by ond or ac without a stipulated subject. The relevant forms are, of course, hym , hine, hit, hire, heo (accus, fem.), and hem. There are no thorn-forms in the Peterborough Chronicle. The gross figures are given in Table 1. Table 1. Pronouns in the Peterborough Chronicle

Main Subordinate Conjunction Totals

Preverbal objects 32 4 13 49

Postverbal objects 26 4 45 75

Total 58 8 58 124

On the basis of these figures one conclusion seems reasonable: since postverbal pronominal objects are significantly more common than preverbal ones (by a ratio of 3:2), decliticisation must have been well underway. Rather we should say: pronominal objects were no longer clitics, since they had no positions unique to themselves in a coherent grammar. While they are undoubtedly more common than nominals in the preverbal position, that position can no longer be specified as a matter of grammar. It has already become a relic position, archaic, a slot open mainly to various stylistic and pragmatic pressures, including formulaic usage: e.g. all the instances of the formula "to receive someone with honour" as in Se kyng him underfeng mid micel wurdscipe (1128:19) have preverbal pronominal

What happened to Old English clitic pronouns and why?

293

objects, and this formula alone accounts for about a quarter of all the preverbal objects in the data. A fact that needs to be given special consideration is that the postverbal pronominal objects show up with increasing frequency in the Final Continuation (from 1124, though not written down in this form until soon after 1154). In the earlier years of that document, where the existing record was presumably just copied by the scribe, the pre- and post-verbal objects are split fairly evenly. Then in the later years, when the final continuator was presumably composing his own materials, not copying the work of earlier scribes, the increase of postverbal objects is startling. The sharp rise is shown in Table 2. The last two columns at the right are the ratio of preverbal pronouns per line of text (Ratio-1), and the ratio of postverbal pronouns per line of text (Ratio-2). They are rounded off trivially. The picture they present is a steady-state (with normal blips) distribution of pre- and post- verbality through 1131, and then a sharp clear rise of postverbal pronouns from 1132 onwards (which really means from 1154, when the final continuator was writing). Table 2. Preverbal and postverbal objects in the Peterborough Chronicle Preverbal Postverbal Number of lines Ratio-1 Ratio-2 9 6 1 90 1123 .6 2 54 1124 5 1 .3 2 1125 3 38 .7 .5 2 3 1 1126 17 1.75 6 8 1127 73 .8 1 4 1 1128 30 1.3 .3 1129 2 1 42 .5 .2 2 0 22 1130 0 1 6 7 42 1.4 1131 1.6 - beginfinalcontinuator 0 1 1132 13 0 .7 1 2 1135 25 .4 .8 6 1137 10 87 .6 1.1 1138 0 0 5 0 0 4 28 84 1140 .5 3.3 1 2 1154 19 .5 1 641 49 75 .7 Totals 1.1 Here are a few examples from the entries prior to the final continuator's work, chosen to show that it was not only that scribe who was apparently comfortable with postverbal pronouns.

294

Robert Stockwell and Donka Minkova

Preverbal (3)

1123 1123 1127

he hem hit wolde typian 'and he them it was willing to grant' se kyng him geaf done cercebiscop rice 'the king him gave the archbishopric' Sodfeste men heom kepten on nihtes 'Truthful men kept watch on them at night'

Postverbal (4)

1123

1123 1123

2.2.

se kyng alihte dune of his hors & alehte hine betwux his earmes. 'the king got down from his horse & lifted him in his arms' & brohten him toforen se kyng 'and brought him before the king' se pape swedolode & gaf him his pallium 'the pope relented and gave him his purple robe'

Orm (around 1200, Northern)

Orm consistently has ///-forms of they and their, but he mixes them and hem with no rationale visible to us, positional or otherwise. His object pronouns are about 50-50 preverbal/postverbal, and it seems clear that decliticisation was underway.9 An important issue might be: Were the preverbal object pronouns usually //-forms, and the post-verbal object pronouns usually I>-forms? If so, a theory of Scandinavian superstate triggering of the change would be supported. If not, such a situation might be possible, but it remains speculative. In the first 3380 lines (including the Dedication, Preface, and Homilies), the numbers of tokens are: peyym - 10 instances, hemm - 80 instances. Due to the metrical nature of the text, only about one eighth of these tokens can legitimately be informative with respect to syntactic positioning. Of the total of 11 lines allowing metrical freedom of positioning hem on either side of the verb there are 7 instances of OV (Ded. 153, 329, Horn 145, 273, 387, 655, 1342) vs. 4 instances of VO (Horn 335, 851, 1955, 2372). Within that portion of the text there is not a single attestation of them whose position cannot be accounted for metrically. Elsewhere in the text, however, such examples exist:10

What happened to Old English clitic pronouns and why? 295 (5)

6499 10153 15344 16218

He {>eyym forrbasd to turrnenn. He Jjeyym forrbaed to takenn ohht patt he Jjeyym yaffpurrh Haliy Gast & he J)eyym yaff annswere onnycen

The overall number of relevant instances is rather small, yet a survey of the positions of the jj-pronouns reveals no striking overlap between the borrowed form and the VO order. Clearly, thefc>-formshave not established a new positional beachhead - they do not suddenly show up in the postverbal position. Thus, this text is (a) on the way to decliticisation, and (b) the "new" J>pronouns do not exhibit any special properties that set them apart from the remaining object pronouns. 2.3.

Richard Rolle of Hampole

Rolle was born around 1300, and died nine years after Chaucer was born, in 1349. He provides the first unmixed testimony that a change has taken place, fully and almost completely. He was a Yorkshireman, which may be an important fact - it tends, at least, to support the foreign influence theory that is discussed below, because Yorkshire was much more heavily Scandinavian than the south. While he wrote some lyric poems, the largest part of his English output (he wrote a lot in Latin also) is prose, in fact prose meditations and accounts of mystic visions. His writings were extremely popular in his own day and for many years thereafter, which is attested by the extraordinary number of copies of his works that have survived, including both Latin and English. A tally of the positioning of all pronominal objects in the Prologue and Psalm 3 of his English Psalter, about 200 lines of text)11 attests to a startlingly modern pronominal word order: 19 instances of the type shown in (6): (6)

Pr-75 Ps3-8

for it leres vs to love God with glad chere Multiplyed ere that angers me

but extremely rarely (only once in these 200 lines) we find: (7)

Ps3-38

and he me herd fra his haly hille.

2.4.

Ayenbite of Inwyt 'The Prick of Conscience'

Written in 1340, in the Kentish dialect, of interest mainly because of the dialect, this text has completely conservative placement of object pro-

296

Robert Stockwell and Donka Minkova

nouns, showing that the wave of the north had not yet hit the deep south. There are few examples of the postverbal placement of pronouns which have by this time become the norm in the north, as we have seen above. 2.5.

Chaucer

Object pronouns had totally ceased to be proclitics in Rolle before the second half of the 14th century; and by the second half of the 14th century, in the midlands and in London, as attested in Chaucer and the Wycliffites - i.e., preverbal pronominal objects had virtually disappeared from the prose (though they are not uncommon in verse). In Chaucerian prose, the pronouns are free. We have found only a few instances of pre-verbal object pronouns in main clauses in his prose texts. In the verse, some instances of pronominal inversion appear to be metrically motivated, but most are simply archaic order. In (8) we illustrate our claim that the pronouns are free in main clauses in his prose: 12 (8) ParsTale 95 sith I have declared yow what thyng is Penitence ParsTale 115 By the fruit of hem shul ye knowen hem ParsTale 284 Andfor as muche as jhesu Crist yeveth us thise yiftes ParsTale 315 And therof doon hym hommage WicliffProdSon 281.9 pis citisein sente him into his toun WiciiffProdSon 281.17 hisfadir sawe him WicliffProdSon 281.26 and pe bigunne to feede hem We take Chaucer to be nearly modern with respect to pronoun positioning, but still archaic with respect to some pronominal forms while modern in others (hem vs. they). 3. Explanations Thus the pronouns somehow became decliticised, and indeed, except for the Ayenbite in Kent, the change is attested from our earliest Middle English texts. For a categorial change to occur ("X is a special clitic" > "X is not a special clitic"), a hypothesis is needed to explain why the change took place. (E.g., the creation of a class of AUX verbs, following the Lightfoot scenario, or the reanalysis of impersonal verbs.) The hypothesis should be of the type we find in other syntactic change, namely a structure in the primary linguistic data which is ambiguous in the sense that it is open to more than one parsing, or more than one categorial assignment. Or

What happened to Old English clitic pronouns and why?

297

there must be some foreign input, some outside unbalancing force which makes for the possibility of reanalysis. The history of such hypotheses is necessarily recent, because the claim that Old English pronouns were special clitics originates with Kemenade 1987. She of course saw the problem of decliticisation in Middle English, but her explanation of the reasons for it was less persuasive than the rest of her important book. We will not elaborate on her hypothesis here because it concerns subject pronouns and because it links their history to the loss of verbal inflexion and V-2.13 What follows in Section 3 is a skeletal exposition of some ideas regarding the causality of decliticisation; it is also a target inviting further comments, critiques, and elaboration. It seems to us that there are three reasonable hypotheses to be evaluated: (1) Scandinavian influence, (2) inconsistencies in the primary linguistic data which might cause a child not to make the correct abductions about archaic pronominal positions, and (3) the drift to SVO as the analogical basis for the change of pronominal positions. 3.1.

The Scandinavian scenario

As is well known, the 3rd plural pronouns of English were borrowed in the north from Scandinavian, with they spreading to the south first, and ultimately to the entire country. Kroch and Taylor have suggested that the borrowing of they "could well have altered the syntactic character of the entire pronominal system" (1997: 314). We are puzzled how to argue in favour of this position other than take the "Scandinavian" scenario in a very broad sense. The crucial text cited by Kroch and Taylor is the Rule of St. Benet. But this text is from the first quarter of the 15th century (Warner 1997). It is far too late to give us useful information about decliticisation, since pronouns had decliticised in both the north and the south well before this date. For reasons we do not know, they did not cite Rolle, who one hundred years earlier displayed all the pronominal properties they were looking for. The almost total modernity of Rolle may be used as an argument for the view that introduction of the /A-pronouns somehow correlates with freeing them from the clitic positions, but the co-occurrence may be accidental - it has no intrinsic "explanatory" value. The juxtaposition of the northern and the southern evidence (especially Rolle's prose vs. the Ayenbite) makes it likely that the change did indeed originate in the north. It took another 25-50 years to spread through the south of England, but like so many other northern morphological features, it eventually did.

298 Robert Stockwell and Donka Minkova Assuming that the correlation between the northern origin of object pronoun decliticisation and the Scandinavian influence there is not accidental, the question which we have been unable to answer is which component or components of the "external" grammar can be isolated as triggers of the change. Starting with the hypothesis that borrowing of they "could well have altered the syntactic character of the entire pronominal system", we note some serious data problems with it. For it to work, there should be a much stronger correlation between the presence of th- pronouns and decliticisation - and we did not find that. The Peterborough Chronicle, our earliest text with significant evidence of decliticisation, employs the pronominal forms of Old English. Neither did we find any evidence of differential treatment of native vs. borrowed pronouns in the text that mixes them - the Ormulum. Moreover, we found, in two approximately contemporaneous texts, The Moral Ode (southern, 1180 - date from Pearsall 1977) and Orm, the following pair: (9)

PM 65 3256 3258

[Per me scal\... and geuen us ure swinkes lan 14 Ormulum [&ta shall ure Laferrd Crist /Att ure lifess ende\ uss yifenn ure swinnkess Icen

More than a century later, the versions of Cursor Mundi (a distinctly northern composition, dated c. 1300), show the same uncertainty:

(10)

CM 110: Scho sal J)am 3eld a hundreth fald Ho wil ham 3ilde an hundre falde. Scho wil him 3eilde an hundreth fald,

MS Cotton15 MS Faifax16 MS Göttingen17

CM4981: Pis time will I do [jam pat grace Pis time wil I. Do ham pat grace

MS Cotton MS Faifax

Clearly, there is no metrical reason for the positioning of either the thor the Α-pronouns in the various copies. Patterns like these confirm our doubts about the correlation between decliticisation and the introduction and spread of the 3rd person plural pronouns from Scandinavian.

What happened to Old English clitic pronouns and why?

3.2.

299

Ambiguity in the primary linguistic data

There are two questions: (1) Did ambiguity of positioning exist? (2) If so, why did it exist? In Section 2 we have documented the fact that preverbal and postverbal pronominal objects competed for acceptance from the middle of the 12th century until well into the 14th century. It did exist, from earliest Middle English; the question is why it existed. We have shown variation from the Peterborough Chronicle up to Rolle, as between preand post-positioning of the object pronouns. The roots of this variation are deep into the pre-Conquest period. We believe that the end-result of early 14th c. decliticisation is the consequence of a variety of factors. These include the following: 3.2.1. The VO!SI construction The construction was limited to only a few verbs in Old English, the causative and perception verbs. Latin influence, as well internal changes connected with the loss of inflections in late Old English /early Middle English contributed to the adoption of the ACI construction with ' expect'type verbs (wish, think, consider, declare), see Fischer (1990: 136-137). While in Old English the pattern was considered 'foreign', a situation corroborated by the fact that "[the ACI construction] would not appear in any text that was not a literal translation from the Latin", in Middle English it clearly became part of the native syntax of English, with borrowed verbs such as desire, suffer, grant, save, added to the native deem, trow, will, know, let, etc. With the loss of the distinction between dative and accusative NPs - full nouns as well as pronouns - verbs which had previously tolerated only a dative object (the 'persuade' verbs), join the pattern illustrated in (10) (Fischer 1990: 83-84,109). The signals that these verbs send to the learner would make the acquisition of pronominal clitics more complex. (II) 18

Andb&d hem alle ihere And makede hem alle knightes And oueral made hem forto calle

Horn 1262 Horn 520 Havelok 3 8

3.2.2. Syncretism ofpronominal forms The loss of distinction between 3rd person accusative and dative pronouns: hine vs. him, hit vs. him, hi, heo vs. hir(e), hi vs. hom, heom, is often ac-

300

Robert Stockwell and Donka Minkova

companied by the use of prepositions for the indirect object and thus increases the proportion of instances of prep+object pronoun, and these cannot be clitics by definition. (12)

CM 5215 CM 1750 JN 4 JN 9

Drightin him come in awision Cotton MS God cam to him in vision Göttingen MS Al pat our lauerd him forwith badd. Cotton MS Al pat our lorde to him bad. Fairfax MS In this same time our lord shewed to me a ghostly sight19 I speake of hem ..., for in this time God shewid me none other

The breakdown of the morphological case system often produces a proliferation of options of prepositional object marking not available in earlier times. The point is illustrated by the history of the verb 'to trust', presumably from OE *trystan, but reinforced from Old Norse. In Middle English this verb is used (often with no change of meaning) with in, of, on, to, unto, upon, as well as with a direct object. Obviously, the bulk of the signals that a learner is getting would disallow a clitic placement of the pronominal object: (13)

Prov.Alf20: lokepat hepe be mide..& on himpu maistpe tresten. A Mayde Oistes:21 Monnes luue nys buten ο stunde pat him triste^)... Meditations on the Passion 120: pat I may... hope and trist to J>e and in pe. Troilus 1: 601: But lest thow deme I truste nat to the ...

3.2.3.

"Quirky " subjects

Somewhat more tentatively, we might suggest that further ambiguity in the primary linguistic data is created by the survival of the "quirky" subjects well beyond the establishment of the SV order; the opaqueness is due to the fact that at a time when any NP preceding a verb was functionally a subject, a morphological object in that slot was a special, semantically

What happened to Old English clitic pronouns and why? 301 marked case. This association between the "object" pronominal form and the pre-verbal subject position would tend to drive away other, "normal", pronominal objects from that position. SW178:22

(14)

He easkej) ham e/ham biluuej) toheren him ae hwile. ParsTale 84-85 ... for which hym oghte to pleyne. ParsTale 123-4 ... hym semeth that it is to him Moost sweete of any thyng;

3.3.

Analogical pressure from generalisation of SVO

The other possible explanation is simple analogy. It is well known that by about 1200 the dominant word order was head-initial, i.e. SVO (still retaining the V-2 constraint). Such a change in the nominal placement rules of the language cannot but have had a strong influence on pronominal placement, the simplest kind of regularisation and analogical cleaning up of the structure. In fact the pronominal development to which we have devoted most of this paper is evidence for the solid takeover of SVO. The change to SVO may itself be the principal cause of postverbal pronominal placement, and possibly no more elaborate theory is needed. In conclusion, all three factors may be seen as related to morphological changes from Old to Middle English, which, in their turn, are related to the settlement of Scandinavian speakers in the northern parts of the country. However, the precise nature of the linkage between the decliticisation we traced in the texts, and these other grammatical changes, remains to be determined. Notes 1. This paper is largely programmatic; after some years of researching mostly phonological and prosodic issues in the history of English we return to the more general topic of word order, which was the basis of our first joint syntactic piece prepared at Dieter Kastovsky's invitation for what we informally and affectionately remember as "The Kellner Conference". We are very happy to dedicate this paper to Prof. Dr. Dieter Kastovsky, the scholar, the able organiser and efficient editor, and to Dieter, our friend, our host, and our energetic tennis partner. 2. The term is from Zwicky (1977); the properties of these special clitics are defined and discussed in Klavans (1982).

302

Robert Stockwell and Donka Minkova

3. This category seems unclear to us; fortunately our argument does not hang on it. 4. We refer the reader to Koopman for extensive exemplification. We include only those types of exceptions which concern verbal objects, i.e. we have not included Koopman's subgroup of pronouns as complements of prepositions (1992: 62), because they represent a different problem. 5. We use the term "free" to refer to personal pronouns when they appear in regular nominal positions, i.e., in all positions except peculiarly clitic positions. All of Koopman's positions listed above would naturally be "free" positions. 6. The relevance of V-2 is that as a rule it interacts closely with pronouns and provides crucial information about whether the personal pronouns are clitic or free. Subordinate clauses may be more prone to show Op ron V because of the archaic hangover of verb-finality in subordinate clauses. Koopman (1992: 53, 62) points out that the clitic status of object pronouns in subordinate clauses is seriously obscured by the fact that full nominals can also occupy the first position in the VP or be to the left of the VP. 7. "Internal evidence shows that this part of the Chronicle [1132-1154] was not written down until after the accession of Henry II" (Earl and Plummer II.xxxv), but soon thereafter. 8. See also Stockwell and Minkova (1991: 379-381). 9. Admittedly, the picture in this text may be somewhat obscured by local metrical considerations, i.e. the pronouns are weak monosyllables, while many of the verb forms to which the object attaches are disyllabic prosodically strongweak forms, which can privilege the positioning of the pronoun to the left of the verb (hemm lufenn), while a to + inf. (to clennsenn hemm) or a pron. subj. + verb (he tallde peyym) would favour postposition. The point is clearly illustrated by the following pairs of almost adjacent lines where any syntactic argument would be overridden by the metrical considerations: 699 Patt Drihhtin shollde lesenn hemm vs. 702 Patt Godd hemm wollde lesenn. 1501/ patt tatt tu Jjeyym shaewesst vs. 1514 ά yiffpu shaswesst hemm off Godd Still, there are numerous clear cases where the positioning would not affect the meter, as in: 5278 Patt illke wise badd hefceyym(comp.: patt Hike wise he Jjeyym badd) 6379 & he Jjeyym droh to rune (comp.: & he droh Jjeyym to rune) 10. Note the positioning in: 8841 Patt he Jjeyym wollde yifenn lif 8842 A butenn ende i blisse, 8843 Yiff patt teyy wolldenn follyhenn himm where the selection of the form peyym seems random and goes against the prediction of northern influence.

What happened to Old English clitic pronouns and why?

303

11. The English Psalter is believed to be the first extended document Rolle wrote in English. 12. In the 1080 lines of the Parson's Tale in the Robinson edition, there are hundreds of conforming examples - i.e., examples where the object pronoun follows its head (either verb or preposition), and fewer than ten - we would say archaic - instances where it precedes the verb (none where it precedes the preposition). In the two Wycliffite texts we examined there are no nonconforming example at all. Given these facts it would be absurd to try to argue that either of these writers had anything different from modern English rules of object-pronoun placement. The Wycliff texts are quoted from the Moss6 Handbook of Middle English by page and line number. 13. Her theory asserts that in Wycliff, "due to the loss of verbal morphology, subject pronouns cease to be interpreted as clitics" (1987: 221). In her view, then, the subject pronouns led the change in Wycliff by not moving, i.e., by not undergoing subject inversion after topic phrases (thus Wycliff has "Now he said . . . " as compared with Chaucer "Now said he . . ." Further, "as a result, surface patterns in root clauses cease to be interpreted as V-2" in Wycliff. This seems to say no more than the following: that subject pronouns, on the analogy of full nominals, came to be treated like full nominals. But that is simply a restatement of the problem: Why did the reanalysis take place? Why did the exact same reanalysis not take place in Anglo-Saxon times, for exactly the same reasons? The loss of verbal morphology cannot be a valid argument because Wycliff and Chaucer both use a full suite of verbal inflections, not distinguishable from the Old English, or the earlier Middle English system, -e, -est, -eth (sg.), -en (pi.) in the Present, weak past in -idelede, etc. Her hypothesis for Chaucer is the inverse of her hypothesis for Wycliff: In Wycliff, the subject pronouns ceased to be clitics but stayed to the left of their verbal heads, thus destroying V-2 (i.e., Topic + Pro-Subj + VP became possible). In Chaucer, where V-2 remained alive, it's also the subject pronouns that led the way - but by moving, i.e. undergoing typical V-2 inversion. The pronominal subjects came to be felt as full nominals and also to behave like full nominals by moving to the right of the verb when the clause started with a Topic (adverb or the like). But this again is a restatement of the problem, in the sense that it merely restates the facts. Pronouns were no longer clitics, and so they behaved like nominals. 14. The position of the pronoun is the same in both versions - Lambeth and Trinity. The line is cited from Hall 1920. 15. First half of 14th c. 16. Second half of the 14th c. 17. First half of 14th c. 18. Examples cited from Saito and Imai (1988). 19. Cited from the Middle English Corpus Electronic Edition of A Revelation of Love by Julian of Norwich, ed. Marion Glasscoe (1976), University of Exeter.

304

Robert Stockwell and Donka Minkova

20. Proverbs of Alfred (112/376). al275 (?cll50) (Trin-C B.14.39) Cited from the MED. 21. Jes-0 29: 56. Citation from the MED al300. 22. Early 13th c. Here cited from The MEC (U of Michigan). Sawles Warde : an Early Middle English Homily, edited from the Bodley, Royal and Cotton mss by R.M. Wilson Leeds: Leeds School of English language texts and monographs; no. 3. 1938.

References Allen, Hope Emily 1931 English writings of Richard Rolle, Hermit of Hampole. Oxford: Clarendon. Burrow, John A. and Thorlac Turville-Petre 1996 A Book of Middle English. Oxford: Blackwells. Earle, John and Charles Plummer 1892 Two of the Saxon Chronicles parallel. Reissued 1952 with notes by Dorothy Whitelock. Clarendon Press: Oxford. Fischer, Olga 1990 Syntactic change and causation. Developments in infinitival constructions in English. Amsterdam: Amsterdam Studies in Generative Grammar 2. Hall, Joseph 1920 Selections from Early Middle English: 1130-1250. Oxford: Clarendon. Kemenade, Ans van 1987 Syntactic case and morphological case in the history of English. PhD thesis. University of Utrecht. Dordrecht: ICG Publishing. Kemenade, Ans van and Nigel Vincent (eds.) 1997 Parameters of morphosyntactic change. Cambridge: Cambridge University Press. Klavans, Judith 1982 Some problems in a theory of clitics. Bloomington: Indiana University Linguistics Club. Koopman, Willem 1992 Clitic pronouns. Some remarks. In: Fran Colman (ed.), Evidence for Old English, 44-88. Edinburgh: Edinburgh University Press. Koch, E. A. 1902 The rule of St. Benet. EETS Original Series 120.

What happened to Old English clitic pronouns and why?

305

Kroch, Anthony and Ann Taylor 1997 Verb movement in Old and Middle English: dialect variation and language contact. In: Ans van Kemenade and Nigel Vincent (eds.), 297-325. Pearsall, Derek 1977 Old English and Middle English poetry. London: Routledge and Kegan Paul. Saito, Roshio and Mitsunori Imai 1988 A concordance to Middle English metrical romances. Frankfurt: Peter Lang. Stockwell, Robert and Donka Minkova 1991 Subordination and word order change in English. In: Dieter Kastovsky (ed.), Historical English syntax, 367-409. Berlin, New York: Mouton de Gruyter. Warner, Anthony 1997 The structure of parametric change, and V-movement in the history of English. In: Ans van Kemenade and Nigel Vincent (eds.), 380394. Zwicky, Arnold 1977 On clitics. Bloomington: Indiana University Linguistics Club.

Tall As He Was: On the meaning of complements-constructions 1 Gunnel Tottie

1. Introduction In spite of its theoretical importance and its concomitant popularity among linguists, true ambiguity is a rare phenomenon in actual language use. One such case concerns constructions such as Tall as he was, which can have diametrically opposite meanings, in that as can be paraphrased either as concessive 'though' or causal 'because'. This is demonstrated by (1): (la)

Tall as he was, he was not able to reach the top shelf.

(lb)

Tall as he was, he was able to reach the top shelf.

In this paper I wish to discuss what factors condition the use and interpretation of this construction. I will follow Kjellmer (1992) in using the term complement-os-construction. I propose to take a quantitative approach, something that previous writers on the subject have mainly refrained from, and finally, I will discuss the origins of the complement-asconstruction, arguing that the causal reading is likely to have had a different history from the concessive construction and suggesting a possible origin for the causal construction.2 My study will be based on a sample of 53 complement-os-constructions taken from works of fiction, journalism, and scholarship, published between 1964 and 1989. 2. Previous work Both Jespersen (1927) and the OED (second edition) discuss complementas-constructions and agree that they derive from an earlier construction with a correlative so or as preceding the complement in the os-clause, as in (2) and (3), quoted from OED2, s.v. Β I 3 b: (2)

To brynge VJ, so fre as we be, in tofyl seruage. (1297 R. Glouc. 47)

308 (3)

Gunnel Tottie I took leave of... Antwerp, as late as it was, embarking for Brussels. (1641 Evelyn Mem. (1857) I. 37)

Jespersen and the OED are also unanimous in stating that constructions having an initial as or so are now obsolete. Thus Jespersen states (1927: 9.41): "The force of these clauses generally is concessive, bad as this is meaning 'however bad this is...' In former times it was usual to have here as...as ; this is still found in Scott, but is now obsolete." And the OED agrees: "In parenthetical clauses forming an extension of the subject or predicate, the antecedent (so, as) formerly present is now omitted, and the relative has acquired somewhat of a concessive force = Though, however." However, as will be clear from the discussion of Quirk et al. (1985) below, Jespersen and the OED are overstating the case of the demise of the as.. .^-construction. Quirk et al. (1985) do not explicitly discuss the homonymy of the construction in (1) above but treat complement-as-constructions under two separate rubrics: Concessive adverbial clauses in 15.39 and "circumstantial" clauses (a subtype of clauses of reason) in 15.47.3 Quirk et al. refer to the concessive use of as (as well as that in, e.g., Fool that he was...) as "exceptional" and also characterise it as especially British, adding: ...concessive clauses sometimes have unusual syntactic orderings when the subordinator is as or though. In a rather formal style, the predication in the concessive clause may be fronted: Genius though she was, she was quite unassuming. Naked as I was, I braved the storm. ['Even though I was naked...'] Interestingly, Quirk et al. find it necessary to gloss their os-example but not the one with though. They further remark (in note 15.39d): ...non-initial placement of concessive as is perhaps to be related to as....as in the following example of an infrequent type of concesssive clause: As widespread as the effects may be, the Midwest still bears the brunt of the recession ['Even though the effects may be widespread,...'] The noninitial as may therefore be regarded as a correlative to an omitted initial as. This construction with a preceding correlative as is obviously not quite as defunct as Jespersen and the OED would have it.4 Concerning the causal reading of complement-as-constructions, Quirk et al. (1985: 15.47) merely state:

On the meaning of complement-as-constructions

309

When as is a circumstantial subordinator, the predication may optionally be fronted; Writing hurriedly as she was, she didn't notice the spelling errors. Tired as they were, they went to bed as soon as they came back. Interestingly, the complements in Quirk et aV s examples are both verb participles. As I shall show below, that may influence the reading in cases like these. Complement-as-constructions like those in (1) in contemporary English have been subjected to a perceptive analysis by Kjellmer (1992), based on a study of 46 instances taken from the Brown and LOB corpora. He discusses several factors, the majority of them semantic, that contribute to the interpretation of the as-clause as either causal or concessive. He begins by pointing out that a subordinate adverbial clause such as "Old as he was... " can have diametrically opposite meanings depending on the matrix clause (1992: 339). Thus either (4a) or (4b) is a perfectly fine sentence, where, as in (1) above, the complement-as-construction is disambiguated by the following matrix clause: (4a) (4b)

Tired as he was, he fell asleep immediately. (Κ 19)5 Tired as he was, he felt obliged to finish the chapter. (K 20)

However, in (4c), we have ambiguity: (4c)

Tired as he was, he watched the news programme on TV. (K 21)

The interpretation of (4c) "depends on our presuppositions, our attitude to news programmes on TV: as boring or relaxing or as strenuous and engaging." (Kjellmer 1992: 340). A wider context will disambiguate the sentence. Moreover, Kjellmer shows that there can also be merger or indeterminacy of meanings. We have a case of merger in (5): (5)

She's determined to get married and leave, young as she is. (LOB, Κ 24).

Kjellmer also points out (1992: 342) that the causal/concessive dichotomy is not exhaustive: there are cases where there is a third possibility of interpretation, and where the content of the as-clause is not presented as cause or contrast, but as merely concomitant, as "knowledge shared by writer and reader", as in (6):

310

(6)

Gunnel Tottie

...unable as he [President Eisenhower] was himself to say his running was best for the country, unconsciously he had placed his party before his country. (Brown, Κ 25)

In (6), "[c]oncessiveness and causality are almost irrelevant to the interpretation of the sentence; the information in the subclause is presented with the implication that it is already known to the reader." (Kjellmer 1992: 342). Following Quirk et al (1985: 15.23), Kjellmer also points out that while concessive and causal clauses stand in a restrictive relationship to the main clause, the clearly concomitant clauses are non-restrictive. Kjellmer then proceeds to discuss factors that condition the interpretation of complement-as clauses as either concessive or causal. He shows that the presence of incompatible semantic elements in the matrix and the subclause leads to a concessive reading. Sometimes (but not very frequently) this contrast is formally and morphologically expressed (cf.7): (7)

Capable as their minds may be in some directions, these guardians of the nation's security are incapable of learning or even observing. (Brown, Κ 28)

Other factors that lead to a concessive reading are the presence of the semantic factor [ANOMALY] in the complement as in (8), the presence of a correlative conjunct such as still, yet, nevertheless in the matrix clause as in (9), or the presence of may in the as-clause as in (7) above and (8) and (10) below: (8) (9) (10)

Absurd as it may seem, I didn't want you to think badly of me (LOB, Κ 37) Divided as the Northern Rhodesians are...they still have far greater strength... (Brown, Κ 42) Yet the family unit, virtually indestructible as it may be, is often capable of improvement... (LOB, Κ 51)

Kjellmer then shows that causal readings tend to occur where there is identity of some semantic components) in subclause and matrix clause, as in (11), where he posits the shared semantic component [SPECTACULARITY]: (11)

The Ommeganck was an extremely colourful affair, dominated as it was by the triumphal cars carrying elaborate enactments of tableaux... (LOB, Κ 62)

On the meaning of complement-as-constructions

311

Finally, Kjellmer also mentions that in the Brown and LOB corpora, concessive as-clauses tended to occur initially or medially (24/30), whereas causal os-clauses were more evenly distributed: 5 initial and 6 final.6 (The number of concomitant clauses in Kjellmer's material was low, only five, which made it difficult to generalise concerning their placement.) 3.

The present study

3.1.

Material and method

At about the same time as Göran Kjellmer was carrying out his study of complement-ay-constructions based on the Brown and LOB corpora, I was engaged in a similar study, noting instances of complement-osconstructions in my leisure reading. Unaware of Kjellmer's work until just before its publication, I assembled a sample about the same size as his, consisting of complement-os-constructions drawn from contemporary novels, biographies, newspapers and magazines. I ended up with a total of 53 examples, 27 American and 26 British.7 My aim was also to establish factors that might favour either the concessive or the causal reading of asconstructions. Inspired by Kjellmer's published work, I subsequently also introduced the concomitant distinction. Like Kjellmer, I examined the importance of the position of the as-clause and the presence of a conjunct in the matrix clause, and I also examined one semantic factor related to Kjellmer's [ANOMALY] factor, viz. whether the complement was "emotionally loaded" or not. I also coded my material for a few other factors of a formal nature. I thus examined the following factors: A

Factor group: 1. Regional variety 2. Text type 3. Position of as-clause 4. Complement semantics 5. Type of complement 6. Subject of ay-clause 7. Word-order in as-clause 8. Verb type in as-clause

Factors: British, American Fiction, non-fiction Initial, medial, final Emotional, not emotional Adjective, past or present participle Noun, pronoun Normal, inverted BE/Other

Factor groups 1, 3, and 4 replicate some of Kjellmer's types, whereas factor groups 2 and 5-8 seem not to have been previously examined in the

312 Gunnel Tottie literature. Factor groups 1-3 are self-explanatory; some comments will be appropriate for the rest. That factor group 4, having to do with complement semantics, is equivalent to Kjellmer's ANOMALY factor (see example (8) above) is demonstrated by (12): (12)

Incredible as it seems, hunger may be completely absent during even an extendedfast. (Cott 61)

Factor group 5 was motivated by my initial impression that although all the complements I included in this study were adjectival, there seemed to be a corrrelation between at least past participle complements and the causal reading, as in (13):9 (13)

If you must tell her something, tell her that used as I am to simple wholesome country food and early hours, the luxuries of the city were too much for me. (Lurie 1986a: 140)

I therefore coded all complements according to whether they were bona fide adjectives, present or past participles. Factor group 6 was inspired by the observation that most as-clause subjects appeared to be realised by pronouns. It therefore seemed interesting to investigate whether a nominal or pronominal subject could be correlated with either a concessive or a causal reading. Thus examples like (14) with nominal subjects were thin on the ground but seemed to have mostly concessive readings: (14)

But, it turned out, serious as this problem was, it was not what was worrying Archer and Herbert out of their minds. (Cross 25)

Factor group 7 was motivated by the observation that some examples had subject-verb inversion, like (15). Again it seemed interesting to check whether this could be linked to a particular reading. (15)

A Chancellor who, grudging as was the acknowledgment he received for it, everyone knew to have saved his party. (Gowers 124)

In most examples, BE was the main verb, as in examples (13), (14), and (15) above, but there were some exceptions, like (12) above and (16) below. Impressionistically, they seemed unlikely to have a causal reading.

On the meaning of complement-as-constructions 313 (16)

And, to be frank I suspected also that she never wrote about me, and that I might feel hideously offended, stupid as that sounds. (Cross 131)

3.2.

Results

Table 1 shows the distribution of concessive, causal and concomitant readings over British and American examples in my sample. Table 1. The distribution of concessive, causal and concomitant readings present sample. Row percentages except in totals column. Concessive Causal Concomitant Regional variety 11 44% 10 40% 4 16% British 25 18 64% 8 29% 2 7% American 28 29 55% 18 34% 6 11% Totals average 53

in the Total 47% 53%

We see that the distribution of concessive/non-concessive readings was somewhat different in the two varieties, with a higher proportion of concessive readings in the American part of the corpus, but a chi-square test revealed that the difference was not significant. In any case, the concessive complement-as-construction is definitely not a British construction, as Quirk et al. claim (1985: 15.39). My findings clearly support Kjellmer's observation that British and American data are similar. Just as in Kjellmer's study, there were fewer instances with a causal reading (34%) than with a concessive reading (55%) - and a very low proportion of instances with a concomitant reading, only 11%. Table 2 shows that the proportions of the three different readings were fairly similar in fiction and non-fiction, with a clear predominance of concessive readings in both. A larger sample would of course be desirable here, as well as spoken data; this will have to await further research. Table 2. The distribution of the three different readings over fiction and nonfiction. Row percentages except in totals column. Genre Fiction Non-fiction Totals

Concessive 20 57% 9 50% 29

Causal 11 31% 7 39% 18

Concomitant 4 11% 2 11% 6

35 18 53

Total 66% 34%

The results of factor group 3, the position of the complement-asconstruction, are presented in Table 3.

314

Gunnel Tottie

Table 3. The distribution of readings over sentences with complement-asconstructions in initial, medial, and final positions in the sentence. Row percentages except in totals column. Position of as-clause Initial Medial Final Totals

Concessive

Causal

21 5 3 29

6 2 10 18

75% 42% 23%

21% 17% 77%

Concomitant 1 5 0 6

4% 42% 0%

Total 28 12 13 53

53% 23% 25%

These results parallel the distributions found by Kjellmer for concessive readings (1992). There is a strong link between concessive reading and initial position. In my material there is also a connection between causal readings and final position. This distribution is highly significant (p < 0.001, chi-quare 26.99, 4 d.f.) Emotional complements, which parallel Kjellmer's [ANOMALY] factor, could also be shown to have a significant correlation with the concessive reading, as appears from Table 4. This distribution is also highly significant (ρ < 0.025, chi-square 8.97, 2 d.f.). Table 4. The distribution of readings over complement-as-construction with emotional/non-emotional complements. Row percentages except in totals column. Causal Concomitant Concessive Total Complement semantics 0 0 9 17% 9 100% + Emotion 44 20 45% 18 41% 6 14% 83% - Emotion 29 18 6 53 Totals

Let us turn next to factor group 5, having to do with the category of the complement: bona fide adjective, present participle, or past participle. Establishing these categories is a tricky undertaking because of the frequent use of both kinds of participles as adjectives, e.g. interesting, charming, tired, old-fashioned, etc. I decided to make an extremely conservative classification in order not to prejudice the results. Thus, I classified all four -/wg-participles in my sample (tempting, grudging, interesting, and frightening) not as adjectives but as participles. One instance was quoted in (15); the others follow as (17>—(19): (17)

Tempting as it is to dispose of a complex and disturbing subject by means of ad hominem ridicule, such a method of argument is unworthy of scholars. (Crews 1975: 6)

On the meaning of complement-as-constructions

315

(18)

Only one more week of nonstop preparation, lecturing, tutoring, marking. Interesting as it was, you could only keep up the pace for so long... (Lodge 1989: 257)

(19)

But, frightening as it is, the market collapse is the alarm, not the fire. (New York Times Editorial, Oct 20, 1987.)

I only counted two participles that were coordinated with bona fide adjectives as adjectives, as in (20): (20)

"Shall we go in? " Kate asked. "I want to ask endless questions, of course, but think perhaps I had better seize this social chance to meet a few more of the faculty, delightful and challenging as they all are." (Cross 57)

Furthermore, I classified three past participles as adjectives, viz. tired, old-fashioned, absent-minded, again bending over backwards in order to avoid skewing the data. In spite of this very conservative method of classification, the distribution, as displayed in Table 5, came out as highly significant (p < 0.001, chi-square 26.93, 4 d.f.) Adjectival complements thus strongly favour the concessive reading, and past participles a causal reading. All four present participle examples had concessive readings, something which could be due to their adjectival character in the present contexts. I will return to these findings in the final discussion. Table 5.

The distribution of concessive, causal and concomitant readings over complement-as-constructions with bona-fide adjectives, present and past participles as complements. Row percentages except in totals column.

Complement type Adjective Past participle Presort participle Totals

Concessive 74% 25 0 4 100% 29

Concomitant Causal 6 18% 3 9% 12 80% 3 20% 0 0 18 6

Total 34 64% 15 28% 4 8% 53

The remaining factor groups-subject type, inversion, and verb type are closely related, in that inversion only occurred with nominal subjects and BE as main verb, as in (15) above and (21) below: (21)

Distinctively resonant as is the voice of Hill's poems, they are consistently impersonal. (The Norton Anthology, p. 2421)

316

Gunnel Tottie

For these three factor groups, I pooled the causal/concomitant readings to improve the odds for significance, but this was not achieved at the 0.05 level; p. was 0.1 for all three factor groups. Table 6 shows the distribution of nominal and pronominal subjects, demonstrating that nominal subjects are indeed in the minority, only 10/53 or 19%. However, although the proportion of nouns is higher among the examples with a concessive reading (28%) than among those with a causal or concomitant reading (8%), the difference is not significant. A larger dataset may well produce significance here. Table 6. The distribution of concessive vs. non-concessive readings over complement-ay-constructions with nominal or pronominal ay-clause subjects. Column percentages only. Subject of ay-clause Noun Pronoun Totals Table 7.

8 21 29

28% 72%

Causal + concomitant 2 8% 22 92% 18

Total 10 43 53

19% 81%

The distribution of concessive vs. non-concessive readings over complement- «^-constructions with nominal or pronominal ay-clause subjects. Column percentages only.

Subject of ay-clause +Inversion -Inversion Totals Table 8.

Concessive

Concessive 3 26 29

10% 90%

Causal + concomitant 0 24 100% 24

Total 3 50 53

6% 94%

The distribution of concessive vs. non-concessive readings over complement-as-constructions with BE vs. other verbs in as-clause. Column percentages.

Subject of ay-clause +BE -BE Totals

Concessive 26 3 29

90% 10%

Causal + concomitant 24 100% 24

Total 50 3 53

6% 94%

At least in the case of subject-verb inversion this could indicate that we have to do with a knockout constraint, and that the grammar of English does not allow inversion in causal complement-as-constructions.

On the meaning of complement-as-constructions

317

4. Discussion It is clear that the interpretation of as as a concessive or causal conjunction is conditioned by a number of different factors. One factor investigated by Kjellmer yielded stronger effects in the present study, viz. the position of the complement-as-construction. It could be shown that a concessive reading was favoured if the complement-as-construction was initial or medial, and a causal reading if it was final. However, the most interesting result of my investigation is certainly the finding that the category of the complement is correlated to the meaning of the ay-clause. Adjectives thus clearly favour concessive readings, whereas past participles favour causative and/or concomitant readings. I would like to suggest that this has to do with the history of complement-asconstructions, which must originally derive from two quite distinct constructions that have now merged and can therefore cause ambiguity. There can be no doubt that the concessive complement-as-constructions are truncated versions of correlative as...as constructions, as maintained by OED2 and Jespersen. However, the causative constructions can hardly have the same source. It seems to me that they have developed from originally concomitant constructions, probably mostly non-finite clauses of the types exemplified in (22), adapted from the original sentence in Quirk et al (1985: 15.47) quoted in section 2 above. (22a)

Tired, they went to bed as soon as they came back.

(22b)

They went to bed as soon as they came back, tired.

Clearly, both (22a) and (22b) must have concomitant or causal readings. The (a)-example definitely sounds more natural than the (b)-example, but notice that if an as-construction is added to (22b), it sounds better, probably because it conforms to the normal pattern of end-weight in English: (22c)

They went to bed as soon as they came back, tired as they were.

Constructions such as those in (22c) with sentence-final complementas-constructions having concomitant/causal readings would thus appear, giving rise to the present situation with ambiguous complement-asconstructions. It is then also easy to imagine that they will begin to be used in sentence-initial position, but that the original position will remain the most frequent one. This hypothesis is supported by my findings concerning the tendency of end-position and past-participle complements to favour

318

Gunnel Tottie

causal readings of complement-as-constructions and to disfavour concessive ones. Constructions with present participles as complements remain problematic. We have seen that when the participles are used adjectivally in the present sample, such constructions all had concessive readings. However, when the present participle is clearly used as a verb, as in (23) - also adapted from Quirk et cd. (1985: 15.47) - it can also have a causal reading: (23a)

Writing hurriedly, she didn't notice the spelling errors.

(23b)

She didn't notice the spelling errors, writing hurriedly.

Notice that here too, an as-clause fits well in final position: (23c)

She didn't notice the spelling errors, writing hurriedly as she was.

We need much more research based on a wide variety of data from different periods of English to arrive at a full understanding of complementas-constructions, but on the basis of available evidence, it seems to me entirely plausible that a process like the one I have sketched here can account for much of the ambiguity of these constructions.

Notes 1. I would like to thank Göran Kjellmer and Andreas Fischer for valuable comments on an earlier version; they are in no way responsible for my conclusions, however. I also thank Morton D. Paley, Sebastian Hoffman and Hans-Martin Lehmann for interesting discussions and the latter two also for technical assistance. 2. Kjellmer (1992) does give the distribution of concessive, causal and concomitant readings in his corpus but not of the different factors conditioning those readings. 3. However, they do point out (in their discussion of comparative clauses in 15.71) that Cautious as the rest of the family was could be either a "concessive or circumstantial clause with fronting of the subject complement;" however, the example they give is followed by the matrix clause she seemed willing to give an immediate reply to my question, which would only permit a concessive reading. 4. Cf. also the following example from Fasold (1990:26): As tedious as the completion of the questionnaire no doubt was, the results show that the respondents, for the most part, completed it carefully. My impression is that the full as...as construction is more current in American English than in British English, but I have as yet no quantitative data to support this observation.

On the meaning of complement-as-constructions 5.

6.

7. 8.

9.

319

References in parentheses indicate the numbering of examples in Kjellmer (1992) if taken from that work. When applicable, Kjellmer's source is indicated as either the Brown or LOB Corpus. When examples derive from my own corpus, the name of the author or periodical precedes a page reference. As indicated by ellipses, I have sometimes abbreviated Kjellmer's examples. Kjellmer also observes that there was little difference between the American and British corpora. This is interesting, as Quirk et al. (1985: 15:39) mark the concessive as-clause as a Briticism. The British data include one example from the Australian author Patrick White, who grew up in England. I found no instances of merger or indeterminacy of meaning in my data. I use the terminology and method of coding employed by variationists; cf. Rand and Sankoff (1990). However, because of the number of knockout constraints on the factors involved, I will refrain from a variable rule analysis and limit myself to chi-square tests. Complements can also be nouns or adverbs, but I will not be concerned with those very unusual types here.

References Primary sources for complement-as-constructions Barnes, Julian 1985 (1984) Flaubert's parrot. London: Picador. Broad, William and Nicholas Wade. 1985 Betrayers of the truth. Fraud and deceit in science. Oxford: OUP. Bradbury, Malcolm 1984 (1975) The history man. London: Arena. Christie, Agatha 1948 Three blind mice and other stories. New York: Dodd, Mead & Co. Cott, Allan 1975 Fasting: The ultimate diet. New York: Bantam Books. Crews, Frederick 1975 Out of my system. Psychoanalysis, ideology, and critical method. New York: Oxford University Press. Cross, Amanda 1985 (1984) Sweet death, kind death. New York: Ballantyne. Bell, Clive 1972 Virginia Woolf. A biography. New York: Harcourt Brace Jovanovich. Gilbert, Sandra and Susan Gubar 1979 The madwoman in the attic. New Haven & London: Yale University Press.

320

Gunnel Tottie

Gowers, Sir Ernest 1986 (1954) The complete plain words. Third ed., revised by Sidney Greenbaum and Janet Whitcut ed. London: Her Majesty's Stationery Office. Kitson Clark, George 1967 The critical historian. London: Heinemann. Lively, Penelope 1983(1977) The road to Lichfield. Harmondsworth: Penguin. Lodge, David 1989 (1988) Nice work. London: Penguin. Lurie, Alison 1985 (1984) Foreign affairs. New York: Avon. 1986a (1962) Love andfriendship. London: Abacus. 1986b (1974) The war between the Tates. New York: Avon. Millett, Kate 1978 (1970) Sexual politics. New York: Ballantine. Mortimer, John 1988 Summer's lease. London: Penguin. The Norton Anthology of English Literature 1981 5th ed., vol. 2. New York: Norton. Pym, Barbara 1980 (1978) The sweet dove died. London: Grafton Books. Rendell, Ruth 1970 (1968) The secret house of death. New York: Ballantine. Raskin, Barbara 1988 (1987) Hot flashes. New York: St. Martin's Press. Woolf, Leonard 1964 (1963) Beginning again. New York: Harcourt Brace Jovanovich. White, Patrick 1973 (1970) The vivisector. Harmondsworth: Penguin. The New York Review of Books, The New York Times, Time Magazine, The [London] Times.

Other primary sources Fasold, Ralph 1990

The sociolinguistics of language. Oxford: Blackwell.

Secondary sources Jespersen, Otto 1927 A Modern English grammar on historical principles. Part III. Heidelberg: Carl Winters Universitätsbuchhandlung.

On the meaning of complement-as-constructions

321

Kjellmer, Göran 1992 Old as he was: A note on concessiveness and causality. English Studies 74: 337-350. OED 1989 The Oxford English Dictionary. 2nd edition. Oxford: Oxford University Press. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman. Rand, David and David Sankoff 1990 GoldVarb. Version 2. A variable rule application for the Macintosh. Centre de recherches mathömatiques, Universiti de Montr6al.

Eyes and ears Werner Winter

The vocalism of the Germanic forms for 'eye' such as Gothic augo, Old Norse auga, Old High German ouga, Old English eage does not correspond veiy well with that of words such as Old Church Slavic oko, Latin oculus, Tocharian Β ek, Tocharian A ak 'eye', Greek osse 'eyes'. While occasional attempts have been made to explain the Germanic diphthong as a regular development (cf. Seebold 1999: 65), the more common solution (cf., e.g., Buck 1949: 225, Pokorny 1959: 777) is to consider it as reflecting the influence of the term for 'ear' which did contain an original diphthong as shown, e.g., by Greek ous, Latin auris, Gothic auso, Old Church Slavic uxo (cf., e.g., Pokorny 1959: 785). The assumption of an analogical reshaping is eminently plausible as it is well known that the principal locus of such reshaping is the paradigm, be it grammatical or semantic (with even the grammatical paradigm in essence one characterized by shared semantic properties; for a more detailed discussion in the context of a tradition established by Hermann 1931 see Winter 1969 and the evaluation in Anttila 1977: 73). The paradigms may be seen as configurations in mental storage which seem to occur alongside storage by form only, but this question cannot be further pursued here. Semantically delimited paradigms may be relatively open-ended, such as the numerals in very many languages, or they may be closed sets with possibly just two members. In a number of Indo-European languages such sets may undergo a development that strikes one at first as rather odd: if in a speech community the identity of, say, a second member of a pair of terms can be surmised with sufficient certainty once the first member is overtly given and the use of a dual form indicates that a pair and not an individual is referred to, the second member of the pair does not have to be named: "X[dual]"can thus be interpreted as a set of two semantically related items "X and Y"; such elliptic duals (and plurals) differ significantly from "normal" duals where "X[dual]" stands for "X and X, 2 χ X". A concise yet highly informative discussion is found in Schwyzer (1950: 50-51 with notes); the author describes a reducible "X and Y" sequence as an "ungleichartiges,... natürliches und ständiges Paar". Notable examples are Vedic pitärau 'father[dual]' and Vedic mätärau 'mother[dual]', both de-

324

Werner Winter

noting 'parents'; parallels are Latin patres 'father[plural]' = 'parents' or Gothis berusjos 'those who have borne' = 'motherfplural]' = 'parents' (cf. Lehmann 1986: 66-67 with further references and discussion). Modern cases in question are such rather formal expressions as the William Joneses where the plural form does not refer to more than one William Jones, but to Mr. William Jones and Mrs. William Jones. A sequence "eye and ear" would seem to qualify as such a nonidentical, yet natural pair and hence a semantically delimited paradigm; it makes sense to consider it a natural locus for the operation of analogical reshaping so that the postulated Germanic transfer of the diphthong in 'ear' to 'eye' falls well within the range of the expectable. However, so far no evidence has been found to indicate that the "X and Y" of 'eye' and 'ear' was ever reduced to "X[dual]" or "Y[dual]". It seems worthwhile to have, in this connection, a closer look at the Classical Armenian terms for 'ear' and 'ears' which form a highly irregular paradigm: the singular is Armenian imkn (ownkn), the plural Armenian akanjk'. Armenian unkn has long been considered to owe its -kn to the influence of Armenian akn 'eye'; the remaining un- can rather easily be explained as a form inherited from Proto-Indo-European (for a summary of relevant proposals cf. Olsen 1999: 175-176). Concerning Armenian akanjk', Olsen states that it "remains enigmatic, and the relation between sound change and analogy in this case is still an open question". Greppin (1983 [1985]), in a contribution not mentioned by Olsen, says (p.264): "The origin of this plural of unkn 'ear' is unclear" and goes on to discuss suggestions made by Pisani (1950) and Dumezil (1947); neither of them comes up with satisfactory answers. Short of giving up hope for ever finding a way out of this dilemma, it seems appropriate to reconsider the problem of Armenian akanjk' in the context of the material mentioned above. Armenian ac'k' 'eyes' has traditionally been considered a (pluralized) continuation of a dual form parallel to Greek osse (cf. Hübschmann 1897 [1962]: 413, with references); Olsen (1999: 171), who finds it difficult to accept the claim that it was the reflex of Proto-Indo-European *-yE that triggered the palatalization of the underlying labiovelar in the Armenian term for 'eyes' (cf. Winter 1965: 104), considers it "quite possible to imagine that the paradigm was influenced by the neuter 'ear', Arm. pi. akanjk', -aca not very convincing proposal as it attempts to explain obscurum per obscurins. The plural form Armenian ac'k' occurs only when the noun 'eye' is used in its basic function - whenever it denotes 'spring' or 'gem', the plural is derived from the full (-«-bearing) singular stem, which results in

Eyes and ears 325

forms such as Armenian akank\ akans. A hypothetical dual form of a stem akan- would be akanj-*; a pluralization would yield attested Armenian akanjk'. The all-too-obvious difficulty at this point is that the Armenian form does not mean '(pair of) eyes', but '(pair of) ears'. Is there a way out? A paired expression 'eye and ear' would contain (or have contained) a sequence Armenian akn plus imkn (or their possible antecedents *ak' and *uri). Ellipsis would reduce the sequence to either 'eye[dual]' or 'ear[dual]'. If dualization was applied to the singular stem, 'eye[dual]' would become *akan-[dual], which means - provided the sequence of *-n-yE resulted in Armenian -nj (cf. Olsen 1999: 176) - that secondarily pluralized Armenian akanjk' would have had the meaning 'eye and ear' to start with. In spite of its developmental history, Armenian akanjk' was, however, no longer necessarily identified as 'eye[dual]' > 'eye[plural]', since the dual (> plural) of 'eye' was firmly established in the inherited form Armenian ad'k'. As the meaning 'eye plus ear' could equally well be conveyed by 'eye[dual]' and 'ear[dual]' the form Armenian akanjk' could be interpreted, by default as it were, as elliptic 'ear[dual]', which opened the door for a use of Armenian akanjk' as a nonelliptic dual (> plural) form 'ears'. One may speculate that such a development, if it took place in the way assumed here, was facilitated by a massive difference between the more regular earlier forms of 'ear' and 'ears': *un underlying attested Armenian imkn and its dual counterpart *uhanj showed a pattern without obvious parallels. Compared to an analogical reshaping in the context of a semantically delimited paradigm (as assumed for the development from *un to attested Armenian unkn 'ear' under the influence of the term for 'eye'), the postulated semantic change from '*eyes' to 'ears' appears to be much more radical. To make it plausible at all, the special conditions of elliptic reduction of closed sets must have existed. The interpretation of Armenian akanjk' proposed here then provides indirect evidence, not surprising in itself, of paired nouns so closely related to one another that upon introducing the first one, the second could be predictable to such an extent that it could be omitted. The [noun etc.] type can thus be assumed with a great deal of confidence to have existed in the history of Armenian alongside copulative compounds such as Armenian elewejk' 'ascent and descent' and ert'ewek 'coming and going' (cf. Winter 1998: 631).

326

Werner Winter

References Anttila, Raimo 1977 Analogy. (Trends in Linguistics 10). The Hague - Paris - New York: Mouton. Buck, Carl Darling 1949 A dictionary of selected synonyms in the principal Indo-European languages. A contribution to the history of ideas. Chicago - London: University of Chicago Press. Dumözil, Georges 1947 S6ries ötymologiques armöniennes. Transactions of the Philological Society. 64-75. Greppin, John A. C. 1983 [1985] An etymological dictionary of the Indo-European components of Armenian. Bazmavep 141: 235-323. Hermann, Eduard 1931 Lautgesetz und Analogie. (Abhandlungen der Gesellschaft der Wissenschaften zu Göttingen, Philologisch-historische Klasse, Neue Folge 20.3). Berlin: Weidmann. Hübschmann, Heinrich 1962 Armenische Grammatik 1: Armenische Etymologie. Darmstadt: Wissenschaftliche Buchgesellschaft. First published Leipzig: [1897] Lehmann, Winfred Philip 1986 Α Gothic etymological dictionary. Based on the third edition of Vergleichendes Wörterbuch der gotischen Sprache by Sigmund Feist. Leiden: E.J.Brill. Olsen, Birgit Anette 1999 The noun in Biblical Armenian. Origin and word-formation - with special emphasis on the Indo-European heritage. (Trends in Linguistics, Studies and Monographs, 119). Berlin - New York: Mouton de Gruyter. Pisani, Vittore 1950 Studi sulla fonetica dell' armeno. Ricerche linguistiche 1: 165-193. Pokorny, Julius 1959 Indogermanisches etymologisches Wörterbuch. Volume 1. Bern München: Francke. Schwyzer, Eduard 1949 Griechische Grammatik. Auf der Grundlage von Karl Brugmanns Griechischer Grammatik 2: Syntax und syntaktische Stilistik. München: C. H. Beck'sche Verlagsbuchhandlung. Seebold, Elmar 1999 Kluge: Etymologisches Wörterbuch der deutschen Sprache. 23rd edition. Berlin - New York: Walter de Gruyter.

Eyes and ears

327

Winter, Werner 1965 Armenian evidence. In: Werner Winter (ed.), Evidence for laryngeals. (Janua Linguarum, series maior 11), 100-115. The Hague - London - Paris: Mouton. 1969 Analogischer Sprachwandel und semantische Struktur. Folia Linguistica 3: 29-45. 1998 Cases of cross-over between finite verb forms and nouns in Armenian. In: Mark Janse (ed.), Productivity and creativity. Studies in general and descriptive linguistics in honor of E.M. Uhlenbeck. (Trends in Linguistics, Studies and Monographs, 116), 629-632. Berlin - New York: Mouton de Gruyter.

Subject index abbreviations 62, 64,200 abstract meanings 1,4 abstract noun 49, 51,282-283 abstraction(s) 4,28-29,49, 145, 150 accord 144, 250-254, 258-259 according to 41, 184,249-260 adjective(s) 8, 37-38, 55, 59-64,6872, 77,84, 138, 180-182,213, 280-284,314-315 - attributive 46-49, 52, 206 - bona fide adjectives 206,312, 314-315 - gradable 55 -non-attributive 48,51-52,206 - non-gradable 55 agent suffix 1,66 agglutinating languages 70-71 agglutinating morphology 59 alliteration 12 ambiguity 224, 299-300, 307, 309, 317,318 analogical pressure 301 analogical reshaping 323, 324, 325 anti- 36-41 archetype(s) 146-147,179 as-clause 307-319 attribute 11,48 attributive 40 see adjective(s) beck 87-91 benefactive datives 292 bleaching 10,27 body 11,45-54,103,170 bona fide adjective see adjective(s) borrowed vocabulary 249 by-names 8, 87 cardinal point(s) 134-137

case 7-8,70,72, 106-107, 111-116, 119-121, 127, 138, 153,183-184, 232,300 -common case 112-113 change from above 249, 259 clitic(s) 112,114,121,177-181, 184,289-292,296-303 - special clitics 289,291,297, 301 cognitive 29, 60, 144, 146, 149, 179, 183 colligation(s) 45-49, 57 collocation(s) 46, 66, 78, 85, 100, 156,231 colloquial speech 236-237 colour verbs 75-85 combining form(s) 36,281 combining word(s) 279,281 common case see case complement semantics see semantics complement-ay-construction 307, 309,313,314,317 complement category 314, 317 compositional semantics see semantics compound premodifier(s) see premodifier(s) compound(s) 9, 10,35-36,41,62, 99, 102, 115, 189, 198,207,212, 214,280-283,286 - copulative compounds 325 - hyponymic compound(s) 280 compounding 10, 59-60, 105,200, 205 concessive 307-319 concrete meanings 1,4 conditionality continuum 221,223 +/-confined 132-133 connotation(s) 165-167, 172-173

330

Subject index

constituent order 180 conversion^) 9-10, 12,76 co-ordinated clauses 291 co-ordinating conjunction 281 copulative compounds see compound(s) corpus 36,45,48^19, 54, 56, 80-81, 208,238,243,246,269,273,313, 318-319 court records 159 -cum- 277-284 decliticisation 290-301 delimited paradigm(s) 323-325 denotation(s) 8, 11, 166-173 denotative meaning 167 denotatum 53, 168-169, 171 deontic necessity 220 derivation(s) 8-10,14, 59,62-63, 76-77, 165-169,173,200 - expressive derivation 165 - prototypical derivation 59 derivational - derivational class 69-72 - derivational microclass 68 - derivational paradigm 68 desemanticisation 19,28-29 determinative phrase 111, 121 determinatum 189,198,280,284 determiner(s) 7-8, 10, 14, 114-116 dictionary/-ies 91,96, 138, 141, 147-150, 188,277,279-284 diminisher(s) 55 diminutive(s) 10, 71,165-173, 199, 286 - diminutive derivation 166-167, 170 direction 125-127, 130, 135-138 distinctive feature(s) 96, 127 Domesday Book 87 double modals see modals

duilem 1-4 dummy subjects 29 dynamic lexicology 189,200 -(e)d 236,244 ellipsis 104-105,280 emotions 49-51, 157, 172 empathy 166 -en 76-77, 303 enclitic 121 epistemic modals see modals eponym(s) 9, 196-197 -(e)s 70, 112,235-244,263-273 -(e)th 235-243, 263-266,273 existential sentences 22-23 expressive derivation see derivations) expressive language 170 extension(s) 144 - semantic extension 168-171 field of terms 99 folk etymology 105 formal style see style(s) fossilisation 101 functional approach 60 general case see common case genitive(s) 111-122,180-182,235236,241,244-245 - classifying genitive 115-116, 121

- group genitive(s) 113-114 genre(s) 205-212,249,252,257259 grammaticalisation 20,26-29,212213,249-253,257,258,260 group genitive(s) see genitive(s) head-parameter 179, 182-183 Aw-construction 112, 114, 117, 120

Subject index historical semantics see semantics hyphenation 206-210,214 hypocoristics 165,167,171-173 hyponymic compounds see compound(s) hypostatisation 286 idea(s) 142-150, 184 impropriety 54 inferential feature(s) 45, 56 inflecting-fiisional languages 59, 7071 inflection(s) 59-61,71,111-114, 177,235-238, 242,299,303 - inflectional morphology 62, 70, 244 - nominative genitive and plural inflections 237 - noun inflections 178 - third person present plural 263 - third-person singular 235-236, 238,243,244 see also -(e)d, -en, -(e)s, -(e)th information condensation 286 institutionalisation 197,200,284 intension 144 it - existential it 23-25, 30 juxtaposition(s) 284,297 lexeme 45,47, 190,238,242,269, 286 lexical - lexical diffusion 242 - lexical enrichment 59,60 - lexical operation 285,286,287 -lexical unit 115-117, 121, 190, 285-287 lexical semantics see semantics

331

lexicalisation(s) 67, 104, 167, 197, 200,284 lexicography 141, 149-150 lexicology 187,259 lingua Adamica 141, 146 loan-processes 189 loanword(s) 63, 98, 102, 106,257, 258 location 27-29, 125, 128-129, 132, 136-138, 176 locative there see there ludic view or function 195 mark(s) 143,319 maximiser(s) 54 meiosis 166 mental 28, 144, 149, 323 - mental lexicography 147, 149 - mental representation(s) 143, 146 - mental space 28-29 metaphor(s) 45-46,49,56, 103, 166, 189,197 metaphorical extension(s) 28-29, 97, 172 metonymic shift 285 metonymy 101-104,189, 197 modal phrases 220 modals - double modals 217-231 - epistemic modals 220,224 morphological complexity 71 morphological richness 59, 64-67, 70 morphology 59-60,71-72, 76-77, 125,137, 175,235,263,269,273, 303 morphophonemic 235-236,239,245 morphopragmatic 166 morphosemantic 59-60, 68, 166 motivation 60, 187

332

Subject index

name(s) 7-15,39,41,47, 57, 65, 80, 87, 88,91, 106, 116, 142, 145149, 166,171-173,187-200,246, 278,319 -nickname(s) 10-11,13,192 - proper name(s) 72, 165, 171, 190, 192 naming 11-12,35,142, 147, 206207,284,286 Natural Morphology 59, 61 negation 217,219-220,231 neologisms 63-64,277,279 news 188,205,208-213,280 newspaper 117,119,205-212 nominal genitive and plural inflections see inflection(s) non-attributive see adjective(s) non-gradable see adjective(s) Northern Present Tense Rule 264266,269-273 notional features 7 numismatic 12, 15 object pronoun(s) see pronoun(s) see pronoun(s) object(s) - prepositional objects 292 - pronominal objects 289,291293,295,296,299,301 observational linguistics 194 o/-phrase 117-119 onomasiology 142, 146,149 onomastic(s) 11-12,87,187,189, 192,200 orthoepists 236-238,241,244 orthographic evidence 238 ostensive definition 46-47 paradigmatic relation(s) 59, 97

past participle 244, 312, 314,315, 317 person 7, 85, 153, 156-157, 178, 263, 266,298,299 personal pronoun(s) see pronoun(s) phrase marker 112,114,119,120 politeness 154-160,224-225,230 - politeness theory 156 polysemy 64,65, 101,223 positioning of pronoun(s) see pronoun(s) possessive 112, 113, 117,120, 121 possessive pronoun(s) see pronoun(s) postposition(s) 113, 121, 179, 184, 292,302 power semantics see semantics pragmatics 153,165, 172,189,230 prefix(es) 36,40,134,281 premodifier(s) - compound adjectival premodifier(s) 205-212 preposition(s) 37,112, 121,134— 135, 138,179,250-260,279-280, 300, 302-303 prepositional use 252,255-257,259 prepositionalisation 249, 253,255, 258 present participle 206,250,252,311, 314-315,318 prestige 99, 100 privative 64,69 pro- 35-41 productive rules 60 productivity 59,61-63,70-71,77, 161,184,207-208,211-213 prolexematisation 286-287 pronominal positions 289,297 pronominalisation 286

Subject index pronoun(s) 8,29, 153, 155-160, 178, 182,184,263-270,289-299, 302303,311-312 - demonstrative prounoun(s) 132 - object pronoun(s) 290-296,299, 302 - personal pronoun rule 264 - personal pronoun(s) 153-154, 263-266,269-273,289,291, 302 - positioning of pronouns 290 - possessive pronoun(s) 112,120 - pronoun(s) of address 153,159 - relative pronoun(s) 243,271 prototypical meaning 112,117,121, 249 prototypical derivation see derivation^) proximity 129,131 quantifiers 181-182 quirky subjects 292 reanalysis 26-27,29,112,161,235, 253-254,260, 272,281,296, 303 referent(s) 7, 9, 11,26,45,47,97, 100-101, 103,127,129,142, 190, 193,226,250-251,259,267,284 referential function 60, 141-142, 145 registers) 102,176,180,236,238, 242 relative clause(s) 180,182,184, 205-206 relative construction 206 relative pronoun(s) see pronoun(s) re-motivation 197 representations 8, 126, 145, 146 root(s) 168,171,220-222,231-232, 257,299,303 rule inversion 244

333

's phrase 115-122 semantic - semantic change(s) 4, 95, 96, 99-107,168,325 - semantic feature(s) 126-127, 130, 134, 144, 156-157,224 - semantic field(s) 96, 98, 106, 167,169-171, 173, 198 - semantic narrowing or restriction 155 - semantic range 78,101 - semantic recategorisation 286 - semantic reconstruction 100 - semantic relation(s) 20, 37, 95, 117,141,283 - semantic shift 285 semantics 19-20,22,27-29, 76, 78, 98, 107,141-144, 146, 149-150, 154, 165, 170, 172,217-218,221, 228,230,269 - complement semantics 312 - compositional semantics 145 - historical semantics 1, 153,285 - lexical semantics 9, 66, 96 - power semantics 154 - prototype semantics 96,99 - solidarity semantics 154 semasiology 146 semiotic(s) 60, 95, 141, 146, 190, 195,199 sentential head 218 sign(s) 60-61,95,97,141, 143-145, 148-149, 166, 190, 195,200 solidarity semantics see semantic(s) special clitics see clitic(s) specialised vocabularies 78 specification 101-103,285 speech act(s) 165,187,200,223 spelling evidence 236 +/-stationary 130-135,138

334

Subject index

style(s) - formal style(s) 236,308 stylistics 100,205 subject 19, 22,24—29,46,48-49, 118,178,221,226,254,264-266, 268-273,286,291-292,297, 300, 303,308,312,318 subjectification 20, 26,27,29 subjective relation 117 suffices) 1, 2, 10, 36-38,62-65, 67, 69, 70-72, 76-77, 120, 129, 136, 167,169,172-173, 195,198-199, 212-213,235-246,270,281 surface syntax 218 surrogationalism 142 syncopated -s in the present indicative 237 syncope 236,238,241-242,244 synonym(s) 80,95,98-101 synonymy 66-61, 70 syntactic - syntactic invisibility 289 - syntactic recategorisation 286 - syntactic reduction 205 syntagma(s) 190,192 system adequacy 60-61, 72 terminologies 102 text cohesion and condensation 207 the which 243 there existential there 19 locative there 19 thou 153-161,256,263

translation^) 1,2, 96-97,99, 101, 111, 114,147,253,256,266,299 transparency 59, 105, 125, 137 typological adequacy 59-61 typological change 175 unitary head-position principle 183 universal preferences 59,61 variation 155, 180, 184,208,219, 220,231,235,239,246,263,269, 273,299 - dialect and register variation 236 verb second constraint 22, 30 vertical direction see direction visual sign(s) 195 VO/SI construction 299 vowel deletion 235-237,244 vowel insertion 235,244 word classes 7-8 word family/-ies 68 word order 111-112,122, 142,253, 291,295, 301 word-formation 1,35-38,59-71, 115, 187, 189-190, 192, 195, 197, 198-200,206,213,277,279,281, 284-286 - definition of word-formation 284 - fiinction(s) of word-formation 72,206,277,284-287 - word-formation nest 68 word-geography 87