English Historical Linguistics and Philology in Japan 9783110808773, 9783110157918

366 118 14MB

English Pages 473 [476] Year 1998

Report DMCA / Copyright


Polecaj historie

English Historical Linguistics and Philology in Japan
 9783110808773, 9783110157918

Table of contents :
English historical linguistics and philology in Japan 1950–1995: A bibliographical survey
On identifying Old English adverbs
On the construction I was go walked
Reflexive verbs in Chaucer
MS Cotton Nero A.x. poems once again: A study of contracted negative forms
An approach to the language of Criseyde in Chaucer’s Troilus and Criseyde
(Pro-) Nominal reference in Old English and the origin of the that-clause
Some notes on the grammatical category of gender in Vices and Virtues
Verbal gerund and its historical development in English
What is the point? Manuscript punctuation as evidence for linguistic change
Finite and non-finite clauses in the English of Alfred’s reign: A study of syntax and style in Old English
Middle English Breaking
Syntactical revision in Wulfstan’s rewritings of Ælfric
On double auxiliary constructions in Medieval English
The development of Middle English ī in England: A study in dynamic dialectology
A new rhyme concordance to Chaucer’s poetical works
Old English verbs of possessing
On the inseparable nature of verb-auxiliary combinations in Old English
The gerund in Chaucer, with special reference to its verbal character
Archaism in the vocabulary of Ælfric
The syllable structure and phonological processes in the history of English
The helle sequence in Old English poetry
Some etymological and semasiological notes on girl
On the functional motivation of phonological changes in English
On the productivity of the suffixes -ness and -ity: The case of Chaucer
Index of names

Citation preview

English Historical Linguistics and Philology in Japan


Trends in Linguistics Studies and Monographs 109


Werner Winter

Mouton de Gruyter Berlin · New York

English Historical Linguistics and Philology in Japan

edited by

Jacek Fisiak Akio Oizumi

Mouton de Gruyter Berlin · New York


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

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

Library of Congress


English historical linguistics and philology in Japan / edited by Jacek Fisiak, Akio Oizumi. p. cm. - (Trends in linguistics. Studies and monographs ; 109) Includes bibliographical references and index. ISBN 3-11-015791-8 (alk. paper) 1. English philology - Old English, ca. 450-1100. 2. English philology - Middle English, 1100-1500. 3. English philology - Japan - Bibliography. 4. English language Grammar, Historical. 5. Manuscripts, Medieval - England. I. Fisiak, Jacek. II. Oizumi, Akio, 1935- . III. Series. PE60.J3E54 1998 429—dc21 98-17771 CIP

Die Deutsche Bibliothek —


English historical linguistics and philology in Japan / ed. by Jacek Fisiak ; Akio Oizumi. - Berlin ; New York : Mouton de Gruyter, 1998 (Trends in linguistics : Studies and monographs ; 109) ISBN 3-11-015791-8

© Copyright 1998 by Walter de Gruyter GmbH & Co., 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: Werner Hildebrand, Berlin. Binding: Lüderitz & Bauer, Berlin. Printed in Germany.


English historical linguistics and philology have been flourishing in Japan for many years but only the names and works of relatively few of a large number of scholars have been known in the western world, notably of those who have published their works in Europe and the United States. Numerous valuable contributions, however, have not reached the international community of English scholars because they have been published either in the Japanese language or in some local journals of very limited circulation. The bibliographical survey of English historical linguistics and philology in Japan while opening the present volume gives good evidence of the development of the field since the 1950's and its dynamic expansion. An ever increasing number of monographs and papers on a wide range of topics from orthography and phonology to semantic, pragmatic and sociolinguistic issues has been published to significantly supplement the achievements of scholars from the remaining regions of the world. Japanese scholars have been known for the meticulous handling of details and great respect for language data which has resulted in superb lexical studies, concordances and dictionaries. The present volume includes twenty-four papers with the bibliography covering 1950-1995. The idea of bringing Japanese achievements to the attention of other scholars came into being during Prof. Fisiak's visit to Japan in 19951}, in discussions with Prof. Akio Oizumi of Doshisha University (Kyoto) and personal contacts with other Japanese scholars at their universities as well as at national conferences in Kyoto and Tokyo. Of course, Prof. Oizumi's active involvement in the project was at the heart of the successful completion of the project. Most of the papers in this volume have already been published in local journals in English or Japanese but have been revised and brought up to date for publication here. Some though have not appeared in print before. All of them basically cover Old and Middle English phonological, grammatical, semantic and lexical issues although a few papers at least address broader topics both synchronically and thematically. The importance of some contributions goes far beyond English historical linguistics and philology.



It is hoped that the volume is fully representative of the work done in the field in Japan and will introduce readers to a linguistic and philological community heretofore unknown to them but constantly advancing and offering new facts for study and fresh interpretations and reinterpretations of already established views. It may encourage readers to further explore Japanese work on English historical linguistics and philology. The editors of the volume have the pleasure to extend their words of thanks and appreciation to all those who were ready to help at all stages of the project. We owe a special gratitude to Prof. Werner Winter, series editor, for his guidance and critical remarks, to Miss Magdalena Grzelak, M.A., Mr. Jaroslaw Michalak, M.A. and Mr. Micha! Jankowski, M.A. for their work on preparing the manuscript for printing. Last but not least our thanks goes to Miss Aleksandra Hans, M.A. and Miss Sylwia Scheuer, M.A. who have proofread the final version of the manuscript. Poznan - Kyoto, June 1996

Jacek Fisiak and Akio Oizumi

Note u

Prof. Fisiak would like to express his gratitude here to the Japan Society for the Promotion of Science for a generous fellowship which allowed him to come to Japan in 1995 to undertake the project. Part of it was completed during his stay in Kyoto, Tokyo, Hiroshima and Fukuoka.




English historical linguistics and philology in Japan 1950 - 1995: A bibliographical survey Akio Oizumi On identifying Old English adverbs Yasuaki Fujiwara


On the construction I was go walked Masayuki Higuchi


Reflexive verbs in Chaucer Eiko Ito


MS Cotton Nero A.x. poems once again: A study of contracted negative forms Yoko Iyeiri


An approach to the language of Criseyde in Chaucer's Troilus and Criseyde Akiyuki Jimura


(Pro-) Nominal reference in Old English and the origin of the ί/ιαί-clause Koichi Jin


Some notes on the grammatical category of gender in Vices and Virtues Shigeaki Karakida


Verbal gerund and its historical development in English Osamu Koma


What is the point? Manuscript punctuation as evidence for linguistic change Tadao Kubouchi




Finite and non-finite clauses in the English of Alfred's reign: A study of syntax and style in Old English Kazumi Manabe


Middle English Breaking Toshio Nakao


Syntactical revision in Wulfstan's rewritings of iElfric Hiroshi Ogawa


On double auxiliary constructions in Medieval English Michiko Ogura


The development of Middle English Tin England: A study in dynamic dialectology Mieko Ogura, William S.-Y. Wang andL. Luca Cavalli-Sforza


A new rhyme concordance to Chaucer's poetical works Akio Oizumi


Old English verbs of possessing Shigeru Ono


On the inseparable nature of verb-auxiliary combinations in Old English Katsuhide Sonoda


The gerund in Chaucer, with special reference to its verbal character Matsuji Tajima


Archaism in the vocabulary of iElfric Schin 'ichi Takeuchi


The syllable structure and phonological processes in the history of English Michiko Terajima The helle sequence in Old English poetry Jun Terasawa





Some etymological and semasiological notes on girl Yoshio Terasawa


On the functional motivation of phonological changes in English Norio Yamada


On the productivity of the suffixes -ness and -ity: The case of Chaucer Hiroshi Yonekura


Index of names


English historical linguistics and philology in Japan 1950 - 1995: A bibliographical survey* Akio


1. Introduction The present survey does not pretend to be a complete one. In general, I have tried to list important works in a book or monograph form written in English by Japanese scholars, but I have not hesitated to include items in Japanese when they offered something uniquely valuable. Publications in Japanese are identified by asterisks after the names of contributors. Each item is accompanied by a brief description of its content and/or by a list of reviews. The items are numbered; in 50.1, 50 refers to 1950, the year of publication of the cited book. An index of Contributors (Authors and Editors) includes their dates of birth (and death), with the aim of providing a little more information about the authors. Our survey begins with Tadao Yamamoto's monumental work entitled Growth and system of the language of Dickens (published in 1950), which was the first international contribution from the Japanese academic world to English linguistics. For its excellence the Japan Academy Prize was awarded to Dr. Yamamoto in 1953, the first time ever presented to a scholar of English language and literature.

2. List of publications Abbreviations Rev. Review / reviews AEB Analytical & enumerative bibliography Anglica Anglica (The Anglica Society of Kansai Univ., Osaka) ArchL Archivum Linguisticum AS American Speech ASNS Archiv für das Studium der neueren Sprachen und Literaturen CHum Computers and the humanities EA Etudes Anglaises


Akio Oizumi

ES English Studies FoLH Folia Linguistica Historica GRM Germanisch-romanische Monatsschrift JEGP Journal of English and Germanic Philology MAE Medium Aevum MESN Medieval English Studies Newsletter (Tokyo) MLN Modern Language Notes N&Q Notes and Queries MLR The Modern Language Review NM Neuphilologische Mitteilungen Outline of English Linguistics Eigogaku Taikei [Outline of English Linguistics] (edited by Akira Ota) RES The Review of English Studies SAP Studia Anglica Posnaniesia SELit Studies in English Literature (The English Literary Society of Japan, Tokyo) SELitEN Studies in English Literature. English Number SLang Studies in Language SN Studia Neophilologica YWES The Year's Work in English Studies 50.1 Yamamoto, Tadao. 1950. Growth and system of the language of Dickens: An introduction to 'a Dickens lexicon'. Osaka: The English Philological Society of Kansai Univ. ii + 508 pp. Originally D.Litt. Thesis, Univ. of Tokyo, 1946. [This work was originally planned to be a general introduction to his unpublished Dickens lexicon.] 52.1 Yamamoto, Tadao. 1952. Growth and system of the language of Dickens: An introduction to 'a Dickens lexicon'. Revised edition, with An index by Chiaki Higashida & Michio Masui with Supplementary notes & corrections. Osaka: Kanasai Univ. Press. 508 + 8 + 4 + 70 pp. Rev.: Sten B. Liljegren, SN 25 (1953): 183-84; Hillis Miller, MLN 69 (1954): 439-42; Noel E. Osselton, ES 35 (1954): 29-31. 52.2 Yamaguchi, Hideo (ed.). 1952. Bibliographia lingvistica & anglistica being a classified list of books chiefly on the English language based on Prof. Shizuka Saito's collection. Compiled & edited with a guide for reading. Tokyo: Shinozaki Shorin. χ + iii + xiii + 117 pp. ['Foreword' and 'Guide for reading' are written in Japanese. A portion of this bibliography was earlier published by Sanki Ichikawa (ed.), Catalogue of the Library of Sanki

English historical linguistics and philology in Japan 1950 - 1995


Ichikawa. Part I: English and Comparative Philology (Tokyo, 1924. Privately printed, viii + 194 pp.).] 53.1 Sato, Masaru. 1953. A grammar of the dialect of West Riding: Historical and descriptive. Tokyo: Azuma Shobo. 5+3+2+151 pp. Originally B.A. Thesis, Univ. of Tokyo, 1945. 56.1 Taniguchi, Jiro. 1956. A grammatical analysis of artistic representation of Irish English with a brief discussion of sounds and spelling. Tokyo: Shinozaki Shorin. 2 + ν + ii + xii + 292 pp. Revised & enlarged edition with an Appendix, Tokyo, 1972. iii + xiii + xii + 419 pp. 58.1 Araki, Kazuo et al. (eds.). 1958. Studies in English Grammar and linguistics: A miscellany in honour of Takanobu Otsuka. Tokyo: Kenkyusha. χ + 419 pp. Rev.: Theodor S. Dorsch, YWES 39. 1958:18; Margaret M. Bryant, AS 34 (1959): 290-93. [T. Otsuka, Professor Emeritus of Konan Univ.; formerly Professor of Literature and Science of Tokyo Univ., Kwanseigakuin Univ., Kansai Univ. of Foreign Studies, Kyoto Univ. of Foreign Studies. This Festschrift is a miscellany of papers by thirty-one former students on the occasion of his sixtieth birthday. It also contains a bibliography of his own studies, most of which are in the field of English grammar, linguistics, and history of the language.] 61.1 Yamaguchi, Hideo. 1961. Essays towards English semantics. Tokyo: Shinozaki Shorin. iv + iii 479 pp. D.Litt. Thesis, Univ. of Tokyo, 1961. Second revised & enlarged edition with an Appendix, Tokyo, 1969. ν + iv + 684 pp. Rev.: Simeon Potter, ArchL 15 (1863): 223-25; Stephen Ullmann, Lingua 31 (1973): 71-73; Kikuo Miyabe* SELit 40,2 (1964): 256-58. 62.1 Maejima, Giichiro. 1962. Studies in the English reflexive verbs. D.Litt. Thesis, Kyoto Univ. Privately printed. 254 pp. [Related publications: "Musings on the English passive", Anglica 2, 4 (1956): 31-45; "Some notes on English medio-reflexive verbs", Anglica 3, 3 (1958): 101-26.] 63.1 Ikegami, Yoshihiko. 1963. ME 'dight': A structural study in the obsolescence of words. (Proceedings of the Department of Foreign Languages and Literatures, College of General Education, Univ. of Tokyo XI, 4). Tokyo: Univ. of Tokyo Press. 63 pp. Rev.: Erwin Mayer, Anglia 85 (1967): 423-27; Ladislav Zgusta, Linguistics 41 (1968): 128-29. 64.1 Kobayashi, Eichi. 1964. The verb forms of the 'South English Legendary'. (Janua Linguarum, Series Practica 15). The Hague: Mouton & Co. 87 pp. Originally Ph.D. Diss., Univ. of Michigan, Ann Arbor, 1961. Rev.: Michael C. Seymour, Lingua 16 (1966): 437-38; Kenneth R. Brooks, ES 51 (1970): 448-49. 64.2 Masui, Michio. 1964. The structure of Chaucer's rime words: An exploration into the poetic language of Chaucer. Tokyo: Kenkyusha. xxii +


Akio Oizumi

371 pp. Originally D.Litt. Thesis, Univ. of Tokyo, 1961. Third impression with corrections, 1989. Rev.: Guy Bourquin, EA 18 (1965): 301-302; Thomas A. Kirby, ES 51 (1970): 450-51: Tauno F. Mustanoja, NM 66 (1965): 262-63; Eric G. Stanley, N&Q n.s. 12 (1965): 389-90. Cf. item 88.2. 65.1 Araki, Kazuo. 1965. A study of Shakespeare's use of simple relative pronouns. Studies in English Literature. English Number 1965: 59-76. D.Litt. Thesis, Tokyo Univ. of Education, 1967. 65.2 Matsuda, Tokuichiro. 1965. A transformational analysis of the Old English 'Pastoral Care'. Ph.D. Diss., Indiana Univ. viii + ν + 301 pp. 65.3 Fukumura*, Torajiro. 1965. Eigo-Tai no Kenkyu [Studies in English Voice]. Tokyo: Hokuseido. iii + 391 pp. Originally D.Litt. Thesis, Univ. of Tokyo, 1961. Rev.: Kikuo Yamakawa*, SELit 43,1 (1966): 123-24. [The monograph is an enquiry into the grammatical and psychological concept of 'voice' in IE and Japanese, the nature of verbal action, the active and the passive voice in English and their interrelations. It ends with a historical survey.] 66.1 Ichikawa, Sanki. 1966. Collected writings of Sanki Ishikawa. Edited by Kotaro Ichibashi, et al. for the Intitute for Research in Language Teaching. Tokyo: Kaitakusha. χ + 239 pp. [This volume was compiled to celebrate the eightieth birthday of Dr. Ichikawa, Professor Emeritus of the Univ. of Tokyo and Chairman of the Administrative Board of the Institute for Research in Language Teaching.] 66.2 Matsunami, Tamotsu. 1966. "Functional development of the present participle in English. Part I: Native syntactic functions of the OE present participle, (I) & (Π)". (I): Collected papers in commemoration of the 40th anniversary of the faculty of letters, Kyushu Univ.: 315-48; (Π): Literary studies (Faculty of Letters, Kyushu Univ.) 63: 1-81. Originally D.Litt. Thesis, Univ. of Tokyo, 1965. [This monograph consists of five chapters: 1. Introduction; 2. The Germanic present participle; 3. Native syntactic functions of the OE present participle, I: Nominal function; 4. Native syntactic functions of the OE present participle, Π: Predicative function; 5. Syntactic trend of the OE present participle.] 66.3 Oizumi, Akio. 1966. A classified bibliography of writings on English philology and Medieval English literature, based on Akio Oizumi's collection. Tokyo: Nan' un-do. xiii + 108 pp. Revised & enlarged edition, Tokyo, 1968. xvi + 182 pp. Rev.: Reinard W. Zandvoort, ES 47 (1966): 482. 66.4 Ueda, Minora. 1966. A Study of the order of clause elements in the later parts of the Peterborough Chronicle, with special reference to constructional types. Ph.D. Diss., Univ. of Texas at Austin, viii + 311 + 2 pp.

English historical linguistics and philology in Japan 1950 - 1995


66.5 Yamaguchi, Hideo (ed.). 1966. A reader's guide to 'Anglia': A classified general index to volumes I-XLV with explanatory notes & quotations. Tokyo: Shinozaki Shorin. xiv + 320 pp. 67.1 Suzuki, Shigetake. 1967. The language of the 'Ancrene Wisse'. (Journal of Social Sciences and Humanities 57). Tokyo: Society of Social Sciences and Humanities, Tokyo Metropolitan Univ. viii + 147 pp. Rev.: Richard M. Wilson, YWES 48. 1967: 34; Fumio Kuriyagawa, Anglica 6, 4 (1968): 35-42; Kikuo Miyabe*, SELit 45,1 (1968): 138-39. 69.1 Ono*, Shigeru. 1969. Eigo Ho-Jodoshi no Hattatsu [The development of the English modal auxiliaries]. Tokyo: Kenkyusha. xi + 303 pp. D.Litt. Thesis, Univ. of Tokyo, 1971. 71.1 Ito, Eiko Tamano. 1971. Sentential complementation in Middle English and early Modern English: A study of linguistic change. Ed.D. Diss., Univ. of Michigan at Ann Arbor, iv + 125 pp. 72.1 Nakao*, Toshio. 1972. Eigoshi Π - Chu-Eigo [History of English Π: Middle English]. (Outline of English Linguistics 9). Tokyo: Taishukan. xli + 656 pp. Part of: D.Litt. Thesis, Tokyo Univ. of Education, 1975. Cf. item 78.1. Rev.: Haruo Iwasaki* SELit 50,1 (1973): 177-80; Eichi Kobayashi* English Linguistics 9 (1973): 119-32. [This book consists of ten chapters: 1. general survey; 2. handwriting, alphabet, spelling; 3. graphemics; 4. phonology; 5. dialects; 6. morphology; 7. syntax; 8. lexicology; 9. metrics; 10. stylistics.] 73.1 Miyake, Ko. 1973. A syntactical study of Chaucer's English. Tokyo: Chuo Koron Jigyo Shuppan. Privately printed, χ + 103 pp. Originally B.A. Thesis, Univ. of Tokyo, 1951. 73.2 Niwa*, Yoshinobu. 1973. Kodai-Eigo Doshi Settoji 'Ge-' no Kenkyu [Studies in Old English preverbal 'ge-']. Tokyo: Shohakusha. xviii + 223 pp. D.Litt. Thesis, Nagoya Univ., 1975. Rev.: Tsunenori Karibe*, SELit 51,1 & 2 (1974): 267-72. [Related publications: "The Proverb ge- added to niman in the OE gloss to the Lindisfarne Gospels", SELitEN 1966: 65-79; "On the collective meaning of OE Proverb ge-\ SELitEN 1974: 155-67.] 73.3 Takahashi, Sakutaro. 1973. The state of 'i'-umlaut in Early West Saxon. Ph.D. Diss., Northwestern Univ. ν + 183 + i pp. 74.1 Kato, Tomomi. 1974. A concordance to the works of Sir Thomas Malory. Tokyo: Univ. of Tokyo Press, xi + 1659 pp. Rev.: Robert W. Ackerman, CHum 9 (1975): 44-5; Yuji Nakao, SELitEN 1975: 121-31. [This computer-assisted concordance is based on the second edition of Vinaver's Works of Sir Thomas Malory (Oxford, 1967).] 74.2 Miyabe, Kikuo (ed.). 1974. A Middle English prose reader. Tokyo: Kenkyusha. lxxiii + 304 pp. [This reader contains specimen pieces beginning with Caxton's Preface to Eneydos and ending with the Peterborough


Akio Oizumi

Chronicles, in which the usual chronological order is reversed. It also provides notes, and a glossary with keys to pronunciation and brief etymologies, as well as a sixty-five page "Introduction to Middle English."] 75.1 Tsuchiya, Tadayuki. 1975. A concordance and glossary to the General Prologue of the 'Canterbury Tales'. Tokyo. Privately printed. 252 pp. 76.1 Ando, Sadao. 1976. A descriptive syntax of Christopher Marlowe's Language. Tokyo: Univ. of Tokyo Press, xxvi + 721 pp. D.Litt. Thesis, Nagoya Univ., 1978. Rev.: Reinard W. Zandvoort, ES 58 (1977): 77-79; Masanori Toyota*, SELit 55,2 (1978): 498-502. 76.2 Maeno, Shigeru. 1976. A Melville Dictionary. Tokyo: Kaibunsha. xi + 277 pp. Cf. item 84.3. [This is a dictionary of mythological, classical, Biblical, literary, religious, historical, and geographical references and allusions, and of capitalized hard words in the works of Herman Melville, designed to serve as a useful companion for the readers of Melville, especially for those readers whose mother tongues are not English and whose cultural traditions are quite different from Melville's.] 77.1 Fujiwara, Hiroshi. 1977. Collected papers on word order and infinitive in English. (Gakushuin Series of Treatises 2). Tokyo: the Gakushuin Educational Foundation, vii + 287 pp. D.Litt. Thesis, Sophia Univ. [This is a collection of five essays on Medieval English syntax originally written from 1963 to 1977. Contents: (1) Some aspects of word order in Old English; (2) Some aspects of word order in Old English (Π); (3) On the word order in phrases with the to-infinitive during the transitional period from OE to ME; (4) On the infinitive in the interlinear gloss of Aelfric's Colloquy·, (5) What was the original language of the Ancrene Riwlel] 78.1 Nakao, Toshio. 1978. The prosodic phonology of Late Middle English. Tokyo: Shinozaki Shorin. ν + 206 pp. Originally D.Litt. Thesis, Tokyo Univ. of Education, 1975. Cf. Item 72.1. [This is the first book to attempt an explicit and comprehensive explanation of the general principle which governs the stress pattern of Late Middle English. The first part of this book exploits a most reliable body of accentual data from alliterative and rhyme poetry in the framework of generative metrics. The second part discusses stress contours from the point of view of the 'standard' theory of generative phonology.] 78.2 Ukaji, Masatomo. 1978. Imperative sentences in Early Modern English. Tokyo: Kaitakusha. xi + 193 pp. Originally D.Litt. Thesis, Tokyo Univ. of Education, 1976. Rev.: Kazuo Araki, SELitEN 1981: 118-23; Liliane Haegeman, ES 62 (1981): 85-88; James Monaghan, Anglia 100 (1982): 15660. 78.3 Yamaguchi, Hideo. 1978. Studies on English style: A prelude to stylistics. Tokyo: Shinozaki Shorin. xi + 580 pp. [This book opens with an

English historical linguistics and philology in Japan 1950 - 1995


introductory survey of classical and medieval rhetoric and a chapter on the directions for modern style-study, followed by a study of rhetorical traditions in the early English Renaissance period. Style in the language of romance is reviewed in King Horn and Sir Perceval of Gales in the second part. The third and last part is devoted to the study of style in the language of faith in Wulfstan, Richard Rolle of Hampole, M. Kempe, the English Bible, and G.M. Hopkins.] 79.1 Irie, Keitaro. 1979. In search of Dryden's language. Hiroshima: Keisuisha. 118 pp. [This book consists of three parts: 1. Dramatic English; 2. Sentence-structure in John Dryden's Earlier Prose; 3. Newer method for the description of sentence-structure.] 79.2 Manabe, Kazumi. 1979. Syntax and style in Early English: Finite and non-finite clauses c. 900-1600. Tokyo: Kaibunsha. vi + 184 pp. Rev.: editors, AS 55 (1980): 72-73; Kikuo Yamakawa, SELitEN 1981: 102-11. 80.1 Ito, Hiroyuki. 1980. The language of 'The Spectator': A lexical and syntactic approach. Tokyo: Shinozaki Shorin. vi + 332 pp. D.Litt. Thesis, Horishima Univ., 1987. Rev.: Fritz Rau, GRM 65 (1984): 366-68. 80.2 Ono*, Shigeru - Toshio Nakao* 1980. Eigoshi I - Ko-Eigo [History of English I: Old English]. (Outline of English Lignuistics 8). Tokyo: Taishukan. xxxiii + 710 pp. [This book consists of nine chapters: 1. general survey; 2. alphabet, spelling; 3. phonology; 4. dialectology; 5. morphology; 6. syntax; 7. lexicology; 8. metrics; 9. stylistics.] 81.1 Terasawa, Yoshio et al.* (eds.). 1981. Eigo no Rekishi to Kozo Miyabe Kikuo Kyoju Kanreki Kinen Ronbunshu [The history and structure of the English language in honour of Professor Kikuo Miyabe]. Tokyo: Kenkyusha. vi + 438 pp. Rev.: Haruo Iwasaki*, et al. SELit 60,2 (1983): 33035. [K. Miyabe, Professor Emeritus of the Univ. of Tokyo. This Festschrift contains seventeen essays, seven written in English, and ten in Japanese.] 81.2 Nakashima, Kunio. 1981. Studies in the language of Sir Thomas Malory. Tokyo: Nan' un-do. xi + 411 pp. D.Litt. Thesis, Nihon Univ. Rev.: Norman F. Blake, ES 63 (1982): 558; Tomomi Kato, SELitEN 1985: 101-108. [The author presents a detailed description of syntax, based on the second edition of Vinaver's Works of Sir Thomas Malory (Oxford, 1967).] 81.3 Ogura, Michiko, 1981. The syntactic and semantic rivalry of 'quoth', 'say' and 'tell' in Medieval English. (Intercultural Research Institute Monograph 12). Hirakata, Osaka: KUFS Publication: Kanasai Univ. of Foreign Studies, iv + iii + 136 pp. Originally D.Litt. Thesis, Senshu Univ., 1980. 82.1 Nakano, Hirozo et al. (eds.). 1982. Studies in Linguistics Change in Honour of Kazuo Araki. Tokyo: Kenkyusha. vii + 397 pp. [K. Araki, Professor Emeritus of Nagoya Univ. This Festschrift contains twenty-six essays in three


Akio Oizumi

parts: 1. sound change; 2. syntactic change; 3. semantic change and other themes.] 82.2 Oda, Takuji. 1982. A concordance to the riddles of the Exeter Book. Tokyo: Gaku Shobo. vii + 293 pp. Rev.: Andre Cr6pin EA 37 (1984): 183-84; Eric G. Stanley N&Q n.s. 30 (1983): 1. 82.3 Otsuka*, Takanobu - Fumio Nakajima* (general editors). 1982. Shin Eigogaku Jiten [The Kenkyusha dictionary of English linguistics and philology]. Tokyo: Kenkyusha. xv + 1582 pp. Rev.: Sadao Ando*, SELit 61,2 (1984): 395-404. [This is a completely revised & enlarged version of Sanki Ichikawa* (ed.), Eigogaku Jiten [The Kenkyusha Dictionary of English Philology], Tokyo: Kenkyusha. Twelfth edition with New Addenda, 1953. xv + 1188 pp.; First edition, 1940. xiii + 1165 pp. Under the general editorship of the two of Dr. Ichikawa's earlier disciples, ninety-nine scholars have participated in the compilation of this monumental work, which lists a total of 1752 items alphabetically, and gives for them brief or detailed explanations depending upon their importance. The fields it covers range from linguistic theories to metrics, rhetoric and stylistics.] 83.1 Hirooka, Hideo. 1983. Thomas Hardy's use of dialect. Tokyo: Shinozaki Shorin. xxxi + 486 pp. D.Litt. Thesis, Kansai Univ. [The dialectal expressions and words would be some obstacle to foreign readers and even to native speakers who are not accustomed to the south-western dialect. The author gives an exhaustive description of the dialect paying much attention to the rationality of dialect grammar. The volume consists of three parts: 1. pronunciation; 2. grammar; 3. glossary.] 83.2 Terasawa*, Yoshio (ed.). 1983. Chusei Eigo Eibungaku Kenkyu Gyoseki Rosuto [A Bibliography of Publications on Medieval English Language and Literature in Japan 1983]. Tokyo: Centre for Medieval English Studies, College of Arts and Sciences, Univ. of Tokyo, ix + 259 pp. [This Bibliography has been supplemented by three English versions which have listed only the works voluntarily reported to the Centre. The 1986, 1988 and 1990 versions published under the general editorship of Yoshio Terasawa and Tadao Kubouchi as the special issues of the Medieval English Studies Newsletter cover the scholarship from April 1983 to March 1989.] 84.1 Araki*, Kazuo - Masatomo Ukaji*. 1984. Eigoshi ΙΠ A - Kindai-Eigo [History of English ΙΠ A: Modern English]. (Outline of English Linguistics 10). Tokyo: Taishukan. xxxi + 620 pp. Rev.: Hayashi Ono, SELit 63,2 (1986): 378-84. [This book consists of five chapters: 1. general survey; 2. phonology; 3. morphology; 4. lexicology; 5. syntax.] 84.2 Ikegami, Masa Tsuda. 1984. Rhyme and pronunciation: Some studies of English rhymes from 'Kyng Alisaunder' to Skelton. (Hogaku-Kenkyu-Kai,

English historical linguistics and philology in Japan 1950 - 1995


Keio Univ., Extra Series 5). Tokyo: Hogaku-Kenkyu-Kai, Keio Univ. xv + 402 pp. Originally M.Litt. Thesis, Univ. of Oxford, 1981. Rev.: Eric G. Stanley, N&Q n.s. 33 (1986): 532-33; Alan Ward, MAE 56 (1987): 116-17; Kazuo Araki*, SELit 62,2 (1985): 399-402. 84.3 Maeno, Shigeru - Kaneaki Inazumi. 1984. A Melville lexicon. Tokyo: Kaibunsha. xi + 523 pp. Rev.: Christine Gerrard, RES n.s. 37 (1986): 437-38; Harold Beaver, MLR 83 (1988): 169-70. Cf. item 76.3. 84.4 Ono, Shigeru et al. (eds.). 1984. Eigogaku Kenkyu - Matsunami Tamotsu Hakase Kanreki Kinen Ronbunshu [Studies in English philology and linguistics in Honour of Dr. Tamotsu Matsunami]. Tokyo: Shubun International, vi + 478 + iii pp. [T. Matsunami, Professor Emeritus of Tokyo Metropolitan Univ. This Festschrift includes thirty-seven papers, twenty-six written in English, and eleven in Japanese.] 84.5 Minakawa, Saburo - Michio Yoshikawa (eds.). 1984. A Thomas Hardy dictionary. Based on the Original MSS first edited by Mamoru Osawa, Yoshinoshin Goto, Takashi Iijima, J.O. Bailey, R. Morrell, F. Pinion & H. Orel. Tokyo: Meicho-Fukyu-Kai. xxv + 652 pp. [This dictionary was compiled by the members of the Thomas Hardy Society of Japan, to give explanations of all Hardy's words in the light of their use in context.] 84.6 Ono*, Hayashi. 1984. Eigo Jikan-Fukushisetsu no Bunpo [A grammar of English temporal adverbial clauses]. Tokyo: Eihosha. xvii + 744 pp. Originally D.Litt. Thesis, Nagoya Univ., 1981. Rev.: Keitaro Irie*, SELit 66,1 (1989): 170-75. 84.7 Yamaguchi, Hideo. 1984. Language and poetry. Tokyo: Meirin Shuppan. ix + 205 pp. [The author presents some aspects of interactions between form and meaning in poetry, in six sections: 1. The word; 2. Lexical fields; 3. The use of the word; 4. Syntax in English poetry; 5. The poetry of faith; 6. The poetry of vision: Emily Dickinson's Songs.] 84.8 Yoshino, Yoshihiro. 1984. Poetic syntax in the Old English 'Meters of Boethius': A comparative study of verse and prose. Ph.D. Diss., Univ. of Pennsylvania, ν + 174 pp. 85.1 Irie, Keitaro. 1985. The sentence-structure in John Dryden's 'An essay of dramatic poesy'. Hiroshima: Keisuisha. 215 pp. [The author attempts to identify the characteristics of the sentence-structure in An Essay of Dramatic Poesy.] 85.2 Nakao* Toshio. 1985. On' in-Shi [English Historical Phonology]. (Outline of English Linguistics 11). Tokyo: Taishukan. xxx + 656 pp. Rev.: Kazuo Araki*, SELit 63,1 (1986): 203-207. [This volume is divided into two parts: 1. How to reconstruct past sounds & theories of sound change; 2. Segments and phonological processes.]


Akio Oizumi

85.3 Ono, Hayashi. 1985. Studies in idiomatic expressions of eighteenthcentury English - Including some examples from authors of the seventeenth century. Tokyo: Taimeido. iv + 198 pp. 85.4 Tajima, Matsuji. 1985. The syntactic development of the gerund in Middle English. Tokyo: Nan'un-do. xii + 154 pp. Originally Ph.D. Diss., Univ. of Ottawa, 1983. Rev.: Witney F. Bolton, Diachronica 2 (1985): 259-62; Bernhard Diensberg, ASNS 224 (1987): 387-89; Willy Elmer, Anglia 105 (1987): 423-26; Kenji Kondo, SELitEN 1987: 126-34. 85.5 Terajima, Michiko. 1985. The trajectory constraint and 'irregular' rhymes in Middle English. Tokyo: Shinozaki Shorin. vii +159 pp. Rev.: Donka Minkova, FoLH 8,1-2 (1987): 481-501. 85.6 Terasawa*, Yoshio - Akio Oizumi* (eds.). 1985. Eigoshi Kenkyu no Hoho [The methods of English historical linguistics]. Tokyo: Nan'un-do. 246 pp. Rev.: Hiroshi Ogawa*, SELit 63,2 (1986): 384-9. [This is a volume of the proceedings of the Round Table Conference on English Historical Linguistics held at Doshisha Univ. in Kyoto on 3-4 December 1983, containing seven essays and comments.] 85.7 Wakatabe*, Hiroya. 1985. Eigoshi ΠΙ Β - Amerika-Eigo [History of English ΠΙ B: American English]. (Outline of English Linguistics 10). Tokyo: Taishukan. xvii + 269 pp. Rev.: Hiroshi Matsuda*, SELit 63,1 (1986): 200203. 85.8 Yonekura, Hiroshi. 1985. The language of the Wycliffite Bible: The syntactic differences between the two versions. Tokyo: Aratake Shuppan. xviii + 525 pp. Rev.: Knud S0rensen, ES 67 (1986): 78-80. 86.1 Ogura, Michiko. 1986. Old English 'impersonal' verbs and expressions. (Anglistica 24). Copenhagen: Rosenkilde & Bagger. 310 pp. + 24 tables. Rev.: Fran Colman, YWES 67.1986: 91-92; Hiroshi Fujiwara, SELitEN 1988: 124-27. [This monograph aims at investigating OE 'impersonal' verbs and expressions and to find out possible causes of their falling into disuse before ME.] 87.1 Nakao, Toshio (ed.). 1987. Historical studies in Honour of Taizo Hirose: A memory volume dedicated with respect and affection by His students and friends. Tokyo: Kenkyusha. Privately printed, vi + 204 pp. Rev.: Fran Colman, YWES 68.1987: 95-97. [T. Hirose, Professor Emeritus of Tokyo Univ. of Education. This memorial volume contains ten essays on syntax and seven on phonology.] 87.2 Ogura, Mieko. 1987. Historical English phonology: A lexical perspective. Tokyo: Kenkyusha. viii + 240 pp. D.Litt. Thesis, Tokyo Metropolitan Univ., 1988. Rev.: Bernhard Diensberg, FoLH 9 (1988): 301307; Anna Ewert, SAP 22 (1989): 202-207. [The author investigates the

English historical linguistics and philology in Japan 1950 - 1995


implementation (how) and actuation (why) of a selected portion of the phonological changes in English.] 88.1 Kubouchi, Tadao - William Schipper - Hiroshi Ogawa (eds.). 1988. Old English studies from Japan 1941-1981. (Old English Newsletter, Subsidia 14). Binghamton, New York at Binghamton. iv + 100 pp. Rev.: Kenneth Turner, et al., YWES 69.1988: 85-86. [This is a collection of seven essays originally published in Japanese.] 88.2 Masui, Michio. 1988. A new rime index to 'The Canterbury Tales' based on Manly and Rickert's text of 'The Canterbury Tales'. Tokyo: Shinozaki Shorin. xvi + 396 pp. Rev.: Mechthild Gretsch, Anglia 112 (1994): 193-99; Eric G. Stanley, N&Q n.s. 36 (1989): 369-70; Lynda C. Mugglestone, ES 71 (1990): 446-47. [This Index is the sequel to The structure of Chaucer's rime words (Tokyo, 1964) (item 64.2). The focus here is on written forms of the words themselves rather than their endings.] 88.3 Masui, Michio. 1988. Studies in Chaucer's language of feeling. Tokyo: Kinseido. vii + 138 pp. Rev.: Eric G. Stanley, N&Q n.s. 36 (1989): 369-70. 88.4 Saito, Toshio - Mitsunori Imai. 1988. A concordance to Middle English metrical romances. Two volumes. Programmed by Kunihiro Miki. Frankfurt am Main, etc.: Verlag Peter Lang. Vol. I: The Matter of England, vi + 770 pp.; Vol. Π: The Breton Lays, ν + 779 pp. Rev.: Mechthild Gretsch, Anglia 112 (1994): 193-99; Dieter Mehl, ASNS 230 (1993): 169-71; Hisashi Takahashi*, SELit 67,1 (1990): 129-32. 88.5 Tajima, Matsuji. 1988. Old and Middle English language studies: A classified bibliography 1923-1985. (Amsterdam Studies in the Theory and History of Linguistic Science. Series V: Library and Information Sources in Linguistics 13). Amsterdam/Philadelphia: John Benjamins, xxxiii + 391 pp. Rev.: John R.J. North, SLang 13 (1989): 518-21; Daniel Donoghue, Speculum 66 (1991): 482-83; George Jack, Diachronica 6 (1989): 151-54. [This bibliography seeks to fill the gap between A.G. Kennedy's Bibliography of Writings on the English Language from the Beginning of Printing to the End of 1922 and the 1980s.] 88.6 Oshitari, Kinshiro et al. (eds.). 1988. Terasawa Yoshio Kyoju Kanreki Kinen Ronbunshu [Philologia Anglica: Essays presented to Professor Yoshio Terasawa on the Occasion of His sixtieth birthday]. Tokyo: Kenkyusha. ix + 508 pp. Rev.: Mitsunori Imai, et al: SELitEN 1990: 169-77. [Y. Terasawa, Professor Emeritus of the Univ. of Tokyo. The Festschrift contains forty-two articles in four parts: 1. history of English; 2. Medieval English literature; 3. The English Bible; 4. lexicography and lexicology. Thirty-two essays written in English, and ten in Japanese.]


Akio Oizumi

89.1 Manabe, Kazumi. 1989. The syntactic and stylistic development of the infinitive in Middle English. Fukuoka: Kyushu Univ. Press. 206 pp. Originally D.Litt. Thesis, Tokyo Metropolitan Univ., 1988. Rev.: David Denison, Diachronica 8 (1991): 285-92; Willy Elmer, Anglia 109 (1991): 119-21; Mitsu Ide, SELitEN 1992: 136-43. 89.2 Ogawa, Horishi. 1989. Old English modal verbs: A syntactical study. (Anglistica 26). Copenhagen: Rosenklide & Bagger, xii + 319 pp. Originally D.Litt. Thesis, Tokyo, Metropolitan Univ., 1985. Rev.: Robert D. Fulk, JEGP 90(1991): 546-49. 89.3 Ogura, Michiko. 1989. Verbs with the reflexive pronoun and constructions with 'self in Old and Early Middle English. Cambridge: D.S. Brewer, xiv + 113 pp. + 8 tables. Rev.: John D. Burnley, ES 72 (1991): 48687; Daniel Donoghue, N&Q n.s. 38 (1991): 92-94; Lilo Moessner, Anglia 110 (1992); Seiichi Suzuki, SELitEN 1991: 74-78. 89.4 Oizumi*, Akio - Haruo Iwasaki* (eds.). 1989. Chaucer no Eigo Kenkyu no Kadai to Hoho [Studies in Chaucer's English: Some Problems and Methods]. Tokyo: Eichosha-Shinsha. vii + 233 pp. [This is a volume of the proceedings of the Kyoto Conference on Chaucer's English held at Doshisha Univ. in Kyoto on 3-4 October 1987, which consists of 'Preface' and eleven articles in three parts: 1. the keynote lecture by Ralph Elliott; 2. research projects; 3. lectures and 'Chaucer's English: A Select Bibliography'. Elliott's contribution is written in English, and the rest in Japanese with English summaries.] 89.5 Ono, Shigeru. 1989. On Early English syntax and vocabulary. Tokyo: Nan'un-do. χ + 301 pp. [This is a collection of fifteen articles in English, in three parts: 1. The 'Modal' Auxiliaries and the Infinitive; 2. Old English Vocabulary; 3. Review Articles.] 89.6 Terasawa, Jun. 1989. A metrical study of Old English compounds. Ph.D. Diss., Brown Univ. vii + 174 pp. [See item 94.6] 90.1 Fujiwara*, Yasuaki. 1990. Ko-Eishi Inritsu Kenkyu [Studies in Old English metrics]. Hiroshima: Keisuisha. iv + 380 pp. D.Litt. Thesis, Univ. of Tsukuba. Rev.: Jun Terasawa*, S£L/f 68,2 (1992): 382-87. 90.2 Mizobata, Kiyokazu. 1990. A concordance to Caxton's own prose. Tokyo: Shohakusha. ix + 620 pp. Rev.: Ingrid Tieken-Boon van Ostade, ES 73 (1992): 571-72. [This is a computerized concordance of the Caxton text in Norman F. Blake's Caxton's Own Prose (London, 1973).] 90.3 Ogura, Mieko. 1990. Dynamic dialectology: A study of language in time and space. Tokyo: Kenkyusha. χ + 303 pp. Rev.: Sheila M. Embleton, Language 67 (1991): 657-58; Sheila M. Embleton, Diachronica 8 (1991): 105118. [This volume consists of six chapters: 1. Theoretical preliminaries; 2.

English historical linguistics and philology in Japan 1950 - 1995


Spatial distribution of the Great Vowel Shift in England; 3. The development of the Indo-European languages; 4. Language change in China; 5. Language contacts in the history of English; 6. The acquisition of phonology.] 90.4 Oizumi, Akio - Toshiyuki Takamiya (eds.). 1990. Medieval English studies: Past and present. Tokyo: Eichosha. Published for the Centre for Medieval English Studies, Univ. of Tokyo, viii + 310 pp. Rev.: Bernhard Diensberg, MESN 24 (1991): 13-16; Eric G. Stanley, N&Q n.s. 39 (1992): 427. [This volume in two parts is intended to show the wide range of Medieval English scholarship. The first part is devoted to a survey of Medieval English studies in particular universities or regions in the past as well as at present. The second part consists of obituaries of distinguished medievalists who have died during the past ten or so years, recording aspects of their lives and the contributions which they have made to Medieval English studies in this century.] 90.5 Jin, Koichi et al. (eds.). 1990. Eigobunkengaku Kenkyu - Ono Shigeru Hakase Kanreki Kinen Ronbunshu [Studies in English philology in honour of Shigeru Ono]. Tokyo: Nan'un-do. 515 pp. [S. Ono, Professor Emeritus of Tokyo Metropolitan Univ. This Festschrift contains thirty-two philological contributions in all. Twenty-three essays are written in English, and nine in Japanese.] 90.6 Yamashita, Hiroshi - Masatsugu Matsuo - Toshiyuki Suzuki - Haru Sato (eds.). 1990. A comprehensive concordance to 'The Faerie Qveene 1590'. Tokyo: Kenyusha Books, xiii + 1216 pp. and two computer disks. Rev.: William P. Williams, AEB n.s. 5 (1991): 109-110; Charles Butler, N&Q n.s. 40 (1993): 538-40; Helen Cooper, RES n.s. 44 (1993): 569-70. [This concordance is designed to supply unsophisticated but informative raw data about the 1950 quarto of The Faerie Qveene, as an aid to more intensive textual study than has so far been given to the first edition of Spenser's poem.] 91.1 Awaka, Kiyoshi. 1991. Λ concordance to the 'Ancrene Wisse' based on J.R.R. Tolkien's text. (A special Issue of Philologica). Tsu, Mie: Association of English Studies, Mie Univ. Privately posthumously printed, ν + 669 pp. 91.2 Kawai, Michio (ed.). 1991. Language and style in English literature: Essays in honour of Michio Masui. Hiroshima: The English Research Association of Hiroshima, Hiroshima Univ.:; published by Eihosha. χ + 708 pp. [M. Masui, Professor Emeritus of Hiroshima Univ. This collection of essays was published to mark the occasion of Dr. Masui's seventy-seventh birthday. It contains forty-one essays and Masui's "Visiting Northern Europe in June", 1956, and "Observing and Thinking in Japan in Summer, 1990".] 91.3 Niwa, Yoshinobu. 1991. The function and development of prefixes and particles in three Early English Texts - The beginning of the phrasal verb -.


Akio Oizumi

Vol. One. Tokyo: Kinseido. xi + 213 pp. [See item 95.3] [The author tries to determine how prefixed verbs are directly or indirectly related to phrasal verbs. This is a continuation of his Studies in Old English preverbal 'ge-' (item 73.2).] 91.4 Oizumi, Akio. 1991. Λ complete concordance to the works of Geoffrey Chaucer. Ten volumes. Programmed by Kunihiro Miki. (Alpha-Omega, Reihe C: Englische Autoren 1). Hildesheim / Zürich / New York: Olms-Weidmann. Vol. I: A KWIC Concordance to 'The Canterbury Tales', Part I. xi + 841 pp.; Vol. Π: A KWIC Concordance to 'The Canterbury Tales', Part Π. vii + 8431694 pp.; Vol. ΙΠ: A KWIC Concordance to 'The Canterbury Tales', Part ΙΠ; A ranking word-frequency list; a reverse word list; a hyphenated word list, vii + 1695-2554 pp.; Vol. IV: A general word index to 'The Canterbury Tales', and word indexes to each tale from 'The Canterbury Tales', vii + 2555-3595 pp.; Vol. V: Concordances to 'The Book of the Duchess', 'The House of Fame', 'Anelida and Arcite', 'The Parliament of Fowls', ix + 708 pp.; Vol. VI: A Concordance to 'Boece', ix + 980 pp.; Vol. VII: A Concordance to 'Troilus and Criseyde'. ix + 1105 pp.; Vol. VIE: Concordances to 'The Legend of Good Women', The Short Poems, 'Poems Not Ascribed to Chaucer in the Manuscripts', and Ά Treatise on the Astrolabe', xi + 1049 pp.; Vol. IX: A Concordance to 'The Romaunt of the Rose', ix + 885 pp.; Vol. X: An integrated word index to the Works of Geoffrey Chaucer. Xii + 718 pp. [This computer-generated Complete Concordance is based upon The Riverside Chaucer published in 1987 under the general editorship of Larry D. Benson. It supersedes the Tatlock and Kennedy manually-produced work of 1927. Related publications: (1) A Rhyme Concordance to the Poetical Works of Geoffrey Chaucer (item 94.5); (2) Towards a Comprehensive Concordance to the Works of Geoffrey Chaucer (Hildesheim / Ziirich / New York: OlmsWeidmann, 1995).] 92.1 Fukushima, Osamu. 1992. An etymological dictionary of English derivatives. With a Preface by John G. Griffith. Yamanashi: Nihon Tosho. 884 pp. [This is the first English dictionary of etymology compiled from a purely derivational point of view.] 92.2 Nagashima, Daisuke. 1992. A historical study of the introductory 'there'. (Kansai Univ. of Foreign Studies. Intercultural Research Institute Monograph 22). Hirakata, Osaka: The Intercultural Research Institute, Kansai Univ. of Foreign Studies, vii + 106 pp. [This monograph contains three parts of Nagashima's article "A Historical Study of the Introductory There" published from 1972 to 1976. It also contains General Introduction, Bibliography and two appendices.]

English historical linguistics and philology

in Japan 1950 - 1995


92.3 Saito, Kurumi. 1992. Nominal modification in Old English prose. Ph.D. Diss., International Christian Univ., Tokyo, vi + 190 pp. 93.1 Ito, Hiroyuki. 1993. Some aspects of eighteenth-century English. Tokyo: Eichosha, χ + 200 + i pp. [This is a collection of ten essays on eighteenth-century English from The Spectator to Richardson's novels.] 93.2 Iwasaki, Haruo. 1993. The language of Lajamon's 'Brut'. Tokyo: Kenkyusha. χ + 203 pp. [The author gives a detailed descriptive analysis of the two tests of La3amon's Brut, intending thereby to throw light on some aspects of the transition from OE to ME, and on the relationship of the two extant manuscripts.] 94.1 Blake, Norman Francis - John David Burnley - Masatsugu Matsu Yoshiyuki Nakao (eds.). 1994. A new concordance to 'The Canterbury tales' based on Blake's text edited from the Hengwrt Manuscript. Okayama: University Education Press, vii + 1008 pp. [Norman Blake's edition of The Canterbury Tales by Geoffrey Chaucer Edited from the Hengwrt Manuscript (London, 1980) contains edited and emended parts: punctuations are generally not scribal, but editorial; the scribe's abbreviations are expanded, and the missing letters or words are, though not in toto, restored by the editor.] 94.2 Burnley, John David - Matsuji Tajima. 1994. The language of Middle English literature. (Annotated Bibliographies of Old and Middle English Literature 1). Cambridge: D.S. Brewer, viii + 280 pp. [As well as serving the purpose of a reference for anyone seeking information on the language of a specific author or text, the bibliography may also serve to define existing linguistic approaches to the study of texts and suggest further possibilities. Works written between 1890 and 1990 are covered, including the increasingly important contribution made from Japan.] 94.3 Chiba, Shuji et al. (eds.). 1994. Synchronic and diachronic approaches to language: A Festschrift for Toshio Nakao on the occasion of his sixtieth birthday. Edited by Shuji Chiba, et al. Tokyo: Liber Press, xx + 627 pp. [T. Nakao, Professor of Tsuda College. This Festschrift contains forty-three articles in six parts: 1. Historical linguistics 1; 2. historical linguistics 2; 3. phonetics & phonology; 4. morphology; 5. syntax & semantics; 6. pragmatics.] 94.4 Moriya, Yasuyo. 1994. A metrical analysis of Middle English alliterative poetry. Ph.D. Diss., Univ. of Cincinnati, 1994. viii + 307 pp. 94.5 Oizumi, Akio - Hiroshi Yonekura. 1994. A rhyme concordance to the poetical works of Geoffrey Chaucer. In two volumes. Programmed by Kunihiro Miki. (A Complete Concordance to the Works of Geoffrey Chaucer, Volumes XI & ΧΠ: Supplement Series I & Π). (Alpha-Omega, Reihe C: Englische Autoren 1/2). Hildesheim / Zürich / New York: Olms-Weidmann. Vol. XI: Rhyme concordances to 'The Canterbury Tales', 'The Book of the Duchess',

16 AkioOizumi 'The House of Fame', 'Anelida and Arcite', 'The Parliament of Fowls', xvii + 1-556 pp.; Vol. ΧΠ: Rhyme concordances to 'Troilus and Criseyde', 'The Legend of Good Women', The short poems, 'Poems not ascribed to Chaucer in the Manuscripts', 'The Romaunt of the Rose'; An integrated rhyme word index to Chaucer's Poetical Works, vi + 557-1327 pp. [Related publications: (1) A. Oizumi, "Towards a rhyme concordance to Chaucer's poetical Works", in: Toshiyuki Takamiya & Richard Beadle (ed.), Chaucer to Shakespeare: Essays in Honour of Shinsuke Ando (Cambridge: D.S. Brewer, 1992): 67-73; (2) A. Oizumi & H. Yonekura, "Rhyme structure in Chaucer's Canterbury Tales", in: Danielle Buschinger & Wolfgang Spievak (ed.), Etudes de linguistique et de litterature en l'honneur d'Andre Crepin (Greifswald: Reinecke-Verlag, 1993): 281-88. This computer-generated rhyme concordance is a sequel to the tenvolume Complete concordance (item 91.4). The Concordance, the first of its kind in the field of philology, has great potential as an essential research tool for students of Chaucer's rhyme, English historical phonologists, and historians of the language. The computer-generated rhyme concordance in eleven parts lists all rhyme elements in alphabetical order, followed in each case by a complete listing of the rhyme words of that element. It also contains a ranking rhyme element-frequency list, a rhyme word index, a ranking rhyme word-frequency list, a rhyme scheme list, and a rhyme structure list.] 94.5 Terasawa, Jun. 1994. Nominal compounds in Old English: A metrical approach. (Anglistica 27). Copenhagen: Rosenkilde & Bagger, χ + 138 pp. Originally Ph.D. Diss., Brown Univ., 1989 (item 89.6). Rev.: Eric G. Stanley N&Q n.s. 42 (1995): 419. [This study shows that certain compound patterns (like hilde-wiga and beado-wiga) are avoided with few real exceptions in Old English poetry. The work also contributes to Old English textual criticism by suggesting on metrical grounds alternative editorial reading concerning compounds and to Old English lexicography by providing a metrical criterion to distinguish between a compound and a syntactic phrase.] 95.1 Blake, Norman Francis - John David Burnley - Masatsugu Matsuo Yoshiyuki Nakao (eds.). 1995. A new rime concordance to 'The Canterbury Tales' based on Blake's text edited from the Hengwrt Manuscript. Okayama: University Education Press, ix + 535 pp. 95.2 Jimura, Akiyuki - Yoshiyuki Nakao - Masatsugu Matsuo (eds.). 1995. A comprehensive list of textual comparison between Blake's and Robinson's editions of 'The Canterbury Tales'. Okayama: University Education Press, ν + 520 pp. [This computer-assisted volume consists of two parts and an appendix: 1. lines with different word forms; 2. word form correspondence index; appendix: cross-references for Robinson's edition. Robinson's edition (2nd edition: Boston, 1957) is based on the Ellesmere Manuscript.]

English historical linguistics and philology in Japan 1950 - 1995


95.3 Niwa, Yoshinobu. 1995. The function and development of prefixes and particles in three Early English Texts - The beginning of the phrasal verb -. Vol. Two. Tokyo: Kineseido. ix + 296 pp. [See item 91.3] 95.4 Oizumi, Akio (ed.). 1995. A Bibliography of Writings on Chaucer's English. Hildesheim / Zürich / New York: Olms-Weidmann, xxv + 81 pp. [This volume is the most comprehensive bibliography on Chaucer's English ever compiled. The bibliography, which contains 706 items, is divided into the following ten headings: 1. bibliographies; 2. manuscripts, facsimiles and editions; 3. textual criticism; 4. English linguistic background; 5. medieval rhetoric and poetics; 6. dictionaries and concordances; 7. language phonology and grammar; 8. language - lexicon / word studies; 9. meter and versification / prosody; 10. style and rhetoric.] 95.5 Sassaki*, Tatsu - Kenzo Kihara* (eds.). 1995. Eigogaku Jinmei Jiten [The Kenkyusha biographical dictionary of English Linguistics and Philology.] Tokyo: Kenkyusha Limited, xv + 430 pp. [The biographical dictionary written by 116 authors is a reference book that documents the lives and work of 575 distinguished scholars throughout the world who have concerned themselves down the ages with the study and description of English and related languages. Generally, each entry consists of a biography, an abstract of the scholar's achievements, including his or her influence, and a bibliography.] 95.6 Yamada, Norio. 1995. Phonological changes in English: A functional explanation in historical phonology. Ph.D. Diss., Tsuda College, vi + 152 pp. Unpublished.

3. Index of contributors (authors and editors)

Ando, Sadao (1927- ): 76.1 Araki, Kazuo (1921- ): 58.1; 65.1; 82.1; 84.1 Awaka, Kiyoshi (1937-1988): 91.1 Chiba, Shuji (1942- ): 94.3 Fujiwara, Hiroshi (1920- ): 77.1 Fuj iwara, Yasuaki (1946- ): 90.1

Fukumura, Toraj iro (1914- ): 65.3 Fukushima, Osamu (1943- ): 92.1 Higashida, Chiaki (1910-1992): 52.1 Hirooka, Hideo (1918- ): 83.1 Hirose,Taizo (1914-1978): 87.1 Ichikawa, Sanki (1886-1970): 52.2; 66.1; 82.3


Akio Oizumi

Ikegami, Masa Tsuda (1933- ): 84.2 Ikegami, Yoshihiko (1934- ): 63.1 Imai, Mitsunori (1938- ): 88.4 Inazumi, Kaneaki (1943- ): 84.3 Irie, Keitaro (1932- ): 79.1; 85.1 Ishibashi, Kotaro (1898-1979): 66.1 Ito, Eiko Tamano (1941- ): 71.1 Ito, Hiroyuki (1927- ): 80.1; 93.1 Iwasaki, Haruo (1933- ): 89.4; 93.2 Jimura, Akiyuki (1952- ): 95.2 Jin, Koichi (1941- ): 90.5 Kato, Tomomi (1926- ): 74.1 Kawai, Michio (1929- ): 91.2 Kihara, Kenzo (1919- ): 95.5 Kobayashi, Eichi (1931- ): 64.1 Kubouchi, Tadao (1939- ): 83.2; 88.1 Maejima, Giichiro (1904-1985): 62.1 Maeno, Shigeru (1924-1992): 76.3; 84.3 Manabe, Kazumi (1938- ): 79.2; 89.1 Masui, Michio (1914-1992): 52.1; 64.2; 88.2; 88.3; 91.2

Matsuda, Tokuichiro (1933- ): 65.2 Matsu, Masatsugu (1945- ): 90.6; 94.1; 95.1; 95.2 Matsunarai, Tamotsu (1924-1995): 66.2; 84.4 Miki, Kunihiro (1959- ): 88.4; 91.4; 94.5 Minakawa, Saburo (1906- ): 84.5 Miyabe, Kikuo (1915-1981): 74.2; 81.1 Miyake, Ko (1929- ): 73.1 Mizobata, Kiyokazu (1950- ): 90.2 Moriya, Yasuyo (1953- ): 94.4 Nagashima, Daisuke (1929- ): 92.2 Nakajima, Fumio (1904- ): 82.3 Nakano, Hirozo (1936- ); 82.1 Nakao, Toshio (1934- ): 72.1; 78.1; 80.2; 85.2; 87.1; 94.3 Nakao, Yoshiyuki (1950- ): 94.1; 95.1; 95.2 Nakashima, Kunio (1925- ): 81.2 Niwa, Yoshinobu (1923- ): 73.2; 91.3; 95.3 Oda, Takuji (1940- ): 82.2 Ogawa, Horishi (1942- ): 88.1; 89.2 Ogura, Michiko (1949- ): 81.3; 86.1; 89.3 Ogura, Mieko (1948- ): 87.2; 90.3

English historical linguistics and philology in Japan 1950 - 1995

Oizumi, Akio (1935- ): 66.3; 85.6; 89.4; 90.4; 91.4; 94.5; 95.4 Ono, Hayashi (1927- ): 84.6; 85.3 Ono, Shigeru (1930- ): 69.1; 80.2; 84.4; 89.5; 90.5 Oshitari, Kinshiro (1932- ): 88.6 Otsuka, Takanobu (1897-1979): 58.1; 82.3 Saito, Kurumi (1950- ): 92.3 Saito, Shizuka (1891-1970): 52.2 Saito, Toshio (1933- ): 88.4 Sasaki, Tatsu (1904-1986): 95.5 Sato, Haruo (1948- ): 90.6 Sato, Masam (1923-1952): 53.1 Suzuki, Shigetake (1903-1976): 67.1 Suzuki, Toshiyuki (1944- ): 90.6 Tajima, Matsuji (1942- ): 85.4; 88.5; 94.2 Takahashi, Sakutaro (1942- ): 73.3


Takamiya, Toshiyuki (1944- ): 90.4 Taniguchi, Jiro (1899- ): 56.1 Terajima, Michiko (1945- ): 85.5 Terasawa, Jun (1959- ): 89.6; 94.6 Terasawa, Yoshio (1928- ): 81.1; 83.2; 85.6; 88.6 Tsuchiya, Tadayuki (1943- ): 75.1 Ueda, Minora (1929- ): 66.4 Ukaji, Masatomo (1931 - ): 78.2; 84.1 Wakatabe, Hiroya (1922- ): 85.7 Yamada, Norio (1947- ): 95.6 Yamaguchi, Hideo (1907- ): 52.2; 61.1; 66.5; 78.3; 84.7 Yamamoto, Tadao (1904-1991): 50.1; 52.1 Yamashita, Hiroshi (1944- ): 90.6 Yonekura, Hiroshi (1941 - ): 85.8; 94.5 Yoshikawa, Michio (1932- ): 84.5 Yoshino, Yoshihiro (1940- ): 84.8

Note *

This is a revised and up-dated version of the paper originally published in Language history and linguistic modelling: A festschrift for Jacek Fisiak on his 60th birthday, edited by Raymond Hickey and Stanislaw Puppel, 1997 (Berlin · New York: Mouton de Gruyter).

On identifying Old English adverbs Yasuaki Fujiwara

Introduction The term adverb can be ambiguous and present perplexing cases in which we cannot clearly identify a given word as an adverb in spite of complete morphological and contextual information on the word; this is partly because the term denotes both function and category at the same time, partly because the difference between adverbs and other word classes, especially adjectives, may often be subtle. The situation is even more troublesome in Old English than in Modern English, for the system of inflection and derivation in the former stage of the language was not established enough to avoid such ambiguity. It is the aim of this article to investigate the problems involved in the identification of Old English adverbs and to find criteria for the determination of the lexical categories made by Anglo-Saxon writers.

1. Problems We will first consider some typical examples of the problems which arise when describing Old English adverbs. The romanized words in (la, b, c) are all indeterminate as to whether they function as an adjective or an adverb; worse still, in examples such as (Id, e) an adjective is employed in places where the modern eye would expect an adverb (Quirk and Wrenn 1957: §138; Mitchell 1985: §1108).


a. Dä



pä säeton of er paet llc

The fellow-men sorrowful then sat upon the corpse 'Then the fellow men sat sorrowful upon the corpse'

/€lfric's Lives of Saints 31.212 b. on öäm fiiccum bremlum and in






and thorn-bushes and

netelum, de

päer on

däm westene J)icce stödon

nettles, which

there in






Yasuaki Fujiwara

'in the thick briers and thorn-bushes and nettles, which lay thick there in the desert' The Homilies of/Elfric ii. 156.28 c. bil eal durhwöd I I fiegne fläeschoman sword all penetrated// doomed body 'the whole sword (or the sword entirely) penetrated the doomed body' Beowulf I561b-Sa.' d. Scop hwilum sang // hädor on Heorote minstrel sometimes sang // clear in Heorot Ά minstrel sang now and again in clear voice in Heorot' Beowulf496b-Sa e. past me ys pus torne on mode, // häte on hredre mlnum that me is thus miserable in mind, // hot in breast my. 'that is thus miserable in my mind, hot in my breast.' Judith 93b-4a Mitchell (1985: §1108) evaded making a definite answer to such puzzling problems, stating that it would be imperceptive to speak of words as examples of adjectives used as adverbs or vice versa. However, the ultimate concern of this article is to find a satisfactory answer, or at least obtain a substantive theoretical clue, to the problems involved in the discrimination or identification of the function and category of words commonly used as adverbs. Since the formation and position of Old English adverbs seem to have a direct relation with our question, we will first have to clarify these two linguistic traits. First, we would like to provide a rough morphological outline of Old English adverbs. Among the inherited adverbs (i.e., adverbs others than those formed anew in the Old English period by means of derivational or inflectional endings) employed as such ever since the prehistoric Old English, some are identical in form with conjunctions or prepositions, as shown in (2); however, they never have corresponding adjectival forms except for some rare instances like neah, and hence no ambiguity arises with respect to function and category between the two word classes. (2)

ä 'always', äefre 'ever', äer 'before', eac 'also', eft 'again', back',/ram 'always',/«/ 'very', gyt 'yet', her 'here', hü 'how', hüru 'indeed', in 'in', mid 'together', neah 'near', nü 'now', oft 'often', on 'on', siddan 'since', swä 'so', pd 'then', päer 'there', ponne 'then', under 'under', up 'up', üt 'our', wel 'well', etc.

Ort identifying Old English adverbs


In contrast to the traditional inherited adverbs, a number of adverbs made anew through derivation during the Old English period may sometimes be identical in form to their matching adjectives or nouns. For example, consider the case of derived adverbs obtained through addition of a suffix -e to the stem syllable of adjectives as shown in (3). In this case, the two originally contrastive words may be identical in form, since the suffix -e is employed not only for the formation of adverbs but also some inflected endings such as nominative or accusative plural forms in the strong declension of adjectives. (3)

biter-e 'bitterly',fiislic-e 'gladly', sld-e 'extensively'

Furthermore, when adverbs are made anew from the corresponding adjectives which originally terminate with -e, the suffix is never added again to the stem; consequently, the number of adverbs that come to be the same in form as the matching adjectives, as shown in (4), is not small (Wright 1925: §553; Campbell 1959: §661). (4)

seöele 'noble', blipe 'joyful', breme 'glorious', eerie 'brave', clsene 'clean', defe 'proper', ece 'eternal', enge 'painful', frecne 'dangerous', gedefe 'fit', gehende 'near', Ilde 'mild', milde 'merciful', myrge 'pleasing', ntwe 'new', slide 'savage', swegle 'bright', tynge 'fluent', dicce 'thick', etc.

Thus, it is evident that the ending -e does not serve as a uni-functional marker of Old English derivative adverbs. On the other hand, there are some adverbs which can easily be discriminated from their corresponding adjectives, since the latter have mutated vowels in the stem syllable, as can be seen in (5). (5)

enge 'sad' - ange 'sadly'; lejye 'easy' - eafre 'easily'; sefte 'soft' - softe 'softly'; smepe 'smooth' - smöfje 'smoothly'; swete 'sweet' -swöte 'sweetly', etc. In addition, adverbs with final -e, such as those in (6), are never confused with adjectives since examples of their corresponding adjectival forms are not attested in the extant Old English texts (Campbell 1959: §662). (6)

äedre 'at once', same 'similarly', sneome 'quickly'


Yasuaki Fujiwara

Next, we will turn to the words with the suffix -unga (or -inga), used for deriving adverbs from adjectives or nouns, though not as frequently as the suffix -e. Since the suffix in question was originally part of inflected abstract nouns ending in -ung (or -ing), it functions primarily as an adverbial marker when added to adjectives, as shown in (7a); but when it is suffixed to nouns, the derived forms may sometimes be identical with the inflected forms, as indicated in (7b). (7)

a. dearn-unga 'secretly', eall-unga 'entirely', ierr-inga 'angrily', etc. b. fier-inga 'quickly', höl-unga 'in vain', wen-unga 'by chance', etc.

Here, another suffix -lice that has come to form the so-called -ly adverbs in Modern English must be mentioned. The suffix is formed as a result of compounding two suffixes, that is, an adverb-forming suffix -e and a suffix -lie which was used predominantly in order to derive adjectives from their corresponding nouns, hence, the suffix -lice was originally not attested in adverbs derived from adjectives. However, in the course of Old English period it came to be regarded as a marker of adverbs because it had already begun to be conjoined with many adjectives as well, as shown in (8). (8)

cläsn-lice 'cleanly', cüd-llce 'clearly', eadelice 'easily', snotor-lice 'wisely', strang-llce 'strongly', yde-lice 'easily', etc.

As far as the examples in (8) are concerned, the function of the suffix -lice seems to be discriminated from other lexical categories. But at the same time, we can find a large number of adjectives with the suffix -lice being used side by side with the same adjectives without the suffix, as shown in (9). This fact tells us that not a few adverbs came to be created anew by the addition of the suffix e to the adjectives with -lie. Hence, it would not be easy to ascertain whether a given adverb ending in -lice is derived from an adjective in -lie plus -e or an adjective plus -lice, nor could we safely conclude that the suffix -lice is the only unambiguous marker of derived adverbs in Old English (Quirk and Wrenn 1957: §165; Marchand 1969: 330; Koziot 1972: §475). (9)

a. b. c. d.

sär, sär-lic 'grievous' - sär-lice 'grievously' wräÖ, wräd-lic 'furious' - wräd-lice 'furious' gesäslig, gesäelig-lic 'happy' - gesielig-lice 'happily' mödig, mödig-lic 'proud' - mödig -lice 'proudly'

On identifying Old English adverbs


From the preceding observations, we may conclude that the derivational system of adverb is by no means rigid in Old English so that adjectives (or sometimes nouns) and derived adverbs may sometimes be identical in their forms, and this indeed causes an occasional ambiguity between the two categories within a larger contex, i.e., a clause or a sentence.

2. The position of adverbs within a sentence Next, we will investigate characteristic properties reflected in the position of adverbs within a sentence. As in Modern English, Old English adverbs take various positions and thus render the general descriptive exposition seemingly less feasible. However, directing our eyes to the syntactic behaviour of adverbs which limit the semantic force of verbs by denoting the nature or intensity of an action, state or quality, or circumstances generally, we can state as a general principle that adverbs of this kind usually occur in the position immediately preceding the verb they modify, as shown in (10) (Quirk and Wrenn 1957: §§142-3; Brunner 1962: 82-3). (10)

a. Her sindon dssra manna naman awritene Here are of-the men names written 'Here are written the names of the men' b. he wel cüpe Scyttisc he well knew Scottish 'He knew Scottish well' c. ne mihte; J>äer äräerde not could; there raised 'could not; raised there'

Focussing on the more specific features of such adverbs, we can demonstrate that adverbs of more than one syllable are often placed in a relatively posterior position, as represented in (1 la, b), and that in verbal constructions composed of an auxiliary and a verb, the adverb often occupies the position between the two elements, as illustrated in (1 lc, d). Although each of these rules may suffer from occasional exceptions, it is evident that the position of adverbs in relation to verbs can be stated rather distinctly (Carlton 1970: 56-7; Jacobson 1981: 1051; Mitchell 1985: §1138).



Yasuaki Fujiwara

a .we winnad rihtlice we fight justly 'we struggle righteously' b. /Elfred kyning hated gretan Wxrferd (...) Alfred King sends greetings to-Warferth (...) Luflice ond freondllce kindly and friendly 'King Alfred sends greetings to Warferth, with love and with friendship' c. Normanding weard swi9e gedreht Normandy was very-much afflicted 'Normandy was greatly oppressed' d. deah he geseö Ufes weg, he ne maeg medomllce ongdn though he see life's way, he not can suitably approach 'though he may see the way of life, he cannot properly follow it'

Thus, most of the problems which arise in discriminating between adverbs and adjectives, if such problems should arise at all, would reside more in the position of adjectives rather than of adverbs. This is due to the fact that adjectives used attributively like those in (12a, b, c) usually precede their nominal heads. But, at the same time, not infrequently do attributive adjectives occur after nouns they modify not only in verse but also in prose, as shown in (12d, e, f)· As a consequence, we are presented with some puzzling cases, as already exemplified in (1), especially (lc), in which the distinction between an adjective and an adverb cannot be readily drawn (Fries 1940: 199-208; Quirk and Wrenn 1957: §138; Carlton 1963: 773-8, 1970: 52-3). (12)

a. mid fullum friodome with entire freedom 'with full freedom' b. mJnne stronglican stol; into päsre hälgan stowe my strong seat; into the holy place 'my sturdy throne'; 'into the holy place' c. eall peos mSre gesceaft; and manig ööer göd man all this famous creation; and many another good man 'all this (or this whole) glorious creation'; 'and many another good man'

On identifying Old English adverbs



eallum; freoöoburh faegere; erug to-the-Danes all; stronghold fair; any meghond neor; land tiled relative nearer; land ploughed 'to all the Danes'; 'fair stronghold'; 'any nearer relative'; 'cultivated land' e. päer bid medo genöh; fram psem müpan öteweardum there is mead enough; from the mouth outward 'there is enough mead'; 'from the outward (part of the) mouth' Denum

f . Dä

st öd





Then stood the signature all thereon 'Then all the signature stood thereon'

3. The principle of discriminating adverbs from adjectives Thus far, we have demonstrated that the distinction between adjectives and adverbs is sometimes obscured in Old English owing to their formal identity or the diversity of the positions they occupy. Hence, our next task centers upon an investigation of the ultimate cause that is ascribed to such ambiguity. However, considering the fact that restrictions on word order are weak not only in verse but also in prose, the position of a given word is by no means a hopeful criterion on which to make a distinction between adverbs and adjectives in crucial cases. Therefore, it is necessary to extrapolate other linguistic principles which seem to be relevant to the distinction of the grammatical category and function of Old English words. Fortunately, some principles of verse composition are found by traditional metrics such as Sievers (1893), Schipper (1910), Kuhn (1933). For example, when a word is chosen for alliteration from among many, the choice is made in almost perfect accordance with the order of alliterative priority among lexical categories, without resorting to the meaning of words. However, in the traditional prosodic framework, there are too many exceptions to principles such as this to draw a significant generalization. Thus they are far from being an ideal model for the identification of lexical categories. For this reason, my foremost concern was the extrapolation and systematization of linguistic constraints within which Anglo-Saxon poets composed their verse. The following rules are derived as a result of various sorts of analyses attempting to capture the nature of verse composition devices that poets of the time seemed to follow (Fujiwara 1984, 1985).





a. In each alliterative half-line composed of more than one word the selection of a word for alliteration is automatically determined solely in accordance with the relative hierarchical ordering of lexical categories, regardless of the meaning or grammatical function of words. b. Priority of alliteration in each half-line is given in the following order: i) nouns, adjectives, derived adverbs, ii) inherited adverbs, iii) verbs, iv) function words. A word in a higher rank, irrespective of its position in the half-line, participates in alliteration in precedence over others of lower rank. c. When words constituting a half-line happen to be in the same category, alliterative priority is given to the word in the leftmost position of the half-line. d. Inherited adverbs are divided into three subcategories. Verbs are strictly separated into two classes: predominant, nonfinite forms and inferior, finite forms.

It should be noted here that in the first half-lines with double alliteration, that is, in those including two alliterating words, the above rules are not always strictly observed, especially in the determination of the second alliterating word. However, considering that stronger constraints are naturally imposed upon this type of half-line when selecting an additional word for alliteration, it is possible to claim that this is not due to a defect in our systematization but to the limitations of the principles followed by the poets. On the other hand, in the first half-lines with single alliteration, including one alliterating word, our principles (henceforth referred to as the Principles of Alliterative Hierarchy) are rigidly maintained. Indeed, since this is easily attested not only for a single poem but also for many other Old English verses, our principles seem to have been valid among Anglo-Saxon poets. Considering the fact that these principles are directly related to the discrimination of adverbs from adjectives, they seem to serve as an adequate standard for an analytic model, with respect to the problems we are dealing with. For example, the author of Beowulf sharply distinguishes between derived and inherited adverbs; the distinction is made explicit by two facts: i) when these two types of adverbs co-occur in the same half-line, preference of alliteration is given only to the derived adverbs, as shown in (14a); and ii) when an adverb is employed along with a noun or an adjective within the same half-line, it is the derived adverb that can take part in alliteration, as illustrated in (14b), not the inherited adverb, as can be seen in (14c). All citations in this paper are from Beowulf

On identifying Old English adverbs


edited by Klaeber (1950), unless otherwise indicated; romanized letters represent an alliterating sound or a group of such sounds. (14)

a . Nö







No here oftener come began shield-warriors 'No shield-warriors have ever attempted to come here more often' Beowulf244-5a

b. J>e is wide cüö who is widely known 'who is widely known' c.






he who formerly long time set





manna / / f i i s




at shores far looked 'he who formerly looked far, being ready, for the dear men for a long time at the shores' Beowulf 1915-6 Incidentally, recent studies on Old English adverbs, especially those focusing on their position within a sentence, such as Carlton (1970), Shores (1971), Jacobson (1981), and Mitchell (1985), have never made a distinction between derived and inherited adverbs. However, the difference between the two has a deeper significance to the complete linguistic description of Old English, because it often reflects some important facts. For example, most inherited adverbs are monosyllabic by nature, while derived adverbs are polysyllabic in their own right, and this numerical difference tends to be reflected in the difference of position of adverbs, as already suggested in (1 la, b). What must be pointed out next are the similarities among derived adverbs, their corresponding adjectives, and sometimes nouns, as well. Two reasons might be adduced in support of our classification of these three lexical categories into one group with the same status of alliterative hierarchy: one is, as shown in (14b), that they all share the same properties of alliteration, just as words in the same lexical class would, and the other is the fact that a derived adverb can construct a half-line by itself without being accompanied by any other word, just like nouns or adjectives do, as illustrated in (15), while none of the inherited adverbs can. And, it might be possible to add, as the third reason, the fact that the three word classes bear a close relation to each other with respect to derivation.



Yasuaki Fujiwara

ydellce 'easily' (1556a), earfodttce 'with difficulty' (1636a), ellenllce 'boldly' (2122a)

The preceding discussion being briefly summarized, derived adverbs evidently present several striking similarities to their corresponding adjectives or nouns in form, word order, and Principles of Alliterative Hierarchy. It would therefore be reasonable to assume that the AngloSaxon poets did not recognize such a substantial difference among the three word classes, although their functions were discriminated; this point will be taken up again later.

4. The subcategorization of inherited adverbs The subcategorization of inherited adverbs requires a more comprehensive explanation. As stated in (13d), the classification is a tentative one based upon the criterion of alliteration, that is, whether the inherited adverb or the verb is selected for alliteration when the two occur side by side within the same half-line. As the first step of analysis, we include, as topics of our investigation, the entire half-line in which all adverbs, i.e., adjuncts, disjuncts, and conjuncts which bear no direct, close relation of modification to verbs, are used along with a verb. The inherited adverbs, such as ford in (16a), belong to Class 1, and are always superior to verbs with respect to alliteration. On the other hand, adverbs in Class 3, like pä in (16b), are invariably inferior to verbs. Adverbs in Class 2, like äsr in (16c, d) do not show such clear behaviour; they are sometimes superior and sometimes inferior to verbs. (16)

a. ponne dü ford scyle, //metod-sceaft seon when you forth must, doom see 'when you must (go) forth to see the appointed doom' Beowulf 1179b-80a b. pä me past geläerdon leode mine Then me that taught people my 'Then my people taught me that' Beowulf 415 c. swd pü ser dydest as you formerly did 'as you did in the past' Beowulf 1676b

On identifying Old English adverbs


d. pass ne wendon aer witan Scyldinga It not expected before wise men Scyldings 'The wise men of the Scyldings did not expect it before' Beowulf 778 The following is a list of words in each class, in order of frequency. (17)

a. Class 1: ford 'forth', inne (or innan) 'within', feor 'far', feorran from afar', in 'in', äsfre 'always', üt 'out', up (or uppe) 'up', nider 'down', α always, hider 'hither', nean 'from near', etc. b. Class 2: ser 'before', eft 'again', oft 'often', wel (or well) 'well', pier 'there', gyt (or git) 'yet', seoööan (or syddan) 'afterwards', söna ' s o o n \ f e l a 'much', panon 'thence', etc. c. Class 3: pä 'then', ponne 'then', swd 'so', ne 'not, tela 'well', tö 'thereto', etc.

It is interesting to note that when an inherited adverb occurs along with another inherited adverb within the same half-line, selection for alliteration from among these two words is made rather mechanically in accordance with the subcategorization of inherited adverbs. In other words, in halflines containing two or more adverbs which are hierarchically distinctive, as shown in the examples of (18), it is the adverb of the higher rank that alliterates, irrespective of its position in the half-line, as shown in (18a, b). On the other hand, when two or three adverbs in the same subgroup occur side by side in the same half-line, it is the leftmost adverb that takes part in alliteration, as illustrated in (18c, d). These facts are in complete conformity with our principles that are valid for other lexical categories as well. (18)

a. se pe äer feorran com he who before from-afar came 'he who formerly came from afar' Beowulf825b b. oööe ä syddan esrfoöpräge II preanyd polaö or ever afterwards time-of-stress misery endures 'or afterwards (he) should endure the time of stress, misery, for ever' Beowulf 293 Adi c. hider üt aetbxr II cyninge minum hither out carried to-king my




'(I) carried (it) out hither to my king' Beowulf 3092-3a d. nean ond feorran from near and from far 'far near and from far' Beowulf 1174a feorran ond nean from far and from near 'from far and from near' Beowulf 839b

5. Adverbial nouns and adjectives Another perplexing question presented by adjectives or adverbs is closely connected not only with derived adverbs but also with case forms of adjectives or nouns used adverbially. In inflecting languages like Old English, all case forms except for the nominative may sometimes have an adverbial function just like inherited and derived adverbs. However, full inflection in such languages does not necessarily imply a complete inflectional system in which every case form can be fully discriminated from each other. Consequently, there are often cases in which the function of a given word is not easily identified; or, since case forms and derived adverbs are sometimes formally identical, its category may be ambiguous. It is commonly established as one of descriptive principles of linguists that basically, a derivational process results in a new word, whereas an inflectional process leads up to a different form of the same word, though the distinction is not always clear-cut (Adams 1973: 13, 27-8; Matthews 1974: Chapter 7). However, in earlier descriptions of Old English adverbials, we often come across the fact that words like wundrum, pass, hwilum, ealles, miclum, püsendmäelum are included in the category of adverbs, although they are, at first sight, undoubtedly inflected forms of adjectives or nouns; two distinct lexical classes are indeed given the same linguistic description on the ground that they are merely functionally, but not categorically, identical. As already shown, several obvious differences are recognized between inherited and derived adverbs. A new question now arises as to whether case forms used as equivalents of adverbs proper can really have the same linguistic properties as inherited and derived adverbs. Thus, our primary concern is whether the current distinction

On identifying Old English adverbs


between derived and inherited adverbs really corresponds to that of category. In this case also, we will use Old English verse and the Principle of Alliterative Hierarchy as an invaluable aid for the discrimination of lexical categories in the same way as the preceding analysis of derived adverbs. First of all, we will consider inflected forms of the adjective eall 'all' which often present puzzling grammatical questions since their functions or even lexical categories are sometimes ambiguous, as shown in (lc). Among the inflected forms of eal, the two varieties eal(l) 'quite, entirely' and ealles 'in every respect' are generally recognized as having an adverbial function in several instances in Beowulf. However, after the application of the Principles of Alliterative Hierarchy to all half-lines which include such instances, it turns out that the form eal in instances like (19) has the property of an adverb not only as a function but also as a lexical category. Hence, if the word in question in (19) is an adjective used adverbially, it would have priority of alliteration over the other word classes. (19)

Dü scealt tö frqfre eworpan // eal langtwldig leodum jjinum You must as solace be quite long-years people your 'You must be granted for quite long years as a solace to your people' Beowulf 1707b-8

Furthermore, when our principles are applied to the word under consideration, it also becomes clear that the other inflected forms ealle and ealne also permit the same interpretation as eal(l) and ealles. But in the case of the word eal of (lc), we cannot identify its category because the leftmost bil has the highest priority of alliteration, even if the second word eal might belong to the same highest group of words as the first one. Next, we will investigate inflected forms of micel 'great, large, much' used adverbially. Klaeber (1950) allows adverbial function for the following two inflected forms of the adjective micel in the whole body of Beowulf: genitive singular micles 'much, far' (only one instance) and instrumental singular micle 'much' (three examples). However, as a result of our analysis, it was found that for one instance of micles (694b) and two of micle (1283a, 2651a), adjectival properties are not observed, and that the remaining singular instance of micle in (20) apparently functions as an intensifier of oftor 'more often', like other common adverbs. Hence, if it is an adjective, alliteration should properly be realized in micle irrespective of its position, because its alliterative priority is higher than the preceding


Yasuaki Fujiwara

word ofior, on the contrary, if micle is an adverb, it will offer no problem at all not only with its grammatical function but also with the Principles of Alliterative Hierarchy. Moreover, as our principles are strictly maintained in the first half-line, it would be less reasonable to ascribe the case in (20) to exception. (20)

öära Jje he geworhte tö West-Denum II oftor of-the which he made to West-Danes oftener micle Sonne on Snne sld much than on one time 'of the which he had made upon the West-Danes, far more often than on that one occasion' Beowulf 1578-9

From these facts and discussions it may be safe to conclude that in the Old English period, micle, and probably micles also, were thought to be adverbs in the proper sense of the word, i.e., in both function and category. Incidentally, as already pointed out in (11), since Old English adverbs may sometimes occur after the words they qualify, micle would cause no trouble with respect to syntactic positions it occupies, if it is regarded as an adverb.

6. The function and category of some nouns used as adverbs Next, we will investigate the problem involved in the relation between function and category of some inflected forms of nouns which may sometimes be used as adverbs. As a notable example, we take hwllum, which is often cited not as a noun used adverbially, but as an adverb proper despite the fact that it is apparently an inflected form of the noun hwll. Among the inflected forms of hwll '(space of) time', the two forms hwlle and hwllum are often used adverbially in Old English with the meaning 'sometimes, formerly, etc.'; but in Beowulf only hwilum is used as an equivalent of an ad verb. Hence, it follows that our analysis is directed to the case form in question; and our investigation based upon the Principles of Alliterative Hierarchy concludes that in as early as the Old English period, the case form hwllum was granted the same linguistic status as inherited adverbs. The following discussion is devoted to the presentation

On identifying Old English adverbs


of several pieces of evidence which motivate our conclusion. First, we present the fact that in all examples of half-lines in which the case form in question is used along with a noun or an adjective in the same half-line, as shown in (21), not a single instance is found in which hwilum takes part in alliteration. (21)

a. Hwilum flitende Sometimes quarreling 'Sometimes in contending' Beowulf 9\6& b. Hwilum for (d)ugude Sometimes before nobles 'Sometimes before the tried warriors' Beowulf2020a c. hwilum sylllc spell sometimes strange story 'at times a strange story' Beowulf 2109b

Still more interesting is the fact that in half-lines such as (22a, b) with single alliteration, the word hwilum is superior to the finite forms of a verb and hence alliterates, but sometimes inferior to adverbials and does not take part in alliteration; meanwhile, in half-lines like (22c) with double alliteration, hwilum may sometimes be one of the alliterating elements together with a noun or an adjective within the same half-line. All these facts converge to suggest that hwilum has the same status as inherited adverbs in spite of its ostensive case ending. (22)

a. hwilum dydon sometimes did '(they) sometimes did' Beowulf 1828b b. hwilum eft ongan sometimes again began 'at times (he) began again' Beowulf 2\\\a c. Hwilum heaporqfe Sometimes brave 'Sometimes brave (men)' Beowulf 864a




As the next step, with the aid of concordance by Bessinger and Smith (1978) we attempt the same analysis as before, of 120 examples of hwilum that are attested in the whole body of Old English poems. The result was that we could not find in any example of hwilum the characteristic properties that a noun would exhibit, although Klaeber (1950: 225) cites one example in (23) with a remark that this is probably the one instance of hwilum which is not used as an adverb in all Old English poems. (23)

nihtes hwilum of-night time 'at the time of night' Beowulf3044a

A large number of case forms of nouns or adjectives other than hwilum are of course often used adverbially, as shown in (24). However, a close one-by-one examination of these words discloses that no single instance of such case forms falls under the category of the adverb although they are undoubtedly exhibiting an adverbial function. (24)

-es: -φ: -um:


dseges 'by day', nihtes "by night' häm 'homewards', lyt 'littele, not at all', wlde-ferhö 'for a long time, for ever', jyrenum 'wickedly', gepyldum 'steadily', lustum 'with joy', stundum 'time and again', wundrum 'wonderfully' weorce 'painfully', wihte 'in any way, at all'

As a matter of fact, most inflected forms, when used as adverbs, do not change their lexical category. Thus, a new question immediately comes to mind: why do specific inflected forms of nouns or adjectives such as hwilum, micles, and ealles only acquire the same linguistic properties as inherited adverbs? To furnish a satisfactory solution to this newly-born question, we will have to take a new look at the details of inflection and derivation.

7. The relationship between derivation and inflection As far as the preceding investigation is concerned, derived adverbs and case forms of nouns or adjectives used adverbially usually fall under the

On identifying Old English adverbs


same lexical group, but are discriminated sharply from inherited adverbs, although all of them perform identical functions; Anglo-Saxon poets seem to have intuitively regarded the former two types of words as members of the same category, though their functions may sometimes be different from those of their matching adjectives or nouns. Indeed, these facts seem to be a reflection of the linguistic intuition of those who had recognized such a great difference in category between nouns or adjectives and their derived or inflected forms used adverbially; in contrast, our modern linguists tend to distinguish sharply between the grammatical categories of the two word groups. Thus, the residual problem is why only a small number of inflected forms exhibit characteristic properties common to inherited adverbs. However, we must stop here to note the fact that some forms like hrade that are commonly regarded as derived adverbs seem to have acquired properties exclusive to inherited adverbs. Thus, in the following discussion, we will have to investigate the relationship between derivation and inflection in terms of these types of words, together with the linguistic status of case forms like hwTlum. First, we will take up hrade (or hrsede), which is commonly regarded as a derived adverb composed of the adjective hrad 'quick' and the suffix -e. Indeed, Campbell cites the word as the first instance of the formation of a derived adverb by means of the suffix -e (Campbell 1959: §661; Luick 1964: §161.1, §392.1; Sievers 1968: §201). However, when we conducted an analysis in accordance with the Principles of Alliterative Hierarchy, it became clear, as shown in (25), that in all examples of Beowulf the word is inferior to nouns and adjectives, and thus does not take part in alliteration as in (25a, b, c); in fact, it was sometimes found to be inferior even to finite verbs and inherited adverbs as shown in (25d, e). (25)

a. hrape heo aepelinga ännahaefde// fasste befangen quickly she of-nobles one had fast seized 'quickly she had seized one of the noble men fast' Beowulf 1294-5a b. Hrape waes tö büre/ Beowulf fetod Quickly was to bower Beowulf fetched 'Beowulf was quickly fetched to the bower' Beowulf 1310 c. Hrsepe wearÖ yÖum/ mid eofer-spreotum Quickly was in waves with boar-spears '(He) was quickly (pressed) with the boar-spears in the waves' Beowulf 1437


Yasuaki Fujiwara

d. ac he geßng hraöe/ forman slöe// släspendne rinc but he seized quickly first time sleeping warrior 'but he quickly seized a sleeping warrior as a beginning' e. Heo him eft hraöe// andlean forgeald She him again quickly retribution yielded 'She quickly yielded him a recompense again' Beowulf 1541 Such above-mentioned metrical peculiarities are common to all inherited adverbs but are never observed in derived ones. It is therefore reasonable to assume that although the adverb hraöe ends in -e and seems to have formal and semantic correspondences with the adjective hrad, the whole lexical form hraöe had already acquired its linguistic status as an independent inherited adverb, and was not regarded as a form due to derivation in the historic Old English period. Several pieces of evidence strongly support our hypothesis. First of all, as Campbell admits, the adjectival form nearly always has a /d/ sound, whereas the adverbial form almost invariably indicates [δ]. Second, as suggested earlier, the final -e is not always a marker of Old English derived adverbs. Third, a relatively new suffix -lice is often added to the adjective ending in /-d/ to form a new adverb fradITce, but the suffix is never employed with adjectives ending in [-Ö]; indeed no single instance of *hraö-Uce is evidenced in Old English. Fourth, hradllce always takes part in alliteration in Old English poems, while hraöe does not expect in such specially motivated occasions as half-lines composed of function words and hraöe. Thus, it is evident that these two forms were clearly distinguished as early as the Old English period.

8. Concluding remarks The preceding investigation leads us to assume that, whether derived or inflected, words, which were used as adverbs long before the advent of the earliest Old English documents, continue to be employed in the same way as inherited adverbs in later periods, without radical modification of their original forms. This assumption is supported in part by some sporadic remnants such as rathe and whilom in Modern English whose forms are directly traceable back to Old English hraöe and hwilum, respectively; indeed, the presence of whilom cannot be explained unless we assume that the entire lexical form hwilum was taken to be an adverb proper, not as a

On identifying Old English adverbs


case form used adverbially, considering the fact that other adverbial case endings have almost entirely disappeared from the stock of Modern English words. As a conclusion of the preceding discussion, we can assert that it is not always adequate to apply to Old English adverbs the traditional principle of description, that is, derivation changes grammatical category while inflection does not. The facts obtained from the present analysis enable us to assume that derivational suffixes in almost all adverbs and also inflectional endings in some nouns and adjectives were never regarded by the Anglo-Saxon poets as being capable of changing the lexical category itself of the word to which they were added; they were merely signs indicating that the nouns or adjectives had an adverbial function. Indeed, most of the earlier linguistic analyses or descriptions of Old English have been made under the assumption that Old English is a synchronic existence and both inflection and derivation are regarded as static phenomena. However, the preceding analyses and the following well-known facts vividly suggest that more dynamic approaches and new analytic criteria are required for the identification of Old English adverbs: i) some derived adverbs have evidently attained the same linguistic status as inherited adverbs, ii) since reduction of inflection had already started as early as late Old English, it is supposed that an inflectional ending itself was never a static entity even in those days, iii) the fact that a new suffix -lice came to be used for deriving adverbs suggests to us that the derivational system itself was not regarded as a fixed one.

References Adams, Valerie 1973

An introduction to Modern English word-formation. London: Longman.

Bessinger, Jess Belsor, Jr. - Philip H. Smith, Jr. (eds.) 1978

A concordance

to the Anglo-Saxon poetic

records. Ithaca and London:

Cornell University Press. Brunner, Karl 1962

Die englische


Ihre geschichtliche


Vol. II: Die

Flexionsformen und ihre Verwendung. Tübingen: Max Niemeyer. Cambell, Alistair 1959

Old English grammar. Oxford: Oxford University Press.

Carlton, Charles 1963

"Word order of noun modifiers in Old English prose", JEGP 62: 778-783.


Yasuaki Fujiwara 1970

Descriptive syntax of Old English charters. (Janua Linguarum: Series Practica 111.) The Hague: Mouton.

Dobbie, Elliott Van Kirk (ed.) 1953 Beowulf and Judith. (The Anglo-Saxon Poetic Records. Volume IV.) New York: Columbia University Press. Fries, Charles Carpenter 1940 "On the development of the structural use of word-order in Modern English", Language 16: 199-208. Fujiwara, Yasuaki 1984 "Koeigo no goihanchu [Lexical categories of Old English]", Tsukuba Eigaku Tenbo 4: 23-30. 1985

"Koeishi no seiritsu to inritsuron [The growth and metrics of Old English verse]", Gengobunka Ronshu 18: 1-29.

Jacobson, Sven 1981 Preverbal adverbs and auxiliaries: A study of word orderchange. (Stockholm Studies in English 60.) Stockholm: Almqvist & Wiksell International. Klaeber, Friedrich (ed.) 1950 Beowulf and the Fight at Finnsburg. Boston: Heath. Koziol, Herbert 1972 Handbuch der englischen Wortbildungslehre. Heidelberg: Carl Winter. Kuhn, Hans 1933 "Zur Wortstellung und -betonung im Altgermanischen", Beiträge Geschichte der deutschen Sprache 57: 1-109.


Luick, Karl 1914-1940 Historische Grammatik der englischen Sprache. Volume I. Part I, Part II. Stuttgart: Bernhard Tauchnitz. [ 1964] [Reprinted Oxford: Basil Blackwell] Marchand, Hans 1969

The categories and types of Present-day English word-formation. (2nd edition.) München: C.H. Beck'sehe. Matthews, Peter Hugoe 1974 Morphology: An introduction to the theory of word-structure. London: Cambridge University Press. Mitchell, Bruce 1985 Old English syntax. 2 vols. Oxford: Oxford University Press. Pope, John Collins 1967-1968 The homilies of/Elfric. (EETS 259, 260. London: Oxford University Press. Quirk, Randolph - C. L. Wrenn 1967 An Old English grammar. London: Methuen.

On identifying Old English adverbs


Schipper, Jacob 1910 A history of English versification. Oxford: Clarendon Press. Shores, David L. 1971

A descriptive syntax of the Peterborough Chronicle from 1122 to 1154. (Janua Linguarum: Series Practica 103.) The Hague: Mouton. Sievers, Eduard 1893 1903

Altgermanische Metrik. Halle: Max Niemeyer. An Old English grammar. (Third edition.) [Translated and edited by Albert S. Cook.] New York: Ginn & Company. [1968] [Reprinted New York: Greenwood Press] Skeat, Walter William 1881-1900 jElfric's lives of saints. (EETS OS 76, 82, 94, 114.) London: Oxford University Press. Wright, Joseph - Elizabeth Mary Wright 1925

Old English grammar. London: Oxford University Press.

On the construction I was go walked Masayuki Higuchi A "curious"1 idiom occurs in The Book of the Duchess (hereafter abbreviated as BD) and The Canterbury Tales (hereafter CT): (1) (2)

(BD 387) I was go walked for my tree, (CT ΠΙ, 1778) His felawe was go walked into toun Forth with his knave,

(1) and (2) seem to be the only extant examples in the English language of the construction go walked (their equivalents included). No other instances have been recorded in the OED, MED or Visser. A discussion on how and why this rare construction was formed will contribute to (a) reading Chaucer's texts correctly, and (b) offering a generalized explanation of ME constructions hitherto unrelated.

1. Previous studies On the construction go walked, as Chaucer puts it, "diverse men [have] said diverse things" (cf. CT Π, 211). Here is a list of explanations I have •i

collected: Gräf (1888: 94) Das Part. Perf. scheint nur durch Analogie veranlasst zu sein in folgender Stelle, wo es zudem pleonastisch steht. Duchesse 387 I was go walked for my tree. [The first half of this explanation is unsubstantiated because Gräf does not mention with what form walked is on analogy. Cf. note 7.] Skeat (1900, 5: 272f.) The simple explanation [of goon a-blakeberied] is that, by a grammatical construction which was probably due (as will be shown) to an error, the verb participate, in such a manner as to give the participle the force of a verbal substantive. In other




words, instead of saying 'he goes a -hunting,' our forefathers sometimes said 'he goes a-hunted.' ... The explanation of this construction I take to be this; the -ed was not really a sign of the past participle, but a corruption of the ending -eth (A.S. ad) which is sometimes found at the end of a verbal substantive.... In D. 1778, we even find go walked, without a. Kenyon (1909: 4fn.) I suggest that the use of this expression [was go walked] may also have been assisted by the analogy of the expression go walk, felt as a single word. To our present feeling, at least, such expressions as go get are essentially equivalent to single words with two phases of meaning. Go get is synonymous with fetch, which often has the double sense. It would then be easy, perhaps with the assistance of the verbal in -ej?, -ed, to feel go walked as part participle of go-walk (cf. walked is, A 2368), just as goth walketh, G 1207, could be regarded as the imperative of it. Robinson (1957) Walked is probably for a-walked, like a-blakeberyed, Pard Prol, VI, 406 (p. 707). The construction of walked appears to correspond to that of the past participle in German ("kam gelaufen", etc.). But there may be involved a confusion with nouns in -ed, earlier -eth, -ath (p. 775).4 Mustanoja (1960: 582) was go walked would seem to exemplify a semantic weakening of go ... and to be roughly equivalent to was walked. ... Connection with German er kam gelaufen, current since MHG, is less likely. Baugh (1963: xxxvi)5 What is possibly an analogical extension of this construction [e.g., Hath doon yow kept] is ["] I was go walked for my tree["] BD 383. Visser (1966: 1248) If this conjecture [-ed has developed from OE -ad and -ep] should be correct one might further suppose that a has disappeared in 'go walked',

On the construction I was go walked


Peters (1980: 101) Some Chaucerian constructions contained redundant words, as ... ME go in I was go walked Ί had walked'; Kerkhof (1982: 138) In the first group [/ go walked] the usual meaning of goon has become so weakened that it only serves to denote the beginning of the action.6 The explanations quoted above may be grouped into four basic views: a) b) c)


an "expletive view", which takes go as a syntactic expletive (Graf, Mustanoja, Peters, Kerkfhof); a "single word view", which takes go walked as a past participle of go- walk (Kenyon); a "remnant-of-OE view", which regards go walked as derived from go a-walked and takes the -ed as a corrupt form of OE eth\ an "attraction view", which takes walked as attracted to the preceding perfect was go ['gone'] (Baugh).7

Among these views the expletive view is not supported by syntactic observations since it says nothing about how or why go was followed by walked.8 Though the remnant-of-OE view is most widely accepted,9 the explanations given by Kenyon and Baugh sound equally valid. In what follows, I shall first examine Kenyon's explanation, secondly, arguing against the remnant-of-OE view, compare go walked with seemingly similar constructions, and lastly, supplement what Baugh left undocumented with instances relevant to go walked.

2. Past participial form of GON Judging from the expressions "felt as a single word" and "past participle of go-walk", Kenyon seems to consider go walk a compound word (cf. Visser 1966: 1248). This view, however, is not evidenced by similar constructions: there are no such compound forms as *he go-walketh/*gorunneth/*go-creepeth/*come-walketh or *he go-walked/*he has gowalked/*he was go walking/*he saw her go-walking. Another problem


Masayuki Higuchi

with Kenyon's explanation is that he does not discuss whether go is the stem, or the past participle, of GON; if by "a single word" he means a compound, go should be a stem or imperative since the first verbal element of a compound is usually a stem or imperative, as in whetstone, runway; forget-me-not, pastime·, if he does not mean that, go must be a participle since it follows was, an auxiliary for forming the perfect. Now it is our task to clarify the form of go in (1-2). In Chaucer's English the past participle of GON can be gon(e), y-go(n) or go(o),i0 the last of which occurs not only in rime position: (3)


(CT Π, 1006) This senatour is to kyng Alia go [: two] To feste, Also in: CT Vffl, 907; LGW 1656; 2656; RR 2423; etc.but in mid-position as well: (HF 434) thoo sawgh I grave how to Itayle Daun Eneas is goo to sayle;

Examples (3-4) show that when GON is used with BEN, the past participle form can be go. Also in other ME works go is used as past participle of GON (notice go in (5-6) occurs with is):11 (5) (6) (7)

(CA I, 64) For it is siththe go noght longe, Also in: CA V, 6420; 7793; VI, 1821; 2102, etc. (Orfeo 194) J)e King into his chaumber is go And oft swoned opon J>e ston (King Horn 1176) Die habbe go mani mile,

The examples above suggest a high probability that, counter to Kenyon's suggestion, the go in was go walked is a past participle.12

3. T h e go α-begged


Saying "the examples of this use [i.e., go a-hunted] are at least seven", Skeat (1900: 273) cites the following (the citations below are supplemented from texts given in the references):13 (8)

(PP1 C. ix. [sic] 138) folk that gon a-begged,H Also in CT V, 1580.

On the construction I was go walked

(9) (10) (11) (12) (13) (14)


(CT VI, 406) Though that hir soules goon a-blakeberyed\ (PP1 C. VIE, 246) In somer for his sleuthe he shal haue defaute And go a-bribeth and a-beggeth and no man beten his hunger. (CT ΠΙ, 354) But forth she wole ... To shewe hir skyn and goon a-caterwawed. (CA V, 145) This Prest was drunke and goth astraied, (CA I, 2030) Wher inne he wolde ride amaied (Robert of Gloucester) As he rod an honteth, and parauntre his hors spurnde.

Other examples recorded in the OED, MED and Visser (1973: 2027, see also 1973: 1910) are: (15) (16) (17)

(Jonh 21: 3) Ic wylle gan on fixaö [AV: go a fishing]. [OED, s.v. On 19] (St. John, Southern Legendary 20) J)are he saigh tweie brej>ren [DO ... afischeth gon [MED, Visser] (Miller, The Soft Talkers (Penguin) 69) that didn't go unnoticed

Also in Sillitoe, Key to the Door (Pen Bks)23. [Visser] Out of this list go unnoticed in (17) cited by Visser should be excluded because it is of the GON ytressed type (see the following section) and because it cannot be paraphrased as *go (by/for) unnoticing. A glance at examples (8-16) reveals that all the words in -ed or -eth begin with a(n)-. Was go walked is different from these instances in that walked is not preceded by the prefix. In other words, the remnant-of-OE view is not supported by parallel constructions such as *go begged/crept/danced.15 Another difference between was go walked (1-2) and go a-begged (8-16) is that, while the former are in the perfect aspect, the latter are all in the simple tense (cf. 5). More importantly, a decisive difference between them is that, though go walked denotes 'manner' (i.e., go by walking), go α-begged denotes 'action' and can be paraphrased with -ing or to/into (cf. OED, s.v. A prep.113; Visser, 1973: 1910).16 Thus, despite the wide acceptance of the remnant-of-OE view, was go walked is not parallel to go a-begged.


Masayuki Higuchi

4. The go ytressed type A construction which may seem parallel to go walked is "GON/COMEN + predicative adjunct" such as go ytressed ['tressed'] (the go walked type and go ytressed type are mentioned without distinction in Hinckley 1907: 165f. and Visser 1966: 1247ff.): (18)

(TC V, 810) To gon ytressed with hire heres clere Doun by hire coler at hire bak byhynde, Also in: CA I, 2218 (gon despuiled); Π, 2489 (bejaped wente); ΙΠ, 830 (go unteid); IV, 3333 (cam desguised).

Instead of GON this type may take COMEN: (19)

(CA IV, 3333) Mercurie, which was al affiled ['prepared'] This Cow to stele, he cam desguised,

Between the go walked type and go ytressed type, however, there are three differences: (a) walked in the former type derives from an intransitive verb, while the -ed form in the latter type derives from a transitive verb; (b) walked in the former type denotes a "manner" of going, while the -ed form in the latter type denotes a "state" of the subject when (s)he goes; and (c) the subject's role in the former type is an agent of walked, while the subject's role in the latter type is a patient to the -ed form. Go walked is thus not parallel to go ytressed.

5. The formation of was go walked How, then was the idiom was go walked formed? To my mind, two factors contributed to its formation: (1) GON verb combination and (b) attraction to the preceding BEN. It should first be noted that in Middle English GON is, like hear say, let go, make believe, go see or come dine, often followed directly by another verb: (20)

(CT I, 1838) He moot go pipen in an yvy leef; Also in: CT I, 3547 (go gete); 3685 (goon slepe); 4094 (go knede); III, 108 (go swlle), etc. Cf. TC III, 560 (come soupen); IV, 654 (Come speken).

On the construction I was go walked


The second verb may be a verb of motion; that is, GON can be followed not only by a present participle, as in (21), but also by a bare infinitive, as in (22):

(21) (22)

(CA Π, 758) Wher thei go walkende on the Stronde, Also in: CA ΠΙ, 364; cf. V, 4957; Vffl, 1162. (CT X, 721) Thise been the newe sheepherdes that leten hirsheep wityngly go renne to the wolf that is in the breres, Also in: CT I, 2760 (go ber); 4250 (go crepen). Cf. CA I, 350 (cam ride); Octavian 196 (com fly) [Visser].

When a mutative verb in ME forms a perfect, it usually requires BEN as the auxiliary: (23) (24)

(CT I, 1413) To Atthenes is he goon the nexte way. (CT I, 2368) Arcite unto the temple walked is

We can argue from the examples (20-24) that GON, when followed by a bare infinitive walk, derives go walk, which in turn derives BEN gone walk, if used in the perfect. This means that the "correct" form of was go(ne) walked is was go(ne) walk. (The lack of a recorded instance of was


walk seems to be accidental since the possibility of this construction can be deduced from the examples (20, 22) and (3-6, 23).) Our task now is to consider why walked is used instead of walk. The above-mentioned combination of GON verb was so close that the second verb was inflected in the same way as the preceding verb (cf. Kenyon 1909: 4, Visser 1969: 1398). Compare the following examples:17 (25) (26)

(CT VI, 201) Go bryng hire forth, and put hire in oure warde (CT 10: 2186) Gooth bryngeth forth the vessels (taken from Blake's text)

More relevant to was go walked is the attraction observed in the type I have herd (NP) said (including synonymous words of SAIN). As pointed out by Higuchi (1984: 119), while the infinitive form say appears irrespective of the tense and aspect of HEREN, the past participle said appears only when HEREN is in the perfect (except for an explainable "exception").18



Masayuki Higuchi

a. b. c. d.


a. b. c. d.

(CT V, 603) thus herde I seye. Also in: CT m, 24; VE, 518; TC I, 876; ΙΠ, 1659, etc. *I hear/heard said, (no occurrences) (RR 2952) as ye have herd say, Also in; CT Π, 613; ΠΙ, 1675; V, 1602; VII, 1964, etc. (LGW 1167) as I have herd seyd. Also in: CT 1,4129; IV, 278; 1637; TC I, 197, etc. (CT V, 953) ye heere me seye, Also in: CT I, 953; 2210; 3364; 3817, etc. *I hear/heard NP said {said, has infinitival force), (no occurrences) (CT VI, 836) as ye han herd me seye. Also in: CT ΠΙ, 1670; VH, 1551; 1697; 3038, etc. (CT V, 1547) as ye han herd me sayd; Also in: TC 1,1009; CA I, 3153; V, 1623; 7609.

The nonexistence of hesr (NO) said (where hear is nonperfect) suggest that the occurrence of said in HAVEN heard (NP) said is due to attraction to the preceding HAVEN heard".

The same kind of attraction can also be observed in the context with a causative verb:19 (29) (30)

(31) (32)

(CT Π, 171) Thise marchantz han doon fraught hir shippes newe, (CT I, 1913) And northward ... An oratorie ... Hath Theseus doon wroght ['Theseus has caused his men to make an oratory'] in noble wyse.20 Cf. (CT I, 1905) He estward hath ... Doon make ['has caused his men to make'] an auter and an oratorie; (TC ΙΠ, 304) Ο ['one'] tonge ... Hath mad fill many a lady bright of hewe Seyd 'Weilaway, the day that I was born!' (CT IV, 1991) Which wolde han lat hym storven in the place (Globe edition)

An example from Chaucer's contemporary: (33)

(CA Π, 1799) Riht so behinde his brother bak With false wordes whiche he spak He hath do slain, Cf. also CA IV, 249 (he hadde wold His time kept).

On the construction I was go walked


Returning to the construction was go walked with this in mind, we can easily explain why was go is followed by walked: walked is attracted to the preceding auxiliary was. The construction I was go walked can thus be explained, without resorting to an OE remnant, as a legitimate grammatical phenomenon common in Middle English. The parallelism between was go walked and have heard (NP) said/have made NP said/has done slain has not been explicitly pointed out up to now (except for Baugh's short note) because sufficient attention has not been paid to the tense and aspect of the main verb.

6. W h y not *go


A problem hitherto left undiscussed is the reason why *BEN go(ne) begged/crept/danced did not come into use. This is probably because, go begging/creeping/dancing being the "correct" construction, BEN go(ne) begged was felt ungrammatical. Why, then, was it possible, though sporadically, to say I was go walked? A speculation is that the combination go walk was felt stronger than other combinations such as go beg/creep/dance because GON in Middle English was synonymous with WALKEN. In the combination go walk, as suggested by the expletive view, the meaning of GON was submerged in WALKEN, while in other "GON + verb" combinations GON kept its meaning 'proceed, move, pass".

7. Conclusion We can now safely conclude that walked in the curious idiom took the participial form through attraction to the preceding auxiliary possibly assisted by the superficial resemblance to the go α-begged type and go ytressed type, and by the synonymity of GON and WALKEN.

Notes 1. Skeat (1899,1:474). 2. Other abbreviated titles: HF = The House of Fame; LGW = The Legend of Good Women; PR = The Romaunt of the Rose; TC = Troilus and Criseyde. Quotations from


Masayuki Higuchi Chaucer are taken from The Riverside Chaucer ed. by L. D. Benson, 3rd ed. (O.U.P., 1988).

3. Some of the explanations quoted here are given by Nakao (1972: 329). 4. This note is repeated by Benson (1988: 969). 5. Cf. also Baugh's note to CT F, 1580 (1963: 484fh.). 6. By "the beginning of the action" Kerkhof seems to mean the inchoative aspect (cf. Kerkhof 1982: 88). 7. 'Attraction' is "generally a purely mechanical process, being the result of simple contiguity, by which a word is made to agree with another word with which it would otherwise not be connected grammatically, as in the opinion of several eminent lawyers were in his favour, where were, although grammatically connected with the singular noun opinion, is put in the plural as if it were governed by lawyers" (Sweet 1891: 44). 'Attraction' should not be confused with 'analogy', which is a "process or result of grammatical and lexical forms changing under the influence of some other regular pattern of the language. Thus hisn for his may be used on the pattern of my:mine, knowed for knew by analogy with mow:mowed, brang for brought by analogy with sing: sang" (Hartmann - Stork 1972: 13). 8. It may possibly be argued that go was expletively inserted merely to satisfy meter. However, meter is not the main cause for the anacoluthon because, whether "was go walked" is pronounced [wäz gat note place (Sir Gawain 2092).

The investigation of contracted negative forms in the four poems in question and in some other Middle English texts strongly suggests that at least the author of Pearl can be differentiated from the author(s) of the other poems in the manuscript. The inference is based upon the following table which displays how frequently relevant examples show negative contraction in Pearl, Cleanness, Patience, and Sir Gawain.4 Table 1. Contracted and uncontracted forms in the Cotton Nero A.x. poems Pearl



Sir Gawain











Although relevant examples are not abundant, the tendency in the four poems is interesting to note. As Table 1 shows, Pearl provides uncontracted forms more commonly than contracted ones, while relevant examples occur almost always contracted in Cleanness, Patience, and Sir Gawain. Since examples of contracted forms in Pearl are both in the form of nis (nys), the general tendency in the text is, as a matter of fact, in favour of uncontracted forms, when nis (nys) is excluded. The uncontracted forms in the text are given below: (9) (10) (11) (12)

I ne wyste in J>is worlde quere J>at hit wace (Pearl 65) J)ou ne woste in worlde quat on dot3 mene (Pearl 293) Of {>e way a fote ne wyl he wryt>e (Pearl 350) And also Jjer ne is neuer ny3t (Pearl 1071).

Outside Pearl, examples of uncontracted forms are virtually limited to the single example in Cleanness:5



Yoko Iyeiri

Jjat he ne wyst on worde what he warp schulde (Cleanness 152).

Thus the overall situation of the Cotton Nero A.x. poems may be summarized in the following manner. First of all, Pearl provides two examples of nis (nys), but apart from them, it shows a strong tendency to yield uncontracted forms. Secondly, the tendency or rather the norm in Cleanness, Patience, and Sir Gawain is to the contrary, in that they employ contracted forms almost entirely. There is only one counter example here. The difference of usage between Pearl and the other three poems is outstanding in the above table. The question is whether the difference of usage of negative contraction is meaningful enough to constitute evidence for multiple authorship of the poems.

3. Negative contraction in Middle English From the investigation of certain other Middle English texts, I am much inclined to advocate that the differences in usage of negative contraction can be at least one of the keys with which to discuss the issue of authorship. It is most unlikely that the same poet would employ contracted forms almost exclusively in one text and uncontracted forms predominantly in another text. As far as Middle English is concerned, whether relevant verbs occur in contracted or uncontracted forms is notably consistent even in verse, although in verse texts, we are generally tempted to expect some irregularities due to particular verse forms. The issue is dialectal to a large extent, as Levin (1958) points out.6 In Northern texts, relevant examples exhibit a strong tendency to occur uncontracted, while on the contrary in Southern and West Midland texts, contracted forms are favoured. As for Northern texts, Levin investigates Cursor Mundi (pp. 1-100 of Morris' edition), Songs of Laurence Minot, and Barbour's Bruce (Π. 12,000), and finds sixteen relevant examples, all of which stay uncontracted. Negative contraction was unavailable as far as his investigation is concerned. Even though Fulk (1992: §147) points to the existence of the contracted forms nart and nere in other parts of Cursor Mundi,1 it is at least true to say that the phenomenon is exceptionally rare in the North. I have made a slightly extended study of MS Cotton Vespasian of Cursor Mundi (Π. 1-10, 122), which again provides no contracted forms, although there are 52 relevant examples in this portion of the text. My investigation

MS Cotton Nero A.x. poems once again



of English Metrical Homilies, also localized in the North, yields seventeen relevant instances, all of which occur uncontracted. Not only Levin's analysis but also mine certifies that the situation is regular even in verse. Relevant examples are constantly contracted in the North. In fact, the area where uncontracted forms prevail seems to spread to the south of the Humber. I have investigated The Alliterative Morte Arthure, localized in Lincolnshire,9 which provides only example of negative contraction as against eleven examples of uncontracted forms. The same tendency is found with The Destruction of Troy, localized in Lancashire.10 The text provides only one instance of negative contraction as opposed to seven instances of uncontracted forms. Contracted forms are extremely limited in broadly northern areas of England, and the tendency is seen with near-consistency. Among the four alliterative Cotton Nero A.x. poems, only Pearl shows an inclination which is relatively close to this situation. Turning to the Southern and the West Midland areas of England, Levin (1958) studies some ten Middle English texts and observes only two uncontracted examples among them, while he comes across as many as 495 examples of negative contraction. Relevant examples are almost always contracted in the South and West Midlands. Though in this case many of Levin's texts are from relatively early periods of Middle English, LALME also finds a number of contracted forms in the South at least, and confirms the abundant occurrence of the phenomenon in late Middle English as well. I have also investigated a part of MS Corpus Christi College 145 of The South English Legendary, localized in Berkshire,11 and found 626 examples of negative contraction as opposed to only three examples of uncontraction. It seems that the occurrence of negative contraction was almost regular in the South and the West Midlands in Middle English, and that this was the case even in verse. Unfortunately, the West Midland area is often excluded from the purview of LALME, as far as negative contraction is concerned. For example, it does not reveal the situation of the phenomenon in Cheshire, where the manuscript of the poems under discussion is localized.12 However, it is at least clear that Cleanness, Patience, and Sir Gawain, where negative contraction highly commonly occurs, certainly belong to the group where the phenomenon of contraction is especially frequent. However this is not the case with Pearl In view of the strongly systematic occurrence of either contracted or uncontracted forms in Middle English dialects,13 it is highly unlikely that a single poet, who of course utilizes the same dialect, would employ contracted forms consistently in one text and uncontracted forms


Yoko Iyeiri

predominantly in another text.14 In fact, the difference in usage of negative contraction between Pearl and the other three poems is large enough even to suggest a slight difference of locality. It is at least reasonable to conclude that the author of Pearl is different from the author(s) of the other three poems.

4. An additional comment Finally, a small comment can be made upon the two contracted instances of nis (nys) in Pearl, which are repeated below: (14) (15)

Nis no wy3 worJ)e J)at tonge bere3 (Pearl 100) {>at nys to yow no more to mene (Pearl 951).

As discussed above, relevant examples usually stay uncontracted in Pearl, whereas these two examples illustrate negative contraction. Among the Middle English texts that I have investigated, the Laud MS of Havelok displays a similar state of affairs. In the manuscript of Havelok, there are 59 relevant instances, of which as many as 56 (94.9) stay uncontracted. The remaining three examples, which display negative contraction, all illustrate the form nis:15 (16) (17) (18)

Weilawei! nis it no korn, t>at men miete maken of bred? (Havelok 462-463) And [)er nis he nouth to frie {Havelok 1999) Jn J)is middelerd nis no knith Half so strong ne half so with (Havelok 2245-2246).

A possible conjecture is that contracted forms are more wide-spread geographically than others and that only wide-spread forms are attested around border areas in terms of the phenomenon of negative contraction. Nis is certainly a widely-attested form,16 and therefore can be found even in texts where negative contraction is not normally common. Both Pearl and Havelok are examples of this kind, although their respective languages are considerably different. The manuscript of Havelok is localized in Norfolk.17

MS Cotton Nero A.x. poems once again


5. Conclusions As discussed so far, the different use of contracted negative forms in Pearl hints at the possible existence of a different author for this poem. In Middle English, contracted forms of negation were employed almost certainly even in verse in the South and the West Midlands excluding the area touching the North. Cleanness, Patience, and Sir Gawain indeed show negative contraction quite regularly. Pearl, however, provides a notable number of exceptions. In fact, uncontracted forms outnumber contracted forms in the text. The only two examples of negative contraction in Pearl both illustrate the form in nis (nys), which was supposedly more widespread than other contracted forms, and therefore tended to be observed even in texts where uncontracted forms were usual. The wide gulf between Pearl and the other three poems in this respect suggest multiple authorship for the Cotton nero A.x. poems.

Notes 1. In contrast to Tajima, however, McColly & Weier argue that St Erkenwald


belongs to the same group as Cleanness and Patience. The present paper will not cover the issue of St Erkenwald. 2. The editions that I have used for the present analysis are: Pearl, ed. Ε. V. Gordon (1953); Cleanness, ed. J.J. Anderson (1977); Patience, ed. J.J. Anderson (1969); and Sir Gawain and the Green Knight, ed. J. R. R. Tolkien and Ε. V. Gordon, and revised by N. Davis, 2nd edition (1967). 3. The present paper deals with negative contraction that occurs with these verbs only. The contraction of ne and if ( j i f ) as follows is excluded from the discussion: "Nif he nere scoymus and skyng and non scaj>e louied" (Cleanness 21); 'Wy/oure lorde hade ben her lodejmon..." (Cleanness 424); and "M/Mare of hir kny3t mynne" (Sir Gawain 1769). The form is available in other Middle English texts as well, according to the Middle English Dictionary. As Taguchi & Yokoyama (1992: 79) point out, / » / i n


(21) is blurred in the manuscript and is rather difficult to distinguish. 4. In determining the frequencies, I have followed the manuscript reading of each text. Nas in the following quotation has, for example, been introduced by the present-day editors in place of was in the manuscript, and therefore has been excluded from the table: "For werre wrathed hym not so much j^at wynter nas wors" (Sir Gawain 726). Also, the example below, where I interpret the ne before wyst as the conjunction ne rather than the


Yoko lyeiri adverb ne, is not included: "As wy3 feat wolde of his wyte, ne wyst quat he myjt" (Sir Gawain 1087). It is usually the adverb ne that provides negative contraction.

5. The Tolkien and Gordon edition of Sir Gawain also provides the following example, where the adverb ne has been inserted in the process of their emendation, however: "I ne wot inwarld whederwarde to wende hit to fynde" (Sir Gawain 1053). Davis also follows their reading. Indeed, the meaning of the clause requires negation in some way or another, but the general absence of contracted forms in Sir Gawain makes it uncertain whether the insertion of the adverb ne is the best way of emendation. 6. Although various conditions are involved, dialectal factors are certainly strong. I have also discussed some other conditions in lyeiri (1992: 177-235). See also lyeiri (1995). 7. Fulk refers to the glossary by Morris, EETS, o.s. 99 (1892). Morris' edition of Cursor Mundi is a parallel text of four different manuscripts. According to the glossary, nart occurs in the Fairfax MS, localized in Lancashire, and nere also occurs in the Trinity MS, now localized in Staffordshire. However, I would assume that Levin investigated MS Cotton Vespasian, localized in the North, rather than MS Fairfax or MS Trinity of the same text, although he does not mention this in his work. For the localizations of the manuscripts, see LALME (I: 65, 108, & 149). 8. This is another Northern manuscript of Cursor Mundi, localized in LALME (I: 88) in Yorkshire. 9. Mcintosh (1967: 231) localizes the manuscript in Lincolnshire. LALME (I: 98) also goes for the localization. 10. The localization of the text in Lancashire is due to LALME (I: 89). 11.1 have investigated the first volume of D'Evelyn & Mill (1956), since the language of the text edited in the first volume is localized by LALME (I: 62) in Berkshire. MS Corpus Christi College 145, edited by them, includes a part whose linguistic features are those of Hampshire, and the second volume shows the mixture. 12. The localization is thanks to LALME (I: 106). 13. It is in fact the case with Old English as well. See Levin (1956 & 1958). 14. The East Midlands show the intermediate situation, providing both contracted and uncontracted forms in a mixed manner. Even here, however, the mixture of both forms occurs in each text, rather than contracted forms alone in one text and uncontracted forms only in another text. 15. The examples are cited from Havelok, ed. G. V. Smithers (1987). 16. LALME (IV: 218) finds nis (nys) in 24 counties, while it attests nam (< ne am) in only five counties (Essex, Gloucestershire, Oxfordshire, Worcestershire, and Warwickshire). Indeed, some contracted forms are more widely observed in Middle English than other contracted forms, though I presume that the information in LALME is slightly exaggerated. I have found nam, for example, in the Fairfax MS of Confessio Amantis, which displays linguistic characteristics of Kent and Suffolk (Samuels & Smith 1981: 301). The form is also found in Corpus Christi College 145 (Cambridge) of The South

MS Cotton Nero A.x. poems once again


English Legendary, which is mainly localized in Berkshire (LALME I: 62) and in the autograph manuscript of Sir Ferumbras, which comes from Devonshire. Apparently, nam was present in some counties other than Essex, Gloucestershire, Oxfordshire, Worcestershire, and Warwickshire. Similarly, I have come across the negative contraction of the preterite-tense of will in a number of Middle English texts, although LALME observes the form in only four counties. For details, see Iyeiri (1992: 177-235). 17. See Mcintosh (1976: 33).

References Anderson, John (ed.) 1969 Patience. Manchester: Manchester University Press. 1977 Cleanness. Manchester: Manchester University Press. Clark, John Williams 1949

"Observations on certain differences in vocabulary between Cleanness and Gawain", Philological Quarterly 28: 261-273. 1950a "Paraphrases for 'God' in the poems attributed to 'the Gawain-poet", Modern Language Notes 65: 232-236. 1950b '"The Gawain-poet' and the substantival adjective", JEGP 49: 60-66. 1951 "On certain 'alliterative' and 'poetic' words in the poems attributed to 'the Gawain-poet'". Modern Language Quarterly 12: 387-398. Davis, Norman - Charles Leslie Wrenn 1967

English and Medieval studies presented to J. R. R. Tolkien on the occasion of his seventieth birthday. London: George Allen & Unwin LTD. D'Evelyn, Charlotte - Anna J. Mill (eds.) 1956-1959 The South English Legendary. 3 vols. EETS o.s. 235, 236, & 244. London: Oxford University Press. Fulk, Robert Dennis 1992

A history of Old English meter. Philadelphia: University of Pennsylvania Press. Gordon, Eric (ed.) 1953 Pearl. Oxford: Clarendon Press. Hamel, Mary (ed.) 1984 Morte Arthure: A critical edition. Garland Medieval Texts, 9. New York: Garland. Heritage, Sidney J. (ed.) 1879

The English Charlemagne romances, I: Sir Ferumbras EETS e.s. 34. London: Oxford University Press.


Yoko lyeiri

Iyeiri, Yoko 1992

Negative constructions in selected Middle English verse texts. [Unpublished Ph.D. dissertation, University of St. Andrews.]


"Negative contraction and syntactic conditions in Middle English verse", English Studies 76: 424-433. Kjellmer, Göran 1975 Did the "Pearlpoet" write "Pearl"? Göteborg: University. Mcintosh, Angus - Michael Louis Samuels - Michael Benskin 1986

A linguistic atlas of Late Mediaeval English (=LALME). 4 vols. Aberdeen: Aberdeen University Press. Levin, Samuel R. 1956 Negative contraction with Old English verbs. [Unpublished Ph.D. dissertation, University of Pennsylvania.] 1958 "Negative contraction: An Old and Middle English dialect criterion", JEGP 57:492-501. McColly, William - Dennis Weier 1983 "Literary attribution and likelihood-ratio tests: The case of the Middle English PearZ-poems", Computers and the Humanities 17: 65-75. Mcintosh, Angus 1967

"The textual transmission of the alliterative Morte Arthure", in: Norman Davis & Charles L. Wrenn (eds.), 231 -240. 1976 "The language of the extant versions of Havelok the Dane", Medium i€vum 45: 36-49. Madden, Frederic (ed.) 1839

Syr Gawayne: A collection of ancient romance poems by Scottish [sic] and English authors relating to that celebrated knight of the Round Table. London: Taylor. Morris, Richard (ed.) 1864 Early English alliterative poems in the West-Midland dialect of the fourteenth century. EETS o.s. 1. London: Oxford University Press. 1874-1893 Cursor Mundi: A Northumbrian poem of the XlVth century. 7 vols. EETS o.s. 57, 59, 62, 66, 68, 99, & 101. London: Oxford University Press. Panton, G.A. - D. Donaldson (eds.) 1869-1874 The "Gest Historiale" of the destruction of Troy: An alliterative romance translated from Guido de Colonna's "Historia Troiana". 2 vols. EETS o.s. 39 & 56. London: Oxford University Press. Samuels, Michael Louis - Jeremy John Smith 1981 "The language of Gower", Neuphilologische Mitteilungen 82: 295-304. Schofield, William Henry 1909 "Symbolism, allegory, and autobiography in The Pearl", PMLA 24: 585-675.

MS Cotton Nero A.x. poems once again


Small, John (ed.) 1862 English metrical homilies. Edinburgh: Paterson. Smithers, Geoffrey Vistor (ed.) 1987 Havelok. Oxford: Clarendon Press. Suzuki, Eiichi 1989 The Language and style of Middle English alliterative poetry. Tokyo: Gakushobo. Taguchi, Mayumi - Shigeki Yokoyama (ed.) 1992 Cleanness: With Japanese translation. Tokyo: Eichosha. Tajima, Matsuji 1975 1978

"The Gawain-poet's use of con as a periphrastic auxiliary", Neuphilologische Mitteilungen 76: 429-438. "Additional syntactical evidence against the common authorship of MS

Cotton Nero A.x.'\ English Studies 59: 193-198. 1989 "Authorship and syntax: The case of Cotton Nero poems" (in Japanese), in: Eiichi Suzuki (ed.), 98-135. Tolkien, John Ronald Reuel - Eric V. Gordon (eds.) 1967

Sir Gawain and the Green Knight. (2nd edition.) Revised by Norman Davis. Oxford: Clarendon Press. Wells, John Edwin 1916

A manual of the writings in Middle English 1050-1400. New Haven: Yale University Press.

An approach to the language of Criseyde in Chaucer's Troilus and Criseyde* Akiyuki


Recently, studies of regional dialects in Middle English writings have progressed remarkably. One of the representative achievements is A. Mcintosh et al. (eds.), A linguistic atlas of Late Mediaeval English (1986). Another is M.L. Samuels and J. J. Smith's The language of Chaucer and his contemporaries (1988), in which they discuss regional dialects in "a linguistic community" created by the fourteenth century writers and contemporary scribes. Much attention has thus been paid to regional dialects, but sociolinguistic studies of Middle English have been insufficiently developed. Though there have been a few outstanding researchers, such as M. Schlauch (1952), V. Salmon (1975), N.F. Blake (1981), and D. Burnley (1983), the sociolinguistic approach has not been applied to Middle English comprehensively. In particular, we still await a study of women's speech in the language of the female characters in Chaucer's works. In this paper we will deal with the language of Criseyde in Chaucer's Troilus and Criseyde, focusing mainly on her vocabulary. Comparing Criseyde's vocabulary with those of Troilus and Pandarus, we will concentrate on the characteristics of women's language in the fourteenth century upper class society, few though they may be. In this study, we have used a data-based text of Troilus and Criseyde and statistical data analysis to investigate Criseyde's language, making use of a personal computer. Table 1 shows the number of the lines and words in each Book and the number of words used by the characters in each.1 First of all, let us examine some general characteristics of women's speech in Troilus and Criseyde. When we read the scene of the CriseydeAntenor hostage exchange arranged between Troy and Greece in Book IV, we encounter Criseyde's upper class women friends. Their language might be considered to reflect contemporary Middle English women's speech. Quod first that oon, "I am glad, trewely, Bycause of yow, that shal youre fader see." Another seyde, "Ywis, so nam nat I, For al to litel hath she with us be."


Akiyuki Jimura

Table 1. Summary Table of Word Counts Book I

Book II

Book III

Book IV

Book V



















































T: Troilus, C: Criseyde, TC: Troilus & Criseyde, P: Pandarus, N: Narrator

Quod tho the thridde, "I hope, ywis, that she Shal bryngen us the pees on every syde, That, whan she goth, almyghty God hire gide!" Tho wordes and tho wommanysshe thynges, She herde hem right as though she thennes were; (4.687-95) The narrator comments on the women's speeches in this passage as being "wommanysshe" (the OED's earliest citation: "2. Characteristic of or proper to a woman or women; womanly, feminine."). The women commonly use the intensive adverbs such as "trewely" and "ywis", and they ask for boons from God, such as "almyghty God hire gide" (which leads to the use of swear words); they also use short or abridged sentences in succession. We also see an aspect of women's colloquial language in Antigone's words of love after she sings her love sonnet. "Madame, ywys, the goodlieste mayde Of gret estat in al the town of Troye, And let hire lif in moste honour and joye." (2.880-82) The characteristics of women's language are shown in the intensive adverb "ywys", the self-conscious expressions such as "gret estat," and "moste honour," the adjective "al", and the superlative adjectives such as "goodlieste" and "moste". These characteristics of women's speech are also reflected in the language of Criseyde. In this paper, we would like to investigate Criseyde's speech, using Jennifer Coate's classification (1986), and that of Robin Lakoff (1975). The aspects of language considered here are as follows: (1)

An approach to the language of Criseyde


Pronunciation, (2) Grammar, (3) Vocabulary, (4) Swearing and taboo language, (5) Literacy and (6) Verbosity.

1. Pronunciation In the "General Prologue" of The Canterbury Tales, Chaucer gives an ironic touch to the Prioresse's manner of speech, but we do not find such a subtle description to Criseyde's. (Further, in "The Reeve's Prologue and Tale" of The Canterbury Tales, Chaucer, changing his London spelling, gives northern dialects to the students and visualizes some rural features of the contemporary pronunciation, but we do not find such a visualized rural dialect in Criseyde's speech.) Chaucer rather favours her speech. In this section we quote the passages referred to only Criseyde's speech to see how the others evaluate her "speche", "word", and "voice". ...ne of speche / A frendlyer,... (1.884-85) And goodly of hire speche in general, (5.822) And with hire goodly wordes hym disporte Shegan,... (3.1133-34) Withpitous vois, (1.111) With broken vois, al hoors forshright,... (4.1147) Herde I myn alderlevest lady deere So wommanly, with vois melodious, Syngen so wel, so goodly, and so cleere (5.576-78) It seems that Criseyde speaks in a compassionate tone of voice, never in a harsh voice, for, as Burnley states, "a woman was expected to have a gentle and sweet tone" (1986: 27). Her speech is referred to by the favourable adjectives such as "frendly" or "goodly". The adjective "melodious" is also used when Troilus remembers Criseyde's beautiful voice in Book V.2

2. Grammar Now we would like to consider the grammatical differences between men and women. In the first half of the section, we will deal with some grammatical items; in the second half, some problems of grammatical constructions will be considered.


Akiyuki Jimura

2.1. Ellipsis, repetition, interrogative sentences, etc. It seems that men used more formal language than women, because formerly men were more likely to have received an education than women (Coates 1985: 24). Criseyde, however,seems to have received a fair education judging from her courtly speech (1969: 147). In this section are presented examples of ellipsis, repetition, and interrogative sentences from Criseyde's speech.

2.1.1. Ellipsis Criseyde sometimes cuts short and leaves out sentences. Some of the instances are as follows: What! bet than swyche fyve? I! Nay, ywys! (2.128) ...I not nat what ye meene." (2.133) "And whi so, uncle myn? Whi so?"...(2.136) "Which hous?" ...(2.1189) "I? no," ...(2.1470) "Horaste! Alias, and falsen Troilus? (3.806) ... "O, mercy, God! Lo, which a dede! (4.1231)

2.1.2. Repetition Sometimes Criseyde repeats words and phrases. Some of the instances are as follows: "And whi so, uncle myn? Whi so?" (2.136) But harm ydoon is doon, whoso it rewe: (2.789) Do wey, do wey,...(2.893) Welcome, my knyght, my pees, my suffisaunce!" (3.1309)

2.1.3. Interrogative sentences Criseyde often uses interrogative sentences. This may mean that she always consults with others for their judgements, and is obedient to them. Though it is a quite feminine style of expression, in Book IV we find that she persuades Troilus of her honesty, pledging herself to come back to Troilus without fail. When she makes up her mind to do something, however, she is not obedient to others at all. Some of her interrogative sentences are as follows:

An approach to the language of Criseyde


Sey ye me nevere er now? What sey ye, no?" (2.277) "Now em," qoud she, "what wolde ye devise? What is youre reed I sholde don of this?" (2.388-89) Ye seyn, ye nothyng elles me requere?" (2.473) "Why, no, parde; what nedeth moore speche?" (2.497) "Kan he wel speke of love?"... (2.503) ... "Who yaf me drynke?" (2.651) ... "It reyneth; lo, how sholde I gon?" (3.562) "What, which wey be ye comen, benedicite?" (3.757) Why doth my deere herte thus, alias?" (3.843) And ye therwith shal stynte al his disese? (3.884) 2.2. Parataxis or hypotaxis Traditionally speaking, the grammatical distinction between parataxis and hypotaxis is used as the criteria to differentiate men and women's speech. Both written language and men's speech are likely to use hypotaxis, while both spoken language and women's speech tend to use parataxis (Coates 1985: 26). We would, therefore, expect to find more sentences involving coordinate clauses than sentences involving subordinate clauses in Criseyde's speech. We will regard "and", "but"3 and "or"4 as representative of coordinate sentences and "if' and "which" as representative of subordinate sentences. Since all of these words except "if' play important role in the narrative parts of this work, the narrator uses these conjunctions most frequently. However, when we compare Criseyde's use of these conjunctions with that of the other characters, we do not detect any striking difference. We do, however, find the conjunctions "and" and "or" in Criseyde's speech used in Books Π, ΙΠ, and IV more than in the other charcetrs' .5 There does seem to be some relationship between the choice of parataxis and hypotaxis and the language differences between the sexes. In particular we find a striking difference between Troilus and Criseyde in their longest speeches. Troilus uses the conjunction "and" 19 times in his speech: Book IV 958-1078 (121 lines), while Criseyde uses "and" 45 times in her speech: Book IV 12541414 (161 lines). In order to build up extended speech, Criseyde seems to need the conjunction "and".


Akiyuki Jimura

3. Vocabulary In this section, (1) intensive adverbs, (2) adjectives, and (3) nouns which show the salient features of Criseyde's language are dealt with. 3. L Criseyde's use of intensive adverbs As Lakoff (1975: 54-55) indicates, intensive adverbs seem to show an aspect of women's language. They tend to be used among the women more than the men in the Middle English period. Taylor (1969: 148-49) also points out that they have a high frequency of use in Criseyde's language. Her choice of adverbs, however, depends on the kinds of adverbs. Here, we would like to investigate the differences between the sexes in the use of adverbs such as "trewely", "certes", and "iwis". The adverbs "trewely" and "trewelich(e)" are used 12 times in Criseyde's speech, compared with only 8 in Troilus's speech and 6 in Pandarus's. The adverb "trewely" and its variants are used mainly by Criseyde. One of the instances is as follows: "Myn honour sauf, I wol wel trewely, And in swich forme as he gan now devyse. Receyven hym full to my servyse, (3.159-61) In this speech, Criseyde faithfully accepts Troilus's declaration of love. Criseyde makes balanced use of the adverb "trewely" from Book Π to Book V. Ironically, she uses this adverb most often in Book V, when she truly submits herself to Diomede's will. The adverbs "certes" and its related words are used mainly by Troilus: 14 in Troilus's speech, 5 in Criseyde's, and 6 in Pandarus's. One of Criseyde's uses of "certes" is given below. It is significant that she skilfully uses the negative expressions rather than the adverb "certes." "But certes, I am naught so nyce a wight That I ne kan ymaginen a wey To come ayeyn that day that I have hight. (4.1625-27) In addition, the adverbs "sikerly" and "sikirly", which are of Old English origin and at present mainly used in Scotland and other northern dialects, are never spoken by Criseyde.

An approach to the language of Criseyde


The adverbs "iwis", "iwys", "ywis", and "ywys", which are archaic in Present-day English, are of Old English origin. It should be noted that, though Troilus uses these adverbs as many as 24 times, and Criseyde 26 times, Criseyde uses them most often. This is especially the case in Book ΙΠ, where Troilus uses these adverbs 2 times, Criseyde so 100 times. As can be seen in the examples below, Criseyde uses adverb when she states her feelings definitely. These adverbs, however, do not show clear differences between Troilus and Criseyde, as Taylor has pointed out (1969: 148-49). Troilus uses them when he persuades himself in his monologue: "For in hym, nede of sittynge is, ywys, / And in the, nede of soth; and thus, forsothe," (4. 1034-35) and "that right as whan I wot ther is a thyng, / Iwys, that thyng moot nedfully be so;" (4. 1073-74). Criseyde, on the other hand, speakes to her friend quite at ease: "Iwis, so wolde I, and I wiste how, Ful fayn," quod she. "Alias, that I was born!" "Iwys, my deere herte, I am nought wroth, Have here my trouthe!" ... (3.1102-11) This is an example of colloquial speech and may be taken as characteristic of not only Criseyde's speech but also that of her upper class women contemporaries. 3.2. Criseyde's use of adjectives Jespersen regards adjectives such as "pretty" and "nice" as exemplars of women's language, and Lakoff sees empty adjectives such as "divine", "charming", and "cute" as characteristic of women's language (Coates 1985: 18-19). Here, we would like to check the intensive adjectives which show the differences between Troilus and Criseyde. The following items are discussed: (1) the adjective "verray", (2) the superlative of the adjectives, and (3) the collocation of "al" and "my(n)". 3.2.1. "ver(r)ay", "verrey" The adjective "ver(r)ay" or "verrey" are usually collocated with the nouns "God" and "lord", and are used by Troilus, Criseyde, and Pandarus. However, the collocation of "verray" and nouns meaning pleasure such as


Akiyuki Jimura

"joie" is used only in Criseyde's speech. We see the noun "slouthe" attached to the adjective "verray" in Pandarus's speech. Criseyde also uses the noun "tene" with the adjective "verray". We find 5 examples of "verray" in Criseyde's speech, 3 in Troilus's, and 6 in Pandarus's. 3.2.2. The superlative of adjectives Some superlatives of adjectives are mainly used by Criseyde. They are "gentileste", "thriftieste", and "worthieste", the last of which is used by Criseyde three times and only once by Pandarus where Pandarus, facing Criseyde, refers to Troilus (3.781). The superlative "goodlieste" is twice used by both Criseyde and Antigone respectively. This adjective may symbolise women's pride. On the other hand, the superlative "wofulleste" is used only by Troilus. Pandarus and Narrator use "beste" more often than the other characters. The following are some examples. Comments are in parentheses: "I thenke ek how he able is for to have Of al this noble town the thriftieste To ben his love, so she hire honour save, For out and out he is the worthieste, Save only Ector, which that is the beste; (2.736-40) And though that I myn herte sette at reste Upon this knyght, that is the worthieste, And kepe alwey myn honour and my name, (2.760-62) (Here "the thriftieste" means the woman who is most "worthy, worshipful, estimable, respectable, well-living," but indirectly suggests Criseyde herself. "The worthieste" is Troilus in the passage.) I am oon thefaireste, out of drede, And goodlieste, who that taketh hede, And so men seyn, in al the town of Troie. (2.746-48) (Criseyde regards herself as "the faireste, ...goodlieste.") For I have falsed oon the gentileste That evere was, and oon the worthieste! Yet prey I God, so yeve right good day,

An approach to the language ofCriseyde


As for the gentileste, trewely, That evere I say, serven feythfully, And best kan ay his lady honour kepe." (5.1056-77) (Both "the worthieste" and "the gentileste" refer to Troilus's qualities. The word "best" in line 1077 of Book V is an adverb which describes the manner of Troilus, who had saved Criseyde's "honour" prudently and substantially.) 3.2.3. "al + my(n)" The idiomatic expression "al + my(n)" is used 16 times by Troilus, 15 by Criseyde and 8 by Pandarus. The frequency of Criseyde's use of the expression is not so different from that of Troilus's, but the nouns collocated with this expression differ between the characters. As Taylor indicates, "al is a general intensifier which acts both as an adverb with the adjective, my, "entirely my knight," and an adjective with the noun, knyght, "my complete knight" (1969: 144). Criseyde uses this expression only in Books ΙΠ (9 times) and IV (6 times), while Troilus uses it mainly in Book ΙΠ (5 times) and V (6 times). The examples in Book V show Troilus's sorrowful state of mind as he remembers happy times he spent with his sweetheart Criseyde. Criseyde, however, never uses this kind of sorrowful expression. Criseyde uses the following nouns in collocations: "knyght", "trist", and "estat", none of which are used by Troilus or Pandarus. The nouns "trist" and "estat" may reflect Criseyde's selfconsciousness, since she always tries to look about herself. The instances are as follows: al my knyght (3.176, 3.996) al my myght (3.178,4.940) al my trist (3.941, 3.1023) al my peyne (4.903,4.942) al my kyn (4.1331) al myn herte (3.1001, 3.1304,4.1313) 3.3. Criseyde's use of "estat", "honour",


The nouns "honour", "estat", and "name" may characterise Criseyde. Taylor regards "estat" as one of Criseyde's favourite words, stating that this word shows "her concern" for social status and wealth" (1969: 16365), while Shirley regards "honour", "estat", and "name" as the key words ofCriseyde (1978: 50-55).


Akiyuki Jimura

3.3.1. "estat" The noun "estat" often collocates with the pronoun of the first person genitive singular "myn" in Criseyde's speech. The other examples also tend to suggest Criseyde's present state as well as her rank. She uses it 8 times: 5 in Book Π and 3 in Book IV. This fact may indicate that she needs this noun "estat" when she faces dangerous situations. It should be noted that Narrator uses "estat" when Criseyde begins to feel Diomede's "gret estat" in Book V, as in "Retornyng in hire soule ay up and down / The wordes of this sodeyn Diomede, / His grete estat, and perel of the town" (5.1023-25). And thus she seyde, "Al were it nat to doone To graunte hym love, yet for his worthynesse It were honour with pley and with gladnesse In honestee with swich a lord to deele, For myn estat, and also for his heele. (2.703-7) 3.3.2. "honour" The frequency of the noun "honour" (including "honoure") is very high in Criseyde's speech (16 times), while it is equally distributed between Troilus and Pandarus (8 times respectively). The noun "honour" always means the magnificent high moral virtue, when collocated with the adjective "sauf', as well as the verbs "kepe" and "have". Moreover, "honour" reveals the reverence of human beings for each other. It suggests the virginity of women, in this case, Criseyde's. The frequency of this important value word diminishes with the development of the story. Interestingly, or perhaps ironically, the diminishing use of "honour" may be connected with Criseyde's "slydynge" behaviour. Ever since her father Calkas had to escape from Troy in the first stage of Book I, leaving her in a dangerous situation which threatened her "honour", Criseyde needs to continually guard her honour. Her hardship is shown in the following Hector's speech: "And al th'onour that men may don yow have, As ferforth as youre fader dwelled here, Ye shul have, and youre body shal men save, As fer as I may ought enquere or here." (1.120-23)

An approach to the language of Criseyde


Since she sincerely and honestly asked the nobles around her for help, she was in fact saved by the "honour" of these who tried to take care of her. This situation leads to the last stage of the work. When talking of Criseyde, Narrator, Pandaras, and Troilus all make much of her "honour". Thus, her "honour" is preserved well not only by Criseyde herself but also by the nobles around her. As we have stated above, Criseyde's use of "honour" diminishes with the development of the story. We find only one example of "honour" in Book V, where she states that she was well loved and cared for by Troilus when she was in Troy. Yet prey I God, so yeve yow right good day, As for the gentileste, trewely, That evere I say, to serven feythfully, And best kan ay his lady honour kepe." (5.1074-77) Does she not carry her "honour" in this situation? We do not know whether or not this kind of attribute of "honour" is applied to Chaucer's women contemporaries (in fact, a certain woman in this work uses the noun "honour"), but it should be noted that this speech is very suitable for Criseyde herself.

3.3.3. "name" The noun "name" is used in the same context as the "honour". It is used differently from "honour". The noun "name" is used as a rhyme-word, and though "name" is rhymed with the noun "fame" in The House of Fame, it is interesting that it is also rhymed with the noun "shame", which has a meaning contrary to that of "fame" in this work. The frequency of this noun is as follows: 6 times by Narrator, 3 by Troilus, 5 by Criseyde, and 7 by Pandarus. One of Criseyde's examples is: And though taht I myn herte sette at reste Upon this knyght, that is the worthieste, And kepe alwey myn honour and my name, By alle right, it may do me no shame." (2.760-63)


Akiyuki Jimura

4. Swearing and taboo languages In the Middle English period, women generally tended to avoid ignoble or bawdy words. Coates has commented that in fabliaux women liked dirty deeds but disliked indecent and vulgar words (1985: 20). Since Troilus and Criseyde does not belong to the genre of fabliaux, we cannot apply Coates' excellent idea of women in the Middle Ages to Criseyde. She does not use dirty words; rather, she controls her emotions as much as possible. In this section, we will deal with swearing which might determine the sex differentiation between men's speech and women's, as well as interjections such as "o" or "alias". Criseyde uses more moderate and less varied swear words than men (Kawai 1983: 196). 4.1. Criseyde's use of interjections 4.1.1. "o" The inteijection "o" is "a natural (or what now seems a natural) exclamation, expressive of feeling," according to the OED. As a matter of fact, Criseyde uses this interjection in this work, but in her speech we do not find any more such interjections than in Troilus's speech. In Book V, Troilus anaphorically repeats the inteijection "o" when he sees Criseyde's empty house. One of the reasons why Criseyde does not use the inteijection "o" as often as Troilus may be because she is always modest, leaving almost everything to the judgement of others. The following is one example which shows Troilus's anaphoric use of the inteijection "o". Than seide he thus: "O paleys desolat, Ο hous of houses whilom best ihight, Ο paleys empty and disconsolat, Ο thow lanterne of which queynt is the light, Ο paleys, whilom day, that now art nyght, "O paleis, whilom crowne of houes alle, Ο ryng, for which the ruby is out falle, Ο cause of wo, that cause hast ben of lisse! (5.540-50)

An approach to the language of Criseyde


4.1.2. "alias" Criseyde uses the interjection "alias" almost as many times as Troilus: 40 times as compared with Troilus 42. While Criseyde uses this interjection 9 times in Book Π, 12 in Book ΙΠ, 10 in Book IV, and 9 in Book V, Troilus mainly uses it in the latter half of the story, 12 times in Book IV and 15 in Book V.6 This fact may show that Criseyde can say the interjection "alias" more easily than Troilus, even though she is not in a sorrowful state of mind. Criseyde speaks little in Book V but uses "alias" 9 times, and this high frequency which may show that Criseyde finds her situation difficult in Greece: She seyde, "Alias, for now is clene ago My name of trouthe in love, for everemo! "Alias, of me, unto the worldes ende, Shal neyther ben ywriten nor ysonge No good word, for thise bokes wol me shende. O, rolled shal I ben on many a tonge! Thorughout the world my belle shal be ronge! And wommen moost wol haten me of alle. Alias, that swich a cas me sholde falle! (5.1054-64) 4.2. Criseyde's use of oaths 4.2.1. "by God" and other oaths We find the expression "by God" 22 times in this work: 3 times by Troilus, 10 by Criseyde, and 3 by Pandarus. This swear word is mostly used with the rhyme-phrase "by my trouthe" which may be considered as a keyexpression in this work. Criseyde uses this swear word most often. One example is as follows: ""I, what?" quod she, "by God and by my trouthe, /1 not nat what ye wilne that I seye."" (3.120-21) As for the pagan gods, Criseyde prefers to use "Jove" as an address form rather than "Venus" (which she only uses in Book IV, and which Troilus uses often). Unlike Troilus and Pandarus, Criseyde never invokes "Mars", "Neptunus", "Fortune", "Furies", "Mercurie", "Imeneus", "Latona", or "Minerve" (although she uses the proper noun "Pallas" 3 times). This may show that Criseyde speakes a polite, non-violent language and does not make use of the variety of swear words used by the male characters.


Akiyuki Jimura

Using the swear expression "on or for + noun," as well as "by + noun," Criseyde invokes various pagan gods. Saturnes doughter, Juno, thorugh hire myght, As wood as Athamante do me dwelle Eternalich in Stix, the put of helle! "And this on every god celestial I swere it yow, and ek on ech goddesse, On every nymphe and deite infernal, On satiry and fawny more or lesse, That halve goddes ben of wildernesse; And Attropos my thred of lif tobreste If I be fals! ...(4.1538-57) Using forms of address for the goddess "Juno" (this proper noun may not be a swear word but rather an invocation to the goddess "Juno"), the god "deite", and the other goddess such as "nymphe", "satiry", and "fawny" who live in the woods, she pledges her truth. Though she is usually cool and stable, here her heart is a little too subtly exalting. When she pledges never to have had improper connection, she uses swearing one after another: But,/or my devoir and youre hertes reste, Wherso yow list, by ordal or by oth, By sort, or in what wise so yow leste, For love of God, lat preve it for the beste; (3.1045-48) The noun "devoir" means "to do one's devoir = to do one's best" and the noun "ordal" means "an ancient mode of trial regarded as the immediate judgement of the Deity." Both of them are used only once by Criseyde in this work. In this passage, we find the following impersonal constructions: "yow list" and "you leste." This kind of impersonal speech is very suitable for Criseyde, because she leaves almost everything to the judgement of others.

An approach to the language of Criseyde


5. Literacy Educated women in the Middle English period would have likely acquired French, Latin, and Greek, as representatives of the upper class. While it is as a matter of course that upper class men knew these languages, they may have been restricted to a few upper class ladies such as Criseyde who were able to learn and use them. As Taylor indicates, Criseyde sometimes uses academic and fashionable words such as the derived words from Latin; (Taylor 1969: 156) i.e., the "goodly" words when we cite Chaucer's language. This fact may show Criseyde's wide learning and her careful usage concerning the contemporary fashion. Now we would like to enumerate the words which Criseyde uses only once in this work, though other characters (including Narrator) do not use them. We will exclude words with variant spellings. Verbal nouns and the adjectival use of past participles are regarded as the independent words. An asterisk (*) shows that the word is cited in the OED. abstinence, amphibologies*, angwissous*, bakward*, bille, bisshop, bridlede, brotel, busshel*, byword*, byquethe, carie, causyng, cave, chartres, chekmat*, chep, cherisynge, childissh*, conceyved, constellecioun*, continuance*, cors, court, convenable, coveyteth, coye*, crowned, debat, depeynted*, deprive*, devoir, disseveraunce*, dissmulyng*, doubleth, dowves, dronkenesse, drynkeles*, enchaunten*, entrecomunen*, excusable*, fawny*, ferventliche, floureth, forlong, fox, future*, gentily, gnat, greyn, grucche, habundaunce, handle, harmyng, heleles, hemysperie*, kynrede*, leful, lesyng, letuarie*, likkere, lustinesse*, maisterfull*, maistresse*, marcial*, me-ward, mencioun, misericode, mocioun*, morter*, muable, mysbyleved*, nobleye, noriture*, novelrie*, nymphe, office, ordal*, papir*, plukke, poeplissh*, pompe, rave*, rebounde, refut*, regioun, religious, remenant, remuable*, repentaunce*, repressed*, repressioun*, resistence*, reyneth*, rooteles*, salve, satiry, saufly, scrit*, shove*, skilfully, slyvere*, sourmounteth*, sours*, sovereignete*, spie, sporneth, spotted, stoppen, stormy*, suffrant*, thewes, threteth, torney*, transitorie*, tribulacioun*, twynnyng*, underserved*, ungiltif*, unhappy, unshethe*, unstable, unteyd*, unthonk, unthrifty*, untriste*, venym, voluptuous*, wanteth, weddynge*, wether, wildernesse, wolf, worthily*, wyvere*, ypleynted. deite*, deyte; honeste, honestee; queme*, quemen.


Akiyuki Jimura

Looking up the vocabularies of the above list, we will summarize as follows: (1) (2)


(4) (5) (a) (b) (c) (d) (e)

There are many words quoted as the earliest citation in the OED. Those are italicised in the list. There are many words derived from Old French, Anglo French, and Latin: e.g., in the above list most words are derived from Romance languages, except for some words which originated in Old English such as "backward", "lustinesse", "mysbyleved", "rooteles", "unshette", "unteyd", and "worthily", and some words originated in Old Norse such as "unthrifty" or "untriste" whose prefix "un-" comes from Old English. We find some scholarly and scientific terms such as astronomy, astrology, and rhetoric: "constellecioun", "hemysperie", and "amphibology", The stem with the OE prefix "un-" originated in Old French, Old English and Old Norse. Other features: The meaning of some words is explained as "figurative" in the OED: "enchaunten", "sours", "stormy", and "wyvere". The meaning of some words is explained as "loosely" in the OED: "busshel". Some words have a pejorative meaning: "depeynted". Some words are quoted in only one citation in the OED: "mysbyleved", "poeplissh", and "suffrant". We find a literally translated word from Latin to English: "byword", which is modelled on Latin pro-verbium.

6. Verbosity The Wife of Bath is a very talkative woman, as Chaucer says: "In felaweshipe wel koude she laughe and carpe (1(A) 474)." Criseyde seems not only talkative at times like the Wife of Bath, but also silent at other times like Griselda in "The Clerk's Tale." In order to investigate whether or not Criseyde is verbose, we have compared the length and the total words of Criseyde's speech with those of the other characters. Please see Table 1 which shows the total words of each character. The lineage of the longest speech of each character is as

An approach to the language of Criseyde


follows: Troilus 121 lines (4.958-1078), Criseyde 161 lines (4.1254-1414), and Pandaras 105 lines (3.239-343). The frequency of each character's speech is as follows: Troilus 153 times (the total words of Troilus's speech: 11,355 words), Criseyde 172 times (8,985 words), and Pandarus 227 times (14,223 words.) Thus we have two kinds of facts concerning Criseyde's verbosity. First, we find that she has the longest speech of any of the three characters. Second, she tends to be less wordy, though she speaks more often than the other characters. We, however, must not jump to the conclusion that these facts show her talkativeness, but it is also a matter of fact that this kind of statistical analysis helps us to see at least one aspect of Criseyde's speech. We have investigated Criseyde's language from various points of view. Her language, closely related to her critical situation in this work, tends to be influenced by her surrounding. Her uncle Pandarus, playing the part of go-between; the natural phenomena encompassing the lovers Troilus and Criseyde, especially their deep love covered with the rain fallen by chance; the political dealing such as the hostage exchange; her obedience forced on her by her father Calkas: all these happenings may indicate that Criseyde is unable to make up her mind, and leaves almost everything to the judgement of others. She may find it convenient that her "honour" is guarded by others, especially Troilus and Pandarus, since this allows her to live life to the full. From this study we are unable to say exactly what kind of relationship exists between the language of Chaucer's women and that of Criseyde and what kind of language the main character Troilus uses. These points will be the subjects of a further study. From the study we are able to state the following. As Masui states "Chaucer's language becomes dramatic in fact, since he dynamically connects the speech with character and that such a dynamic language is reflected in Criseyde's speech and Pandarus's from a courtly point of view" (1962, 19732: 203), Criseyde's language shows at least one aspect of women's language in the fourteenth century courtly society of England where Criseyde, created and characterised by the poet Chaucer, lives with flesh and blood as well as her contemporary women.

Notes *

The present paper is a revised and abridged version of the papers originally published in Essays on English Language and Literature in Honour of Michio Kawai edited by


Akiyuki Jimura

Nobuyuki Yuasa (Eihosha, 1993), English and English Teaching: Festschrift in Honour of Hisashi Takahashi and Jiro Igarashi edited by Kichiro Nakatani et al. (Faculty of School Education, Hiroshima University, 1993) and Bulletin of the Faculty of School Education, Part II, Vol. 16 (Hiroshima University, 1994). I am sincerely obliged to express my heartfelt thanks to Professor Jacek Fisiak and Professor Akio Oizumi, who have given me constant encouragement and valuable suggestions. 1. All Chaucer citations are from L.D. Benson (ed.), The Riverside Chaucer, 3rd edition (Boston: Houghton Mifflin, 1987). Italics in the passages are mine. All the data in this paper are also based on Benson's edition. We would like to express our appreciation to Oxford University Computing service. The computer program used in this study was developed by Masatsugu Matsuo, Hiroshima University. 2. We find the expression "wikked speche" (5.1610) in Criseyde's letter. This is used when Criseyde says that one of the reasons why she cannot return from Greece to Troy is because she is extremely conscious of the evil speech around her. This shows Criseyde's self consciousness which makes her aware of the un-courtly speech around her and consequently leads her to using the courtly speech. 3. Here we include the word "but" meaning "unless", because we have collected the examples mechanically, using a personal computer. 4. Here we include not only the paratactic constructions of sentences but also those of words and phrases. 5. We notice that Troilus uses "which(e)" (61 times) more often than Criseyde does (35 times). 6. The frequency of Troilus's use of the inteijection "alias" becomes proportionally higher as the total frequency of "alias" increases in this work.

References Benson, Larry D. (ed.) 1987 The Riverside Chaucer. (3rd edition.) Boston: Houghton Mifflin. Blake, Norman F. 1981

Non-standard language in English literature. (The Language Library) London: Andr6 Deutsch. Burnley, David 1983 A guide to Chaucer's language. London: Macmillan. 1986 "Courtly speech in Chaucer", Poetica 24: 16-38. Coates, Jennifer 1993 Women, men and language. (2nd edition.) London: Longman.

An approach to the language of Criseyde


Date, Toshihiro (ed.) 1990 Festschurift for Kazuso Ogoshi, Ohtani Women's College. Kyoto: Apollpnsha. Elliott, Ralph W.V. 1974 Chaucer's English. (The Language Library) London: Andri Deutsch. Jespersen, Otto 1922 Language and its nature, development and origin. London: George Allen & Unwin. [1947] [Reprinted London: George Allen & Unwin.] Jimura, Akiyuki 1990 "Chaucer's use of northern dialects in The Reeve's Prologue and Tale", in: Toshihiro Date (ed.), 159-183. Kawai, Michio 1983 "Modes of swearing in eighteenth-century drama", in: Hiroshige Yoshida (ed.), 191-197. Lakoff, Robin 1975 Language and women's place. New York: Harper & Row, Publishers. Masui, Michio 1964 The structure of Chaucer's rime words: An exploration into the poetic language of Chaucer. Tokyo: Kenkyusha. 2 1962, 1973 Studies in Chaucer (in Japanese). Tokyo: Kenkyusha Mcintosh, Angus et al. (eds.) 1986 A linguistic atlas of late mediaeval English. 4 vols. Aberdeen: Aberdeen University Press. Oizumi, Akio (ed.) 1991 A complete concordance to the works of Geoffrey Chaucer, programmed by K. Miki. 10 vols. Hildesheim, etc.: Olms-Weidmann. Ringbom, Η ikon (ed.) 1975 Style and text: Studies presented to Nils Erik Enkvist. Stockholm: Spräkforlaget Skriptor AB. Robinson, Fred N. (ed.) 1957 The works of Geoffrey Chaucer (2nd edition). Boston: Houghton Mifflin. Roscow, Gregory H. 1981

Syntax and style in Chaucer's poetry. (Chaucer Studies vi.) Cambridge: D.S. Brewer. Salmon, Vivian 1975 "The representation of colloquial speech in The Canterbury Tales", in: Häkan Ringbom (ed.), 263-277.


Akiyuki Jimura

Samuels, Michael L. and J.J. Smith 1988 The English of Chaucer and his contemporaries. University Press. Scholar, Claes

Aberdeen: Aberdeen


Some types of narrative in Chaucer's poetry. (Lund Studies in English 25) Lund: C.W.K. Gleerup. Schlauch, Margaret 1952 "Chaucer's colloquial English: Its structural traits", PMLA 67: 1103-1116. Shirley, Charles G., Jr. 1978 Verbal texture and character in Chaucer's Troilus and Criseyde [Unpublished Ph.D. dissertation, University of South Carolina.] Taylor, Davis 1969 Style and character in Chaucer's Troilus. [Unpublished Ph.D. dissertation, Yale University.] Yoshida, Hiroshige (ed.) 1983

Festschrift for Michiko Masui. Tokyo: Kenkyusha.

(Pro-) Nominal reference in Old English and the origin of the i/wzi-clause.1 Koichi Jin

1.1. Nature of reference Nouns and pronouns belong to a class of words in the lexicon, which in the first place refer in terms of vocables to objects we observe or to those which we believe exist in the world around us. The element pro- implies that pronouns are substitutes of nouns. A referent is a particular object, real or imagined, to which a noun refers, when the latter occurs in actual speech. The particularnessof the referent is reflected linguistically in three stages of definiteness: A proper noun, the + a class noun, a [certain] + a class noun. At the expression level, definiteness decreases in forward movement, but increases in backward movement, while the particularness of reference remains unaffected by this process: I met an English lawyer I met the [English] lawyer - I met John Smith. The pronominal system corresponds to the three stages in nominal reference. an / some English lawyer - someone the lawyer - [the one] John Smith - he I met someone from England. I met the one (i.e., the lawyer) from England. I met him.

1.2. Anaphora - mistaken identity Anaphora in Greek means 'carrying up, tracing up (=back)'; figuratively all it has to convey is that when one says 'X', one is talking about X (which may or may not have been mentioned earlier). Thus when for example we say 'Plato' and use 'the man' to refer to the same person, the Greeks would have said 'Plato' anapheretai eis Platonem, ton philosophon which literally means 'The name Plato is to be traced back to Plato, the



philosopher. Similarly autos 'ipse' or ho aner 'the man' to refer to Plato himself. Another context in which the verb occurs is 'to report an event', which is 'to refer to it.' Without doubt reference is then made to the occurrence itself. In Greek and Latin, non-use of personal pronouns in the third person suggests that the referent was to be supplied from the situation. Derivation of the personal pronoun autos 'he' from autos '(him) self is thus natural enough. The 'omission' of the third person pronouns is also frequent in Old Norse. Here too, the absence of a pronoun rather proves the presence of a referent. It is curious, when one comes to think of it, that one should speak of anaphora only with respect to the third person. Obviously one has been misled into this belief by the historical fact that in many dialects the pronouns in the third person are 'weakened' demonstratives, e.g., Lat ille, French il. Anaphora, to sum up, is reference to an actual object or event by means of linguistic signs but - this is very important - signs themselves Eire not anaphors of other signs.2 The function of a linguistic sign is, be it a noun or a pronoun, to refer to an extralinguistic element in order to thematize it. But Huddleston's idea of the antecedent runs counter to this (obvious) fact, because he expressly states that a pronoun is an anaphor of a noun (=the antecedent). We may for example ask what will happen if we say the chairperson, instead of the Empress? ??The chairperson has not arrived yet but he or she should be here any moment. What is the use of an antecedent, if it is incapable of providing the pronoun with the information on gender? So far as the object is identified, the sign is said to be definite - cf. 'Miss Macpherson is the chairperson' / 'She is the chairperson' / Ά woman professor from Harvard is the chairperson.' We will say a proper noun may alternate with a personal pronoun but we will never say the latter is an anaphor of the former. On the other hand, a noun or pronoun is identified if there are more than one potential referent to pick out from. 'Anybody', 'everybody', 'each', 'all', 'both', 'either' are indefinite in this very sense. If we read in the AHD that SOMEBODY is 'an unspecified or unknown person,' neither of these adjectives is a synonym of indefinite. 'Somebody' in futuristic context3, however, may denote an indefinite person or thing, e.g., 'Something will turn up.' 4 Only Historical Statement makes somebody or something definite. For example, 'Somebody told me you are going to quit. She was very concerned about you.' This will further imply that the personal

(Pro-) Nominal reference in Old English


pronoun retrieves or is an anaphor of, not a noun, but the referent itself. It may be an actual object in direct observation or a psychic image created in the mind of people who understand the language. Huddleston's second example 'She looks remarkably old for a tightrope walker') reflects this state of affairs correctly. But despite his assertion, there is only one rule, not two, to be considered. His first rule with its 'antecedent' should drop out. Figure 1.

Alfred (Nom1) - the king (Nom2) - a king (Nom3) Referent he (Pron1) - [the one] (Pron2) - somebody (Pron3)

The referent in this figure is a particular person, or an object, though it may also be an idea.5

1.3. The particle öe and the relative clause The relative clause as we define it is a noun equivalent in clausal form, formally introduced by the relativizing particle de. Since it is substantival, it may occur in apposition with nouns and (personal) pronouns. Theoretically the following combinations are possible: Nominal Nom + t>e (Alfred + pe) Nom2 + pe (se cyning + pe) Nom3 + J)e (sum cyning + pe) 1

Pronominal Pron1 + J>e (he + pe) Pron2+ pe (se + pe) Pron3 J)e (sum + pe)

And all of these combinations are attested in the extant literature of Old English. Nom1 pe /ECHom I, 3 46.10... and dauides maerpe pass maran cyninges and salomones wuldre pe gode paet maere tempel arasrde, /ECHom (Thorpe) 88,10 pyllic waes herodes forpsip pe manufullice ymb pass heofonlican aepelinges tocyme syrwde,



Nom2 pe Being the commonest of all patterns, this is essentially the particle be preceded by a definite common noun. All the definite nouns are results of rightshifting: /ECHom I, 2 34.9 Drihten waes acenned on paere byrig pe is gecweden betheleem. Se haelend waes gefaestnod on anre rode seo halga rod pe so haelend on gefaestnod waes. Nom3 - pe /ECHom (Thorpe), I, 58. 31 pa asende he hine on wraecsip to anum igeope pe (to a certain island which) is Papmas geciged, /ECHom (Thorpe) I, 62.1 Soplice min lareow Crist sumne cniht pe (a certain youth who) gewilnode pass ecan lifes pysum wordum laerde..., Pron1 pe /ECHom I, (Thorpe) 464.33 du pe asendest dinne ancenndedan Sunu..., jECHom Π, 43 324, 188 paer we ealle beop gegaderade pe her lif underfongen, Jn 4, 26 Ic hit eom pe wip pe sprece (hit refers to messias pe is genemned crist mentioned in the preceding verse), Mr 7, 11 Eornustlice nu ge pe yfle synt cunnan gode sylena eowrum bearnum syllan... The personal pronoun may be repeated after pe: Christ 22 Huru we for pearfe pas word sprecap ... pe we in carcerne sittap (see Bosworth, AS Die pp. 1038-9). Pron2 pe What is reported in the clause se de withstent deofles lare for example presupposes that there is someone who opposes the Devil's teaching. In the process of relativization 'a certain man' becomes 'the man (se)\ who then is identified as 'the one who opposes the Devil's teaching', namely Christ. /ECHom I, 366 32 Se is lybbende God pe haefd lif and wununge durh hine sylfne, butan anginne. /ECHom, (Godden), I, 57 and swa deah pone liflican waestm abaer se de is sod biscop and ure salwa alysend (=God), Dan 416 Pa cwaeÖ se de waes cyninges raeswa, wis and wordgleaw, J η 9.8 Hu nis pis se de saet and wddlode?, /ECHom I, 29 424.14 deme gehwa pass wurdmyntes wyrpe sy: se pe geworht is: odde se pe ealle ping gesceop, (General Statement) Lk 14.23 Se pe me hatap hatap minne Faeder. The feminine form in the following must be due to ellipsis: /ECHom I, 24 346. 274 Ac seo (sawol) pe bedxled is pam godnyssum, heo geomrige, 'Let such a soul as is deprived of those excellences mourn (adapted from Thorpe)' cf. 1.25 Wa daere sowie de orhlyte hyre lif adrihd paera haligra mihta. It may also occur with seo: /ECHom Π, 10 Heo pa gewende ongean..., seo pe

CPro-) Nominal reference in Old English


blind gelaed waes, With past: Jη 1,4 past waes lifpe on him geworth wps & past lifwaes manna leoht (past...pe = that... which). Pro3/* /ECHom I (Pref) 6.11 ... He gemacode pa paetfyr... and forbaemde eall his seep ut on felda. And pa hyrdas samod buton anum pe hit him sypan sceolde, Guth A 709 eom ic para twelfa sum pe he getreowste under monnes hiw mode gelufade. In a number of instances pe with the value of 'one who' or 'those who' 6 occurs independently as Mt 7, 13 & swype manega synt pe {Li. & Ru. has pa pe for Latin qui) purh pone weg farap, Lk 7, 20 Eart pu pe to cumenne eart pe (Li. & Ru.: sepe) we sculon opras onbidan, Beo 138 pa waes eapfynde, pe him elles hwaer gerumlicor raeste [,sohte]. The fact remains that pe is found both in prose and verse and interchanges with se de, possibly for external reasons such as meter, rhythm and the like. In other words, the pronoun se (Pron2) tends to combine with de to form a relative cluster. It gains in frequency in later prose (older de = later se de). Thus earlier se... de... occurs as se.. se de in later language - note that both are attested in Beowulf. In predication, the former develops into dy... de + clause, the latter into dy... daet(te) + clause.

1.4. Tokens and types The structure of reference in term of which language operates may be envisaged as being trigonal. First, we posit an object in extra-linguistic reality to be referred to via a linguistic sign. Such an object nay be termed res ipsa. A linguistic sign of this kind is a name corresponding one-to-one to the particular individual in question (proprium). In the triangular system just mentioned a proprium is linked with two appellativa to complete the picture. appellativuml) proprium appellativum2)


Koichi Jin

Proprium then, alias known as token is opposed to the other two, which are types, in the terminology of philosophy. We find it defined by the OED in the following manner: Token 1. 'something that serves to indicate a fact, even, object, feeling, etc.; a sign, a symbol 3. something serving as proof of a fact or statement, an evidence' At the grammatical level, the one type in our triangle is a noun or a noun phrase, the other takes the form of a clause headed by a relative. Our picture appears now in a slightly different form: Β (reference to A by means of a class noun or noun phrase) A (res ipsa) C (reference to A by means of a relative clause)

1.5. The Referential Triangle - its implications for syntax. A token indicates an individual object to which sometimes a proper name is given (persons, places &c.) Most tokens are 'anonymous', however. John l(the car)


the boy - the one who lives in Manchester. the car - the one that I bought the other day.

A nameless token can be referred to indirectly by means of type® or typeb with the aid of a determiner 'the'. We shall designate them by Β and C respectively. There are a number of levels at which a token occurs. At the one extreme of the scale we have a token as the representation of res ipsa, e.g., John is a token of the boy. It is extremely important not to lose sight of this 'dual' nature of class nouns: Ί was introduced to Jim Turner, the British painter' / Ί was introduced to the British painter, Jim Turner' / Ί was introduced to a British painter.' In defiance of the rule generally taught at schools, these nouns or noun phrases, definite or indefinite, have the same referent (Jim Turner = the British painter = a British painter). The three items are interrelated with each other in such a way, that they form what might be termed REFERENTIAL TRIANGLE, in which the token plays the leading role. token

type3/ typeb

(Pro-) Nominal reference in Old English


Type3, which is a class noun is said to be in appositional relation with type , a class noun in clausal form. By apposition is meant the relation between the token and the type8 on one hand; between the token and the typeb on the other. Ut aliter dicam, the two types are equally distanced from the token. For the sake of brevity, we shall call the token Λ, the types Β and C in the order mentioned. a) b)

John lives in Manchester. The boy (or the boy who lives in Man.) [=John] joined the choir.

The noun and the noun clause in b) do business for a proper noun and is interchangeable with he1 from the pronominal series (see the Figure on p. 77). He and the boy, however, are not exact synonyms at the syntactic level. The personal pronoun can never occur as Β or type2. While he is a full-size equivalent of a proper noun in the sense that they share a common referent (res ipsa), the Old English se and type1 (the + a class noun) have only indirect reference to the object itself ('the aforementioned one / the thing οin question). In this sense, too, se is an equivalent of se + a common noun. This also explains the non-occurrence of he (&c.) in combination with pe to form a relative sentence (se de). Thus, our TRIANGLE occurs either as A—Β [apposition]C :the relative clause9 A—C [predication]B :the content clause APPOSITION: Oceanus (A) is se micle sone pe we Garsecg hatap. Even where we have Β and C within a sentence complex, the apposition B/C may not be realized, as when a subject-complement relation holds between them. That(C) the earth is round is an established truth (B). It must then be concluded that Β and C belong respectively to different sets of apposition. We may compare: 'the truth(B) that(C) is now established (relative clause)' / 'the truth(2?) that (independent C) the earth is round (non-relative clause).' PREDICATION: God fyrenpearfe (transposedB) ongeat, past (C) hie drugon aldorlease.



God perceived a [certain] dire state of affairs, [namely] that they had suffered from an interregnum [< that they had suffered from an inter regnum (C) was a dire state of affairs (fi)].10 Predication is possible under one condition, namely: C quotes a message: We know [the fact] that (Q the earth is shaped like a ball. Since I have discussed the relative clause elsewhere, the following pages will be dedicated to the origin and the development of the content clause in OE with occasional glimpses into comparable phenomena in present-day English. Some modern examples may be introduced to illustrate this point: Darwin's Origin of Species forcibly states a (certain) harsh truth: only the fittest survive - That only the fittest survive is a harsh truth. The empty coffin in he center of the crypt had a single horrifying meaning: Dracula had left his tomb to stalk the village streets in search of fresh blood11 - That Dracula had left his tomb ... a single horrifying meaning. If A chairman (J?) resigned is a transform of Mr. X is a chairman, then we should be able to say (and we actually do say): A harsh truth is stated in his book, i.e., transformed from That only fittest survive is a harsh truth. By the same token: The chairman (B), Mr. X(A) resigned / The harsh truth (B) that (A) only the fittest survived is stated in Darwin's book. In the same manner that A singer is going to hold a concert is ultimately related to John Smith is a singer, the classifying noun 'a harsh truth' may be carried over to the left of the subject clause: Darwin tells a harsh truth: only the fittest survive. But this leftward transposition is only apparent. What happened is that you speak of a singer who is going to hold a concert, you are referring to a particular individual whose existence, name or identity you already known, which enables you to say: By la singer' I mean John Smith - hence 'a singer' = 'a certain singer'. The real movement is therefore not from right to left but, as is the way with natural speech, from left to right: 'Darwin tells a certain harsh truth - I know what it is - it is that only the fittest survive.' a) b) c) d)

That only the fittest survive12 is a truth. The truth is that only the fittest survive. Darwin tells a harsh truth: Only the fittest survive!13 (as above). Darwin tells a (certain) harsh truth, namely, that only &c. NOT., a harsh truth that only the fittest survive.

(Pro-) Nominal reference in Old English



The harsh truth is: only the fittest survive.

In b) the thematic fronting has inverted the order14 - cf. 4The lesson to be learned is, in fact, that it is only by such relatively insensitive tools as this that associative meaning can be systematically studied (G. Leech).' C) is the only choice if the common noun cannot constitute the complement to the that clause: *4That he passed the exam at a first try is a feat.' The 'correct' sentence should be: Tom achieved a feat: he passed the exam at a first try.15 The sentence in d) represents the pattern Pron3 - Pron2, of which the following is another example: 'This upholds what I said earlier...: that...associative meaning is less stable, and varies with the individual's experience (G. Leech).' As we shall see later the same was true of Old English usage. See Beo 2-3, 3-6 discussed below.16

2.1. Old English To begin with, we have paet introducing a noun clause, if the preceding noun (Β) enters into a subject-complement relation with that clause. Nouns such neop and pearf may serve as Β in the capacity of the complement to the paet clause (Λ): Beo 14

FyrenJ>earfe ongeat, paet hie aer drugon aldorlease. God perceived α dreadful pearf: [id est] that they had suffered from an inter regnum for a long while' - That they had suffered ... was a dreadful pearf}1

Our reading agrees essentially with that of Kock's, except for the idea of the TRIANGLE. Klaeber's rendering 'the' distress (B) which (C) the Danes had suffered points to pafyren pearfe}% rather than to fyrenpearfe, which is the one actually found in the text.19 We may dismiss as erroneous Klaeber's attempt to change the well-established solution of the MS abbreviation into paet,20 See also Heyne-Schücking: Beowulf 2. Teil 16f. The old rule that the identifying relative clause requires the preceding noun to have a determiner is as valid today as it was in the poet's days compare: Ί lost the camera I had bought in London', i.e., (I bought α camera in London) I lost the camera / the one that I had bought in London - 4I lost a camera I had bought in London' would be the explanatory use in



both old and new. Only predication can explain the apparent juxtaposition of a non-determinate noun and a past clause reporting a historical occurrence. The modern practice of handling apposition is exemplified by 'My father is an old man who believes in discipline / My father is the old man who wrote the other day to the Minister of Education.' M. Swanton's modern text 'He had perceived the cruel distress they had once suffered when for a long time they lacked a king' is an ingenious one and gives excellent sense and style in MnE but his deviation from the original wording is obvious. Swanton retains the MS reading (past in 1.15) but his interpretation of the sentence structure does as much injustice to our poet as Klaeber's (see the next Footnote). The poet describes as pearf the historical fact that the people had long suffered from lack of a king. a) b)

God perceived the distress they had suffered - relative clause God perceived a most distressing state of affairs: [that is] that they had long lacked a king.

It is curious the Beowulf scholar of our age should have missed that much. The position of the complement may be filled with nouns, such as a matter (of concern or distress), circumstance, fact, state of affairs; a piece of news &c. Some of these nouns can stand in apposition with a that clause as the circumstance / fact / that while others cannot: *the matter that... It is therefore largely a matter of modern English usage whether one should choose distress or a distressing circumstance to fit into the place of the Old English word. For the AngloSaxons at any rate pearf could represent Y in That they had long suffered (=X) was Y. A similar pattern is found in, Beo 426

...Ic J)e nu f)a, brego Beorht-Dena, biddan wille, eodor Scyldinga, anre bene, past t>u me ne forwyrne, wigendra hleo, freowine folca. Ί will now ask you a [certain] favor, Prince of the Scyldings, [which is] that you will not refuse me...' ('That you do not refuse me would be a gracious - favor to me').21

The insertion of past indicates that our poet saw predication, not apposition, in the relation between the noun and the past clause - see above.

(Pro-) Nominal reference in Old English


Similarly, Beo 700

...Sop is gecy^ed, past mihtig God manna cynnes weold widerferhj).22 Christ 373 Us is lissa £>earf, past J)u us ahredde, ond us haelogiefe.23

Examples of predication from other works include: Andreas 158 wass him neod micel past hie tobrugdon blodigum ceaflum fira flasschoman him to foddorpege, ('There was a matter of great urgency for them, namely, that they should touch with bloody jaws the flesh of men', Phoen 189 ...bip him neod micel past he pa yldu ofestum mote purh gewites wylm wendan to life feorg geong onfon, Gen 2054 ...him wass pearf micel past hie on twa healfe grimme gupgemot gystum eowdon heardne handplegan. It is worth taking note that neod (&c.) is the subject of be: 'There was (a certain) dire need, namely that ....' As stated passim, the pronoun past stands generally for a word like the circumstance. The normal order with respect to X and Y would yield in modern English 'That they had long suffered... was a needy circumstance.' Similarly for pearf: Gen 664 Unc is his hyldo pearf WHom (Napier) Π 16.4 as us is peah my eel pearf, past, we αα habban rihtne geleafan on god aslmihtigne.. This construction too is subject to the restraint that the noun must be such as goes into a subject-complement relation with the that clause. Pronominalization Under certain circumstances, the noun {B in our TRIANGLE) may be replaced by past/hit. An example of this will be Beo 705 past ('the fact' or 'circumstance') wass yldum cup / past hie ne moste...pa Metod nolde/se scynscapa under sceadu bregdan. The first past may represent any noun in the lexicon which is said to be known (the report, the news, the fact &c.). The same treatment may be accorded to Beo


Ne waes J>aet24 wyrd pa gen past he ma moste manna cynnes picgean ofer pa niht.


Koichi Jin

Here {)aet takes the place of 'particular course of events',25 corresponding to X in X is wyrd; 'The particular course of events was not wyrd yet (i.e., was not pre-determined), viz. 'that he should be allowed to partake of more human flesh for the rest of the night. We may restore the 'basic' order: 'That he should be allowed to partake of ... would have been a dreadful course of events.' In like manner we have past for p&t word cwedan pxt... Word is still 'report, message' in 'to send word'. For further discussion on this issue, see 2.8.

2.2. Non-occurrence of past Under certain conditions, pxt does not occur at all. Beo 1

Hwaet we Gar-Dena in geardagum Jjeodcyninga prym gefrunon, hu J>a ae{>elingas eilen fremmedeon. Oft Scyld Scefing sceaj>ena J>reatum mogegum maegjmm meodsetla ofteah.

The noun prym corresponds to Β in our REFERENTIAL TRIANGLE; the following statement then constitutes a verbal equivalent of A. But we have no relative clause to go with B. We heard of chieftains 'prym, how they in days gone by performed deeds of valor.' In a formula, we can represent it this way: B-A If the meaning of prym is 'greatness, glory, or mightiness', then it will not run parallel to α fact or a circumstance, usually expressed by p&t clause (our C). •That they in days gone by performed deeds of valor is prym. Or in ModE (see above), Tom achieved a feat: he passed the exam... •That Tom passed the exam... is a feat.

(Pro-) Nominal reference in Old English


The structure repeats itself in the next few lines. hu aej>elingas eilen fremmedon. Oft Scyld Scefing sceapena preatum mogegum maegpum meodsetla ofteah. Here too, their dragging away of the mead benches is eilen but the fact or the circumstance that Scyld dragged away the mead benches &c. is not, because eilen is an action, something one performs (cf. eilen fremman). Hence the absence of past: They performed deeds of valor: they dragged away &c. This rule appears to be observed meticulously in Old English prose and verse. Eceptions exist but they may be caused by anomaly as /ECHom I, (Thorpe) 528.31 'Seo is soö lufu, past gehwa his friend lufie on gode, and his feond for gode' (lit. 'the aforesaid one is true love, love which consists in the circumstance that everyone love his friend well, and his foe for his good - adopted from Thorpe.' Here, it seems, two different concepts have gone into one construction.

2.3. Past introducing an adverbial clause (= swa past) The pronoun past appears when it is used adverbially in the sense of 'so that': Beo 20

Swa26 sceal geong guma gode gewyrcan, fromum feohgiftum on faeder bearme, past hine on ylde eft gewunigen wilgesijias, J>onne wig cume, loede gelassten.

It will be the easier to understand the meaning of theses lines, if we 'reduce' it to HISTORICAL STATEMENT: 'The king as a young man acted in such and such a way, that when he came of age good companions stood by him and, and lent him aid to the people when war came.' A comparable situation develops in, Beo 64

t>a waes Hrojigare heresped(Β *)gy fen, wiges weorjjmynd, past him his winemagas georne hyrdon, o^aet seo geogoj) geweox magodriht micel.


KoichiJin Translate: 'Success in war was granted him, glory in battle, so that his friends and kinsmen obeyed him [one after another], until...'

These lines imply that he emerged as winner out of his struggle with his friends and kinsmen. Thus we read, Beo 893

Haefde aglasca eine gegongen, pdst he beahhordes brucan moste selfes dome.

Ellen refers back to his fight with Grendel described in the preceding lines: 'He had thus performed a deed of valor, so that he was now in a position to dispose of the treasure at his own will.' Understandably, that he could dispose of the treasure was not a deed of valor, it, the circumstance, was on the contrary the result of such a heroic deed. So 1769 ff.

2.4. Omission of Β In supplying the missing items, we have tried to find words that can serve both as the object of the finite verb and as the complement of a copulative sentence of which the />aei clause is the subject. Some of the words will be judged uncertain. Beo 86

{>a se ellengaest earfojplice Jjrage [?]27 geJ>olode, se J>e in Jjystrum bad, f??et he dogora gehwam dream gehyrde hludne in healle.

The Old English verb polian occurs with two types of meaning. a) b)

to receive an more or less injurious action such as a blow or punishment (drype / wite poliari) - passive. to go through, experience, an unpleasant or irritating situation or period.

I suggest torn, but with considerable hesitation. Torn denotes either anger, irritation, any 'roused' state of mind or a grief, trouble, afflication, distress

(Pro-) Nominal reference in Old English


(Bosworth). Take the Modern English word trouble means anything that causes trouble and that it could be a situation or circumstance as well as an object ('The trouble is that my old car breaks down almost at every ten miles'). A further implication of our argument is whether or not we can assume anything analogous for the Old English torn and its like. Torn is certainly 'a situation' in, Gen 976

Brego engla beseah, ...Caines ne wolde tiber sceawian. past waes torn were hefig at heortan.

God paid no attention to, positively ignored, Cains offerings and that was torn hefig at heortan [an aggravating situation] for the (wicked) man. But this is a rare word occurring only a few times in the verse and for this reason our conclusion must remain pending. Compare sagu discussed below. Beo 175

Beo 388

Hwilum hie geheton aet haergtrafum rj wigweorJ>unga, [lissa, miltsa] wordum baedon paet him gastbona geoce gefremmede wij) |)eodJ)reaum. Gesaga him [ spell] eac wordum, past hie sint wilcuman Deniga leodum.

The phrase spell secgan is attested: Widsip 54 Forpon ic maeg singan ond secgan spell·, also followed by a paet clause: Maldon 50 ...sege pinum leodum miccle lapre spell / pact her stynt unforcup eorl mid his weorde... It should also be noted that secgan does not seem to occur with paet word... paet + clause (acc.), whereas it is a common idiom with cwepan. Beo 591

Secge ic J>e Csage] to so^e, sunu Ecglafes, past nseere Grendel swa fela gryra gefremmede,..

Saga / sagu may co-occur with a past clause as exemplified by pin saga bip geswultod... pact he minne sunu ansundne araere (see Bosworth, under saga). Saga secgan is also attested: 'Fabulae' sind pa saga pe menn secgap ongean gecynde (ibid. sagu). Sagu can summarize a previous statement as Hi saddon pam kinge paet he haefde swype agylt wip Crist... pa lasg se king and aswaeartode eall mid pare sage (ibid. sagu).



Beo 595

'Ac he hafa[)[?] onfunden past he J)a f^ehjie ne {jearf,'

That onfindan can occur with past is illustrated by: Beo 2219 past sie peod {on fand) / b(ig)folc beorna / past he gebolge(n) wass, Beo 2628 Ne gemealt him se modsefa, ne his masges laf / gewac ast wige. past se wyrm onfand, Beo 2300 ..he past onfand / past haefde gumena sum goldes gefandod, heahgestreona. So Judgl 73-75; Riddle 27.9-11 &c. We could possibly supply a word like wyrd in pa he pa [wyrd\ onfunde, \>ast he dead beon scolde (see Bosworth under onfindan). Beo 1512 J>a se eorl ftacn] ongeat Jjast he in nijjhele nathwcum was,.. Then he perceived [evidence, signs, indications &c.] that he was in some hostile hall.' Ongitan 'to perceive' may take as object tacen followed by a past clause: past we py geornor ongietan meaton tacen, past se fugel purh bryne beacnap - Bosworth. Compare also Jn 6.30 hwast dest pu to tacne, past we gelyfonl Beo 625

.., GodeCmiltsa] Jjancode3 wisafajst wordum, pass J)ea hire se willab gelamp, \)asth heo on benigne eorl gelyfde fyrena frofre.

Pasta in J)a?s pe goes with pancodea and is not directly relatable to the following pastb. pass may be repeated as in 1778. The sense is then, 'She thanked God for the fact that3 her willb was fulfilled. It (her will) was thatb she might rely on some warrior for help in suffering = her will that she might ... was fulfilled.' All those variations are made plausible by the underlying structure [X is Y].

2.5. Apparent similarity28 A totally different interpretation would be required for Beo 828: Haefde East-Denum Geatmecga leod gilp gelassted, swylce oncyppe ealle gebette

(Pro·) Nominal reference in Old English


inwidsorge, pe hie xr drugon ondfor preanydum polian scoldon torn unlytel. Swylce goes with the particle pe, which does business for paette, as in some other places of the poem. Lines 830-1 (Swylc oncypepe...pe) represent the classifying relative construction. These lines will thus translate 'Beowulf performed his lofty promise; he removed all such distress as they had endured and had inevitably suffered - [it was] no small tribulation.29 Swylce followed by pe with classifying force is attested in JEifLS I pref. 60 He ne maeg beon wurpful cyning / buton he haebbe pa gepincpe pe him gebyriap / and swylce pening-men pe peawfaestneysse him gebeodeon 'He cannot be an honored king, unless he has conditions a man in his position deserves and such attendants as [will] offer him obedience.' I do not see the point in Skeat's rendering '...unless he has the state which befitteth him and as it were serving - men, to offer him obedience (my italics).' Moreover, it is tempting to change the MS reading py to pe in line. 1797. Beo 1796 se for andrysnum ealle beweotode £>egnes Jjearfe, swylce pe [for py] dogore hea|)oliJ)ende habban scoldon. Oft Scyld Scefing sceajjena Jjreatum mogegum maegjmm meodsetla ofteah. Here too, their dragging away of the mead benches is eilen but the fact or the circumstance that Scyld dragged away the mead benches &c. is not (cf. 'eilen fremman'). Hence the absence of paet: They performed deeds of valor: they dragged away &c.

2.6. Past introducing an adverbial clause (= swa past) The pronoun paet appears when it is used adverbially in the sense of 'so that': Λ 1

Beo 20

Swa sceal geong guma gode gewyrcan, fromum feohgiftum on faeder bearme, paet hine on ylde eft gewunigen wilgesijjas, {)onne wig cume, leode gelaesten.


Koichi Jin

It will be the easier to understand the meaning of these lines, if we 'reduce' it to HISTORICAL STATEMENT: 'The king as a young man acted in such and such a way, that when he came of age good companions stood by him and, and lent him aid to the people when war came.' A comparable situation develops in Beo 64

J)a waes Hrojjgare heresped(B-*) gyfen, wiges weorjjmynd, pazt him his winemagas georne hyrdon, o^aet seo geogoj) geweox magodriht micel. Translate: 'Success in war was granted him, glory in battle, so that his friends and kinsmen obeyed him [one after another], until...'

These lines imply that he emerged as winner out of his struggle with his friends and kinsmen. Thus we read, Beo 893

Haefde aglasca eine gegongen, past he beahhordes brucan moste selfes dome.

Ellen refers back to his fight with Grendel described in the preceding lines: 'He had thus performed a deed of valor, so that he was now in a position to dispose of the treasure at his own will.' Understandably, that he could dispose of the treasure was not a deed of valor; it was on the contrary the result of such a heroic deed. So 1769 ff.

2.7. Absence of concord - summary One thing that prompted Klaeber's decision to change the MS reading in line 15 was without doubt the lack of concord in gender between pearf, which is feminine, and p&t, which Klaeber misread as a relative pronoun. He knew well enough that such correlation does not hold between the subject and the complement. The upshot is that he had started on the wrong track. Furthermore, it must be borne in mind that not any noun can occur in this position. It is only those which can characterize or epitomize a particular state of affairs indicated by a pxt or that clause that are allowed in. For example, tragedy may go in a pair with a that clause but famine

(Pro-) Nominal reference in Old English


may not.32 Basically, this rule applies by and large to Old English as well. See also 5.4.7.

2.8. Apparent RIGHTSHIFTING This is the case if and only if Β and C are in apposition. We have A and Β but not C expressed in Beo 189

Swa J)a maelceare mage Healfdenes singala seaj), ne mihte snotor haelej) wean onwendan; waes J)aet gewin to swyj), laj> ond longsum, J^e on J)a leode becom, nydwaracu nijjgrim, nihtbealwa maest. past fram ham gefraegn Higelaces |>egn god mid Geatum.

This, if will be agreed, is a common example of anaphora by se pronoun: Smith was going to resign [and] I knew it / the fact. By dropping A, and at the same time making it re-appear as C through dislocation, we get the familiar structure paet(B)... past clause (C).33 Beo 290

ic past gehyre, J?ae pis is hold weorod frean Scyldinga. Ί have it from certain sources that this is a troop dedicated to the lord of the Scyldings' J?aet pis is hold weorod frean Scyldinga. past ic gehyre 'This is a troop dedicated to the lord of the Scyldings. So I have heard / That is what I have heard.'

A careful observer will notice at once that we are again dealing with predication, not apposition: 'That their engagement had been cancelled was a somber message,' / Ί received from him the somber message that their engagement had been cancelled.' Compare 'The somber message (B) [which -C] I received from him was that their engagement had been cancelled.' Our discussion of the issue, I believe, will demonstrate clearly enough that the 'anticipatory' it really represents the complement of the sentence of which the that clause is the subject. The process may be illustrated as follows.


KoichiJin A certain message(compl.), namely that ...(subj.) the message Ccompl.) that...(subj.)

We are told by the complement how the report conveyed by the that clause should be characterized (news, message, truth, tragedy &c.) Needless to say, it(B) is the complement, the that clause (C for A), the subject of the sentence: it (=the circumstance in question) is that... preference to the report). In formulaic representation: past, id est past. The type Y (=the NOUN) is X (the past clause) is illustrated by Beo 1240 Waes peaw hyra, past hie oft wasron anwiggeare. This, in essence, will correspond to the modern 'The harsh truth is that only the fittest survive', represented by the formula 'Determinate Β followed by C (i.e., a J)£et clause). Characteristically, peaw used in this way is invariably accompanied by a determiner, not infrequently a possessive pronoun. This accords well with the modern locution discussed above: Or (Bately) 24,17.., \>eah pe hiora peaw wasre past hi eallehiora cyningas hetan Pharaon, Or (Bately) 29, 29 sippan wass hiera peaw, past hie aslce geare tosomne ferdon.

2.9. Non-determinate Β as the complement Non-determinatepeaw, on the other hand, is found in conjunction with hit/ past, to which it is the complement.34 Gen 50,3 (Bosworth) Feowertig daga hit wass peaw past man sceolde wepan xlcne deade mann, Or (Bately) 17,31 & pact is mid Estum, peaw past pasr sceal aelces gepeodes man beon forbserned. In brief, Β in apposition with C assures the relative status to the latter, whereas predication transforms C into a past clause with the restrictions mentioned above. Further examples will illustrate our point, Beo 632

Beo 278

Ic past hogode, Jja ic on holm gestah, sasbat gesagt mid minra secga gedriht, past ic anunga eowra leoda willan geworhte,.. Ic pass Hrojjgar masg rasd gelasran, hu he frod ond god feond oferswyjjeJ).

Beowulf offers to give HroJ>gar advice as to (pass) how he should overcome his enemy. 5

(Pro-) Nominal reference in Old English


Beo 2027 ond past raed talaf) past he mid [)y wife waelfashjm dael saecca gesette. 'He considers it/the idea as a good solution ['raed'], that by means of this woman he should settle their share of slaughterous feuds (adapted from M. Swanton).'36 Although Modern English employs much less of the correlation (it that), it has not entirely disappeared: I took it for granted that Mr Smith would be elected chairman rightshifted from Mr Smith would be elected chairman. I had taken it for granted but to my surprise ... Beo 1778 ...pass sig Metode Jjanc ecean Dryhtne, pass pe ic on aldre gebad, I assume this is derivable from ?/c swa on aldre gebad, pass sig Metode panel (cf. Beo 586, with pass referring to the preceding mentioning of a fact.) The second pass must have been analogically carried over from the first one: Ί have lived long enough and I thank God for my longevity / for it' or Ί thank God for the circumstance that I have lived long enough &c.' 38 Our idea of TRIANGLE justifies us in speaking of the omission of B. This takes place in all probability when an anaphoric past emerges in correlation with a clausal past. In general, one may venture the remark that nouns resist omission because of the fullness of their lexical content as compared with an anaphoric pronoun, hit or past. Perhaps exceptionally in a sustained piece of writing either in prose or verse, the member A in our TRIANGLE may appear 'unprocessed' in the place of C, as for example in The harsh truth (Β) Darwin's book reveals is: only the fittest survive (A for Q. Christ 600 past is J)ass wyrj)e, Jjaette wer-J>eode secgen dryhtne J)onc dugu|>a gehwylcre t>e us si{) and aer simle gefremmede Jnirh monig-fealdra maegna geryno. he us ast giefep and ashta sped welan of er wid-lond and weder lipe under swegles hleo sunne and mona


Koichi Jin

We could move lines 605ff. before 1.600 without changing the scene materially. T h e circumstance,' says the poet, 'is worthy of a reaction on our side, which is that we thank God for each of His favors He has granted us.'

3. Mistaken identity In the following, the defined class noun seems to have reference to the that clause in the next line thus, forming the arrangement B, C. Beo 1167 gehwylc hiora his ferhpe treowde, f)3et he harfde mod micel,... This, however, is not the case. First, hisfeohpe does not go into predication with a that clause. Second, a defined class noun cannot be separated from the that clause in the predicate - subject relation. This is a rule still meticulously observed - see our earlier discussion on fact and feat. It is of utmost importance to know the fact that our earth is growing warmer every year. *To know the fact will be of utmost importance that our earth is growing warmer every year. The case is made still clearer by the next lines. Beo 1180 ...Ic minne can glaedne Hrofmlf pset he £>a geogojpe wile arum healdan. Obviously a proper noun cannot be the subject of a sentence in which a that clause is the complement. We are dealing here with an apo koinou construction: can + minne Hropulf / + [an abstract noun], pset. It is improbable that cunnan could co-occur with a fjset clause in Old English. Hence, it will necessarily follow that cunnan takes two different types of object in a single sentence.39

Concluding Remarks 2 In the 'normal' relative clause, C, which is analyzable into Ν - Ν (cf. OE se pe), goes with B, which is N \ N 2 then copies the grammatical features IJ

(Pro-) Nominal reference in Old English


(including case relations) from N1 (e.g., tofraemmanne, f?aem pe Johannes waes gehaten). N3, i.e., pe in the above - on the other hand is part of the clause that follows. Removal of Β or Ν'from the cluster B/C leaves N 2 standing side by side with N3.

Notes 1. This article appeared for the first time in Jinbungakuho

or The Journal of


and Social Sciences, No. 263, Tokyo Metropolitan University (1995). It has been slightly revised. 2. For a modern definition of anaphoric reference, see Katie Wales, A Dictionary


Stylistics, pp. 23-24. 3. Future is often a synonym of uncertainty, e.g., 'He'll be sleeping now.' 4. Namely: 'Nobody has yet an idea of what it's going to be.' 5. In our discussion of gender relation, it will be seen, only human beings become our concern. With lower animals, it is of secondary significance. 6. See Bosworth, op cit. fre 1(4). The insertion of a personal pronoun is the more called for because our particle has no oblique cases, e.g., Jye his = whose; Or 2,5 |)a men f)e mon hiora aer ofslog. 7. See Footnote 2. 8. There is a lingering reminiscence of the earlier situation in the use of that in 'The population of London is much greater than that [i.e., the population] of Liverpool.' Disappearance of the characteristically masculine and feminine forms has isolated the neutral that (and the plural those) in later English. 9. Note that the position now occupied by 'the one' used to be reserved for the se pronoun in Old English. 10. Mere spatial transposition of Β does not produce predication, although it often appears to do so: past(B)nis Petrus fjxt(C) cnucaj) as his engel - Petrus(A)nis se mann (B)se f>e (Qcnucap

ac his engel. This construction, based on fronting by the thematization, is

now called most improperly a cleft. 11. M.L. Wadell et al., The Art of Styling Sentences, (Baron, N.Y. 1983), p. 18. 12. Naked quotations (our A), cannot always serve as the subject. It then needs to be processed into C. - compare, 'No, I won't' was the reply. 13. Cf. 'Now, may I introduce to you a well-known artist who's going to hold a concert first time in our town: ladies and gentlemen, John Smith!' 14. Darwin states a truth: the truth in question (=it) is that... The Latin is est reflects this situation. 15. One performs or achieves a feat but one cannot perform 'that one passes the exam at a first try.' Or to put it in other words, 'that' in this position refers to an action, not to a



fact or a circumstance. This will also exclude the type d), which is based on the predication X is Y: *Tom achieved a feat, namely that he passed &c. 16. For e) we can compare: jEChom I, 39 604, 14 pset is gejxjht ure drihtnes willan: us ne gepafap mandaede to gefremmenne. 17. Our reading assumes that dreogan is here intransitive. 18. Cf. Beo 354-5. pa andsware.. pe me se goda agifan pencep, Beo 1482 Swylce pu pa madmas, pe pu me sealdest... 19. Pe in Beo 830 introduces a relative clause with classyfing force: swylce oncypde ealle gebette / inwidsorge, pe hie serdrugon 'they remedied all of such distresses as they had suffered...' 20. See Klaeber: Beowulf, p. 125 (notes to 1.15). Since Searf is feminine, his reading would exclude pset as a relative pronoun. M. Swanton ignores this fact. 21. Or If you do not refuse me, it would be a gracious favor to me, according to the modern idiom. 22. Ά (solemn See.) truth has been made known, namely that God eternally wields the fate of mankind.' Cf. line 14. For the use of the past tense 'weold', we may compare 'He knew everything! (= I have not been aware of it)' said when one suddenly realized a fact. 23. We will see at once that in this example the past clause is futuristic as against (fyrendearf) pset in Beo 1405 with reference to a current situation. Thus Β - liss: C pset... 24. Note that wyrd is a feminine. 25. Cf. OE gelimp may translate 'a course of events': paforhtede pe [sic] biscop for pam fserlice delimpe - Bosworth. 26. Very probably swa in 1.20 does not belong to our construction. In Old English both in prose and verse, HISTORICAL STATEMENT is often followed by GENERAL STATEMENT with swa in- between, to mark the transition. This logical structure is sometimes called esse ad posse. Historical facts are given and checked before general conclusions are drawn from them. A typical example is Beo 2425ff. A step further will bring us to a more subjective arrangement of thoughts. As above, facts are given. The narrator then goes on to give a sermon on the ethical implcations of those facts. Cf. Beo 175-185, where 1 Wei bip psem...' expresses the narrator's ethical and religious persuasion. In Beo 20 the first part(esse) is only vaguely outlined (Beowulf wses breme blasd wide sprang). The first part again, is understood in 1384-9 (Ne sorga snotor guma! Selre bip seghwsem /past he his freond wrece, ponnefela murne...) 27. Cf. Waes seo hwil micel / twelf wintra tid torn gepolode / wine Scyldinga, weanna gehwelcne, sidra sorga (torn = 'an unpleasant circumstance, nuisance'). 28. This is based on the factual observation that pe often has the value of se pe, seo pe, past Pe.

CPro-) Nominal reference in Old English


29. Otherwise I suspect ANACOLUTHON in the last lines. After oncyp pe, ..inwidsorge pe aar drugon, our poet began a new sentence, to ring the changes on the foregoing ides: 'and they had indeed suffered through the dire necessity no little trouble* - adapted from M. Swanton. 30. Although py dogore is an established idiom, the particle may not be an indispensable part of it: 2573 t>aer he Jjy fyrste forman dogore. 31. Very probably swa in 1.20 does not belong to our construction. In Old English both in prose and verse, HISTORICAL STATEMENT is often followed by GENERAL STATEMENT with swa in - between, to mark the transition. This logical structure is sometimes called esse ad posse. Historical facts are given and checked before general conclusions are drawn from them. A typical example is Beo 2425ff. A step further will bring us to a more subjective arrangement of thoughts. As above, facts are given. The narrator then goes on to give a sermon on the ethical implications of those facts. Cf. Beo 175-185, where lWel bip psem...' expresses the narrator's ethical and religious persuasion. In Beo 20 the first part(esse) is only vaguely outlined (Beowulf wses breme bleed wide sprang). The first part again, is understood in 1384-9 (Ne sorga snotor guma! Selre bip scghwsem /past he his freond wrece, ponnefela murne...). 32. That a plane crash killed 50 passengers is a tragedy', that there was scarcely enough food to go round among the villagers was again a tragedy but not a famine. Klaeber should have restricted himself to the B/C cluster, where the rule of concord is binding. 33. That the second pact is shortened from psette is evidenced passim, e.g., Beo 1255 pset gesyne wear£, / widcuf) werum, psette wrecend J>a gyt / lyfde aefer la^um... 34. Hit is anaphora in BlickHom 67,7 pa bseron hie him togeanes blowende palmtwigu; forpon pe hit [= the afore-mentioned practice] wses Iudisc peaw, ponne heora ciningas... 35. One would expect namely to connect pset and the statement beginning with hu \pset, id est: hu he frod &c.] Here we are dealing with a case very similar to that of lines 2-3 or lines 3-5: prym, namely, hu... / eilen, namely: Often... But the determiner pset makes a difference. The use of hu reflects the interrogation implicit in the noun clause, much the same way as the modern Ί had no idea how I should go about it* does in contrast to Ί had no idea that he was coming.' Cf. hwset in line dl72. 36. An example from the prose: JElfLS XXXI 892 J)aet he pone rsed ne cu|)e pset he hine swa hraf>e gebaede. 37. 'Thanks be to Providence, the eternal Lord, that after ancient tribulation I have lived long enough to gaze with my eyes on that bloodstained head! - M. Swanton. Notice that fusion of consonants (-tp > tt> t) did not take place (psette = Q. 38. While a locution like to take it for granted that... is still in use, it is no longer possible to say with the Germans Ί thank God for it that... ('Ich danke Gott dafür [: für den Umstand], daß ich so lange irdisches Leben genossen habe').



39. 'It' is understood in 1246: Waes [hit] Jwaw hyra, / past hie oft waeron anwigearwe,.. Nor do we say 'That they were ready for war., was their custom.' 40. Or in more familiar terms, RES IPSA

the camera (N1) / the one (N2) that (Ν3) I had

bought in London.

References Bosworth, Joseph - Toller, T. Northcote 1921 An Anglo-Saxon dictionary. Oxford: Clarendon Press. Heyne, Moritz - Levin L. Schücking 1958 Beowulf. Paderborn: F. Schöningh. Klaeber, Francis 1950 Beowulf. Boston: D.C. Heath and Company. Murry, J. et al. (eds.) 1933 The Oxford English dictionary. (1 st edition) Oxford: OUP. Soukhanov, Anne (ed.) 1992 The American heritage dictionary. Boston: Houghton Mifflin & Co. Swanton, Michael 1978 Beowulf. Oxford: OUP. Urdang, Laurence 1988 A Dictionary of differences. London: Bloomsbury. Wadell, M.L. et al. 1983 The art of styling sentences. N.Y.: Baron Educational Series.

Some notes on the grammatical category of gender in Vices and Virtues Shigeaki Karakida

0. Introduction Jones (1967a, b) analysed the grammatical category of gender in some late Old English and early Middle English texts. He showed that the originally case- and gender-distinctive markers, -ne, ~(e)s, ~(u)m, -re, are often used as case markers only, and have no gender significance. In 1988 Jones reanalysed the grammatical category of gender in the Lindisfarne Gospels, the Durham Ritual, the Peterborough Chronicle, Lajamon 's Brut and Vices and Virtues. His analysis is based upon localist case grammar.1 I agree with him when he (1988: 2) says that ... the gender reflecting morphology of such attributive words as definite articles, possessive pronouns, "strong" adjectives and the like was not immediately the subject of a "sudden" catastrophic neutralization... But his argument is very confusing because, while he uses the terms denoting grammatical relations, subject, object, etc., and the terms denoting surface cases, e.g., "nominative", "dative", he uses the technical terms of localist case grammar, e.g., "ergative", "absolutive", "locative", to denote deep case relationships without defining them. I think that we should discuss grammatical gender in Vices and Virtues on the premise that the language of this text is archaic, e.g., much more archaic than that of the Ancrene Riwle and the Katherine Group, contemporary with Vices and Virtues. Hall (1920: 443-444) concludes that "the language of the text is older than that current at the time when the copy was made." Karakida (1988) showed that "the language of Vices and Virtues is very archaistic from the point of view of the high frequency of pronominal recapitulation.2 According to the MED, the text was composed about 1200, and the extant MS (MS Stowe 34) was written before 1225. Härtung (1972: 702) says that it may have been composed 1175-1225.


Shigeaki Karakida

Our corpus is based on the text edited by Holthausen. The microfilm of MS Stowe 34 was referred to throughout the study.

1. < ö e > - f o r m While there appears no longer to be any phonologically distinct feminine nominative definite determiner, the historically masculine nominative form se still exists, occurring with the three noun forms of the historically masculine gender and with the five noun forms of the historically feminine gender. Jones is only partly right when he (1988: 221) says, as follows: Clearly the "morphology-free" (öe) form is intruding strongly into absolutive contexts which in the other texts we have examined was characteristically the innovative domain for shapes for nouns of all genders... Table 1 appears to show that the form had ceased to be a case- and gender-distinctive marker: strictly speaking, however, it is true that it gradually began to cease to be a case- and gender-distinctive marker, but it had not completely ceased to be so. Table 1 shows that of 61 examples (41 noun forms) of the (öe form, as many as 33 examples (54%) (20 noun forms: 49%) occur with historically feminine nouns, in the objective (a). Jones does not mention at all why the form is "intruding strongly" into the "absolutive" feminine. It seems to me that öe has two phonologically distinct variants, i.e., de as an unmarked definite article with weak demonstrative force, and de as a marked definite determiner with strong demonstrative force, though these cannot be distinguished from each other in orthography. The "strong intrusion" of the form into the objective (a) feminine is nothing but an apparent phenomenon: in fact, many examples of the form not only intrude into the objective (a) feminine, but also remain there as transformed or unstressed feminine determiners derived from the OE feminine accusative marker pä. This phenomenon is similar to the fact that is almost always preserved with historically neuter nouns in the objective (a).

Some notes on the grammatical category of gender in Vices and Virtues


Table P. The Distribution of the Forms of (öe) and (öat) classified by the Historical Gender Category Μ


se öe öat Obj. (a) öe öa öo

3 (3) 112(30)

5 (5) 22 (14)

2 (2) 12 (9) 4 (4)

2 29 2 2

öat (b) öe öa öo öat

1 (Ό

2 3 1 1

(c) öe (d) öe öa öae

8 (7) 57 (24) 10 (3)

19 (8) 3 (3)

öo öat

1 (1)

1 (1)


(2) (17) (2) (2)




5 (5) 21(11)

2 (2)

5 (1)

7 (6)

1 (1) 2 (1)

2 (1)




(1) 25 (15) (1) (1) (1)

2(2) 3(2)



ι υ

3 (3) 28(15)

2 (2)

11 (8)

1 (1)

7 (3) 2 (2)

1 (1)

1 (Ό 4 (3) 2 (2)

There exists a peculiar form in Holthausen's text, which is the form (pes)e in (pes)e hali mihte (137/7). In MS Stowe 34, pes is written above de, as follows: 1.1.

To alle öo nedes öe man hafö to done. J>es l>anne is öe hali mihte swiöe helpinde. (p. 84, 11.6-7)

Holthausen's emendation is misleading, because the original form de is corrected to pes, and pese (or dese) does not occur elsewhere as a nominative singular marker, as Karakida (1990) showed. At first, the scribe of this part must have written Öe as a definite determiner (not an article) with strong demonstrative force, i.e., the marked form Öe. But the same scribe4 must have become aware that de could not fully convey its demonstrative force to the audience. Probably, Vices and Virtues is not an orally-delivered prose text. If preachers had only read aloud to the audience, this correction would not have been necessary, because they could literally put stress on Öe. But if there had been opportunities for


Shigeaki Karakida

people to read this MS privately, then this correction would have been necessary: hence the scribe corrected the original form de to pes.

2. and Table 2. The Distribution of The Forms of and classified by the Historical Gender Category Μ Obj. (a)

öane öanne dan öisne -ne

Obj. (c) Obj. (d)

dan öan öen öane öene öesen



24(15) 3 (1)


1 (1) 2 (2) 10(10)


1 (i) 13 11) 3 1) 2 (1)


9(5) 1(1)

1 Ο) 3(3)

Table 2 shows that the historical agreement of the marked forms (-ne) and (öan) is almost always preserved. As I have pointed out in the Introduction, the language of Vices and Virtues is very archaistic. It is that same archaism that is the reason why the marked form does not intrude into the "absolutive" feminine, examples of are: 2.1. (1) (2) (3)

(4) (5)

hie brohte dane brihteste angel (5/9) Wolden hie hlesten dane hali apostel (67/14) Si uis perfectus esse, '}if 9u wilt,' he seid, 'bien öurhut god mann, öanne forlat öu öe woreld; 7 3if 3u nelt naht, J>u miht wunijen on öare woreld, 7 ec bien 3ebore3en, 3if öu öinne cristendom wel hal(d)st.' (73/6) [In the Latin-based sentence; cf. Matt. 19:21] He for3af hire dane dead (111/30) [cf. se eche deaö (51/31)] (a) On öelliche wise ouercam Crist, godes sune, dane swikele dieuel öurh öessere iblescede mihte (51/18) (b) Jjurth hersumnesse of öe hali rode ourcam Crist dane ealde dieuel (119/6)

Some notes on the grammatical category of gender in Vices and Virtues


(7) (8) (9)

(10) (11) (12) (13) (14)



(a) hes...underfoö al swa dane hali gast (37/2) (b) faste on me dane gost of strengte (83/22) (c) hafst 3esaent dane froure gost (83/33) (d) rihtne gost newe inne me (83/9) [cf. se haligast (19/22)] Qvia hillarem datorem diligit deus, 'Gladne yuere luueö godd.' (139/28) [In the Latin-based sentence; 2Cor. 9:7] All he 3eald dane harm öe was 3ecumen öurh Adam (117/33) [The harm refers to original sin.] (a) hie makeö dane religiuse man...sari 7 drieri 7 heui (3/12) [This refers to a monk or friar.] (b) alle hie motem disne hali mann Daniele foljin (43/16) We findeö on öe write öat...he scolde habben mildsce, bute 3if hit ware öat he dane prest forhowede (123/17) Miserere mei deus, dane derewurde salm anon he makede (81/26) [cf. Ps. 50:3 (51:3)] wit boöe anne sceppend hadden (95/18) hie fol3iö Noe dane gode stieresmann (43/21) anbidende öa eadi hope 7 dane to-cyme of öare michele blisse of Criste(s), gode(s) sune (31/9) [In the Latin based sentence; Tit. 2:13. There occurs the feminine accusative marked determiner da, which agrees in gender with the head noun hope, reclassified as feminine: see Section 3] (a) fol3in dane rih[t]wise 7 onfald lob (41/17) (b) dane eadie lob f o ^ i n (81/9)

These examples in 2.1. show that the historically masculine accusative marker is preserved in very homiletic phrases or contexts which need archaistic marked modifiers, in the sense that it occurs with important nouns in Christianity, angel, apostel, dieuel, (hali) gast, prest, sceppend, etc., or in Latin-based sentences: the primary function of is, therefore, more stylistic than syntactic. Table 3. The Phonological Environment of and + vowel Obj. (d)

+ consonant 18(72%)

7 (28%)

4 (3%)

144 (97%)


Shigeaki Karakida

I think that Jones is partly right when he (1988: §5.10.i) shows that the form is used as "locative", because it seems to me that the choice of is, though not completely, determined by its phonological environment. Table 3 shows that is used eighteen times (72%) before a vowel, while is used only four times (3%) in the same phonological environment, in the objective (d). Of 16 examples of occurring with historically masculine nouns, 10 examples are used in order to avoid hiatus in the objective (d):5 2.2.


me seal don bi him al swo bi öan asse (93/10) to öan aöe (9/15) (9/16) On öan ilche dai3e öe öu tebreest (89/18) ICH öe bidde for öan ilche hlauerd öe öe iscop (47/8) cumj) to öan unbiliefde manne (45/20) mid öan eadi3e well-streme (103/4) at ten aende (25/3) (89/26) aten ande (33/10)

The examples in 2.3. are stereotyped prepositional phrases to avoid hiatus.6 There is no example of . Moreover, of 9 examples of occurring with historically neuter nouns, 8 examples are used in order to avoid hiatus in the objective (d).7 2.4.

hu öu scalt from öan euele bujen (65/13) in to öan eche fiere (19/31) (25/30) besecheö of öan ilche gode (11/30) to öan eche Hue (25/29: MS dan) (53/4) bringe Jje to öan eche lif (33/21) hit cam eft te öan eche Hue (119/4)

And in 2.5., seems to be used in order to avoid hiatus. 2.5.

mid pan onen of J)ese ei3en (125/17) ne öen enne worpen ouer dan odre (135/3)

3. Table 4 appears to show that the historically feminine dative marker "intrudes" into the historically non-feminine nouns.

Some notes on the grammatical category of gender in Vices and Virtues


Table 4. The Distribution of the Forms of classified by the Historical Gender Category Μ Obj. (d)


öare öaere

79 (29)

Ν 3(2)

M/F 4(3)

F/N 1 (1)

M/F/N 1 (1)

1 (1)

öere öessere


1 (1) 32(13)


1 (1) 1 0) 22 (14)

öesere öesre -re


KD 1(1)



All the examples of the unhistorical use of in the objective (d) are listed in 3.1. 3.1. Μ


(a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k) (1) (m) (n) (o) (p)

belieueö on öessere soöe beleaue (51/33) for 9are euele 3ewune ne öin(c)ö hit hem no misdade (79/17) swinkeö for öessere eadi hope (33/13) For öessere eadi hope (35/3) swinkeö for sumere hope (33/9-10) hauen none strengj>e a3ean öessere ileaue (27/8) habben strengte a3ean öessere gode ileaue (27/11) for öare misbileause (53/7) for öare unwurscipe (53/8) 3iuen 3ew forbisne of mire a3ene wille to donne (15/10) öe 3ew haueö of öessere (michele) wrecchade ibroht (21/17) on öese liue ne on öere oöre (63/1) cumen into öare riche (115/3) Jjebringö to öare riche (129/10) wel ilieue be are tacne (31/25) drinken of öare wine (149/6)

Among them, the examples of unhistorical agreement except 3.1. (b) (o) (p) can be explained by gender change. The morpheme -leaue, which is historically masculine, is pronominalized by the feminine personal pronoun hie in 3.2. 3.2.

Fide(s) sine operibus mortua, 'Ileaue wiö-uten werkes, hie is dead.' (29/24-25)


Shigeaki Karakida

Jones (1988, preface) says that the scope of this book "is strictly limited to a discussion of attributive word morphology in the noun phrase": in other words, from the outset, he fails to take into consideration noun phrase external gender agreement. Therefore he does not take into account the following sentences in (3.), in which the historically masculine noun hope, one of the virtues, is referred to by the feminine pronouns. 3.3. (a) (b)

Ne Jjinche hit te naeure swa bitter, öat pies hope hit ne sw(i)eteö. Swa hie dede alle δο halie martirs. (33/31-32) De faste hope hafö hire stede up an heih, for öi hie is rof and wrikö alle öe hire bied beneöen mid öe scincles of holie J>ohtes {>e sapientia hire flint. (95/6-8)

The nouns ileaue and beleaue belong to the abstract nouns denoting virtues: hence their gender is changed from masculine to feminine, as is discussed in Karakida (1983). The noun misbileaue was coined in the early Middle English period, according to the MED, which cites 3.1. (h) as one of the earliest recorded examples: MED s.v. misbileve 2(a). Because this noun shares the morpheme -leaue with ileaue and beleaue, the gender of misbileaue must have been regarded as feminine. The historically masculine nouns unwurscipe, ajene wille, wrecchade may also have been reclassified as feminine in gender because of the feminization of the abstract nouns denoting vices, as is also discussed in Karakida (1983). Table 5. The Frequency of Historically Feminine Nouns occurring with and Obj. (d)

+ F nouns 23 (10) (22%)

+ F nouns 81 (29) (78%)

Table 6. The Frequency of Historically Non-Feminine Nounso ccurring with and

Obj. (d)













121 (51) (82%)

26 (18) (18%)

Some notes on the grammatical category of gender in Vices and Virtues


Table 5 shows that historically feminine nouns occur with only twenty-three times (22%) (10 noun forms), while they occur with as many as eighty-one times (78%) (29 noun forms), in the objective (d). On the other hand, Table 6 shows that historically non-feminine nouns occur with as many as 121 times (82%) (51 noun forms), while they occur with only twenty-six times (18%) (18 noun forms), in the objective (d). These two tables appear to show that " was still a powerful case form" (Jones 1988: 223): it appears to be more powerful not only than but also than . Strictly speaking, however, is less powerful than Jones seems to think, because as many as 26 examples (32%) of occur with in the objective (d), as Table 7 shows. In short, the frequency of occurring with is higher than that of occurring with other historically feminine nouns. Table 7. The Frequency of and occurring with and Other Historically Feminine Nouns

other F nouns

other F nouns

5 (26%)


Obj. (d)

14 (7) (74%)

55 (28) (68%)

Moreover, Jones (1988: 227) says that the historically neuter noun was reclassified as feminine in gender. If we admit the feminization of , then we should also admit the feminization of the abstract nouns denoting vices and virtues, probably influenced by Latin, e.g., ileaue, beleaue, misbileaue, hope, unwurscipe, ajen-wille, wrecchade. does not "intrude" into historically non-feminine nouns when it occurs with these nouns: in fact, these nouns are reclassified as feminine in gender. This feminization is closely related to the female personification of the abstract nouns denoting Christian vices and virtues. The sennes are described as sisters by the Soul at 3/22. It is also worth noticing that rihte jeleaue is regarded as a sister of fast hope at 29/31, that pudicicia is also regarded as a sister of castitas, i.e., clannesse at 131/8, and that, in this way, the hali mihtes are described by Reason as sisters, whose mother is called discrecio, i.e., sckelewisnesse at 149/10. 3.4. shows that the genitive case of the historically neuter noun bede 'prayer', one of the virtues, belongs to the feminine inflexion. 3.4.

hat [=at] alchere bede aende (141/27)


Shigeaki Karakida

Another explicable form is the following definite determiner Öere. (3.5.)

on öese Hue ne on öere oöre (63/1)

The phrase-structure of 3.5. is similar to that of (3.6.). (3.6.)

aider on öessere woreld 7 ec on öare oöre (35/12)

While the nominal head woreld (F) occurs with the feminine forms öessere and dare in 3.6., the nominal head Hue (N) does not agree with the determiner öere in historical gender in 3.5.. The answer to this problem seems to consist in suggesting the feminization tendency of and the fluctuation in gender of in our text. almost always occurs with the feminine definite determiner in the objective (d), besides 3.6., as follows: 3.7.

on öare woreld (17/19) (39/32) (41/18) (41/33) (67/16) (73/5) (75/2) (77/5) (109/16)

as well as at (29/34; 33/9; 35/7; 39/10; 41/15; 43/20; 45/9; 49/6; 65/27; 73/2; 73/21; 73/22; 73/34; 81/8; 103/21; 143/12; 149/4). On the other hand, always occurs with the ending -es in the genitive singular case in our text, as follows: (3.8.)

alles woreldes blisse (31/27) nanes woreldes blisse (31/29)

as well as at (31/19; 33/15; 35/4; 43/5; 61/25; 67/19; 75/22; 79/33; 81/3; 139/33). In short, fluctuates in gender between feminine and non-feminine.8 It should be noted, moreover, that, even in OE, had the non-feminine genetive ending -es, as follows:9 3.9.

(a) worldes frymöe (Cotton MS) weorldes fremde (Hatton MS) Cf. worulde (Lindisfarne) (Lk. 1:70) (b) suno woreldes öisses filii saeculi huius (Lindisfarne) Cf. Jjysse worulde beam (Cotton MS) J)isse worulde beam (Haton MS) (L£.20:34)

Some notes on the grammatical category of gender in Vices and Virtues


It seems that the genitive ending -es caused the paradigm of to approach that of the historically neuter noun , while the historically feminine gender influenced the gender of . Moreover, as I noted above, 3.5. is similar to 3.6. both in the phrase-structure and in semantic value. The meaning of dese liue and bessere woreld is the earth as opposed to heaven or hell, while that of dere odre and dare odre is heaven.10 Consider the three examples of in 3.10. 3.10.

(a) man to michel mai drinken of dare wine (149/6) (b) for dare euele jewune ne din(c)ö hit hem no misdade (79/17) (c) wel ilieue be are tacne öe he hafö ijiuen me (31/25)

Evidence for gender change in these head nouns cannot be found elsewhere. But if Jones's localist theory is correct, what case relationships do these three examples represent? The occurrence of in 3.10. (a), in which the prepositional phrase of dare wine, which is traditionally called a "partitive genitive", can be explained as a "locative" ("ablative") form from the point of view of Jones's localist case grammar. Because Jones would probably regard the two examples of in 3.10. (b) (c) as cause or reason, he would call them "locative" ("ablative").

4. Conclusion Jones's analyses (1988), to which considerable labour must have been devoted, are very interesting and informative but a number of problems remain to be solved. First, a weak point in Jones's gender theory is that he excludes argument of noun phrase external gender agreement from the outset. It is clear that grammatical gender cannot be discussed enough unless both noun phrase internal and external gender agreement are taken into consideration. Secondly, the terms denoting "deep cases", e.g., ergative, absolutive, locative, are used without any fully strict definitions. The language of Vices and Virtues is very archaistic, considering that it was composed about 1200 A.D. When we discuss the grammatical category of gender in this text, we should always take this fact into consideration. In the first place, the form began to cease to be a case-


Shigeaki Karakida

and gender-distinctive marker, but it had not completely ceased to be so. The reason why the form appears to be intruding strongly into the "absolutive" feminine is that many examples of the form not only intrude into the objective (a) feminine, but also remain there as transformed or unstressed forms derived from the OE accusative feminine marker J?ä. Secondly, the historical gender agreement of and is almost always preserved. It is because of archaism that the marked form does not intrude into the "absolutive" feminine. Because many examples of are preserved in homiletic phrases and contexts, its function is more stylistic than syntactic. is very often used, instead of , in order to avoid hiatus. Thirdly, most examples of unhistorical use of can be explained by gender change of head nouns. appears to be a powerful case form in the objective (d): strictly speaking, however, the main reason why it appears to be so is that the historically non-feminine abstract nouns denoting vices and virtues are reclassified as feminine in gender, probably influenced by Latin, which are very important nouns as key-words for understanding European culture, literature, theology, philosophy, etc., in the Middle Ages. This feminization is closely related to the female personification of those abstract nouns.

Notes *

This is a revised version of the article entitled "A note on grammatical gender in Vices and Virtues," which was published in Studies in English Literature [English Number], 1992, 75-90.1 wish to express my deep gratitude to the late Professor Kikuo Miyabe for his invaluable suggestions on this topic and to Professor Kikuo Yamakawa and Professor Yoshio Terasawa for their useful comments and suggestions. I am also grateful to Dr. Yoko Iyeiri for sending me a copy of Philippsen's dissertation. The responsibility for the views presented in this paper remains my own.

1. Anderson (1971, 1977) 2. What I call pronominal recapitulation is a sentence-structure such as the following example: Se de is of gode, he harkeö bleöeliche godes wordes. (47/22-23) A sentence-initial subject or object is modified by a relative clause is often repeated in pronoun form.

Some notes on the grammatical category of gender in Vices and Virtues


3. The figures refer to occurrences of determiners, and the parenthesized figures refer to noun forms which occur with each determiner. The objective (a) refers to forms which occur as the direct object of the verb, whether accusative or dative in OE. The objective (b) refers to forms which occur as the "adverbial accusative". The objective (c) refers to forms which occur as the indirect ("dative") object of the verb which takes two objects or as the "dative" complement of the adjective. The objective (d) refers to forms which occur after prepositions. 4. The scribe who corrects de to f>es is the original scribe. (Personal information from Mrs. M. P. Brown of The British Library) 5. Six exceptions are as follows: of öan michele brene (119/24) be öan daije (55/4) for öan michele embeöanc (69/14) of öan 3et>anke (19/3) in öan gastliche ofhe (73/29) in to öan pette (109/19) If the semivowel [j] in jefranlce is added to the phonological environment to avoid hiatus, the percentage becomes higher (77%). 6. The same idiomatic prepositional phrase to avoid hiatus can also be found in the Katherine Group (MS Bodley 34): Ker 1960. ed ten ende (M30r / 24: 3623) (HM54r / 8:60) (HM54V /10: 79) There is no example of . For details, see Karakida 1979. 7. The only exception is as follows: of öan holie watere (83/2) If the consonant [h] is added to the phonological environment to avoid hiatus, it can be said that is always used with historically neuter nouns in order to avoid hiatus. 8. In the following example, the definite determiner dare probably agrees with the historically feminine noun scame. for öare worldes scame (61/25) The use of the noun without the definite determiner can also be found in the following examples:


Shigeaki Karakida forlateö ... alle worldes wele 7 blisse (35/4) forlate... alle worldes wele (67/19) luuied worldes luue (139/33)

9. Skeat 1970. 10.1 agree with Philippsen (1911: 80) when he says: ... im dat. sg. tritt einmal femin. genus entgegen in 61/34-63/1: on dese liue ne on dere obre. Dieser fall wäre also ein dritter beleg für die Wirkung des dativ -e. Da es jedoch auffällig bleibt, warum der so häufig belegte dat. sg. liue (9/3, 15/2, 19/12, 21/2, 35/27, 25/29 u.a.m.) nicht auch in andern fällen geschlechtswechsel zeigt, so scheint mir die annahme wahrscheinlicher zu sein, daß hier die Verwendung des dere die folge einer kreuzung des ausdrucke on dere liue and on de obre mit dem gleich gebräuchlichen, im sinne mit dem vorigen völlig übereinstummenden 'on bessere woreld and on dare odre' (35/12) ist.

References Anderson, John M. 1971 The grammar of case, Cambridge University Press. 1977 On case grammar. London: Croom Helm. Hall, Joseph (ed.) 1920 Selections from early Middle English. Oxford: Clarendon Press. Härtung, Albert E. (ed.) 1972 A manual of the writings in Middle English 1050-1500, Vol. 3. Connecticut: Archon Books. Holthausen, Ferdinand (ed.) 1888 Vices and Virtues, Part I: Text and translation. EETS OS 89. London: Oxford University Press. 1921 Vices and Virtues, Part II: Notes and glossary. EETS OS 159. London: Oxford University Press. [ 1967] [Reprinted. Oxford: Oxford University Press.] Jones, Charles 1967a "The functional motivation of linguistic change: A study of the development of the grammatical category of gender in the late Old English period," English Studies 48: 97-111. 1967b "The grammatical category of gender in early Middle English," English Studies·. 48: 289-305.

Some notes on the grammatical category of gender in Vices and Virtues


1988 Grammatical gender in English: 950 to 1250. London: Croom Helm. Karakida, Shigeaki 1979 1983

1988 1990

"The grammatical category of gender in the Katherine Group. (MS. Bodley 34)", Research Report ofKochi University 28: 161-193. "The grammatical category of gender in Vices and Virtues: Evidence for gender change in early Middle English", Studies in English Literature 1983 [English Number]: 83-99. "A note on syntax and style in the Vices and Virtues" in: Kinshiro Oshitari et al. (eds.), 126-139. "The definite determiners (pes) e and deies in Vices and Virtues," Medieval English Studies Newsletter. 35-37.

Ker, Neil R. 1960 Facsimile of MS. Bodley 34, with an introduction by N. R. Ker, EETS OS 247. London: Oxford University Press. Kurath, Hans - Sherman M. Kuhn - et al. (eds.) 1952Middle English Dictionary. Ann Arbor: University of Michigan Press. Oshitari, Kinshiro et al. (eds.) 1988 Terasawa Yoshio Kyoju Kanreki Kinen Ronbushu [Philologia Anglica: Essays presented to Professor Yoshio Terasawa on the occasion of His sixtieth birthday]. Tokyo: Kenkyusha. Philippsen, Max 1911

Die Deklination in den "Vices and Virtues". Erlangen: Κ. Β. Hof- und Universitäts - Buchdruckerei von Junge & Sohn. Skeat, Walter W. (ed.) 1874 The Gospel according to Saint Luke in Anglo-Saxon and Northumbrian Versions. Cambridge: Cambridge University Press. [1970] [Reprinted as The Gospel according to Saint Luke and according to Saint John. Darmstadt: Wissenschaftliche Buchgesellschaft.]

Verbal gerund and its historical development in English* Osamu


1. Introduction Various attempts have been made to describe the historical development of the English gerundive construction. These attempts seem to agree that one of the interesting aspects of this development lies in the acquisition of verbal characteristics in Middle English1 or Early Modern English. It seems to me, however, that most previous studies, including philological and linguistic ones, have not explored two problems. First, most philologists and linguists have focused their attention on the NP-internal structure of the nominal or verbal gerund, as illustrated in Figure l. 2 This biased attitude has led them to believe that the development of the verbal gerund was autonomous; that is, it occurred independently of its external contexts. By external I mean a whole-sentence context where the gerund appears (see Figure 1). Of course I do not mean to claim that it is logically impossible to attribute this change exclusively to the inherent factors of the gerundive construction. But the assumption of such an autonomous development does not seem to agree with the traditional account that the verbal properties of the gerund were borrowed from such verbals as present participles or infinitives (cf. Mustanoja (1960: 566-73)). In general we cannot gain a correct understanding of any syntactic change of a particular construction without taking into account its external as well as internal contexts. (For the significance of the external contextual effects on linguistic change in general, we shall see again in 5 (cf. fn. 3).) In the case of the gerund, the tangible change is apparently confined to the inside of NP structure containing -ing forms, but we must not overlook intangible contextual effects, if any, that might be relevant to the change in question. 3 A second serious flaw in the previous studies is their lack of semantic consideration. The traditional account for the rise of the verbal gerund is as follows: the suffix of the present participle, -inde, was morphologically and phonologically merged into that of the verbal substantive, -ing, with the


Osamu Koma





internal context/ of the nominal gerund external context of the nominal gerund





internal context/ of the nominal gerund external context of the nominal gerund

Figure 1. The internal and external context (or structure) of the nominal and verbal- gerund

result that the former gave the latter some verbal properties.4 In this case, however, the traditional account has not touched upon any semantic relation between these two non-finite forms. We have to admit here that this kind of account is the less persuasive for lack of semantic reasoning or motivation. When two syntactic forms influence each other and bring about a novel structure, I think it natural that some semantic aspects should be crucially involved in this process. Needless to say, semantic aspects of syntactic constructions of the earlier stage are especially difficult to approach. But we could say here that if we can find any clue to the

Verbal gerund and its historical development in English


semantics of the gerund and the present participle, it will prove to be a great step toward the verification of the traditional account given above. In the present paper I will tackle these two problems by resorting to a different angle of observation from that which previous studies have presupposed.

2. Aims and data In this section, keeping in mind the two problems raised above, I will present two aims of this study, and touch briefly upon the corpus on which the present study is based. 2.1. External structure The original -ing nominal in OE acquired the following verbal properties in the course of its development: (1)

a. b. c.

cooccurrence with adverbials taking the object without the intervention of of cooccurrence with passive or perfect morphemes (i.e., be-en or have-en)

In what follows I will concentrate on the second verbal property, (lb), and investigate those external contexts in which this property began to emerge in LME; to put it another way, I will explore the initial locus of the change from Type A or Β to Types C or D: (2)

Nominal gerund: Type A: the refusing of the offer Type B: refusing of the offer Verbal gerund:6 Type C: (the) refusing the offer Type D: (John's) refusing the offer

As is well known, LME is the period in which Types A and Β were prevalent, and in which the less prevalent Types C and D began to show initial sign of growth. What is most interesting to me is to see whether the


Osamu Koma

development from A and Β to C and D was dependent on some aspects of the external environments of the construction. The first aim of this paper is to clarify this point. Now let us refer briefly to the text to be examined in the present study. Among extensive corpus available, I have chosen the Paston Letters of the fifteenth century (henceforth, PL). PL has many linguistic advantages: homogeneity of materials, extensiveness of data, prose data, informal, educated style, etc. The selection of this text is also important from the chronological point of view. Thus, the period later than the fifteenth century is too late for us to identify the beginning stage of Types C and D, since in EModE these constructions were already extended into broader environments. In contrast, the period earlier than the fifteenth century seems too early for us to collect sufficient number of examples of the Type C and D (cf. fn. 5). Thus, both too many and too few examples prevent us from properly recognizing the critical stages of the development of the verbal gerund. To sum up, our first goal is to answer the following question. In which external contexts did the change from nominal to verbal gerund (i.e., from Type A or Β to Type C or D) begin to happen at the earlier stage? Suppose now that the answer to this question is Ex (a particular external context). 2.2. Semantic interface between two verbals After designating the Ex, we will have to consider the connection between the semantics of the present participle and that of the gerund used in the Ex. If there is such a connection, it will lend semantic support to the conventional hypothesis that syntactic properties transmitted from the present participle to the gerund, as a result of phonological and morphological merger of the suffixes of the two verbals that were originally distinct from each other. Thus, the second aim of this paper is to search for the semantic gerund-participle interface.

Verbal gerund and its historical development in English


3. Numerical distribution of verbal gerunds in PL 3.1. Biased


To begin with, let us show some of the typical environments where the verbal gerund in Present-day English can appear: (3)

a. b. c. d. e.

Subject: Watching television keeps them out of mischief. Direct Object: He enjoys playing practical jokes. Subject Complementation: Her first job had been selling computers. Adjectival Complementation: They are busy preparing a barbacue. Prepositional Complement: I'm responsible for drawing up the budget. [Quirk et al. 1985: 1063]

We shall next give the numerical distribution of Types C and D in PL corresponding to each of the five functions, (3a-e). Table 1. The occurrence of Types C and D in the five external contexts 3a

















Interestingly enough, Table 1 shows that the occurrences of Types C and D were limited exclusively to (3e), prepositional complements. Here it should be emphasized that it is possible to find many instances of Types A and Β in broader contexts, as illustrated below:7 (4)



Subject: my lord nay verily knowe that the complysshyng of pe seid appoyntement is nat deferred ner delayed by me. (547.124) Direct Object: ye may justyfye the kepyng of the plase for the pesybyll possessyon that ye haue had in it mor then iij yeer. (342.20-2)


Osamu Koma

c. d. e.

Subject Complement: the hasty purchace of mony and men schall be the getyng and rescu of it (243.47-8) Adjective Complementation: no instance8 Prepositional Complement: wherof the date is the day of makyng ofthiez presentz, (253.7-8)

Especially noticeable is the fact that we can find instances of Types A and Β not only in (3e) but also in (3a-c). One might think that the attested gap in (3a-d) of Table 1 can be partly due to the general tendency of Types A and Β to occur more frequently in (3e) than in (3a-d).9 Although some allowance must be made for this possibility, it remains to be seen why the occurrence of Types C and D shows such a distinct distributional bias toward a particular context, while A and Β can emerge in wider contexts. 3.2. Prepositional


Next let us examine those 46 instances in more detail. According to Quirk et al. (1985: 657), prepositional phrases have the following three functions: (5)


b. c.

Complementation (i) o f V : look at (ii) of A: afraid of Post-modifier in NP Adverbials

Table 2 below shows the numerical distribution of Types C and D in PL corresponding to each of the three functions: Table 2. The functions of the examples of Types C and D in the context of prepositional complement

Type C TypeD




0 1

1 6

2 32

The following is the only example of (5a-i) complementation where a prepositional phrase is closely related to the preceding verb ceace. (6)

we ... commanded yow ... to have ceased of makeinge any assemblye of our people (757.2-5)

Verbal gerund and its historical development

in English


There are 7 instances of (5b), post-modifier-in-NP. In these cases, it is uncertain whether there is any particular restriction on the nouns followed by the PPs containing Types C and D. (7)

a. b. c. d. e. *f.


as for the justificacyon of entryng the place sege layng to the same, (242.88-9) to fynde the meanes how I myght beste be spedde for the licence ofmorteisyng certein lyuelode, (569.34-6) I had no comffortable answere of spedyng the seid paymentes here, (604.22-3) all was oon thyng in substance of puttyng you in knolege of the Kyng your vncles deth, (646A.5-6) if he graunt my brodyr Edmund Clyppysbys son in recompence for takyng my brodyr Edmundys son, (381.55-7) for grete remorse I haue in my soule of the vntrewe forgyng and contryvyng certayne testamentys and last wyll by naked wordes ... (901.31-3)10 my maistyr was right well pleasyd wyth youre feithefull labour infulfellyng the patent for the warrd of A.B.C, (510.3-4)

Finally, our investigation reveals the highest frequency of (5c), adverbial PP. Eight (compound) prepositions are involved in 38 instances of this construction: after, as for, because of, for, in, of, on, and without. A complete list of the examples is given below. (8)




after (I instance) *a. with-in vj dayes aftire the so takyng aswell the names of the sayd prisoners as theire estate degre, or condicioun (396.54-5) as for (3 instances): a. as for callyng vppon the replevyn that the bestys of Drayton were delyueryd by. (183.36-7) b. as for ... yevyng the maner of Castre or hys londes in Norffolk to John Paston ... (895A.12-3) *c. as for the letyng me haue knowlage of the areragys of your lyuelod; 336.32. because of {one instance): a. because of castyng Bradwell and Tychewell yn the Kynges handes. (566.26-7)

Osamu Koma (iv)



for (5 instances): a. but as a drane amonges bees whech labour for gaderyng hony in the feldes (72.38-9) b. ye were not troubled ne vexid for lettyng the execucion of my seyd lordes wille (538.7-8) c. I haue aforetyme ben accused vnto the Kinges highnesse and the Quenes for owyng my pouere gode will and seruice vnto my lord of York and o^er, &c. (575.11-3) d. Other examples: (629.9-10) (626.1-2) in (25 instances): a. I neuir coud fele ner vndirstand hym poletyk ner diligent in helpyng hym-self, (72.37-8) b. Be ware how that ye spend it but in acquityng you ageyn such as ye be in daungere to, (208.11-2) c. Robert Lethum and his men assauted on John Coke of Witton, in brekyng vppe his dorys atte a xj of the cloke in the nyght; (48.80-2) d. I founde Herry Greye, Lomnour, ... ryght weele disposed to you ward at this tyme in helpyng and in gevyng ther goode avice to me for suche maters as I had to doo. (189.85-8) e. my nevew, whome I hartely beseche in executyng and perfomyng this my laste will to do and dispose ... (113.93-5) f. thynkyng you allys mete a man in executyng there comaundement as cowde be chosyn (242.44-5) g. I haue no worde from you of them, ner whethere ye haue yit in yowr kepyng the euydence of Est Bekham owt of hys handys, (243.65-6) h. Butt the Kyng ententyth in eschyewyng all jnconvenyentys to be as bygge as they bothe, (281.5-6) i. I beseke yow to be my good may sty r in purse wyng the seyd ateynte. (511.4-5) j. he my3t do grete ease, as yn disavowyng of it or yn wythdrawyng it owte of the bok. (559.4-5) k. Other examples: (92.22-3) (310.1-2) (572.24-5) (634.8-9) (686.14-7) (686.65-9) (700.5-8) (811A.1-4) (864.14-6) (881.4850) (893.9-10) (894.16-8) (901B.5-7) (901B.17-9) (912.48-50) ο/(one instance): a. James Cook ... hath accused and diffamed me and my wif of settyng vp billes agayn lordis (492.8-10)

Verbal gerund and its historical development in English




on (one instance): a. Ye nede at this tyme rather to have had three solicitours than in ony other terme past this iij yere, on concyderyng the maters hangyng, &c. (601.14-6) without (one instance): a. they treuly pay or agre for, and frendly entrete our seyd cousyns subiectys withowth eny robbyng or exstartyng them in there bodyes ner goodys, (824.14-6)

So far we have given all the individual examples of Types C and D in PL, and classified them from the functional point of view. The total picture of their numerical distribution is summarized in Table 3: Table 3. The total picture of the distribution of Types C and D subject


direct object


subject complement


adjectival complementation

0 complementation


post-modifier in NP prepositional complement

6 + 1* after as for

adverbial adjunct

1 2 + 1*

because of











36 + 2*

1 Total


7 + 1"

43 + 3*

* indicates Type C.

4. Discussion 4.1. Context-sensitive


The result of our detailed investigation of PL reveals many interesting phases of the development of the verbal gerund that previous studies have

162 OsamuKoma failed to observe. As can be readily seen from Table 3, the distribution of Types C and D are severely restricted. Especially important is the scarcity of occurrence in contexts other than prepositional complements. Moreover, the majority of the cases of prepositional complements exhibit an interesting distribution. Thus, they occur more frequently in adverbial PP than in post-modifier-in-NP or complementation. These facts seem to imply, among other things, that the initial stage of the development of the verbal gerund (that is, Types C and D) was more or less sensitive to its external environments. If Types A and Β had acquired verbal properties irrespective of their external contexts, then Types C and D could have occurred in all the positions in which the noun could usually appear. However, so long as we look into PL, the change in question seems to have taken place only in the limited contexts. That is one argument for regarding this change as context-sensitive. This context-sensitive hypothesis leads us to conjecture that the final (i.e., present-day) stage of this development illustrated in (3), was not reached abruptly. I am inclined to think that this development took at least two steps: first, the nominal gerund in the context of prepositional complement began to turn into the verbal gerund (this is what we call initial locus of change); second, the restructured verbal gerund in turn extended its distribution gradually from the initial locus toward other contexts such as (3a-d), and finally reached the present-day stage (this is the so-called extension). The exact nature and chronology of the second process are yet to be seen and out of our concern here, but we could safely assume for the moment that the first process evidenced in this study took place around in the fifteenth century, and the second extension proceeded through the period of EModE.11 4.2. Preposition in and its special status Another interesting implication that the result of our investigation has for the mechanism of the change under consideration is deeply connected with the special status of the preposition in. As is evident from Table 3, among various prepositional usages, the adverbial use of in shows the strikingly high frequency of cooccurence with Types C and D. This fact suggests that this adverbial in-phrase, rather than the prepositional complement, could be an initial locus, in the true sense of the term. In what follows, I will argue that this conspicuous behaviour of in is not simply accidental, and that the preposition in is, so to speak, a key to the development of the verbal gerund.

Verbal gerund and its historical development in English


Let us recall the question posed in 2.1: in which external contexts did the change from the -ing nominal to the verbal gerund begin to emerge at the earlier stage? Now we are in a position to answer this question. A most likely candidate for this answer, Ex (cf. 2.1), is an adverbial phrase introduced by in. Keeping in mind this prospective answer, let us turn to the second question raised in 2.2. The question was whether there is any overlap or similarity between the semantic function of the present participle in LME, on the one hand, and that of the verbal gerund of Types C and D appearing in the context Ex, on the other. If there is such semantic interface between the two, it will substantiate the conventional argument mentioned in 1 that the -ing nominal gained the verbal property from the present participle after the latter gained the morphological property from the former. To answer the above question, let us begin with reviewing the usages of the present participle in ME. We could name several types of constructions with the present participle: (9)

(i) (ii) (iii) (iv)

participial constructions progressive constructions participles following the so-called perception verbs manner participles following such verbs as go, come, lie, stand, etc.

Let us direct our special attention to the semantics of (9i), participial constructions in ME. It should be noted that one of their principal meanings was temporal (cf. Mustanoja 1960: 555-6);12 thus, the semantics of participial constructions is considered to be almost equivalent to that of subordinate clauses preceded by when or while. Interestingly enough, the temporal interpretation holds true of the verbal gerund appearing in Ex, that is, of the phrase in V-ing Object. As is obvious from the definitions of in in the OED and the MED, and from the interpretations of the examples adduced in (8v), this prepositional phrase also conveys the temporal meaning that could be paraphrased by when, while or in the course of. Therefore we might conclude that such semantic overlap between the participial construction and the m-phrase (exactly speaking, in V-ing of Object) enabled the latter construction to acquire the verbal property of taking the object directly, or to turn into in V-ing Object; the semantic factor facilitated the borrowing of the syntactic property from the present participle. This conclusion is consistent with the peculiar fact observed above that the novel verbal gerund (Ving Object) appeared primarily in the


Osamu Koma

restricted context of m-phrases. In this way, I have argued that the semantic similarity is a prerequisite for the initial development of the verbal gerund, although other factors might be relevant to their subsequent development. I hope I have shown that combined with the context-sensitive hypothesis proposed in 4.1, the above consideration as to in serves to semantically justify the traditional account that the present participle partly triggered the development of the verbal gerund.14

5. Concluding remarks: initial locus, reanalysis and extension In the present paper, I hope I have shown that one syntactic change that at first glance appears to have occurred within NP structure turns out to be crucially contingent on its external environments. More generally speaking, I have emphasized the significance of paying more attention to the external context of any linguistic change. If my argument that the development of the verbal gerund was subject to intangible contextual effects is on the right track, then it will shed new light upon the future research on the diachrony of various linguistic phenomena. To recapitulate the system of my explanation of linguistic change, it consists of three essential parts. First, identification of a particular context relevant to the change in question. This is what we call initial locus of change. In the case of the gerund, I claimed that it was prepositional complement (or, more particularly speaking, the complement of in). To identify such an intangible contextual effect is not so simple a task as might be supposed, but it is a subject eminently worthy of more careful study. Second, restructuring or reanalysis in the identified context. Combined with the first part, it constitutes the essential body of hypothesis. In the case of gerund, restructuring is thought to have created the verbal gerund from the nominal gerund. This process should accommodate not only the change in terminal strings (i.e., the insertion of of) but also that in the NPinternal phrase structure (cf. Koma 1980). Of course, to make it possible, further technical elaboration of restructuring will be necessary. Third, extension to other contexts. The newly restructured unit as a result of the second part is supposed to extend its distribution to other contexts than the identified initial locus. Although I have not pursued the detail of this extension process of the verbal gerund in this paper, it is obvious that the restructured verbal gerund extended later to other

Verbal gerund and its historical development

in English


contexts, say, subject, verbal object, predicate complement and so on (cf. fn. 11). I think that the extension is an indispensable part of my explanatory theory in that it constitutes the confirmation of the hypothesis of restructuring. Thus, as we observed in the development of the verbal gerund, some external contexts lagged behind the initial locus. This kind of time lag is the chief motive for my idea of two-step change — restructuring and extension. I believe that the theory of linguistic change presented above could apply to smaller linguistic units as well as larger ones. To make clearer the nature of this theory, let us take one example of smaller unit that the theory could account for: a well-known history of words like apron, umpire, nickname and so on. These words are well known as examples of the socalled metanalysis. Applying the idea of initial locus to these examples, we realize readily how intriguing the idea of metanalysis is. If we don't pay any attention to the external context, we, looking at only the tangible change, would think that the initial nasal sound of napron was dropped, and then a new word apron was coined. But if we looked at the external context of this change, we could reach the interesting theory fo metanalysis. Needless to mention, the initial locus of this change is the position preceded by indefinite article a/an. In this case, the restructuring is simpler than that of the verbal gerund, because it modifies only the constituent structure of sound sequence without any change in the linear order of the terminal string (thus, a napron>an apron). The story of apron probably ends with the extension; thus, the reanalyzed word apron extended its distribution to other contexts than the post-indefinite article position. In this case, I don't know whether the same kind of time lag that we observed in the verbal gerund could actually be confirmed between the initial locus and the extension. It should be noted here that the hypothesis of metanalysis can be confirmed only if such time lag is actually observed. So far we have seen that identification of the initial loci of particular changes could lead to reasonable explanations like metanalysis or reanalysis. But such identification does not necessarily lead to reasonable explanations of linguistic change. There are many cases of linguistic change where although the initial loci and the extension are identified, it remains to be seen why the initial loci are such and such. Thus, Jones (1967), which is concerned with the disappearance of grammatical gender system, suggests that in the transitional period from Late OE to Early ME, an ending -ne, which had originally been a marker of singular/masculine/accusative, encroached upon those cases of NP whose head noun was feminine or neuter. The point here is that this change


Osamu Koma

occurred not uniformly in all NPs, but under a particular contextual condition; thus, the change took place in the context of V_, but not in P_. Apart from the minute differences between my study and Jones', the two share the following angle of observation: an innovation that apparently took place within an NP proved to be constrained by the environments outside the NP. In Jones' case I do not understand why the initial locus is V_, or what it implies. However, I believe that even such findings of the initial loci could be a preliminary step toward a more interesting theory of linguistic change.

Notes *

This is a revised version of the paper "On the initial locus of syntactic change: verbal gerund and its historical development" published in English Linguistics 4: 311-24 in 1987. I am very grateful to Katsuhide Sonoda, Kozo Kato, Masayuki Ohkado, Shinji Uchioke for their invaluable comments and suggestions.

1. Donner (1986) makes the claim that the verbal gerund was not established during the ME period. His conclusion is based upon his own examination of all the citations of the -ing - nominal in the volumes from A through Ο of the Middle English Dictionary. According to his report, some fifteen thousand of them are cited in the MED under some two thousand separate entries. Although his main concern lies in the chronological aspects of the verbal gerund, my present concern is the mechanism of its establishment. Apart from such difference in attitude, there appears to be a significant limitation to his study. That is, his data is based exclusively on the citations of the MED so that he fails to collect enough examples of the verbal gerund. He does not even mention the language of the Paston Letters, where we can, as I will show later in this paper, find a large enough number of examples of the verbal gerund. It is highly questionable whether such use of the MED is really effective is searching for particular constructions like verbal gerunds. Donner mentions Regional Pecock's frequent use of verbal gerund, but he considers it exceptional and attributes it to Pecock's idiosyncratic style. I do not know for the moment whether it is really Pecock's idiosyncratic style as he claims. But at least I could say here that if Donner were to include the Paston Letters in his data, he may have very well changed his conclusion. 2. Among the philological studies whose concern is mainly the internal structure of the gerund, is Tajima (1985). It presents detailed usages of the nominal- and verbal- gerund and other related constructions in ME. This study and others dealing with particular literary works or authors provide us with much information about the chronology of the gerund.

Verbal gerund and its historical development in English


While philologists have accumulated factual data, some linguists have attempted to describe the development of the gerund in terms of generative grammar. Thus, Koma (1980) argued that the acquisition of verbal properties can be described as the innovation of phrase structure rules within the permissible bounds of the X-bar theory proposed by Jackendoff (1977). But this theoretical study also deals only with the NPintemal structure of the gerund. 3. This is one of the most important methodological points that diachronic linguists always have to keep in mind. If syntactic forms of particular categories (such as Ss, NPs, VPs, APs, etc) undergo any changes, we must not exclude, in advance, any possibility that their external contexts, outside the particular categories, might be a crucial condition for the changes. To make this point clear, let us take a typical example. It is generally believed that the word order change from SOV to SVO occurred predominantly in main clauses at the earlier period, and later was extended to subordinate clauses (Lightfoot 1979: 125-8). Speaking from the viewpoint of generative grammar, this change took place within the same category S, since both subordinate and main clauses are members of the S. But taking into account wider surroundings outside the S, we can technically distinguish and capture the notion of embedded-S and root-S. Only by so doing, can we consider such interesting diachronic topics as initial locus or extension of syntactic change. 4. Infinitives, along with present participles, are supposed to have been partly responsible for this change. Cf. Mustanoja (1960: 572). 5. Emonds (1973) also fails to get over these two problems, although his consideration as to the effects of present participles on gerunds is stimulative and suggestive. Incidentally, it seems to me that his failure stems partly from the selection of data from Chaucerian English. For the problem of selecting data, see 2.1. 6. The term verbal gerund usually refers to Type D alone. But in this paper we shall use it to refer to both Types C (the so-called mixed gerund, a marginal construction) and D. Although the definite article in the mixed gerund apparently exhibits a nominal character, it is not our primary concern. As was said above, attention will be focused exclusively upon the verbal property of taking the object without of. I believe nothing crucial hinges on this decision. 7. All the citations given below are from Norman Davis' edition. References are made to letter numbers and lines in that order. 8. Notice here that there is no instance of Types A and Β in the context of adjectival complement. It is not surprising in the light of the fact that even in Present-day English such instances are severely restricted to a few adjectives, such as busy, near, and worth. 9. To elucidate this point, it might be necessary to show exact statistics of Types A and B. For the sake of convenience, however, I have only shown here that the nominal gerund was actually used in all the contexts except (4d). So long as there is no proof against this factual observation, nothing will invalidate the argument presented later.

16 8

Osamu Koma

10. An asterisk preceding an example indicates Type C. 11. For the description of this development in EModE, see Inazumi (1973-5). His detailed investigation of Roger Ascham's works also presents numerical distributions of various kinds of -ing forms, the results of which seem to be by and large consistent with this assumption. 12. Of course, in ME there were other adverbial usages meaning reason, concession, or attendant circumstance. 13. One might think that the development of the so-called be on Ving construction, which originates in OE, is relevant to the change in question. This construction, which was rare in OE, was replaced by be in Ving in ME, and this in turn developed into be α-Ving in EModE. But it was not productive in ME, either (cf. Denison 1993: 387-8). Furthermore, most examples, like "a bird is on flying", seem to lack direct objects. In light of these two facts, I could say here that there are not enough compelling reasons to regard it as a possible cause for the development of the verbal gerund. But I do not mean to claim that it played no role in the development of the gerund. Interestingly enough, the following examples are found in ME: i)

j?e yomen of Schordych, f>at f?ere were in amendyng of here berseles the yeomen of Shoreditch that there were in repairing of their archery-butts (Doc. Brewer in Bk. Lond. E. 234.19) 'the yeoman of Shoreditch who were repairing their archery targets there'


and therwashe kepyng of coort {Paston Letters 338/18) and there was he holding of a court-session [Denison 1993: 388 & 401]

These examples clearly indicate that there was some confusion between the progressive form and the gerund in ME. Therefore, in pursuing a more comprehensive history of the gerund, we will have to take this construction into account. 14. One has to discuss the adequacy of the traditional view that the infinitive, in addition to the present participle, also played an important role in causing the change under analysis (cf. fn. 4). Our analysis based on the semantic interface between the present participle and the in-phrase cannot explain those examples of Types C and D which are preceded by for, such as (8iv), since we cannot find such interface between the present participle and the /or-phrase. One of the basic meanings of the for phrase in LME was probably 'purpose' or 'reason'. To my knowledge, no historical evidence has yet been available to show that the participial construction was used as a purpose adverbial. However, the infinitive, unlike the present participle, is qualified as an adverbial of purpose, and at the same time, has the VP-like internal structure. Consequently, it is not unlikely that the infinitive was also involved in the evolution of the verbal gerund. We will have to examine this possibility in more detail.

Verbal gerund and its historical development in English


15. I am grateful to Kozo Kato for pointing out to me that the idea of two-step change presented here bears a certain similarity to Warner's (1982: 150) account of the extension of what he calls (NP to VP) constructions [the so-called non-Equi infinitival complement/O.K.], which suggests that the construction is introduced in contexts where it is syntactically least noticeable or least salient. However, I have no idea how the notion of syntactic saliency can be defined or how the notion can be applied to the case of the gerund.

References Davis, Norman 1971, 1976 Paston

letters and papers

of the fifteenth century,

Part I-ll.


Clarendon Press. Denison, David 1993

English historical syntax. London: Longman.

Donner, Morton 1986

"The gerund in Middle English", English Studies 67.5: 394-400.

Emonds, Joseph 1973

"The derived nominals, gerunds, and participles in Chaucer's English", in: Braj B. Kachru et al. (eds.), 185-198.

Inazumi, Kaneaki 1973-1975 "Roger Ascham no gengo" [The language of Roger Ascham], Bulletin of Toyama University. Jackendoff, Ray 1977

X-bar syntax. Cambridge, Mass.: MIT Press.

Jones, Charles 1967

"The grammatical category of the gender in Early Middle English", English Studies 48: 289-305.

Kachru, Braj B. et al. (eds.) 1973

Issues in linguistics. Urbana: University of Illinois Press.

Koma, Osamu 1980

"Diachronic syntax of the gerund in English and the X-bar theory", Studies in English Literature (English Number) 1980: 59-76.


"On the initial locus syntactic change: verbal gerund and its historical development", English Linguistics 4: 311-324.

Lightfoot, David 1977

Principles of diachronic syntax. Cambridge: Cambridge University Press.

Mustanoja, Tauno 1960

A Middle English syntax. Helsinki: Societd Neophilologique.


Osamu Koma

Quirk, Randolph - Sidney Greenbaum - Geoffrey Leech - Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman. Tajima, Matsuji 1985 The syntactic development of the gerund in Middle English. Tokyo: Nan'undo. Visser, Frederikus Theodoras 1966 An historical syntax of the English language, Part II. Leiden: E.J. Brill. Warner, Anthony 1982 Complementation in Middle English and the methodology of historical syntax. London: Croom Helm.

What is the point? Manuscript punctuation as evidence for linguistic change* Tadao Kubouchi And there a poynt, for ended is my tale. God sende every trewe man boote of his bale! 'And there is the end; for my take is finished. May God send every honest man the remedy of his suffering!' - Geoffrey Chaucer, The Canon's Yeoman's Tale, 1480-81. The point or punctus was not only used to mark the end of a tale in the medieval English punctuation system. It was also used in many other ways.1 The main uses the punctus is pre-late-tenth-century English manuscripts might be characterised as those developed in the system of distinctiones, a system which was of Greek origin and based on a three-fold division of a sententia 'sentence' by punctus placed at heights of three different grades to indicate the nature of different pauses. According to Parkes (1992: 13), a low point ( . the subdistinctio) was used to indicate a minor medial pause after a comma, that is, where the sense (sensus) is incomplete; a medial point ( · the media distinctio) indicated a major medial pause after a colon, that is, where the sense is complete but the meaning (sententia) is not; and a high point ( * the distinctio) indicated a final pause after a periodus, that is, where the sententia is completed. This system, as we might easily imagine, was difficult to carry out accurately in a minuscule script and it was finally abandoned. It was the system of positurae that was introduced to serve the need which the system of distinctiones failed to fulfil. The positurae served as more than a mere substitute; they were of continental origin and "all derived from the system of ecphonetic notation which originally indicated the appropriate melodic formula to be used in the liturgy" (Parkes 1978: 140). Therefore it is far from a mere coincidence that they began to appear in liturgical manuscripts in England in the late tenth century.


Tadao Kubouchi

The punctus circumflexus ( 7 (on the Continent) or . (in England)) denoted a lowering of the voice at the end of a comma. the punctus elevatus ( . or its variant form I ) an elevation of the voice at the end of a colon, the punctus versus ( j ) a lowering of the voice at the end of a periodus or sententia containing a statement, and the punctus interrogativus ( . ) the inflection of an interrogative sentence at the close of a periodus or sententia containing a question. It is not a far cry from here to punctuation symbols. The forms of the positurae in themselves are noteworthy. The symbols were all formed by combining a neume with a punctum (punctus) as the base.2 The punctus circumflexus was derived in England from a punctum (a low point) only, on the Continent from a punctum plus a clivis ( J ) , the punctus elevatus from a punctum plus a podatus ( ) , the punctus versus from a punctum plus an ancus (clivis) (\/ ), and the punctus interrogativus from a punctum plus a porrectus

The melodic formula contours necessarily and at the ends of phrasal/clausal units, as their punctum (punctus)-based forms suggest. The positurae thus indicated not only "the appropriate melodic formula" but also a pause and therefore a rhythmical and syntactical break which it is the primary function of punctuation to mark. Thus the basic punctuation symbols used in English manuscripts of the eleventh-thirteenth centuries are: the simple point, the punctus elevatus, the punctus versus, and the punctus interrogativus? Capitals became important as litterae notabiliores at the beginnings of sententiae, when the minuscule script became a rule. The virgule (virgula suspensiva) ( / ) was a new addition about the end of the thirteenth century and its value was roughly the same as that of the punctus after a comma, but sometimes it coalesces with the punctus elevatus. The punctus circumflexus does not seem to have been used commonly in manuscripts of other than liturgical texts. The most common mark of punctuation was the punctus or point. Scribes often used it to indicate all kinds of pauses. It might safely be said

What is the point?


that it was only when some emphatic or marked force was required that the punctus elevatus or punctus versus was called upon. The punctus versus was sometimes used to end questions. Passage (1) below is from 'the Passion of St Stephen' in iElfric's First Series of Catholic Homilies (hereafter Stephen) as it is contained in MS. London, British Library, Royal 7 C. xii (ff. 14v-19v); passage (2) is from Wulfstan's Sermo Lupi ad Anglos (hereafter Sermo) as it is contained in MS. London, British Library, Cotton Nero A. i (ff. 1 ltf-l 15r): (1)

Hwa maeig beon rihtlice geciged mannes beam buton criste anum . Jjonne aelc mann is tweigra manna beam buton him anum . 'Who can rightly be called the Son of man, save Christ only, when every man besides him is the son of two persons?' - f. 1672-4 (Thorpe 1844, i: 48/23-25; 49/22-24)4


Hy her V aö 7 hy baernaö · rypa{) · 7 reafiad · 7 to scipe laedaö ; 7 la hwaet is aenig oöer on eallum J)am gelimpum · butan godes yrre · ofer Jjas jjeode · swutol · 7 gesaene j 'They ravage and they bum, plunder and rob, and carry away on board; and indeed, what else is there inall these events but the wrath of God clear and visible towards this nation?' - f . 113Γ/14-18 (Bethurum 1957: XX, 272/126-128; Whitelock 1976 [1939]: 60/129-132; Swanton 1975: 120/13-16)

While in (1) the question beginning with an interrogative pronoun hwa ends with the punctus interrogativus, the similar question in (2) is ended by the punctus versus. This might mean that the punctuation system as it is in Cotton Nero Sermo put more stress on the indication of the melodic formula contour and the pause than on that of grammatical or syntactical function. Thus the point was used to indicate a minor medial pause, or, when placed before a capital, to indicate a final pause, or, when combined with another two (or three) more dots to form a 'triple period' ((I), ( > ) , (..,), (j.), (,j) or (.;.))» to indicate the end of a chapter or a tale. When it was combined with neumes, it was able to indicate the 'appropriate (liturgical)


Tadao Kubouchi

melodic formulae' (where demanded) as well as the nature of different pauses. It was also used to mark the beginning and the end of quotations, to separate words (e.g., .ae. 'law', .äa. 'always', .iiii. 'four'), and to mark abbreviations (e.g., L. 'Leofan men' (CCCC 201, 36/9, 14, 26), .&. (CCCC 419, 95/2)). The punctus placed between we and soplice in the manuscript text of Beowulf 273 in MS. London, British Library, Cotton Vitellius A.xv (f. 137720) is used to divide the one from the other word (Cf. also f. 16779). Other special uses of the point include: the gemipunctus, i.e., 'two full points placed horizontally on the base-line immediately before titles of dignity or office' (Hector 1958: 46), the diacritic point, i.e., a dot to distinguish u (v) and y from η and ν or thorn respectively, and the expunction point, i.e., a dot placed below a letter that the scribe wished to delete. There are other functions of the medieval point we could indicate. For one thing, we could investigate the prose rhythm in the period concerned through observation of the manuscript punctuation. Mcintosh (1949) is an example, and Kubouchi (1983) is an attempt to review the claim made in Mcintosh, and to review the result from the social historical point of view taking it as linguistic evidence revealing linguistic variation in the period of JElfnc and Wulfstan. The rather heavy punctuation that is a distinctive feature of manuscripts associated with Wulfstan might be better explainable, when we take into consideration the audience and the intention that Wulfstan had. In what follows I would like to show that the pointing in elevenththirteenth century manuscripts could give us clues to the changing process of a word or element order whose usage showed a sharp decline during the transition period between late OE and early ME, particularly between iElfric/Wulfstan and Ancrene Wisse/Sawles Warde. I have in mind the element order with the pre-verbal noun object. I hope that the following discussion will encourage manuscript studies in the field of historical syntax and that manuscript text and manuscript punctuation studies will hereafter be included in the discussion of evidence for linguistic change.5 The following examples (3)-(5) are from Part lof the Ancrene Wisse (hereafter AW), an early thirteenth-century West Midland religious prose work, which has been shown to indicate one of the last phases of the syntactical change here under discussion.6 The oldest extant manuscript of AW, whose original manuscript is now lost, are: MS. Corpus Christi College, Cambridge, 402 (MS. A), MS. London, British Library, Cotton Cleopatra C. vi (MS. C), and MS. London, British Library, Cotton Nero A. xiv (MS. N). The production dates given to them are c. 1225-1230, c. 1230,

What is the point?


and c. 1225-1250, respectively. The modern English translations below are from White (1993). Italics are mine. (3) (A)

(C) (N) (4) (A)

(C) (N)

(5) (Α)

f. 6r/2-5 Euchan segge hire ures as ha haueö iwriten ham . & euch tide sunderliche ase forö as 3e mähen seggeö in his time . ear to sone J>en to leate . 3ef 3e ne mahen eauer halde j?e time. 'Let everyone say her hours as she has written them down and say each hour separately, as far as you can at its own time - rather too soon than too late, if you ever cannot keep to the time.' f. 1174-5 3ef 3e/>e \ihte' time Kne mahen' halden f. 5713-14 3if 3ε ne muwen euer holden pe rihte time. f. 6722 Efter euensong anan ower placebo euche niht seggeö hwen 3ε beoö eise . 'After Evensong, say your "Placebo" at once each night, when you are able.' F. 11726 Efter euensong ^ 1 anan Placebo . vhche nicht seggeö V 3ef 3e beoö aise . f. 5730 Efter euesong anonriht siggeö ower placebo eueriche niht hwon 3ε beoö eise . f. 6v/22 Seoue salmes & pus fjeose fiftene seggeö abutεn under . ['Say the Seven Psalms and the Fifteen in this way at about the third hour of the day, i.e., 9 a.m.' \ /

(C) (N)

f. 12710 Seoue salmes & fiftene 1 seggeö abuten vnder. f. 5728 seoue psalmes & teos fiftene psalmes siggeö abuten vndern dei8s .

In example (3), while MSS. A and Ν employ the modem SVO order, MS. C holds on to the archaic S.noun-O.V. (hereafter SOnV) order. On the SOV element order in Parts 6 and 7 of AW, I concluded in my 1975 article entitled "Word-Order in the AW (1975: 28) that "white the εΐεπιεηί order S.pron.-O.V. (herεaftεΓ SOprV) is still found not only in dependent clauses but also in independent ones, the ο^εΓ SonV is found only in dεpεndεnt clauses" and that "ενεη in coordinate clauses after ant and ah [OE ac 'but'] a noun object never prec8des the verb". Rihte and ne mahen in MS. C are interlinear additions made by scribe B. The fact that Scribe B, accepting scribe A's basic eteirent οΜεΓ, confincs himself to adding to the


Tadao Kubouchi

text written by scribe A (i.e., the original scribe of MS. C) only such clarifying words for the intended sense might be taken to show that the above conclusions based on Parts 6 and 7 hold good here as well. Although the element order was archaic, which accounts for the order alteration in the Corpus and Nero texts, it was still not unacceptable in dependent clauses. In 116 examples of dependent clauses with the noun object in Parts 6 and 7, there are at least five examples of the order SOV (4%), although the other 101 examples appear all in the SVO order (Kubouchi 1975: 24-26). Scribe Β is the main reviser of MS. C, whose revisions, according to Dobson (1962: 161-163; 1966: 195, 201-202; 1972: xi, xciii), were incorporated in the text in the revision of which MS. A is a fair copy, and also the scribe whom Dobson takes to be identical with the original author of AW. Scribe B's alterations, especially but not only with regard to element or word order, might therefore be of particular interest from the historical-syntactical point of view. However, he is not alone in providing clues with regard to element order usage in the period here concerned. Not only those deliberate alterations and innovations but also unconscious but telltale slips of the pen such as are found in every related manuscript might be said to provide evidence revealing the state of the word or element order in question in the early thirteenth century. "In ignoring the context of scribal responses", says B.A. Windeatt (1979: 120), "in which medieval texts are preserved to us in the manuscripts, the modern reader may waste a valuable resource." What Windeatt says about Chaucerian manuscripts might be applicable to our manuscripts. Among the scribal responses which we tend to ignore is manuscript punctuation that is in most editions replaced by the editorial and therefore modern punctuation. Those responses might certainly give witnesses revealing the linguistic situation in the transition period. Mitchell (1980) stresses the dangers of syntactical studies based on medieval texts in modern punctuation. Examples (4) and (5) are imperative sentences. In Parts 6 and 7 the order V(0/C) is the rule in such sentences (63 out of 64 examples). The OV order appears only once, but with the pre-verbal pronoun object ((A) 96718 Ba t>eos bihald in me q9 dauid '"Both these behold in me", said David.') (Kubouchi 1975: 23). Out of 74 imperative clauses with the noun object in Part 1 of AW as found in MS. A, there are some ten instances of the OV order with the pre-verbal noun object, among which are the examples above (examples (4) and (5)). Their presence as we see it in the case of MS. A contradicts the claim of 'modernity' for AW. But when we turn to the texts of MSS. C and N, a different picture comes out. In example (4), C

What is the point?


(scribe A) adds a punctus after the noun object Placebo, and N, putting the noun object after the verb sigged, alters the OnV order to the modern Von order. In example (5), while Ν leaves the basic element order unaltered, C (scribe B) adds a punctus elevatus after the noun object seoue salmes & fiftene. On the punctuation of medieval prose texts, Parkes (1978: 138-139) says, "Medieval scribes and correctors punctuate where confusion is likely to arise and do not always punctuate where confusion is not likely to arise.... Elements which may have a similar syntactic function or convey similar meaning, and which are punctuated in one context, need not be punctuated in another when the context ensures that confusion is not likely to arise", and he suggests that "the key to the understanding of medieval punctuation lies not in grammatical theory, nor in the analysis of syntactical or intonation patterns, but in the concern of the scribe or corrector to elucidate the text transmitted to him according to the needs of his own audience."8 This well explains our instances: wherever marks of punctuation are used, the pointing shows where the scribe feared confusion was likely to arise. Witness the pointing after the pre-verbal noun object in examples 4(C) and 5(C) above. It might be said that the scribe felt that confusion might arise there if the reader were to read the OnV construction as it stands. We therefore might be allowed to see there linguistic evidence of the linguistic change in process, or of its terminus ante quem. Statistical studies with reference to the change in that period of the element order in question show that it took place during the period between iElfric/Wulfstan and AW/Sawles Warded Now for the terminus post quem of the change we turn to JElfric and Wulfstan. Mitchell (1964: 124) quotes two instances of the element order from Stephen: Thorpe 50/17: Da reöan Iudei wedende J>one halgan staendon; and Thorpe 50/30: Stephanus soölice gebigedum cneowum Drihten baed £>aet ... Apart from one or two manuscripts which provide unclear readings, e.g., MS. Cambridge, Corpus Christi College, 302 or MS. London, British Library, Cotton Vitellius D. xvii, all the manuscripts containing this part show no sign of this pointing, that is, a pointing after a pre-verbal noun object. The date which N.R. Ker gives in his Catalogue to the manuscripts ranges between s. x/xi and s. xii2. It is of more than ordinary interest to be able to see in what are called 'Worcester manuscripts', which are manuscripts written, kept or corrected in Worcester,10 quite a different facet of pointing. Example (6) is from Wulfstan's DeFalsis Dies [sic], MS. Oxford, Bodleian Library, Hatton 113


Tadao Kubouchi

(MS. T) and example (7) from one of the manuscripts preserving jElfric's original version De Falsis Diis [sic] as is contained in MS. Oxford, Bodleian Library, Hatton 116 (MS. S): (6)

T, f. 58715 (Bethurum ΧΠ/13) {>urh done ealdan deofol · J>e adam' iu aer beswac 'by the old Devil who had long before betrayed Adam'


S, p. 371/10 (Pope XXI/132) for {>an {>e he gefeoht: lufode . 'because he loved battle.'

MS. Τ was written in the eleventh century (third quarter), and MS. S in the first half of the twelfth century. We might possibly say, for the time being, that the point of time that the production date of these two manuscripts suggests could indicate that our terminus post quem might be somewhere near this point. It is also notable that these examples are dependent clauses. Other examples also show that the proposed terminus post quem might not be far from the real one. They further show that this practice was not restricted to 'Worcester manuscripts.' The following examples (8)-(10) are from what we generally call 'jElfric's First Old English Pastoral Letter for Wulfstan' whose full text survives in three manuscripts, i.e., Corpus Christi College, Cambridge, 190 (MS. O; s. xi med.) and 201 (MS. D; s. xi med.), and Oxford, Bodleian Library, Bodley 343 (MS. Oz; s. xii 2 ): 12 (8)

Fehr 12 (O 321/21-22, Oz 13372, D 31/40) Ο 7 ästealde cristendom · 7 claennysse taehte . Oz 7 astealde cristendöm · 7 claennesse taehte . D 7 astealde cristendom on[d] claennesse · silf bisnode 7 taehte · 'and established Christianity and taught chastity [making a model of himself]'


Fehr 37 (O 324/15-16, Oz 134716, D 33/27) Ο 7 hi ealle |)a lare {>e we leorniaö on bocum awriton 7 gesetton Oz 7 alle 3eo lare öe we on bocum awritaen leorniaeö · 7 13

3esetten D 7 hi ealle J)a lare J)e we leorniaö on bocum · awriton 7 gesetton · 'and they wrote and appointed [by God's direction] all the teaching which we learn in books.'

What is the point?



Fehr 40 (O 324/22, Oz 134r /19-20, D 33/33) Ο \>aet hi h(eo)ra aehta ealle beceapedon · Oz \)&t heo heorae aehta alle beceapodon · D Jiaet hi heora aehta · ealle gesealdon · '[so...] that they sold all their possessions.'

In examples (8) and (9) which are of non-dependent clauses beginning with a coordinate conjunction14 and in example (10) which is of a dependent clause, the pre-verbal noun objects are punctuated in MS. D. MS. D is dated in the mid-eleventh century and is one of the manuscripts of Wulfstan's so-called 'Commonplace Book'. The version of the text of the Pastoral Letter that occurs in MS 201 has been extensively revised by Wulfstan. This manuscript is not of Worcester provenance.15 It is hoped, although it is certain that more work has to be done, that these findings amply show that the pointing in 11th-13th century manuscripts could give us clues to the process of a change in element order, i.e., the element order with the pre-verbal noun object. I hope, as I have said before, that the above discussion will encourage manuscript studies in the field of historical syntax and that manuscript text and manuscript punctuation studies will hereafter be included in the discussion of evidence for linguistic change.

Notes *

My thanks are due to the Centre for Medieval English Studies, College of Arts and Sciences, University of Tokyo, whose foundings members include Professor Yoshihiko Ikegami and which houses the microfilms of those manuscripts which I consulted, although at the final stage I checked all the manuscripts at the original libraries which include the Parker Library, Corpus Christi College, Cambridge, the Bodleian Library, and the British Library, to which I am grateful and to whose staffs I would like to express my gratitude for their warm encouragement and kind assistance. Special thanks are due to Dr Bruce Mitchell, Professor Fred C. Robinson, Dr John Scahill, Professor Malcolm Godden, Mr T. C. Graham, Dr Margaret Laing, Professor Jacek Fisiak, Professor Akio Oizumi, Professor Antonette diPaolo Healey and all the staff at the DOE, University of Toronto, and Professor R. I. Page and all the members of the Research Group on Manuscript Evidence at the Parker Library, Corpus Christi College, Cambridge, who helped me at various stages and in various ways. My special thanks are also due to Dr Malcolm Parkes and the late Professor Kikuo Miyabe. As the title shows, which is partly borrowed from that of Dr Parkes's seminar 'What is the point? Medieval


Tadao Kubouchi

punctuation', held at Oxford in the Hilary term of the academic year 1981-1982, I am much indebted to his lectures, works, and kind advice, arid it was Professor Miyabe who first made me interested in manuscript punctuation as in many other topics. 1. I owe the accounts of the following four paragraphs especially to Clemoes (1952), Ker (1960), Brown (1974), Parkes (1978) and Parkes (1992). Also cf. Petti (1977). 2. See Clemoes (1952), Caldwell (1978) and Bischoff (1990). 3. For other symbols see Parkes (1992). Ogura (1988: 94, fn. 1) mentions a hybrid between a punctus elevatus and a punctus versus found in LK 8.49, Jtt 3.10 and 4.7 as in MS. Oxford, Bodleian Library, Bodley 441. Those for verse texts are omitted. 4. References are to folios (recto or verso) or pages and lines of the manuscript; in the case of printed editions or modern translations references are to pages and lines of the edition. All line references are in principle to the line in which the finite verb in the element or word order in question occurs. [ ] indicates words and letters which have been deleted by the scribe by means of crossing out, erasure, or expunction; " an addition to the manuscript text. 5. Rauch (1990) of course does not include manuscript punctuation studies in her discussion. 6. It has been statistically demonstrated in such works as Kubouchi (1975), Kohonen (1978), Kubouchi ([1981], 1982) that AW/Sawles Warde stand at the terminus ad quem in terms of the period in which the sharp decline of the element order with a pre-verbal noun object (SOnV) took place. The discussion in the present article shows that this is well supported by manuscript evidence. 7. I have discussed this more fully Kubouchi (1995). 8. This situation might be better understandable, when we recall the punctuation in Milton's Paradise Lost. One of the reasons for its inconsistency is the fact that the proofs had to be read aloud to Milton and that he intervened only when the reader got it wrong, e.g., by mistaking a preterite for a past participle such as 'advanced'. Cf. Treip (1969), Wright (1957) and Mitchell (1980: 184). I am grateful to Dr Bruce Mitchell for drawing my attention to this fact and giving me references related to it. 9.

See note 6 and Jucker (1990). Kubouchi (1995) is an answer to Jucker (1990).

10. See Franzen (1991) and Mclntyre (1978). 11. The examples are from Kubouchi (1985) and (1988). 12. The examples are from Kubouchi "The Decline of the SOV Element Order: The Evidence

from Punctuation





jElfric and


(forthcoming). 13. In the Bodley text Caroline 'g' is almost regularly used for the OE voiced guttural stop and Insular 'g' (3) for the OE voiced palatal spirant [j] (Irvine 1993: lx, lxix). Bodley 343's orthography

deserves a close study; it provides valuable evidence for

phonological values of the late West Saxon spellings.

What is the point?


14. On the influence of initial ond, ac and ne see Mitchell (1964: 117-120). Non-dependent clauses beginning with a coordinate conjunction are labeled as of type C in Kubouchi (1975: 13). 15. For the provenance of CCCC 201 see Whitbread (1959), Whitelock (1976 4 : 2) and Bishop (1971: xv, fn. 2).

References 1) Manuscripts Manuscripts containing 'The Passion of St. Stephen' in yElfric's First Series of Catholic Homilies, particularly Thorpe 50/17 and 50/30: Cambridge, University Library, Gg. 3.28, f. 1 l v (Ker 15 art. 5; s. x/xi) London, British Library, Royal 7 C. xii, f. 17r (Ker 257 art. 3; s. χ ex.) London, British Library, Cotton Vitellius C. v, f. 24r (Ker 220 art. 5; s. x/xi, x 1 ) Cambridge, Corpus Christi College, 188, pp. 27, 28 (Ker 43 art. 3; s. xi1) Oxford, Bodleian Library, Bodley 340, f. 8 r (Ker 309 art. 2; s. xi in.-xi med.) Oxford, Bodleian Library, Hatton 113, f. 127™ (Ker 331 art. 34; s. xi (3rd quarter)) Cambridge, Corpus Christi College, 198, ff. 9V, 10r (Ker 48 art. 2;s. xi1, xi2) Cambridge, Corpus Christi College, 302, pp. 53, 54 (Ker 56 art. 7; s. xi/xii) London, British Library, Cotton Vitellius D. xvii, f. 30™ (Ker 222 art. 11; s. xi med.) Cambridge, University Library, Ii. 1.33, f. 122v (Ker 18 art. 21; s. xii2) Wulfstan's Sermo Lupi ad Anglos: London, British Library, Cotton Nero A. i, ff. 110 r -l 15r (Ker 164 art. 20; s. xi in.) ^Elfric's De Falsis Diis[sic] and Wulfstan's De Falsis Dies[ sic]: T: Oxford, Bodlein Library, Hatton 113, ff. 58v-61 (Ker 331 art. 16; s. xi (3rd quarter) S: Oxford, Bodleian Library, Hatton 116, pp. 365-373 (Ker 333 art. 21; s. xii1) '/Elfric's first Old English pastoral letter for Wulfstan' O: Cambridge, Corpus Christi College 190, pp. 320-336 (Ker 45B art. 2; s. xi med.) Oz: Oxford, Bodleian Library, Bodley 343, ff. 133r-137r (Ker 310 art. 67; s. xii2) D: Cambridge, Corpus Christi College 201, pp. 31-40 (Ker 49B art. 17; s. xi med.) Ancrene Wisse: A: Cambridge, Corpus Christi College, 402, ff. 1-117 v (ca. 1230) C: London, British Library, Cotton Cleopatra C. vi, ff. 4-199 (ca. 1225-1230) N: London, British Library, Cotton Nero A. xiv, ff. 1-120 v (ca. 1225-1250) 2) Printed and facsimile editions: Ackerman, Robert W. - Roger Dahood (eds.) 1984

Ancrene Riwle: Introduction and Part I. (Medieval and Renaissance Texts and Studies 31.) Binghamton, N.Y.: CEMERS, SUNY - Binghamton.


Tadao Kubouchi

Bethurum, Dorothy (ed.) 1957 The homilies of Wulfstan. Oxford: Clarendon Press. [ 1971 ] [Reprinted with corrections.] Clemoes, Peter A.M. 1956

jElfric's 'Catholic homilies', First Series: The text and manuscript tradition. 2 vols. [Unpublished Ph.D. dissertation, University of Cambridge.] Day, Mabel (ed.) 1952 The English text of the Ancrene Riwle edited from Cotton MS. Nero Α. XIV, on the basis of a transcript by J. A. Herbert. (EETS, os 225.) London: OUP. Dobson, Eric John (ed.) 1972 The English text of the Ancrene Riwle edited from B.M. Cotton MS. Cleopatra C. VI. (EETS, os 267.) London: OUP. Eliason, Norman E. - Peter A. M. Clemoes (eds.) 1966 AZlfric 's first series of Catholic homilies: British Museum Royal 7 C. XII, fols. 4-218. (EEMF 13). Copenhagen: Rosenkilde and Bagger. Fehr, Bernhard (ed.) 1914 Die Hirtenbriefe /Elfrics in altenglischer und lateinischer Fassung. [Reprinted 1966 with a Supplement to the Introduction by Peter Clemoes. (Bibliothek der angelsächsischen Prosa. 9.) Darmstadt: Wissenschaftliche Buchgesellschaft.] Irvine, Susan 1993 Old English homilies from MS Bodley 343. (EETS, os 302.) Oxford: OUP. Lyon, Henry R. (ed.) 1971 A Wulfstan manuscript containing institutes, laws and homilies: British Museum Cotton Nero A. I. (EEMF 17.) Copenhagen: Rosenkilde and Bagger. Morton, James (ed.) 1852-53 The Ancrene Riwle: A treatise on the rules and duties of monastic life. (Camden Society, os 57.) [Repr. 1968, New York: AMS Press.] Napier, Arthur Sampson (ed.) 1967 [1883] Wulfstan: Sammulung der ihm zugeschriebenen Homilien nebst Untersuchungen über ihre Echtheit, mit einem bibliographischen Anhang von Klaus Ostheeren. Dublin/Zürich: Weidmann. Pope, John Collins (ed.) 1967-68 Homilies of /Elfric: A supplementary collection. 2 vols. {EETS, os 259 and 260.) London: OUP. Shepherd, Geoffrey (ed.) 1959 Ancrene Wisse: Parts six and seven. London: Nelson. [1985] [Revised edition. (Exeter Medieval English Texts.) Exeter: University of Exeter.]

What is the point?


Thorpe, Benjamin (ed.) 1844-46 The homilies of the Anglo-Saxon Church. 2 vols. London: £lfric Society. Tolkien, John Ronald Reuel (ed.) 1962 The English text of the Ancrene Riwle: Ancrene Wisse edited from MS. Corpus Christi College Cambridge 402, with an introduction by N.R. Ker. (EETS, os 249.) London: OUP. Whitelock, Dorothy (ed.) 1939 [1976]

Sermo Lupi ad Anglos. (Methuen's Old English Library.) London: Methuen. [Revised edition (4th edition). Exeter: University of Exeter.]

3) Translations Ackermann, Robert W. - Roger Dahood (eds.) 1984 (See above) Salu, Mary B. 1955

The Ancrene Riwle (The Corpus MS.: Ancrene Wisse). (The Orchard Books.) London: Burns & Oates. [1990] [Reprint. (Exeter Medieval English Texts and Studies.) Exeter: University of Exeter Press.] Swanton, Michael James 1975 Anglo-Saxon prose. (Everyman's University Library.) London: Dent / Totowa, N.J.: Rowman and Littlefield. White, Hugh R.B. 1993

Ancrene Wisse: Guide for anchoresses. (Penguin Classics.) London: Penguin Books. Whitelock, Dorothy, et al. 1981 Councils & synods with other documents relating to the English Church. Vol. I: A.D. 871-1204. 2 parts. Oxford: Clarendon Press. 4) Catalogues and secondary sources Anderson, George K. 1941 "Notes on the language of jElfric's English Pastoral Letters in Corpus Christi College 190 and Bodleian Junius 121", JEGP 40: 5-13. Apel, Willi 1958 Gregorian chant. Bloomington / Indianapolis: Indiana University Press. Baba, Akira et al. (eds.) 1995 Essays in linguistics and philology presented to Professor Kinsuke Hasegawa on the occasion of his sixtieth birthday. Tokyo: Kenkyusha. Berglen, Angelika 1994 "On the historical background of English punctuation", Zeitschrift für Anglistik und Amerikanistik 42: 243-250.


Tadao Kubouchi

Bischoff, Bernhard 1990 Latin paleography: Antiquity and the Middle Ages. Translated by Däibhi 0 Cröinfn and David Ganz. Cambridge: Cambridge UP. Bishop, Terence A. M. 1971 English Caroline minuscule. (Oxford Palaeographical Handbooks.) Oxford: Clarendon Press. Brown, Thomas Julian 1974 "Punctuation", The New Encyclopaedia Bntannica. (15th edition.) London and New York: Encyclopaedia Bntannica Company. Macropaedia, Vol. 15: 274-277. [1985]

[The New Encyclopaedia Britannica. Revised edition (16th edition.) London and New York: Encyclopaedia Britannica Company. Vol. 25: 1006-8.] Caldwell, John 1978 Medieval music. London: Hutchinson. Cameron, Angus 1973

"A list of Old English texts", in Roberta Frank - Angus Cameron (eds.), 25306. Clemoes, Peter A. M. 1952


The Liturgical Influence on Punctuation in Late Old English and Early Middle English Manuscripts. (Occasional Papers: No. 1.) Cambridge: Printed for the Department of Anglo-Saxon. [Reprinted as Old English Newsletter Subsidia 4. Binghamton, N.Y.: CEMERE, SUNY - Binghamton.]

Davis, Norman - Charles L. Wrenn (eds.) 1962

English and medieval studies presented to J.R.R. Tolkien. London: Allen and Unwin. Dobson, Eric John 1962 "The Affiliations of the Manuscripts of Ancrene Wisse", in: Norman Davis C.L. Wrenn (eds.), 128-163. 1966 "The Date and Composition of Ancrene Wisse", Proceedings of the British Academy 52: 181-208. Fisiak, Jacek (ed.) 1990 Historical linguistics and philology. (Trends in Linguistics, Studies and Monographs 46). Berlin - New York: Mouton de Gruyter. Frank, Roberta - Angus Cameron (eds.) 1973 A plan for the Dictionary of Old English. Toronto: University of Toronto Press. Franzen, Christine 1991 The tremulous hand of Worcester: A study of Old English in the thirteenth century. Oxford: Clarendon Press.

What is the point?


Fujii, Takako 1992 A study of the word order in The homilies of Wulfstan [Unpublished MA Thesis, University of Tokyo.] Gerritsen, Marinel 1990

"The relationship between punctuation and syntax in Middle Dutch", in: Jacek Fisiak (ed.), 187-225.

Gneuss, Helmut 1981 "A preliminary list of manuscripts written or owned in England up to 1100", Anglo-Saxon England 9: 1-60. 1985

"Liturgical books in Anglo-Saxon England and their Old English terminology", in: Michael Lapidge - Helmut Gneuss (eds.), 91-141. Gradon, Pamela 1983 "Punctuation in a Middle English sermon", in: Eric G. Stanley - Douglas Gray (eds.), 39-48. Gruber, L.C. - R.P. Tripp, Jr. (eds.) 1982 In geardagum IV: Essays on Old English language and literature. Denver, Co: The Society for New Language Study. Handley, Rima 1974 "British Museum MS. Cotton Vespasian D. xiv", Notes and Queries, ns 21, 7: 243-250. Harlow, Christopher Geoffrey 1955 The punctuation of Six of /Elfric's catholic homilies in the earliest manuscripts [Unpublished B. Litt, thesis, University of Oxford]. 1959 "Punctuation in some manuscripts of jElfric', RES, ns 10: 1-19. Hector, Leonard Charles 1966 [1958]77ie handwriting of English documents. London: Edward Arnold. Hicks, Corola (ed.) 1992 England in the eleventh century: Proceedings of the 1990 Harlaxton Symposium. (Harlaxton Medieval Studies 2.) Stamford: Paul Watkins. Hill, Joyce 1992

"Monastic reform and the Secular Church: /Elfric's pastoral letters in context", in: Carola Hicks (ed.), 103-117.

Jost, Karl 1950 Wulfstanstudien. (Schweizer anglistische Arbeiten 23.) Bern: Francke. Jucker, Andreas H. 1990 "Word order changes in Early Middle English: Some evidence against the conservatism of subordinate clauses", Studia Anglica Posnaniensia 23: 31-42. 1991 "Between hypotaxis and parataxis. Clauses of reason in Ancrene Wisse", in: Dieter Kastovsky (ed.), 203-220.


Tadao Kubouchi

Kastovsky, Dieter (ed.) 1991

Historical English syntax. (Topics in English Linguistics 2.) Berlin - New York: Mouton de Gruyter.

Ker, Neil Ripley 1957 Catalogue of manuscripts containing Anglo-Saxon. Oxford: Clarendon Press. [ 1990] [Re-i ssued with a Supplement.] 1960 English manuscripts in the century after the Norman Conquest [The Lyell Lectures 1952-3]. Oxford: Clarendon Press. 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. (Meddelanden frän Stifteisens for Äbo Akademi Forskningsinstitut 38.) Äbo: The Research Institute of the Äbo Akademi Foundation.

Kubouchi, Tadao 1975 "Word-order in the Ancrene Wisse", Hitotsubashi Journal of Arts and Sciences 16,1: 11-28. 1982 "Word order in Richard Rolle's English epistles", in: L.C. Gruber - R.P. Tripp, Jr. (eds.), 19-31. 1983

"A note on prose rhythm in Wulfstan's De Falsis Dies [sic]", Poetica 15/16: 57-106.


"Manuscript punctuation, prose rhythm and S... V element order in Late Old English orally-delivered prose", in: Kinshiro Oshitari et al. (eds.), 71-87.


'Texts of 'Be Cynestole' in Wulfstan's Institutes of Polity", in: Takashi Suzuki - Tsuyoshi Mukai (eds.), 211-217. 1995 "Word order in the Ancrene Wisse revisited", in: Akira Baba et al. (eds.), 573581. Lapidge, Michael - Helmut Gneuss (eds.) 1985 Learning and literature in Anglo-Saxon England: Studies presented to Peter Clemoes. Cambridge: CUP. Lucas, Peter J. 1971 "Sense-units and the use of punctuation-markers in John Capgrave's Chronicle", Archivum Linguisticum, ns 2: 1-24. Mcintosh, Angus 1949 "Wulfstan's prose", Proceedings of the British Academy 35: 109-142. [1990] [Reprinted in: E. G. Stanley (ed.), British Academy Papers on Anglo-Saxon England. Oxford: OUP (Published for the British Academy), 111-144.] Mclntyre, E.A. 1978 Early-twelfth century Worcester Cathedral Priory with special reference to the manuscripts written there [Unpublished D. Phil, thesis, University of Oxford].

What is the point?


Mitchell, Bruce 1964 "Syntax and word-order in the Peterborough Chronicle 1122-1154", NM 65: 113-144. [1990] 1980 [1990]

[Reprinted in: Bruce Mitchell, On Old English: Selected papers. Oxford: Blackwell, 221-242.] "The dangers of disguise: Old English texts in modern punctuation", RES, ns 31:385-413. [Reprinted in: Bruce Mitchell, On Old English: Selected papers. Oxford: Blackwell, 172-202.]

1985 Old English syntax. 2 vols. Oxford: Clarendon Press. Morrison, Stephen 1995 "A reminiscence of Wulfstan in the twelfth century", NM 96: 229-234. Murphy, James J. (ed.) 1978 Medieval eloquence: Studies in the theory of medieval rhetoric. Berkeley: University of California Press. Ogawa, Hiroshi 1989 "Revised syntax in Wulfstan's rewritings of jElfric's prose", in: [Aspects of English linguistics: to the memory of Professor Saburo Ohye], Fukuoka: Kyushu Univ Press, 3-17. 1991

"Syntactic alterations in Wulfstan's rewritings of jElfric: An annotated list", Bulletin of the Department of Foreign Languages, College of Arts and Sciences, Univ of Tokyo 1991: 83-107.

Ogura, Michiko 1988 "Direct or indirect? - pact as a quotation indicator", in: Kinshiro Oshitari et al. (eds.) 88-105. Oshitari, Kinshiro et al. (eds.) 1988 Philologica Anglica: Essays presented to Professor Yoshio Terasawa on the occasion of his sixtieth birthday. Tokyo: Kenkyusha. Parkes, Malcolm Beckwith 1978 "Punctuation, or pause and effect", in: James J. Murphy (ed.), 127-142. 1992 Pause and effect: An introduction to the history of punctuation in the West. London: Scolar Press. 1994 "Punctuation and the Medieval history of texts", in: Atti dei Convegni Lincei 111. Convegno Intemazionale sul tema: La Filologia Testuale e le Scienze Umane organizzato in collaborazione con VAssociazione Intemazionale per gli Studi di Lingua e Letteratura Italiana (Roma, 19-22 aprile 1993). Roma: Accademia Nazionale dei Lincei, 265-277. Petti, Anthony G. 1977 English literary hands from Chaucer to Dryden. London: Edward Arnold.


Tadao Kubouchi

Polom6, Edgar C. (ed.) 1990

Research guide on language change. (Trends in Linguistics: Studies and Monographs 48.) Berlin - New York: Mouton de Gruyter.

Rankin, Susan 1985 "The liturgical background of the Old English advent lyrics: A reappraisal", in: Michael Lapidge - Helmut Gneuss (eds.), 317-340. Rauch, Irmengard 1990 "Evidence of Language Change", in: Edgar C. Ροΐοηιέ (ed.), 37-70. Robinson, Fred C. 1973 "Syntactical glosses in Latin manuscripts of Anglo-Saxon provenance", Speculum 48: 443-475. Scragg, Donald George 1979 "The corpus of vernacular homilies and prose saints' lives before jElfric", Anglo-Saxon England 8: 223-277. Stanley, Eric G. - Douglas Gray (eds.) 1983 Five hundred years of words and sounds: A Festschrift for Eric Dobson. Cambridge: D.S. Brewer. Suzuki, Takashi - Tsuyoshi Mukai (eds.) 1993 Arthurian and other studies presented to Shunichi Noguchi. Cambridge: D.S. Brewer, Treip, Mindele 1969 Milton's punctuation Methuen.

and changing English usage 1582-1676. London:

Whitbread, L. 1959 "MS. C.C.C.C. 201: A note on its characteristic and provenance", Philological Quarterly 38: 106-112. Wilcox, Jonathan 1992 "The dissemination of Wulfstan's homilies: The Wulfstan tradition in eleventh-century vernacular preaching", in: Carola Hicks (ed.), 199-217. Willard, Rudolph 1950

"The punctuation and capitalization of jElfric's homily for the first Sunday in Lent", University of Texas Studies in English 29: 1-32. Windeatt, Barry A. 1979 "The scribes as Chaucer's early critics", Studies in the Age of Chaucer 1: 119141.

Finite and non-finite clauses in the English of Alfred's reign: A study of syntax and style in Old English* Kazumi Manabe In the study of syntactic and stylistic features of Old English prose insufficient attention has hitherto been paid to some of the more significant factors, such as the genre of the prose in question, the 'source conditioning' from the Latin original, the translator's attitude in rendering the original text, the purpose for which the texts are written, the nature of the audience, etc. The purpose of the present paper is to explore Alfredian finite and nonfinite clauses, paying due attention to these factors. The corpus consists of eight texts, four of which have already been examined in the analysis of these types of clause in my previous book (Manabe 1979: 9-23). The present study concentrated therefore on the clauses in the remaining four texts, the editions of which are the following: Dialoge Gregorys des Grossen (GD), ed. Η. Hecht (Darmstadt: Wissenschaftliche Buchgesellschaft, 1965). The first 1000 lines of prose in Book ΙΠ are examined. King Alfred's Orosius (Or), ed. H. Sweet. EETS OS 79. The selected portions comprise the first 1000 lines of prose. King Alfred's Old English version ofBoethius (Bo), ed. W. J. Sedgefield (Darmstadt: Wissenschaftliche Buchgesellschaft, 1968). The first 1000 lines of prose are examined. The laws of the earliest English Kings (Laws), ed. F. L. Attenborough (New York: Russell & Russell, 1963): 'The Laws of Ine,' 'The Laws of Alfred,' 'The Treaty of Alfred and Guthrum,' 'The Laws of Edward,' Ί Aethelstan,' and the first 3 lines of 'Π Aethelstan'. The selected portions comprise 1000 lines of prose. In the course of the discussion, however, I will also refer to the texts analyzed in my previous work, i.e., Bede's Ecclesiastical history (Bede, 1018 lines of prose), Gregory's Pastoral care (CP, 1026 lines of prose), The Parker chronicle (Parker, 974 lines of prose), and The meters of Boethius (Met, 1750 lines of verse).


Kazumi Manabe

The analysis in this study is basically the same as that in my previous work (Manabe 1979: 2-4; Svartvik - Quirk 1970: 395-7), so that the following categories were examined: subject; complement; object, without a subject of the non-finite verb (-S), with a subject of the non-finite verb (+S); adjunct, with non-purposive function (-P), with purposive function (+P); adjective modification. In analysing finite clauses, only dependent finite clauses with subordinators were considered.1 As for non-finite clauses, I examined the frequency and distribution of clauses containing infinitives (Inf), past participles (Ved), present participles (Ving) and verbal nouns in -ing (-ung) (VNing), while excluding certain types of clauses.2 Both finite and non-finite clauses are far less frequent in the historical works, except in Bede. This is in particular true of Parker, as will be discussed in detail later. Or begins with an enumeration of the division of the world and a geographical introduction. The technique for describing the geographical structure of an area is fairly simple: the territories, seas, mountain ranges by which each country is bounded are enumerated, starting from the east and moving clockwise. This procedure leads to a large amount of repetition and simplicity of expression (Deloletz 1971: 255) with few of the finite and non-finite clauses in questions: (1)

Asia ongen 9aem middeldaele on J)aem eastende, J>aer ligeö se muj>a ut on Jjone garsecg J>aere ie J>e mon hateö Gandis. Pone garsecg mon haet Indisc. BesuJ)an J>aem muj)an wi9 Jjone garscecg is se port J>e mon haet Caligardamana, 7 be su|)aneasan {jaem porte is Jjaet igland Deprodane. 7 {>onne be norJ>an Jjaem Gandes mu£>an, J>aer J>aer Caucasis se beorg endad neh f>aem garsecge, J>aer is se port Samera. Be norJ>an {)aem porte is se mu£>a t>aere ie J>e mon nemneö Ottrogorre. I>one garsecg mon haet Sericus. (Or, 10)

The selected portion from Or also comprises the account of the adventures of two mariners, Ohthere and Wulfstan, where characteristic linguistic features are parataxis, a looseness of sentence structure (Duncan 1972: 26), and correlation rather than subordination (Potter 1939: 48-9). The narrative of Ohthere and Wulfstan display, for instance, on overwhelming for />a-forms (Liggins 1970: 292; Enkvist 1972: 94):



t» CO

co "t

CS co




"fr η co





Γwo «Λ






ο CS

CO uo


r* NO





























ο CO




CS 00






47 13.5%













cs CO























_aet. Mid öy he J)urf syndrige noman öeara stafa aefter öaem biscope cuaeö, J>a heht he se biscop him syllabas 7 word forecwedan (Bede, 388)

Bede may be a historical work in a broad sense of the term, but it should be noted that it differs greatly from Or in the distribution and frequency of the clauses (Bately 1970: 435). It is also worth noting that Bede contains a significantly higher proportion of non-finite to finite clauses than Alfred's works (CP, Bo, Met). In this respect, Bede differs greatly from the known works of Alfred. This could be one of the grounds on which Alfred's authorship of Bede might be disputed (Whitelock 1962: 77; Draat 1916: 346; Kuhn 1972b: 206; Kuhn 1972a: 180). On the other hand, Bede and GD are rather similar in the proportion of non-finite to finite clauses (Duckett 1957: 138), though GD contains a slightly higher proportion. This may partly due to the similarity of their subject matter. It is true that many historical events are scattered throughout Bede with detailed information on the early church and early kingdoms in England. However, Bede also includes saints' lives: most of the unhistorical matter in Bede belongs to hagiography (Baugh 1967: 16) which may be said to be characteristic of CD's genre. In fact, a number of holy persons and their miracles are dealt with in GD in the style of the conventional medieval hagiographers (Bonner 1973: 85; Wrenn 1967: 213). The proportion of non-finite to finite clauses is the highest in GD of all the eight texts examined: it is rather similar to that in the early Middle


Kazumi Manabe

English corpus (24.7% : 75.3% in GD against 27.6% : 72.4% in EME). Furthermore, GD has the widest functional range of the clauses in question. It should be mentioned in this connection that we can observe a certain similarity of the distribution of the clauses and of the proportion of non-finite to finite clauses between Bede and GD. As mentioned above, the resemblance may be attributable to the similarity of their subject matter (Duncan 1972: 27-8). Infinitives as objects, for instance, abound in GD as in Bede. The ratio of infinitives to finite clauses in this category rises to 5 : 7 in GD, while it is 1 : 3 in the Alfredian corpus as a whole (EME; almost 1 : l): 6 (10) (11)


J)aet he forseah to donne, J)aet heo hine baed, ac eac forho3ode hit to yehyrenne. (GD, 180); J)a o n j a n h e f a r a n to Constantinopolim J)acre byri3, 7 he becom to cozenthi £>aere mse3Öe. E»a on3an he secan him swa rum hus (GD, 184); he eode ut of J)am 3ebedhuse zemette eenne beran standan beforan J)am durum. (GD, 206).

GD contains as many as thirty-five instances of present participles as adjuncts, which are also fairly frequent in Bede (twelve instances): their strikingly high frequency in GD makes present participles predominate greatly over past participles in this category in the Alfredian corpus (4 : 1 against 3 : 1 in EME). It is worth nothing that more than half (twenty) of the thirty-five instances correspond to the same idioms in the Latin original. Clearly, 'source conditioning' should be taken into account, though the translation of GD shows less literalness than that of Bede (Harting 1937: 283; Potter 1939: 48; Duckett 1957: 106; Matsunami 1958: 161-80): (13)


hi J)a ferdon butu aetsomne to Affricalande, 7 {>a forö yinyndum jsaes cynin3es aöume, J)e hire sunu haefde (GD, 180); *Latin: Perrexere igitur utrique ad Africam. Procedenti autem ... regis genero, qui ejus filium habebat 7 J)a J)a hine his ma3as wepende to him brohton, se drihtnes wer 3eorne acsode, hwaejjer hi haefdon J)one 3eleafan, J>aet he mihte beon 3ehaeled. (GD, 184); *Latin: Cumque hunc propinqui illius flentes obtulissent, vir Domini sollicite requivit an curationis illius haberent fidem.

Finite and non-finite clauses in the English of Alfred's reign




J)a andswarode se arwyrjja biscop Sabinus J>am deacone £>us cwepende: '3a 7 cwej) to J)aere ea: Reblin öu {jyses flodes 3 cyrr to öinum a3enum rihtryne!' Pa se deacon jjaet 3ehyrde, J>a hloh he 7 hine forseah. Pa se drihtnes wer him to 3ehet his writere 3 him dihtode {JUS cwedende (GD, 193); •Latin:... venerabilis vitae Sabinus episcopus respondit, dicens: Vade, et dice ei: Mandat tibi episcopus ut te compescas, et ad proprium alveum redeas. Quod diaconus ejus audiens, despexitet irrisit. Tunc vir Dei, accersito notario, dicta vit, dicens 7 J>a saetiende hi ofslo3on {jone ylcan beran. (GD, 206). •Latin: eumdem ursum insidiantes occiderunt.

In addition to the resemblance in the clause structure mentioned above, it is generally accepted that GD and Bede share a great deal of diction (Whitelock 1966: 77) and they both demonstrate the tendency to use tautological pairs of words. This has led several scholars to remark that the resemblance between Bede and GD is in some respects a striking one (Draat 1916: 320). At the same time, it is also true that the style of GD does show some individual quality. For instance, finite clauses as objects are far less frequent in GD than in Bede, and on the other hand, there appear in GD verbal nouns in subject, complement and object, while there is not a single instance of them in Bede: (17) (18) (19)

J>a 3elamp hit in Jjaere nihte, J)a J)a seo maeste stilnes 7 swiying waes (GD, 184); hit ma waeere bysmrun-5 |>onne efensar3un3. (GD, 180); he manna earduny of J)am huse adrife. (GD, 184).

The pleasing colloquial simplicity with few clauses in conversational parts is quite peculiar to GD (Potter 1940: 117; Wrenn 1967: 212-3), and no demands are made on the reader's intellect there: (20)

361 'sot> is» b®1 t>ysum men 3ehyrdesr; 7 ic {) nu sec3e deo3ollice, J>aet ic J>e aer ne 3esaede. witodlice nu toniht on swefne ic 3eseah rice deman on heahsetle sittan on3aen me, betwyh t>am ic 3eseah £>isne ylcan sittan, 7 £>aet anweald mines rices, J)e ic 3efyrn onfen3, jiurh heora dom hit me is of 3enumen. as acsa f>u, hwaet he sy; witodlice ne wene ic ne



KazMtni Manabe l>ysne wer swa mycelre 3eearnun3e swa swij>e beon [risse worulde man, swa swa we wenaö.' (GD, 181); 'on saelum, la broöru, on saelum! hwaet! 3e wel habbaö 3eworht 7 3ewunnen. blinnaö nu sume hwile.' (GD, 202).

Finite clauses have the highest frequency in CP of all the texts examined. It should be noted in this connection that Alfred's purpose is strictly utilitarian (Brown 1970: 11) and he concerns himself not with rhetorical elegance, but with writing as clearly as possible not only for ecclesiastical rulers but also for secular rulers (Whitelock 1966: 79; Potter 1940: 117-9; Duckett 1957: 132). With the exception of some Scriptural passages that Gregory quotes (Brown 1969: 669), Alfred translates the Cura Pastoralis for the most part by paraphrasing, and his usual practice is dissolving long Latin sentence into combinations of clause in an expansive style (Brown 1970: 19). The high frequency of finite clauses in CP may be accounted for by reference to this method of translation: (22)

Se lareow sceal mid geornfullice ingehygde foreöencean na öaet an öaette he öurh hine nan woh ne bodige, ac eac öaet he nane öinga dset ryht to suiöe & to ungemetlice & to unaberendlice ne bodige, foröaem oft öaet maegen öaere lare wierö forloren, öonne mon mid ungedafenlicre & unwaerlicre oferspraece öa heortan & öaet andgiet edweleö öara öe öaerto hlystaö, ond eac se lariow biö gescinded mid öaere oferspraece, öonne he ne conn geöencean hu he nyttwyrölicost laeran maege öa öe öaerto hlystan willaö. (CP, 95)

In particular, finite clauses as adjuncts (+P) abound in CP: their ratio to infinitives is 15 : 1 in CP, against 4 : 1 in the Alfredian corpus as a whole: (23)

Daette we swa lufian öisne uterran & öisne eorölican fultum öaette we foröaem from öaere wilnunga & from öaere geornfulnesse öaere godcundan lufan ure mod ne awendan, öylaes us weoröe to wope & to elöiodignesse öaes ecean lifes öaet öaet us on öisse elöiodignesse to fultume & to are gelaened is (CP, 389)

Finite clauses as complements have the highest frequency in CP. However, their ratio to infinitives is lower in CP than to overall ratio in the Alfredian corpus ( 1 : 2 against 3 : 5 ) . This is due to the repetitive use in CP of

Finite and non-finite clauses in the English of Alfred's reign


predicative infinitives with the equative verb denoting necessity or obligation: CP is designed to instruct the clergy in performing their duties (Schlauch 1956: 85). (24)

On oöre wisan sint to manianne öa öe öisse worulde lotwrenceas cunno, & dalufigead, on oöre öa medwisan. Da lytegan sint to mannianne öaet hi oferhycggen öaet hie öaer wieton, öa sarwisan sint to manianne öaet hie wielnien to wietanne öaet öaet hie nyton. (CP, 203)

The use of present participles as adjuncts is also frequent in CP (twelve instances). Many of them, however, do not correspond to the Latin models (Manabe 1979: 19; Brown 1970: 12-4): (25)

gedafiende he hit forbireö for öam dome his geöylde. (CP, 27) •Latin: quia quos permittendo tolerat, profecto per judicium reprobationis ignorat.

As mentioned above, CP is written in an expansive style, "as if by someone who understood readers have difficulty in following terse and compressed statements" (Whitelock 1966: 79). Thus, the brief sentence in the original can be expanded into a long sentence containing almost three times as many words (Brown 1969: 674), and it is quite common for one word to be expanded into an entire clause (Brown 1969: 671). Again, Alfred in one instance reworks three Latin clauses into six in the English, as W. H. Brown points out 91969: 671). This may account for the highest frequency of the clauses in CP of all the texts in the corpus. Drowsiness and awkwardness may then result when he packs a complicated Latin period into a single sentence with overextended concatenation of clauses (Brown 1970: 16-7). As is generally accepted, the rendering in Bo is very free (Whitelock 1966: 76); each stage of Boethius's highly subordinated argument is turned into an isolated statement (Payne 1968: 143), while in CP Alfred packs a complicated Latin period into a single sentence heavy with oversubordination. He seems to be content to write so as to be understood by the people, for his audience was not clerics and leaned men, but common men who were struggling to see the light despite the lack of educational opportunity (Donaghey 1964: 44-6). Hence, he makes no attempt to imitate the artificial and involved periods of Boethius (Sedgefield 1900: xxi): in place of the syllogism Alfred uses the single sentence, the saying, the


Kazumi Manabe

maxim (Payne 1968: 138). Thus, finite clauses and infinitives are far less frequent than in CP which is written in an expansive style. This is in particular true of conversational parts: (26)

Hit t>a andwyrde η cwaej): Ic wat f> ic on libbendum men 7 on gesceadwisum eom 7 feah on deadlicum. Da andwyrde se Wisdom 7 cwaeö: Wastu auht ofres bi f e selfum to secganne buton £ t>u nu saedest? Da cwaeö f) Mod: Nat ic nauht oöres. Da cwaeö se Wisdom: Nu ic haebbe ongiten fine Ormodnesse, nu öe self nast hwaet f u self eart, ac ic wat hu fin man tilian sceal. (Bo, 13)

Finite clauses as objects, however, abound in Philosophy's dialogue, which makes them fairly frequent in Bo: (27)

Gif f u fonne heora feawas witan wilt, fonne meaht f u ongietan p hie ne beod nanu men getreowe. Be faem f u meaht ongietan past pu freer nane myrhde on naefdest öa f a f u hie haefdest, ne eft nane ne forlure f a öa f u hi forlure. Ic wende p ic pe geo gelsered haefde f> f u hi oncnawan cuöe, 7 ic wisse p pu hi onscune dest öa öa J>u hi haefdest, öeah f u hiora bruce. Ic wisse past du mine cwidas wid hiora willan oft saedest, ac ic wat p nan gewuna ne maeg nanum men bion onwended f) f) mod ne sy be sumu daele onstyred. (Bo, 15)

Finite and non-finite clauses are about equally frequent in Bo and Met. The total proportion of non-finite to finite clauses is also rather similar in both texts. However, some noteworthy differences in their distribution appear when each of the categories is examined separately. For instance, all types of clause as subject are far less frequent in Bo than in Met: only one instance of verbal noun occur against three in Met: (28)

öe waes fios hwearfung betere foröaem öe öissa woruldsaelöa to wel ne lyste, 7 faet öu J>e eac betre na gelefde. (Bo, 18)

No examples of past participles as adjuncts appear in Bo while Met contains four. On the other hand, finite clauses as objects are much more frequent in Bo than in Met, as I suggested earlier. Moreover, Bo contains seven finite clauses as complements while there is not a single instance in Met:

Finite and non-finite clauses in the English of Alfred's reign

(29) (30)


öaet is p ic begringe eadmodnesse on heofonum (Bo, 18); Eall hire god ic J>e maeg mid feam wordum areccan; t>aet is daet hio is on ealu peawu hiere feeder gelic. (Bo, 22).

There are in Bo six infinitives in the category adjective modification, whereas there is only one instance in Met: (31)

I>yncaö J>e nu swiöe diore 7 swiöe leofe öa ding öa 9e nawöer ne sint ne getrewe to habbanne, ne eac ieöe to forlaetannel {Bo, 16)

More than twice as many present participles occur in Bo: (32)

Da lioö J)e ic wrecca geo lustbaerlice song ic sceal nu heofiende singan, 7 mid swij)e ungeradu wordu gesettan, J>eah ic geo hwilu gecoplice funde; ac ic nu wepende 7 gisciende ofgeradra worda misfo. (Bo, 8)

Thus a comparison of Bo (prose) and Met (verse) reveales a significant difference in the distribution of the clauses in these two texts (Anderson 1962: 279). Laws contains by far the highest proportion of finite to non-finite clauses of all the texts examined. 90% of the finite clauses are, however, clauses as adjuncts (-P). In terms of sentence structure, the great majority of legal sentences there have an underlying formulaic structure which says something like 'if X, and (then) Ζ shall do (be) Y' (Bethurum 1932: 2678). Thus, 76% of the finite clauses as adjuncts (-P) are clauses introduced by gif (Carlton 1970: 161): (33)

Gif he hie oferweorpe 7 mid ne gehaeme, mid χ scill. gebete. Gif he mid gehaeme, mid LX scill. gebete. Gif oöer mon mid hire laege aer, sie be healfum daem Sonne sio bot. Gif hie mon teo, geladiege hie be sixtegum hida, oööe öolige be healfre J)aere bote. Gif borenran wifmen 3is gelimpe, weaxe sio bot be öam were. (Laws, 70)

As is obvious from these passages cited at random, the sentences in Laws consist almost without exception of the main statement with few subordinates (apart from 'if X' clauses repeated monotonously).


Kazumi Manabe

In addition, speculative statements about indefinite and unknown persons are frequently used, whose future actions are approved or condemned in the independent clause; they are in most cases introduced by se pe: (34)


Se öe hlo|)e betygen sie, geswicne se hine be cxx hida oööe swa bete. Se 9e hereteama betygen sie, he hine be his wergilde aliese oööe be his were geswicne. (Laws, 40); Se öe dearnenga beam gestrieneö 7 gehileö, nah se his deaöes wer, ac his hlaford 7 se cyning. Se J>e öeof gefehö, he ah χ scill., 7 se cyning öone öeof; 7 J>a maegas him swerian aöas unfaehöa. {Laws, 44).

Statements of penalties and compensations incurred for various offences are thus very brief and concise, and are often elliptical to the point of confusion. It is quite natural then that the use of the finite and non-finite clauses in question is very scarce in Laws except for finite clauses introduced by gif. The percentage of non-finite to finite clauses is very small in the Alfredian corpus (14.1% : 85.9% against 27.6% : 72.4% in EME), and in fact the actual use of non-finite clauses is very limited. However, a separate examination of each text gives clear evidence that there is a significant difference in the distribution of the clauses in each of the eight texts according to the difference of the genre or the subject matter of the text. Furthermore, the distribution of the clauses in question can often be greatly influenced by 'source conditioning' from the original, the attitude of the translator in rendering the original text, the purpose for which the texts are written, and the nature of the audience. It is evident from the present analysis of Alfredian finite and non-finite clauses that an adequate consideration of the genre of the text as well as the above-mentioned factors is essential in syntactic and stylistic studies in Old English.

Notes *

This is a revised version of the paper originally published in Studies in English Language and Literature 38 (Kyushu University, 1988).

1. Finite clauses introduced by subordinators other than that are disregarded, except in their adjunctive use. Finite clauses of place are, however, excluded from consideration, because there could be virtually no functional contact with non-finite clauses.

Finite and non-finite clauses in the English of Alfred's reign


2. The exceptions are non-finites operating in construction with closed system verbs, participles with no direct relation to an infinitive, the formulaic this/that is to say, verbal nouns which can be replaced by a perfectly regular concrete count noun. 3. The distribution of finite and non-finite clauses in early Middle English is shown in the table. Finite As subject -it +it As complement As object -S +S As adjunct -purpose +purpose Adj mod Total





8 36 3

23 10 18









869 58 4

14 60 31








4. See note 3 above. 5. Sixty-two entries in Parker consist of a single clause (Manabe 1979: 15; Shannon 1964: 44). 6. See note 3 above.

References Anderson, George K. 1949 The literature of the Anglo-Saxons. Princeton, NJ: Princeton University Press. [ 1962] [Rpt. New York: Russell & Russell.] Andrew, Samuel O. 1940 Syntax and style in Old English. Cambridge: Cambridge University Press. [ 1966] [Rpt. New York: Russell & Russell.] Bately, Janet M. 1970 "King Alfred and the Old English translation of Orosius", Anglia 88: 433460. Baugh, Albert C. (ed.) 1967 A literary history of England, Vol. I: The Middle ages. (2nd edition.) London: Routledge & Kegan Paul.


Kazumi Manabe

Bethurum, Dorothy 1932

"Stylistic features of the Old English laws", MLR 27: 263-279.

Blake, Norman F. 1977 The English language in medieval literature. London: Dent; Totowa, N.J.: Rowman and Littlefield. Bonner, Gerald 1973 "Bede and medieval civilization", Anglo-Saxon England 2: 71-90. Brown, William H., Jr. 1969 1970

"Method and style in the Old English Pastoral Care", JEGP 68: 666-684. A syntax of King Alfred's 'Pastoral Care'. (Janua Linguarum, Series Practica 101.) The Hague: Mouton.

Carlton, Charles R. 1970

Descriptive syntax of the Old English characters. (Janua Linguarum. Series Practica 111.) The Hague: Mouton.

Clark, Cecily 1971 "The narrative mode of 'The Anglo-Saxon Chronicle' before the Conquest", in: Peter Clemoes and Kathleen Hughes (eds.) 215-235. Clemoes, Peter - Kathleen Hughes (eds.) 1971 England before the Conquest. Cambridge: CUP. Davis, R. H. C. 1971 "Alfred the Great: propaganda and truth", History 56:169-82. Derolez, Ren6 1971 "The orientation system in the Old English Orosius", in: Peter Clemoes Kathleen Hughes (eds.), 253-268. Donaghey, Brian S. 1964 "The sources of King Alfred's translation of Boethius's De Consolatione Philosophiae", Anglia 82: 23-57. van Draat, P. Fijn 1916 "The authorship of the Old English Bede: A study in rhythm", Anglia 39: 319-346. Duckett, Eleanor S. 1967 Alfred the Great: The King and his England. Chicago & London: Collins. Duncan, Edgar Hill 1972 "Short fiction in medieval English: A survey", Studies in Short Fiction 9: 128.

Enkvist, Nils Erik 1972

"Old English adverbial pa Mitteilungen 73: 90-96.

an action marker?",


Finite and non-finite clauses in the English of Alfred's reign


Foster, Robert 1975 "The use of pa in Old and Middle English narratives", Neuphilologische Mitteilungen 76: 404-414. Fowler, Roger (ed.) 1966 Old English prose and verse. London: Routledge & Kegan Paul. Gneuss, Helmut 1972 "The origin of standard Old English and jEthelwold's school at Winchester", Anglo-Saxon England 1: 63-83. Gordon, Ian A. 1966 The movement of English prose. London: Longman. Greenfield, Stanley B. 1965

A critical history of Old English literature. New York: New York University Press. Harting, P. N. U. 1937

"The text of the Old English translation of Gregory's Neophilologus 22: 281-302. Krapp, George P. (ed.)



The Anglo-Saxon poetic records, Vol. V. New York: Columbia University Press.

Kuhn, Sherman M. 1972a "The authorship of the Old English Bede revised", Neuphilologische Mitteilungen TS\ 172-180. 1972b "Cursus in Old English: rhetorical ornament or linguistic phenomenon?", Speculum 47: 188-206. Liggins, Elizabeth M. 1970 "The authorship of the Old English Orosius", Anglia 88: 289-322. Manabe, Kazumi 1979 Syntax and style in early English - Finite and non-finite clauses c. 900-1600. Tokyo: Kaibunsha. Matsunami, Tamotsu 1958 "On the Old English participles", Studies in English Literature 34: 161-180. Miller, Thomas (ed.) 1890

The Old English version of Bede 's ecclesiastical history of English people. (EETS OS 95 & 96.) London: Oxford University Press. Payne, Frances A, 1968

King Alfred and Boethius: An analysis of the Old English version of the 'Consolation of philosophy'. Madison, etc.: University of Wisconsin Press.


Kazumi Manabe

Plummer, Charles (ed.) 1892

The Parker manuscript of the Anglo-Saxon chronicle, in: Two of the Saxon chronicles parallel, on the basis of an edition by John Earle, Vol. I. Oxford: Clarendon Press.


[Reprinted with Bibliographical Notes by Dorothy Whitelock. Oxford: Clarendon Press]

1902 The life and times of Alfred the Great. Oxford. [ 1970] [Rpt. New York: Haskell House.] Potter, Simeon 1939 "The Old English Orosius", TPS 1939: 44-53. 1947 "The Old English Pastoral Care", TPS 1947: 114-125. Schlauch, Margaret 1956 English medieval literature and its social foundations. University Press.

London: Oxford

Sedgefield, Walter J. (tr.) 1900 King Alfred's version of the Consolations of Boethius. Oxford: Clarendon Press. Shannon, Alice Ann 1964 A descriptive syntax of the Parker manuscript of the Anglo-Saxon chronicle from 734 to 891. (Janua Linguarum. Series Practica 14.) The Hague: Mouton. Sprockel, Cornells 1965The language of the Parker chronicle. (2 vols.) The Hague: 1973 Nijhoff. Svartvik, Jan-Randolf Quirk 1970

'Types and uses of non-finite clauses in Chaucer", English Studies 51: 393411. Stanley, Eric G. (ed.) 1966 Continuations and beginnings. London - Edinburgh: Nelson. Swanton, Michael (ed.) 1975 Anglo-Saxon prose. London: Dent. Sweet, Henry (ed.) 1871 King Alfred's West-Saxon version of Gregory's Pastoral Care. (2 parts.) (EETS OS 45 & 50). London: Oxford University Press. 1958 [Reprinted: EETS] Waterhouse, Elizabeth Ruth 1969 "The theme and structure of 755 Anglo-Saxon chronicle", Neuphilologische Mitteilungen 80: 630-41. Whitelock, Dorothy 1961 The Anglo-Saxon chronicle. (A revised translation.) London: Eyre and Spottiswoode.

Finite and non-finite clauses in the English of Alfred's reign 1962


"The Old English Bede." (Sir Israel Golancz Memorial Lecture.) The Proceedings of the British Academy 48: 57-90. 1966 "The prose of Alfred's reign", in: Eric G. Stanley (ed.), 67-103. Wrenn, Charles L. 1967 A study of Old English literature. London: Harrap.

Middle English Breaking* Toshio Nakao

1. Introduction Middle English Breaking is an epenthetic process first divorced from what is traditionally known as the ME diphthongization processes by Lass Anderson (1975). In this paper we will focus primarily on the details of ME Breaking reckoning with other closely related phonological processes such as ME Smoothing, Pre-Vocalic Shortening, among others. We will argue that there is another epenthetic process that has to be included in ME Breaking proper but that this and the 'original' process should not be collapsed into one rule. As a corollary we will also touch upon the issue of how long the 'long' diphthongs survived in ME.

2. ME Breaking Let us have a look at the examples illustrated in (1), in which high vowels [i, u] are epenthesized in the environments (i) e/e:_? (lai), (ii) α/ο:/ o/o:_x (laii), (iii) α/οJ (l.b). (For further examples see Nakao 1985: 258-259, 261). (1)




OE eahta>ehta>ME eighte 'eight', OE heah>heh>ME heih 'high', OE fashr>faht>ME faught '(he) fought' ii. ON lägr>läh>ME lough 'low', OE brohte>brohte>ME broughte '(he) brought', OE ploh>ME plough 'plough' smal>smaul 'small', all>aull 'all', fallen>faullen 'fall', talken>taulken 'talk', halp-haulf 'half, folde>foulde 'earth', follofoulk 'folk', gold>gould, sold>sowld 'sold', rolle>roulle 'roll', old>ould cf. ille 'ill', ilke 'ilk', eld 'age'

Some comments will be required on the contextual vowels. In the first place, [e, e:] in (lai) are due to LWS Smoothing in terms of which [aea/ae:a] get smoothed to [ae/ae:] before [k, γ, χ] or after [c, sc, g], LWS


Toshio Nakao

ae/ae:-Raising in terms of which [ae/ae:] get raised to [e/e:] in the exactly the same contexts as LWS Smoothing, [a] in (laii) comes either from that which has escaped LWS Smoothing and LWS ae/ae:-Raising, as in faughte, raughte, or from OE Pre-biconsonantal Shortening of [a:] as in taughte, naughte. [o:] in (laii) is raised to [o:] as in OE häm>ME höm, OE äld>ME old. Furthermore, LOE [o] due to OE Pre-biconsonantal Shortening of [o:] and LOE [o:] afford sources for contextual [o, o:] in (laii). [a] in ( l b ) is due to EME ae-Backing in terms of which OE [ae] is backed to [a] as in OE waeter>ME water, OE aet>ME at, while [o] either comes direct from OE [o] or is due to ME Homorganic Shortening in terms of which long vowels become shortened before homorganic consonant clusters as in old>old, lämb>lamb. These are processes comparable to Old English Breaking and are together referred to as Middle English Breaking. Chronologically viewed, ME Breaking was provoked in the fricative contexts in the early 13th century, while it was once again provoked in the [1] context about 1400 or a little earlier particularly south of the Humber. The ME Breaking process is unlike OE Breaking in several significant ways. ME Breaking affects [e, e:] alone of all the front vowels, but has effects upon back vowels [ο, ο:, ο:, a] as well. The segments epenthesized between vowels and [5] or [x/1] are dependent upon whether the preceding vowel is [-back] or [+back]. ME Breaking shares [x], [i] as the environmental consonants, but the last does not need to call in any support from other consonants. [rC] is incapable of Breaking any longer. Thus the phonological grammar of Middle English has to contain a rule that deals with the facts in (la) and (lb).


ME Breaking


+high -mid


b. 0->

+back +high -mid


aback +stress


ahigh amid -long +back +stress

+cont -strid aback +high -voiced +lat 1 +backj

-syl +cons

Middle English Breaking


3. Related processes What counts in this connection is to note the fact that some instances of ME Breaking have to invoke an ME smoothing process immediately after. The process at issue, acting on the long outputs [e : i], [o : u], generates [i:], [u:] before [9], [x] in syllable-final or stem-final position. However, it produces no effect on short [ei, ou] at all; if it did, surface forms with [i:, u:] would be yielded erroneously. The consequence of this is we are led to admit the existence of long vowels [e : i], [o : u] due to ME Breaking (and ME Vocalization, too) at least until the time when ME Smoothing is in full swing, though their disappearance is generally dated immediately after the vocalization of [g, γ, w], i.e., about 1200. It is generally recognized that the Smoothing commenced in the southern border of the Midlands in the later 13th century, spreading to the central Midlands and the central South during the 13-14th centuries. (3)

a. b.

heigh>high, neigh>nigh, theigh>thigh, leih>llgh,fleih>fiigh plöugh>plöügh, böugh>böügh, töugh>töügh,löugh>löügh, inöugh>inöügh

The Smoothing processes that followed ME Breaking can be formalized as the first approximation in the following way. ME Smoothing (1st version)


aback SD: +mid +stress +long


+high -mid


+cont aback +high -strid -voiced


SC: Γΐ

], [aback J

Γ2], [0J

There is another half to this Smoothing. There are cases in which ME Smoothing has to be brought into play to smooth to [i:], [u:] the long

212 ToshioNakao diphthongs [e : i], [ο : u] that are due to Vocalization of [g, γ, w]. In the synchronic phonology of Middle English the rule to deal with this Smoothing follows ä-Raising, Vocalization, but pecedes Pre-vocalic Shortening with the effect of V—>V/_V. The point worth stressing here is that diphthongs whose first element is [ε:] due to EME S-Raising in which OE [ae:] gets raised to [ε:] as in säe>se, cläene>clene or [a:] due to äRaising are disqualified for the input. (5a) contains some representative examples. (5)

a. b.

leien>Uen 'lie', flein>flien 'fly', e^e>le 'eye', teien>tlen 'tie', seien>slen 'say' böwes 'boughs'>«, wögian 'woo'>wöwen>ü, dröwen 'drew' (pl.)>w, inowe 'enough' (pl.)>w

A little more general version of the rule for ME Smoothing will be given as (6). To distinguish between those eligible for the input ([e:, o:]) and those not ([ε:, ο:]) the rule requires a specification [+high] for the first term in its SD. (6)

ME Smoothing (2nd version)



aback +high +mid +stress +long


rf η +high -mid , aback < J


1 aback

+cont +high aback -strid -voiced Γ+syl Ί -stress


3 3

4. ME Breaking as bipartite processes ME Breaking has been captured as consisting of bipartite processes whose effect is to epenthesize [i, u] between stressed vowels and [ς, χ], on one

Middle English Breaking


hand, [u] between stressed vowels and [1], on the other. These bipartite processes, however, cannot be collapsed into one rule because another rule, HS, has to intervene between them. Otherwise it is not possible to arrive at correct forms of say old, gold, etc.: ME Breaking in which (2a) and (2b) are collapsed would alter [o:] into [o:u], which would be changed by ME Smoothing to [u:], which would, in turn, be converted into [u] in terms of HS. The output [u] would have to be lowered to [λ] in early ModE. There arise derivations as shown in (7) in which ME Breaking and other relevant processes are crucially involved. (7) input




/le:g Vn/

M E ä-Raising

/a: Id/ ο:



M E Br (2a)


M E Smoothing (6)






V->V/_V HS


ME Br (2b)



Note *

This paper originally appeared in: Toshio Nakao (ed.) 1987, Historical studies in honor ofTaizo Hirose. Tokyo: Kenkyusha, 179-183.

References Lass, Roger - John M. Anderson 1975

Old English phonology. Cambridge: Cambridge University Press.

Nakao, Toshio 1985

English historical phonology. Tokyo: Taishukan.

Syntactical revision in Wulfstan's rewritings of iElfric* Hiroshi Ogawa The Wulfstan canon contains three pieces which are reworkings of homilies and a pastoral letter by jElfric: Homilies IX (11. 19-106) and ΧΠ (Whom 9 and 12, ed. Bethurum), reworked from ^Elfric's De Septiformi Spiritu (/ESpir, ed. Napier) and De Falsis Diis (/EHom 22, ed. Pope) respectively, and the D version of ^Elfric's Second Pastoral Letter (/ELet 2, ed. Fehr). The pieces are called by Bethurum (1957: 32-33 and passim) 'rewriting' in explicit contrast with mere revisions of ^Elfric's material (Homilies VI and XVIII) on the one hand and on the other texts which are, like Homily IV, only loosely dependent on vElfric's material. In these rewritings Wulfstan, while following vElfric's original syntax fairly closely, still allow himself freedom to revise it and consequently gives us some clues to the most distinctive features of his own syntax. Apart from frequent additions and omissions of material and lexical substitutions,1 there are altogether some 270 alterations which Wulfstan made to the original syntax of ^Elfric's three works as he rewrote them. They encompass a wide variety of features, from inflexional variation to sentence structure, which can be roughly sorted into twenty groups.2 Space does not permit a full discussion of all. The one I take up for examination here as the largest of my groupings consists of alterations that are related in various ways and degrees to coordination/subordination. About 70 examples have been assigned to this group, and several important patterns can be distinguish among them. It is mainly these and a certain related type of alteration that are my immediate concern below. One may first note those cases in which Wulfstan revises Elfric's subordinate clauses using a coordinate structure. This sometimes takes the form of an independent sentence, as in JELet(D) 2.61 Se feorda sinod wees Se sinoÖ fordernde...3 for ./Elfric's Se feorjja sinod wees..., se fordernde..., but, more often, a syndetic paratactic clause. Thus, in /EHom 22.77 Da |3a hi toferdon to fyrlenum landum, and mancynn J)a weox, J>a wurdon hi bepaehte J>urh J>one ealdon deofol...


Hiroshi Ogawa WHom 12.10 Pa syööan toferdon hy wide landes, 7 mancyn Jja sona swyöe weox; 7 öa aet nyhstan wurdon hi bepsehte Jjurh öone ealdan deofol... /EHom 22.74..., of)öaet J>a entas worhtan £>one [wundorlican] stypel aefter Noes flode, WHom 12.7 Ac syööan {>aet gewearö J>aet Nembroö 7 öa entas worhton jjone wundorlican stypel aefter Noes flode, /ESpir 59.1 aelc man biö eadig, haefö {)one wisdom, gif he his agen lif gelogaö mid wisdome. Whom 9.68 selc man biö gesaelig 7 eadig J)e haefö J>asne wisdom t>e of Godes agenre gyfe cymö 7 öurh {)aet his agen lif gelogaö mid wisdome,

vElfric's correlative da J?a ..., pa construction, opdcet clause and gif clause dependent on a relative J?e clause are transformed into a coordination structure with and or ac; for the addition of facet wearp pcet in the second example see §3.1. Somewhat more complicated is the alteration in /EHom 22.118 I>as manfullan menn waeron J)a maeroston godas J)e J>a haet>enan wuröodan, and worhton him to godum; WHom 12.50 t>as manfullan men ... waeron getealde for öa maerostan godas J?a on öam dagum, 7 (>a haeöenan wuröodon hy swyöe t)urh deofles lare. Wulfstan here first changes yElfric's categorical assertion wceron into what is unmistakably a matter of pagan belief by adding getealde 'reckoned, considered' (On this addition see further §5), but this change obviously made it impossible to retain the following relative clause qualifying the superlative mcerostan, just as it would be rather awkward to say in ModE These wicked men were reckoned (by the pagans) the greatest gods that the pagans honoured'; hence, one assumes, Wulfstan's use of and to make the original single sentence into two discrete statements.

Syntactical revision in Wulfstan's rewritings of /Elfric


The last-mentioned case is one of the three examples of the use of an independent sentence or coordinate clause to rewrite Elfric's relative clause,4 though the others (WHom 9.59, with extensive added material intervening before what is to become the antecedent of the relative pa in AZSpir 58.12, and /ELet(D) 2.61 [see §2]) are both comparatively simple alterations and, unlike the above-mentioned one, do not appear to be dictated by semantic considerations. This subgroup as a whole has apparent counterexamples in cases where Wulfstan uses a relative clause for another construction in itilfric. In fact, the two types of alteration, with and without the relative construction, are perhaps not contradictory but point to a certain more general tendency which dominates Wulfstan's style, as we shall see. [It may be seen that the examples given above are corroborative of Wulfstan's inclination to polysyndeton, an inclination which has often been noted in previous studies of his style.5 On the other hand, they are by no means a majority in the collection of syntactic revisions under examination. A greater number of examples in fact display a pattern which exactly reverses the alteration just discussed: Elfric's independent sentences and coordinate clauses are changed into subordinate clauses in Wulfstan. There are two distinct forms Wulfstan employs in this process. The first, with hit wearp/wces (pcet) appended before Elfric's simple sentence, making the latter into a subject clause, is implied in WHom 12.7 for JEHom 22.74 (§2) but is seen in its proper form in AlLet(D) 2.36 Hit geweard cefter ures drihtenes prowunge and cefter his up-stige, pcet se halgast com of heofonum to eorSan in place of the original /Ejfter pees hcelendes prowunge and his upstige to heofenum com se haliga gast of heofenum to eorpan, and, more significantly, in /ELet 2.39 l>a gebugon to fulluhte of J)am folce J)reo Jrnsend manna on anum daege... (D) J)a gewearö hit J>aet gebugon to fulluhte {>reo fmsend manna on anum daege... /ELet 2.45 Hwaet J>a aefter fyrste asprang faerlice ormaete ehtnys ofer J)a cristenan ... (D) Hwaet da aefter J>am hit gewearö siööan öa asprang faerlice ormaete ehtnes ofer öa cristenan wide and side... The original inverted word-order VS is kept unchanged in both examples; the latter example has it preceded semi-paratactically by the new clause hit


Hiroshi Ogawa

geweard siÖdan, which in its turn is inserted after the original temporal phrase which it reinforces (da cefter ... siddari) (On the absence of conj. Pcet see §3.2). In other words, Wulfstan's new main clause seems to have been added not so much for its semantic contribution as to give a greater prominence to the sense of change and movement than in the original sentence; the addition is stylistic rather than semantic. And it is presumably in accordance with this desired effect that Wulfstan employs the verb (ge)weorJ?an in all the examples of alteration being discussed but one (ALLet(D) 2.9, with hit was pcet). Here the verb, taken in its own right, means primarily 'to come to pass, ... to happen' (Bosworth-Toller s.v. weorpan 1(2)), while at the same time it serves to give an additional nuance of change to the new sentence as compared with iElfric's original; the verb is thus used partly as 'implying movement, change of position' (BosworthToller s.v. IV). As we shall see, Wulfstan's use of (ge)weorj?an with these stylistic values is not restricted to the present pattern of alteration.6 The second type of subordinate clause rewriting iElfric's independent sentence occurs when the latter is changed into an object clause dependent on the main clause which is added. The main verb then is usually understandan or witan in the imperative,7 as in: /ELet 2.111 Begen synd on anum hade, se bisceop and se maesse-peost, (D) Understandaö, {jaet beggen sind on anum hade, se biscop and se maesse-preost, MLet 2.161 Beo His calic geworht of ecum antimbre, (D) And witaö, J>aet beo aelc calic ge-worht of myltendum antimbre, jELet 2.185 Ne mot nan peost beon mangere of){)e gerefa. (D) And witaö eac J>aet ne mot mid rihte nan preost beon gitsiende mangere ne world-strutere on gerefscire. By employing these arresting verbs, Wulfstan appeals the more directly than ^Elfric to his audience. Wulfstan's new main clause is here again primarily of stylistic importance, with little to add to the contents of his message; it has the character of an interjection. This semi-parenthetical nature perhaps explains at least partly why in the second and third examples quoted above Wulfstan's main clause is followed by an object clause in which he retains the original inversion, together with the subjunctive verb in the second example. One assumes that the retention of

Syntactical revision in Wulfstan 's rewritings of AZlfric


these features is no more lapse on Wulfstan's part10 but rather argues that he is well aware of what he is doing as a reviser; for he does not forget to abandon the inversion when he uses a verb of full meaning to introduce the object clause: ALLet 2.5 Byö swa-J)eah sum swa onbryrd fcurh God,... (D) Ic truwige J>eah J>aet sum wurde abrird jDurh God,... It is also interesting to note that here apiin Wulfstan has recourse to the verb (gejweorpan to effect his revision. In fact, the retained inversion might even encourage us to speculate that /ELet(D) 2.185 above, for example, was intended as a paratactic construction, to be repunctuated: And witaf? eac pcet. Ne mot mid rihte nan prest..., a possibility which appears somewhat greater in /£Let(D) 2.45 discussed in the last section. The manuscript (CCCC 201), with what looks like a point after siddan,12 suggests that the sentence could be read as: Hwcet da cefter pam hit geweard siddan. Da asprang..., as opposed to the reading of Fehr ([1966]: 91), who puts no punctuation mark after siddan and translates what follows it as a subject clause ('Dann aber (geschah es nach diesem, daß) entstand...').13 The truth is difficult to prove for either of the two sentences, of course. But it seems at any rate important to observe that there is room for such a reading in some of the examples I have grouped as Wulfstan's subordinate clause for jElfric's independent sentence or coordinate clause,14 especially if we are to understand the nature of that group and what Wulfstan intended to achieve through it. There is another type of example which might be mentioned here. In about half of a dozen cases from /ELet(D), Wulfstan adds the parenthetical he cwcep or the like within jElfric's simple sentence (e.g., /ELet(D) 2.27) or complex sentence (e.g., /ELet(D) 2.154), and in case the verb cwcep also functions as the main verb to introduce what is now an object clause preceded by the conj. Pcet, in a way similar to understandan and witan discussed above: ALLet 2.82 and gif he elles do, J>olige his hades. (D) and gif he elles do, se canon cwaej), J>aet he Jjolige his hades. As I have suggested, there is a sense in which Wulfstan's subordinate clauses for ^Elfric's independent sentences or coordinate clauses as a whole are often not far removed from paratactic structure. Even apart from the two 'semi-paratactic' sentences discussed in the last paragraph but one, the


Hiroshi Ogawa

pattern, in both its main types, consists of a new main clause added to the original main clause which now follows it in the form of a pcet-c\a\ise in a way that reminds us of syndetic parataxis. What we have just seen points to the tendency for Wulfstan's sentence structure to become more complex than iElfric's, in pursuit of effects which are more stylistic than semantic. This agrees with a general characteristic of vElfric's style which Nichols (1971: 1-12; 1978: 157-180) discusses in reference to the tradition of the 'brief style' in mediaeval rhetoric. In fact, one of the features which Nichols (1978: 165ff.) mentions as characteristic of this style in ^Elfric also occurs in the three works being considered and is consistently rendered into a more complex structure by Wulfstan: yElfric's appositions and phrases are replaced by finite clauses in Wulfstan. This pattern is the more revealing because it is common, while there are no cases of revision in the reverse direction.15 Examples are: MLet 2.46..., swa t>aet J>a haej^enan on J)one haelend gelyfdon, oft an Jnisend manna, J>aer man aenne ofsloh, jjurh Jra myccla tacna, (D)..., swa Jjaet t>a haeöenan ... on Crist gelifdon. Oft ge {msend manna aetgaedere gelifde, ]>xr man aenne martir ofsloh, J)urh öa micclan tacna, AiHom 22.150 Sum wif hatte Uenus, seo waes Ioues dohter, swa fracod on galnysse J)act hire faeder hi haefde, WHom 12.77 And sum wif hatte Uenus seo waes Ioues dohtor, j seo waes swa fill 7 swa fracod on galnysse J>aet hyre agen broöor wiö hy gehaemde, /EHom 22.73 ... on eallum J>am fyrste aer Noes flode, WHom 12.7 ... on eallum Jjam fyrste J)e waes aer Noes flode. Here Wulfstan shows the full variety of independent sentence, coordinate clause and subordinate clause. In the last category, the relative clause is the rule, as in the last example, with the single exception of a temporal clause in WHom 12.37 ... mistlice entas γ strece woruldmen pe mihtige wurdan on woruldafelum γ egesfulle wceran pa hwyle pe hy leofedon, γ heora agenum lustum fullice fulleodan in place of /EHom 22.101 ...mislice ents and men ... pa pe mihtige wceron on woruldlicum gepincdum, and egefulle on life, peah pe hy [leofodon] fiillice. Indeed, the

Syntactical revision in Wulfstan 's rewritings of Ailfric


relative clause is so dominant in the present pattern that there are few added relative clauses which do not belong in this category. 16 What is common to all these revisions is that they expand a simpler • 17 construction in jElfric's originals; Wulfstan's is an 'expansive' style as against jElfric's 'brief style'. The effect Wulfstan thereby achieves may be seen at its clearest in his version of yElfric's original on Jove and Juno in De Falsis Diis: MHom 22.108 He laefde swajjeah aenne to l[i]fe, J)eah J>e he abite his gebroöra on aer; se waes Iouis gehaten, hetol and {>rymlic. He [afligde] his faeder of [>am foresaedan iglande, and wolde hine acwellan, gif he him come to. Se Ious waes swa swiöe gal, J>aet he on hys swustor gewifode; seo waes gehaten Iuno, swiöe healic gyden. WHom 12.43 He laefde swaj>eah uneaöe aenne to life, J>eah öe he fordyde J>a broöra elles; η se waes Iouis gehaten, 7 se wearJ) hetol feond. He aflymde his agene faeder eft of öam ylcan foresaedan iglande J)e Creta hatte 7 wolde hine forfaran georne gif he mihte; 7 se Iouis wearö swa swyöe gal J>aet he on his agenre swyster gewifode, seo waes genamod Iuno, 7 heo wearö swyöe healic guden after haeöenscype geteald. jElfric has two appositive structures (Iouis..., hetol and frrymlic and Iuno, swiöe healic gyden), both of which Wulfstan expands into a coordinate clause by adding and. Very characteristically, Wulfstan's verb is here again (ge)weorpan in both examples. In view of the obvious absence in jElfric's original version of the sense of contrast between Jove and Juno's earlier and later selves, Wulfstan's revision, the use of (ge)weorpan in particular, leads us to argue either (a) that he intentionally added some sense of change or (b) that the verb has nothing of its original sense of change in these examples, as in a certain example in Wulfstan's Canons of Edgar to which the Glossary of the EETS edition assigns the sense 'to be, remain'. 18 However, this latter argument assumes a very heavy burden of proof, requiring a very unusual sense for the verb 1 because it is wanted for a particular reading. One should rather assume that (ge)weorfjan, as distinct from beon/wesan, probably involves a sense of change in the examples being discussed too, though to what degree can certainly be a matter for discussion. It is possible that Wulfstan used the verb in its full sense of becoming, as is implied in Swanton's translation (1975: 124) ('he was called Jove, and he became a malignant


Hiroshi Ogawa

enemy....sister, who was named Juno; and she became very important goddess'), or, more probably, he may have used it with some weakened nuance. In any event, it is a sense of change, of movement, which is obviously evoked by the addition of and to begin a new clause, for Wulfstan's (ge)weorj?an occurs with it in both examples.20 In other words, what vElfric has condensed into apposition Wulfstan expands into a flow of events and situations linked by and, with the verb (ge)weorjjan to sustain the sense of movement. vElfric's style is static, while Wulfstan tends to favour one which is more kinetic. The use of (ge)weorJjan, as part of that distinction, is largely stylistic, as with the examples of the verb in my second pattern (see §3.1). It is a use that explains in much the same way a third alteration in the parallel passages quoted above between yElfric's Se Iouis wees swa swide gal... (in the last sentence but one) and Wulfstan's7 se Iouis weard swa swyde gal... The passage in Jove and Juno just discussed also contains, in Wulfstan's version, another interesting feature of this homilist's syntax, though it does not directly concern coordination/subordination. The periphrasis weard ... geteald in Wulfstan's last clause (for jElfric's simple phrase) plays no further important part in his rewriting of the three works being examined. But its related form beon + past participle is often used elsewhere to replace jElfric's simple beon, as in WHom 12.50 in place of /EHom 22.118 (§2) and JESpir 58.20 aelc wisdom is of gode, foröam {>e god sylf is wisdom, WHom 9.67 jEIc riht wisdom is cumen of Gode, foröam J>e God sylf is se soöa wisdom. Scholars have drawn attention to these particular examples. For example, Jost (1950: 133) sees in the first example the reviser's wish to emphasize that 'die heidnischen Götter nur Irrgötter seien, ja dass sie überhaupt nur nach der Meinung der Heiden als Götter gelten oder galten', while to Mcintosh (1949: 122) the second example is important because it illustrates the reviser's own rhythmical scheme of two-stress phrases according to which jElfric's rhythmical prose has been reshaped. They may well be right as far as they go. However, that these examples are only part of a general tendency in Wulfstan's syntax is suggested by the presence of related examples ignored by Jost and Mcintosh. Thus, rewriting the Pastoral Letter on clerical duties, Wulfstan adds the

Syntactical revision in Wulfstan 's rewritings ofAilfric


participles genamode, getacnod and getealde without any implication of erroneous beliefs at all: ALLet 2.17 J)a synd leorningc-cnihtas, (D) J)a synd leorningc-cnihtas on bocum genamode, ALLet 2.28 On {)am lendenum is, swa-swa we leornigaj) on bocum, seo fule galnys ... (D) On fam laendenum is getacnad, swaswa we leorniad on bocum, seo fule galnes ... ALLet 2.8a t>reo tyda synt on t)isse worulde: (D) Dry timan sind getealde on J>issere worlde. Nor are these examples mentioned by Jost (1950: 133-148) as among those alterations in the work which are most obviously made for rhythmical reasons. On the other hand, they are syntactically related to another type which occurs in the same work, where the participle is added as appositive to the subject of the clause, e.g., /ELet(D) 2.61 Se feorda sinoö wees six hund biscopa and XXX sacerda, swide widan gegaderode in place of ^Elfric's Sefeorpa sinod wees eft, pcet wees syx hund bisceopa and frrittig sacerda. There is even an example of periphrasis with the present participle in rewriting ^Elfric's simple beon: ASLet 2.60 se haelend is soölice an Crist, on anum hade aefre wunigende. (D) ure drihten is soölice on anum hade aefre wunigende. These various examples combine to suggest that syntactically Wulfstan tends to write in a way that is verb-oriented and concrete, as compared with jElfric, whose expression tends to centre around substantives and consequently to be condensed and abstract. jElfric's tends towards a nominal style, Wulfstan's towards a verbal style - a distinction which Hollowell (1977: 291-293) equates with that between high style and low style. Hollowell herself distinguishes these different levels of style within Wulfstan's works, relating them to various syntactic features which are found in different degrees within them. These include, in addition to the distinction between verbal and nominal styles, another which is concerned with the 'Content Level' of clauses: in a low style, 'while the principal


Hiroshi Ogawa

clauses provide skeletal structure, the subordinate clauses actually carry the burden of the message' (Hollowell 1977: 291) - a feature which in fact has again been found in Wulfstan's rewritings in comparison with iElfric's originals in, e.g., the addition of hit geweard φαί) 2 (see §§3.1, 3.2). If Hollowell's distinction thus applies to Wulfstan, or at any rate to his rewritings under examination as a whole, it is perhaps this comparatively low level of style that underlies his preference for ami-coordination and 'semi-paratactic' accumulation of clauses discussed earlier (see §§2-4); for it is, one assumes, an aspect of what Hollowell (1977: 290 and 295) has called the homilist's 'additive' style, in which 'the basic movement of thought ... is forward... It is a style that keeps the mind focused ahead in anticipation of the rush of words to come.'23 I have examined Wulfstans' syntactic revisions of ^Elfric's originals, with particular reference to those concerned with coordination/ subordination. The results of the examination, focussed as it is on several patterns within this limited group of alterations, cannot of course be expected to represent all the distinctive features of Wulfstan's syntactical revisions, let alone the syntax of his works as a whole. It is also possible that some of the results may prove less decisive in the light of the evidence of other examples of various groups of alterations not treated in this essay. Within these obvious limitations, the present examination has revealed certain characteristic tendencies of Wulfstan's revised syntax sufficiently clearly. They are seen, most importantly, in Wulfstan's preference for syndetic paratactic clauses in place of ^lfric' subordinate clauses and appositions and phrases, and for clauses accumulated in a 'semi-paratactic' way in place of ^Elfric's single clause (simple sentence). Taken together, these tendencies point to a general distinction between the styles of the two writers: Wulfstan's style is expansive as compared with vElfric's 'brief style'; Wulfstan inclines towards expressions which may be called kinetic, while jElfric's expressions are generally more static and condensed. It is, in sum, a distinction that might well reflect different levels of style, as between low (or colloquial) style and high (or literary) style.

Notes *

This is a revised version of the paper originally published under the title 'Revised Syntax in Wulfstan's Rewritings of /Elfric's Prose' in Approaches



Linguistics: Papers in Memory of Saburo Ohye (Kyushu University Press, 1989). I owe

Syntactical revision in Wulfstan 's rewritings of t€lfric


special thanks to Mr Brendan Wilson of the University of Tokyo, who kindly read the earlier version and suggested many stylistic improvements. 1. These changes in Wulfstan's three pieces under examination have been discussed extensively by, e.g. Bethurum (1957: 2, 47, 98, and 306), Fehr ([1966]: LXV-LXIX), and Jost (1950: 117-148). However, these scholars are not much concerned with syntax, and little attention has been paid to the problem of Wulfstan's sentence structure which I shall examine in this essay. 2. They are: (1) inflexional variations; (2) demonstratives added; (3) demonstrative + adjective/demonstrative + adjective + noun; (4) reflexives added; (5) relative pronouns; (6) prepositional variations; (7) subjunctive/indicative; (8) modal verbs; (9) preposition/case, and repetition/omission of preposition; (10) beonlwesan + past participle, etc.; (11) impersonal/personal construction; (12) direct object and indirect object; (13) (demonstrative) pronoun/noun; (14) negation; (15) word order; (16) subject (+verb) repeated; (17) and, fja, peah, etc.; (18) conjunctional variations; (19) participial construction; and (20) coordination/subordination. I discuss these in Ogawa (1990: 85107). 3. Citations are made from the editions mentioned in §1. References are to sentence number of the Pastoral Letter following Fehr's edition but are otherwise to page and/or line. 4. Here, as elsewhere in this study, clauses are classified as independent or subordinate according to the editorial punctuations of the respective texts. For the ambiguous demonstrative/relatives see Mitchell (1985: §§2444-2449). 5. See, e.g., Mohrbutter (1885: 84), Kinard (1897: 20), and Fowler (1966: 14). 6. Similar emphasis on the sense of change might well be also responsible for the alteration of JEHom 22.86 and wurf>odan hy [georne] into WHom 12.22 γ agunnan hy weordian georne. 7. This use of the imperative verb has also been noted by Fehr ([1966]: LXVI) and Jost (1932: 274-278) as one of Wulfstan's favourite formulas. 8. See Jost (1932: 276). 9. For Wulfstan's customary use of parenthesis, see Fowler (1966: 5-6) and Ogawa (1989: § 10. The possibility of error cannot be entirely excluded in other cases, of course; see AiSpir 56.11 Isaias se witega awrat on his witegunge be dam halgan gaste and be his seofonfealdum gifum, which Wulfstan changes into WHom 9.19 Be dam seofanfealdan Godes gyfan Isaias se witega awrat on his witegunge be dam halgan gaste γ be his .vii. fealdum gifum, with some tautology and different inflexional endings for dat. pi. of giefu. But the VS word-order itself is not uncommon in subordinate clauses in OE; see Mitchell (1985: §3934). 11. The verb (ge)weorf>an also plays a part in two other sentences in Wulfstan's rewritings as a whole: WHom 12.37 for &Hom 22.101 (§4), and /ELet(D) 2.186 ne to druncan-


Hiroshi Ogawa

georn wurdan for ^Elfric's ne druncen-georn beort, as well as in the passage on Jove and Juno to be discussed in §4. 12. This observation is based on an examination of the manuscript in microfilm. According to Ure (1957: 111-112), the scribe of the manuscript uses only points as punctuation marks. 13. The parenthesized portion corresponds to Wulfstan's addition. 14. For a similar possibility, see LS 7 (Euphr) 97 Pa [MS Ο lamp if] purh godes mildheortnysse gemette he an para muneca, translated by Skeat as 'Then (it happened), by God's mercy, (that) he met one of the monks'. 15. We have one of Wulfstan's nearest approaches to apposition in And habban us on handa ure leoht-fatu, pcet synd halige weorc for /ELet 2.29 And habban us on handum halige weorc. But even here Wulfstan uses a parenthetical or explanatory formula pcet synd; his aim is anything but 'brief style'. 16. The exceptions are WHom 9.68 for AZSpir 59.1 (§2), and /ELet(D) 2.12 ... clcennesse slif bisnode and tcehte, pa clcennesse pa he lufad on his clcenum penum for jElfric's ... clcennysse tcehte. pa clcennysse he lufad on his clcenum penum. 17. Another notable case of Wulfstan's expansion is seen in WHom 9.72 deofol scewd pcertogeanes unwisdom γ swicdom γ geded swa purh pcet pcet unscelig man wisdomes ne gymed ne wislice his lif ne fadad as compared with ALSpir 59.3 se deofol forgifd pcertogeanes dysig, pcet he wisdomes ne gyme ne wislice ne libbe, and in a series of similar parallel sentences which follow each. Where jElfric has a single verb followed by a noun or ^ / - c l a u s e as object, Wulfstan prefers to insert a second verb, mostly a causative verb (gedon, macian, Iceran), before the pcet- clause. 18. Fowler (1972: Glossary s.v. (ge)weordan). The example referred to occurs in section 15: And we Icerad pcet preosta gehwilc fulluhtes tidige sona swa man his girne, and ceghwar on his scriftscire beode pcet celc cild sy gefullod binnon xxxvii nihtum and pcet cenig man to lange unbiscopod ne wurde. 19. Of course, dictionaries do not recognize the sense 'to be' or the like for the verb except for its use as a passive auxiliary. Bosworth-Toller has a special entry 'weor^an act to be at something' (Supplement s.v. weorpan 111(a)) just for ALLS 32.4 pa wurdon hi cet sprcece oddest dunstan rehte be sancte eadmunde. As the initial pa suggests, however, it could well be an example of what is defined in the dictionary as 'to get into a state of action, to come to be doing something, to fall to an action, to take to' (s.v. III(b)2). 20. The same is true in the example from the Canons of Edgar mentioned above; see n. 18. 21. However, Mcintosh is challenged by Hollo well, who says that in Homily IX 'Wulfstan does not achieve two-stress rhythm at all' (1982: 7). See also Funke (1962: 311-318). 22. Hollowell's skeletal structures (1977: 290) in fact include this clause type, as found in WHom 2.28. 23. Hollowell herself attributes this aspect as found in Homilies V and XX to a high style. But it is generally agreed that features such as coordination by and and addition of

Syntactical revision in Wulfstan 's rewritings of /Elfric


details belong to common speech and colloquial rather than literary style. See, e.g., Brewer (1986: 234-236).

References Bethurum, Dorothy (ed.) 1957 The Homilies of Wulfstan. Oxford: Clarendon Press. Boitani, Piero - Jill Mann (eds.) 1986 The Cambridge Chaucer companion. Cambridge: University Press. Bosworth, Joseph - T. Northcote Toller (eds.) 1898

An Anglo-Saxon dictionary. London: Oxford University Press. (Supplement, by Toller, 1921. Enlarged addenda and corrigenda, by Alistair Campbell, 1972.) Brewer, Derek 1986 "Chaucer's poetic style", in: Piero Boitani - Jill Mann (eds.), 227-242. Fehr, Bernhard (ed.) [1966]

Die Hirtenbriefe Mlfrics. (Bibliothek der angelsächsischen Prosa, IX. Band.) Darmstadt: Wissenschaftliche Buchgesellschaft. Fowler, Roger 1966 "Stylistic features of the Sermo Lupf', JEGP 65: 1-18. 1972 Wulfstan's Canons of Edgar. (EETS 266.) London: Oxford University Press. Funke, Otto 1962 "Some remarks on Wulfstan's prose rhythm", English Studies 43: 311 -318. Hollowell, Ida Masters 1977 "Linguistic factors underlying style levels in four homilies of Wulfstan", Neophilologus 61: 287-296. 1982 "On the two-stress theory of Wulfstan's rhythm", Philological Quarterly 61: 1-11. Jost, Karl 1932 "Einige Wulfstantexte und ihre Quelle", Anglia 56: 265-315. 1950 Wulfstandstudien. (Swiss Studies in English 23.) Bern: Francke. Kinard, James P. 1897

Α study of Wulfstan's homilies: Their style and sources. [Unpublished Ph.D. dissertation, Johns Hopkins University.] Mcintosh, Angus 1949 'Wulfstan's prose', Proceedings of the British Academy 34: 109-142. Mitchell, Bruce 1985

Old English Syntax. 2 vols. Oxford: Clarendon Press.


Hiroshi Ogawa

Mohrbutter, Alfred 1885

Darstellung der Syntax in den vier echten Predigten des angelsächsischen Erzbischofs Wulfstan. Münster diss., Lübeck.

Napier, Arthur (ed.) 1967 Wulfstan. Sammlung der ihm zugeschriebenen Homilien nebst Untersuchungen über ihre Echtheit. (2nd edition with Bibliographical Supplement by Κ. Ostheeren.) Dublin: Weidmann. Nichols, Ann Eljenholm 1971 "jElfric and the brief style", JEGP 70: 1 -12. 1978 "Methodical abbreviation: Α study in Aelfirc's Friday homilies for Lent", in: Paul E. Szarmach - Bernard F. Hupp6 (eds.), 157-180. Ogawa, Hiroshi 1989 Old English modal verbs. A syntactical study. (Anglistica 26.) Copenhagen: Rosenkilde and Bagger. 1990

"Syntactic alterations in Wulfstan's rewritings of iElfric", The Proceedings of the Department of Foreign Languages and Literature, College of Arts and Sciences, University of Tokyo 38, 3: 85-107.

Pope, John C. (ed.) 1967-1968 Homilies of /Elfric. 2 vols. (EETS OS 259 and 260.) London: Oxford University Press. Szarmach, Paul E. - Bernard F. Huppi (eds.) 1978 The Old English homily and its backgrounds. Albany, NY: State University of New York Press. Swanton, Michael (tr.) 1975 Anglo-Saxon prose. London: Dent. Ure, James M. (ed.) 1957 The Benedictine office. Edinburgh: University Press.

On double auxiliary constructions in Medieval English* Michiko Ogura In my article on shal (not) mowe constructions, I started by posing two questions: "where did the construction come from? and does may (mowe) really function as a periphrastic future" (Ogura 1993: 539) as Mustanoja (1960: 496) stated. I gave six examples of the shal not mowe construction found in the Wycliffite Gospels of the Earlier Version, i.e. Mk 3.26, Lk 1.20, Lk 13.24, Lk 16.2, Lk 20.35-36, and Lk 21.15. The Later Version shares the feature except for Lk 16.2. In these six instances, the combination corresponds to Latin possum in various future forms, which can be explained that the futurity was translated by shal and the sense of possum by mowe. I give the six examples again to show that this combination of shal and mowe never goes back to Old English or Old High German. 1 (1) Mk 3.26 Li: WSCp: EV:


[et si satanas consurrexit in semet ipsum dispertitus est et non poterit stare sed finem habet] 7 gif i öeah se wiöerwearda efne ansa on hine sulfne toworpen waes \ bi3 7 ne mceg gestonde ah ende haefeö 7 gif satanas wind ongen hine sylfne he bid to-dailed 7 he standan ne mceg ac haefö ende And if Sathanas hath risen a3eins hym self, he is disparpoilid, and he shal not mowe stonde, but hath an ende. And if Sathanas hath risun ajens hym silf, he is departid, and he schal not mowe stonde, but hath an ende.

Cf. Tatian 62.3

(2) Lk 1.20 Li: WSCp:

Oba Satanas in imo selbemo ziteilit ist, vvuo gistentit thanne sin rihhi? iz ni mag gistantan, ouh enti habet [Et ecce eris tacens et non poteris loqui usque in diem quo haec fiant] 7 heono öu bist suigende 7 ne mcege 3u gesprece odd on doege of öaem öas geworöes And nu J)u bust suwiende. 7 J>u sprecan ne miht. od [Done daeg {je öas öing gewuröaj)


Michiko Ogura EV:


And loo! thou shalt be stille, or doumbe, and thou shalt not mowe speke til in to the day, in which thes thingis schulen be don And lo! thou shalt be doumbe, and thou shalt not mow speke til in to the dai, in which these thingis schulen be don

Cf. Tatian 2.9

(3) Lk 13.24 Li:



(4) Lk 16.2

Inti nu uuirdist thu suigenti inti ni maht sprehhan unzan then tag, in themo thisu uurdent [contendite intrare per angustam portam quia multi dico uobis quaerunt intrare et non uuerdent] geörinsgas \ to ingeonganne öerh nearo gaett f>te menigo ic cuoeöo iuh soecas i biddas to inngeonganne 7 ne mcehton WSCp: efstaö f> ge gangen fmrh f) nearwe get forjjam ic secge eow manega secad f> hig ingan 7 ne magon Stryue 3ε for to entre by the streit 3ate; for I seye to 30U, many men seken for to entre, and thei schulen not mowe [var. mi3ten not; may not]. Stryue 3ε to entre bi the streite 3ate; for Y seie to 30ε, many seken to entre, and thei schulen not mowe.

[quid hoc audio de te redde rationem uilicationis tuae iam enim non poteris uilicare] Li: huietd 5is ic hero from de agef i forgeld rehto groefscire öines uutedlice foröon ne mceht öu gescira \ WSCp: Hwi ge-hyre ic J>is be f>e. agyf J)ine scire ne miht f>u lencg tun-scire bewitan EV: What here I this thing of thee? 3eld resoun of thi fermer, for now thou schalt not mowe [var. no more] holde thi ferme. LV: What here Y this thing of thee? 3elde reckynyng of thi baili, for thou mijte not now be bailli. Cf. Tatian 108.1 zi hiu gihoriu ih sulih lastar fon thir? gib reda thines ambahtes: iu ni mahtu sculdheizo sin. (5) Lk 20.35-36 [illi autem qui dignihabebuntur saeculo illo et resurrectione ex mortuis neque nubunt neque discunt uxores neque enim ultra mori poterint {var. poterunt)] Li: öa uutedlice öaö[e] wyröo habbaö \ wyröe biöon worulde öaem 7 erest from deadum ne singaö \ ne laedeö \ fatas wifo ne foröon leng i ofer f> deadage / magon

On double auxiliary constructions in Medieval English



öa wutudlice öaöe wyröe habbaö i wuröe bioöon worulde öaer 7 on eriste from deaöe ne synnogaö ne laedas \ ne foas wif ne foröon leng deadiga magan WSCp: Da de synt J>aere worulde wyröe. 7 aerystes of deaöum ne giftigeaj) hi ne wif ne laedaö ne ofer f> sweltan ne magon EV: forsothe thei that be [var. schulen be had; ben; schulen be] worthi to that world, and rysing a3en fro deede men, neither ben weddid, nether wedden wyues, nether schulen mowe deye more LV: but thei that schulen be had worthi of that world, and of the risyng ajen fro deeth, nethir ben wedded, nethir wedden wyues, nethir schulen mowe die more (6) Lk 21.15 [ego enim dabo uobis os et sapientiam cui non poterint resistere et contradicere omnes aduersarii uestri] Li: ic foröon sello iuhmuö 7 aec snyttro öaem ne magon hia wid-stonda i 7 wid-cuoeda \ ongeaegn Alle wiöiwordas \ fiondas iura Ru2: ic foröon selo iow muö 7 snytru öaem ne magun hia giondsworia 7 wid-cweoda alle wiöer-worda iowre WSCp: ic sylle eow muö 7 wisdom. {)am ne magon ealle eower wiöer-winnan wiö-standan 7 wid-cwedan EV: for I schal 3yue to 30U mouth and wysdom, to whiche alle 30ure aduersaries schulen not mowe ajenstonde, and ajenseye LV: for Y schal 3yue to 30U mouth and wisdom, to whiche alle 30ure aduersaries schulen not mowe ajenstonde, and ajenseie Cf. Tatian 145.8 ih gibu tu mund inti spahida, theru ni mugun uuidarstantan inti uuidarqucedan alle iuuuere uuidaruuerton These six examples are all in negative constructions. In MED, examples of the affirmative construction (e.g., Orm 3944 shollde mujhenn wel Upp cumenn) are cited as well as those of the negative construction (e.g., Orm 7301 Ne shall peer mujhenn mtzlenn).2 Lilo Moessner (1989: 31) pointed out an earlier example quoted in her book; it is neither in the negative nor in the combination of shal and mowe, (and for this reason I forgot to include it in my article), but exactly a double auxiliary construction, i.e.,3


Michiko Ogura

(7) Poema Morale (Lamb) 332 Mest alle men 3iued drinke. of one deofles scenche he sceal him cunne sculde wel. 3if he him nele screnche (Most all men it gives drink of a devil's draught; He shall be able to shield himself well if he will not shrink.) The Poema Morale (Lamb) (al225(?cl 175), Southwest Midland) can be properly called a work in the transitional period and too late to be called Old English. According to the MED order of citations, Orm precedes PMor (Lamb). This example, however, suggests the possibility of double auxiliary combinations in the last stage of Old English. (This combination of shal and cunne is also found later in Kentish, Ayenbit of Inwit 70/27: Lyerne to sterue; panne sselt pou conne libbe. It is not difficult to find examples of double auxiliary constructions in Middle English. I carefully avoid using the term 'double modals', because in almost all instances one of the auxiliaries represents tense. Here I reproduce the table, which I gave in my previous article, with examples (Ogura 1993: 547).

(a) (b) (c) (d) (e) (f) (g)

AUX 1 shal shal let shal shal wolde shal

AUX 2 con/gan/do do/make/ger/let make wilne dare mot mowe

texts Cursor 2570 Cursor 17434 Cursor(F) 5323 John(EV) 15.7 Cursor 22603 Cursor{T) 10290 Mark(EV) 3.26

(8) Cursor 2570 C: Namar sal J)ou jsam cun rede J)an sterns on light and sand in see F: na mare saltow ham con rede. |>en sternes of heyuen: and sande of see (9) Cursor 17434 C: You iesu all sal ger tru in him F: Iesu shalle make alle to hym bow (10) Cursor 5323 C: t>e king J?an did his lettres write F: J)e kinge lete make his littres write

On double auxiliary constructions in Medieval English


(11) Jn 15.7 [Si manseritis in me et uerba mea in uobis manserint quodcumque uolueritis petetis et fiet uobis] EV: If 3e schulen dwelle in me, and my wordis schulen dwelle in 30U, what euere thing 3e schulen wilne, 3e schulen axe, and it schal be don to 30U. LV: If 3e dwellen in me, and my wordis dwelle in 30U, what euer thing 3e wolen, 3e schulen axe, and it schal be don to 30U. (12) Cursor 22603 C: Saint Petre sal be dumb Jrat dai, f>at he a word ne sal dur speke, F: saint peter sal be doumbe J)at day. J)at he a worde dar no3t speke (13) Cursor 10289-90 F: ffro chirche he went for J>at shame ffor shame ne myjt he go hame T: Fro chirche he went for J>at shome For shame wolde he mot gon home Aux 1 is a tense auxiliary in most instances. Aux 2 in (a) is pleonastic, since 'con/gan/do + Inf functions as a simple infinitive. Aux 2 in (b) and (c) is a causative auxiliary; let is a confusion of causative and preterit auxiliaries. Aux 2 in (d) and (e) are half-modals. As I quoted in Ogura (1991: 65), there is one possible prototype of (d) in Old English, i.e., (14) GD 61.11 [Non ergo voluit Dominus quidquam fieri, et minime potuit; sed quid velle ejus membra debeant, quidve de eis etiam nolentibus fiat, doctrinae magisterio exemplum dedit.] C: nis hit na f>aet, Petrus, J)aet drihten wolde aht swylces beon 7 ne mihte beon, ac he sealde J>a bysene mid J>y lareowdome his sylfes lare, hwaet his leomu sceolan wilnian Jjaet we waeron, oööe hwaet hi sceolan nyllan. Cf.H: eornostlice drihten nolde na, {>aet aeni J)in3 swilces 3ewurde 7 hit na beon ne mihte, ac he sealde bysene mid lareowdome his sylfes lare, hwaet his limu scylon jewilnian oööe hwaet eac be him nellendum jewurdan sceoldon. (It is not, Peter, that the Lord wanted any of such things to happen or it could not be [so], but he gave the parable with the instruction of his own teachings, [telling us] what his limbs should desire for us to be, or what they should not.)


Michiko Ogura

In (f) mot might be interpreted as an infinitive. If (g) can be considered as a caique of Latin, in which Aux 1 denotes the future tense and Aux 2 the meaning, the combination should seldom be found in ordinary English. MED cites, the shal (not) mowe construction from the works later than Orm: al325 (cl250) Gen & Ex 1818 and 2090, cl350 MP Psalter 17.42, α 1375 WPal. 1933 — all came from South-East Midland. Much later it is found in WBible, Trevisa, and Chaucer. The distribution of the construction tells us its widespread use in the region where the Middle English standard variety emerged. MED also suggests the diversion of the sense of mouen: under 10(d) 'shal be permitted to do, be or have sth.)', Orm 3944 and WBible(l)(Bod959) Judg. 11.2 are cited, while the rest of the examples are under 10(b) 'shal be able (to do, be, or undergo sth.)'. I may add more combination, Aux 1 may + Aux 2 will, which is in MED, i.v. mouen 10(a), quoted by Hopper-Traugott (1993: 46) (though they cut the variant reading and insert their translation, instead). (15) 1434 Misyn ML 128/8 No-J)ing to hafe is sum-tyme of need, bot no3t to may will haue is of grete vertew. [Rolle ML; but no-thyng to desire is grete vertu]. Here may can be said to function as a periphrastic future, but I should rather interpret 'not to permit oneself to desire'. The double auxiliary construction seemed natural in Middle English, though the use was limited in some works. The combination of two (or more) auxiliaries appears in Modern American and Scottish English, as well as Old Norse and Modern Dutch. At this point we must compare the element order of auxiliaries of each language.5 ON: MnDu: ME: ScotE: AmE:

munu skulu 'shall be probably' zouden moeten hebben kunnen 'should have been able to' (N.B. The order may vary.) sceal cunne, shal möwe, may will might can/could, must can/could, should can/could, will/would could might could/should/ought/would

The example in Poema Morale is the closest to Scottish should can. The interlocking semantic field of late OE cunnan and magan could make the combination shal mowe possible as well as sceal cunne. The ME shal (not)

On double auxiliary constructions in Medieval English


mowe may not be the direct source of Modern English varieties, because in some combinations Aux 1 does not always represent tense. But the order 'Auxl (tense) + Aux 2 (modal, etc.)' seems natural in Old Norse and Middle English, and might have been natural in Old English, even though we have not found any example. As long as the modal auxiliaries (or preterit-present verbs and willari) had their full meaning, they could hardly be put in apposition. The only possible OE example is (15). The grammaticalization of Aux 1 may probably have started around late 10th to early 11th century, when there would be no appropriate texts to use the combination.

Notes *

This revised, supplementary version does not replace Ogura (1993), but such useful comments have been made on it for these two years that it seems a good time to summarise those comments and findings. 1. I have used the following editions: Eduard Sievers (ed.), Tatian (1872; rpt. Paderborn: Ferdinand Schöningh, 1966), W.W. Skeat (ed.), The Gospel according to St. Matthew, St. Mark, St. Luke, and St. John (1871-87; rpt. Darmstadt: Wissenschaftliche Buchgesellschaft, 1970), and J. Forshall and F. Madden (eds.), The Holy Bible ... by John Wycliffe and his Followers (Oxford: Oxford University Press, 1850). For the abbreviated titles of Old English works I follow Bruce Mitchell, Christopher Ball, and Angus Cameron, 'Short Titles of Old English Texts', in Anglo-Saxon England 4 (Cambridge: Cambridge University Press, 1975), 207-21, and 'Addenda and Corrigenda', ibid. 8 (1979), 331-3. For Middle English works I follow the MED abbreviations.

2. See MED s.v. mouen and R. Holt (ed.), The Ormulum, with the Notes and Glossary of R.M. White (Oxford: Clarendon Press, 1878). 3. I use the edition and translation by R. Morris, Old English Homilies. First Series, EETS OS 29 & 34 (1886; rpt. New York: Greenwood Press, 1962). 4. MED s.v. shulen 16(a). 5. I have taken the Old Norse example from Cleasby-Vigfusson-Craigie (1982), the Modern Dutch example from a private communication with Dr. Willem F. Koopman (Amsterdam), the Scottish English example from Keith Brown (1991) and the American English example from di Paolo (1989).


Michiko Ogura

References Brown, Keith 1991

"Double modals in Hawick Scots", in: Peter Trudgill and Jack K. Chambers (eds.), 74-103. Cleasby, Richard - Gudbrandr Vigfusson - Sir William A. Craigie 1957 [rpt. 1982] An Icelandic-English dictionary, 2nd ed. with a supplement by Sir William A. Craigie. Oxford: Clarendon Press. Hopper, Paul J. - Elizabeth C. Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Kurath, Hans - Shermann Kuhn, et al 1952Middle English dictionary. Ann Arbor: University of Michigan Press. Moessner, Lilo 1989 Early Middle English syntax. Tübingen: Max Niemeyer. Mustanoja, Tauno 1960

A Middle English syntax. (M6moires de la Societe Neophilologique de Helsinki 23). Helsinki: Sociite N6ophilologique. Ogura, Michiko 1991 1993

"Periphrases with OE and ME verbs". Chiba Review 13: 37-70. "Notes: Shal (not) mowe, or double auxiliary constructions in Middle English". Review of English Studies n.s. 44: 539-548. Paolo, Marianna di 1989 "Double modals as single lexical items". American Speech 64.3: 195-224. Trudgill, Peter - Jack K. Chambers (eds.) 1991 Dialects of English. New York: Longman.

The development of Middle English ι in England: A study in dynamic dialectology Mieko Ogura — William S-Y. Wang - L. Luca


1. Theoretical preliminaries The language differences that separate individuals and those that separate dialects must basically be of the same sort. This fact was generally appreciated by nineteenth-century linguists, from Hermann Paul to Hugo Schuchardt. Thus, Schuchardt wrote, in his famous study, "Therefore, everything that holds true for the relationship between dialects on any level also holds true for the relationship between idiolects" (1885; quote from 1972 version, p. 49). Conversely, principles at work within speech communities may also be found to operate dynamically across dialects. Although a great deal of valuable data has been gathered over the past century on a variety of languages, not many attempts have been made to build on the aforementioned insight. Dialect maps have remained rather largely static depictions of the pronunciation of individual words or of the geographic extent of some linguistic trait.1 For lack of an adequate theory with which to interpret these data, it has been difficult to relate the crosshatching isoglosses to larger scale historical developments. As a result, although the comparativists proclaim the exceptionlessness of sound change, the dialectologists insist that each word has its won history, a paradox that has been clarified by Malkiel (1967). There thus exists in our field at present an unfortunate chasm between the study of language in time and the study of language in space. In recent years, several contributions have been made toward solving this problem. In his analysis of the problem, Labov (1981) attempts to sort out various types of linguistic change, according to phonetic structure, and to distinguish those types that operate regularly (in the Neogrammarian sense) from those that do not. Trudgill (1974, 1975) develops a method that incorporates some of Hägerstrand's ideas and applies it to English and Norwegian data. Bailey (1973) proposes a wave model for spatial diffusion and applies it to some data from the north of England.


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Although the present study has benefited from these earlier investigations, it is closer in approach to those of Barrack (1976) and Hashimoto (1983) in its reliance on lexical diffusion. Within this perspective, it is assumed that a phonological change operates on words severally at a time, rather than on the phoneme per se. In this perceptive review of an anthology of works in lexical diffusion, Hashimoto (1981) suggests that "lexical diffusion can explain not only the internal development of languages, but also developments caused by language contact" (p. 190). A specific hypothesis toward this purpose had been proposed by Barrack when he suggested in 1976 that "a phonological change will have progressed furthest in the lexicons of those dialects closest to the point of origin" (p. 151).2 The proposal may be represented by a hypothetical series of changes, i.e., A > Β > C, as shown in Table 1. Here we see the development at five sites through five stages in time. At the initial stage, all the relevant words at all five sites have the A pronunciation, indicated by the 100% within the parentheses. At the next stage in time, 50% of these words have changed to the Β pronunciation at site 3 and 10% at site 2; the other sites remain unchanged. This process continues, at different rates for different sites, through a series of phonological changes. Table 1. Scheme for lexical diffusion Of A>B>C at five sites through five stages in time

tö t1 t2 t3 t4






A(100) A(100) B ai

(m) (h)



Figure 1. Relations among the 17 reflexes of Middle English I

The 39 words selected for our study are as follows, grouped roughly according to the consonant directly after the long ϊ: DRY, SKY, FIRE, IRON, FIVE, HIVE, IVY, WIFE, KNIFE, TIME, NINE, MINE, ICE,


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

ICICLE, MICE, SLICE, LICE, BESIDE, HIDE, SPIDER, SLIDE, FRIDAY, WHITE, WRITING, SCYTHE, EYE, DIED, FLIES, STILE, THIGH, MIGHT, FIGHT, -WRIGHT, RIGHT, LIGHT, NIGHT, SIGHT, LIGHTNING, and DIKE.4 Many factors led to the conclusion, that the words were all pronounced with a long ϊ (as their primary vowel) in ME times, including consideration of their Old English sources. The philological details can be found in Ogura (1987, Ch. 3) and are not presented here. The sites we used from the dialect materials total 311; these are listed in Appendix 1. Figure 2 shows the locations of these sites. For a particular word, if a site shows more than one pronunciation, we choose the first one recorded. If no pronunciation is recorded at a site for that word, we interpolate its pronunciation on the basis of neighboring dialects (5.9% of our database is interpolated). In this way, our database consists of one pronunciation for each of the 39 words at each of the 311 sites, or 12,129 samples, where each sample is one of the 17 reflexes diagrammed in Figure 1. These samples constitute the body of data in the present study. We next tabulated the frequencies of the different vowel reflexes. The most frequent vowel at a site is called the primary vowel site; the next most frequent reflex is its secondary vowel; the next after that is its tertiary vowel.4a A few sites, such as site 98, which is Diddlebury in Shropshire, have only the primary vowel. All 39 words at this site are pronounced with /ai/. A few others, such as site 97, which is Clun in Shropshire, have only primary secondary vowels. Most sites, however, have three or more reflexes. The three most frequent reflexes of the first ten sites are shown in Table 2 for illustration.5 The complete data on the frequencies of the reflexes for each of the 311 sites are presented in Appendix 2.6

The development of Middle English ϊ in England 2 4 1

1·. J 9 IS


• · ·ΙΙ


: »




3 : -4

17 14·, 19 2T

23 15 24 25 2« 45 4· 4} 44 47 « 4 9 44 . . ' . · 27 » 51.. 52« , „ 55 * 54 53 59 • 29 54 57 1 32 B34 42 « 44 41 41 : 2 14 u 45 44 ·&;ίμ·.·'μ 37 4 9 * 3B 71 77 157 15β "· Β · ; ·39 «D 7 3



• #





113 ®




|«W|73 1 7 1 1 7 4 1JUI * m m i a l n 155 B177 m , e *_ w L ' » I I7 'I* .'24J , „ · . . . m w 199 - V " * l l 8 " » .24 " · « * » „ » : Ä W 128 .27 1 M ' r a B 7 » * » 112 129 214 2 „ ID ^ t l 4 I13t2t 122123,_ 2 1 . τ , . , M l II) 137 I3i . ·_" ? l t w ? \ e 0 3 '"114 . «« " 13213] IJ9J4214· ' « 21» 211 219 227 Φ P 12

° I«

235 252 74» ' 23,237 253262 243 271 .··"» Π»24· ^ 255 299 3 · 2 « 244 272 2>1 244245 » * * 2 « 274 ^ . .' » 24* 247 »7 ' ^ ·' " » 24« 294 J 3 , 1 » * 274 »4 249 3H 311 2SB . 294 295 .'·'·.•··•••«'·''' .·' ^

«2 ·

. •' 2· 3"· · Figure 2. The 311 sites examined in the present study


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Table 2. Three most frequent reflexes of Middle English Fat ten sites in Northern England

Primary Site 1 2 3 4 5 6 7 8 9 10

Stage f f f e f i f f e e

Number 23 23 22 22 25 18 21 21 20 20

Secondary Stage i i i i i e i i i i

Number 11 10 13 10 10 9 10 13 10 18

Tertiary Stage e e e a e f e e a a

Number 3 5 3 4 2 6 4 1 7 1

3. Discussion of maps Turning now to Figure 3, we see the distribution of the primary vowels in England. We note immediately that two vowels dominate the landscape. These are the vowels i (=/aif) and k (=/ai/). In our database of 12,129 samples, they make up 2,725 samples (or 22%) and 2,102 samples (or 17%), respectively. The next most frequently occurring vowel is d (=/ai/) which makes up 11%. Although our data consist of 17 reflexes, these three vowels make up half of the total samples. Although the vowels i and k co-occur to some extent in a transitional zone consisting of Lincolnshire and Yorkshire, they are largely in complementary distribution. As Figure 3 shows, i occurs in the north and the west, whereas k is found mostly in central England, extending to the southeast. Given that the regions with vowel i roughly correspond to the border with Celtic-speaking areas, one is tempted to look for explanations having to do with language contact; however, much detailed investigation is needed before anything tangible can be advanced. From an examination of the distribution of primary vowels, we can discern several regions where the change is more advanced, in the sense illustrated earlier in Table 1. One such region is the coast of Essex, as shown in Figure 4. There the primary vowel has reached q (=/öi/). This focal area is surrounded by sites where the primary vowel is ο (=/oi/), which in turn border with sites where the primary vowel is η (=/λϊ/) or k (=/ai/).

The development of Middle English ϊ in England

Figure 3. Distribution of primary vowels in England



Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

The development of Middle English ι in England

Figure 5. Diffusion wave of the primary vowel from Oxford



Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Figure 6. Distribution of secondary vowels in England

The development of Middle English ϊ in England

Figure 7. Distribution of tertiary vowels in England


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Figure 8. The three stages of diffusion for /ai/.

The development of Middle English ι in England




1377 O v e r JO per


)0 - 40 2 0 - 3 0 10 -


Below 10 No information T o w n j t n d Citiej •

— 1,000 pcrsom

- 5.000

Figure 9. Lay population according to tax returns of 1377 (from Darby, 1936, p. 232)


Mieko Ogura - William S-Y. Wang - L. Luca


A similar situation can be seen around the Oxford area, shown in Figure 5. Here again, the primary vowel has reached q (=/öif) in a focal area that is surrounded by sites where earlier reflexes occur. Another focal area revealed by the primary vowels is around Manchester and Sheffield and yet another is around Birmingham. We may refer to such regions as exhibiting a "gradient," which allows us to make inferences concerning historical developments. Typical of such gradients is the observation that the inner areas are more advanced in the change, either in having a later reflex or in having more words, pronounced with a later reflex. This is in part what we tried to suggest in Table 1. We can do a parallel analysis of the secondary vowels, that is, those vowels that are second in frequency at each of the sites. Figure 6 shows the result of such an analysis. In general, the boundary lines here resemble those found for the primary vowels shown in Figure 3. This resemblance is particularly striking with respect to the southwest region, where the lines run roughly vertically into the English channel. We refer to this feature later. Figure 7 is based on the analysis of tertiary vowels. We see that no tertiary vowels are represented in a large region of the lower half of England. This indicates that at the sites of this region the variation is less than elsewhere in the country. The primary and secondary vowels in this region, as shown in Figures 3 and 6, are mostly m (=/dif), η (=/λϊ/), and ο (=/oi/). Whereas the 39 words studied here are usually pronounced with more than two different reflexes elsewhere, they are pronounced with greater uniformity in this region. Our study of the development of ME ς, again shows greater uniformity in this region (Ogura 1990, Ch. 2). Whether the greater uniformity in this region is due to a higher degree of interaction among these sites is a question that should be investigated. Figure 8 represents another type of picture we can construct from the data shown in Appendix 2. It is a map of the most frequent vowel, that is, the reflex i (= /ai/). The P, S, and Τ in the map mark the sites where the reflex i is the primary vowel, secondary vowel, or tertiary vowel, respectively. Again, a certain amount of gradient is to be seen in this type of map. It is clear that the reflex i is primary in four noncontiguous areas, separated from each other by sites at which it is not primary. As the changes we are studying have their origin many centuries ago, it is useful to refer to the conditions found in England during that period. One useful indicator available to us is the map constructed by Darby, which we reproduce as Figure 9. This map gives the population density according to the poll tax returns of 1377, for lay population over 14 years of age. The

The development of Middle English ι in England


two major black areas, where the population was densest, have an interesting relation to the diffusion waves we portray in Figures 4 and 5. The origin of the wave in Figure 4 is near the coast in Essex and that in Figure 5 is near Oxford. Both waves spread to London and to the north. The direction of the change to London in Figures 4 and 5 correlates with the population movement. The historical evidence bears out this hypothesis in a remarkable way. In Domesday Book, Norfolk and North Suffolk were the most densely populated areas. The land was more valuable than the hilly country to the north and west, and in an agricultural age this advantage was reflected in both the number and the prosperity of the inhabitants. From names in taxation lists, we learn that immigration into London was highest from Norfolk and North Suffolk, with Essex and Hertfordshire next in the earlier ME period (Ekwall, 1951). This agrees with the direction of the change from Essex to London.7 Then in the mid fourteenth century there was a change in proportion of the population. The population of Leicestershire, Northamptonshire, and Bedfordshire had increased as Figure 9 shows, the immigrations to London from the latter two counties were equal to those from the East Midlands (Darby, 1936; Ekwall, 1956); and the East Midland coloring of the London dialect was largely eradicated by an influx of Central Midland features.8 Figures 4 and 5 show that the wave to London is stronger from the Central Midlands than from the East Midlands. The wave from Oxford is mainly toward London, but that from Essex is toward the northwest. According to Samuels (1972, Section 8.5), the Norfolk dialect was peripheral and unsuited for the role of lingua franca, whereas the Central Midland dialect, which was the most widely and readily intelligible dialect, fulfilled that role to a greater extent than other Midland dialects. This may explain why the Central Midland influence on the London dialect is stronger than the East Midland influence.9 As mentioned earlier, we have also observed, in our data, diffusion waves from Manchester and Sheffield and from Birmingham. The Modified Standard Type spoken in the larger towns such as Manchester, Sheffield, and Birmingham may become so well established that each of these may form a starting point from where the wave spreads over areas coextensive with their social and economic influence (Wyld, 1936, p.7). The main direction from Manchester and Sheffield is to the north, and that from Birmingham is to the south and west. We suppose that in the north of England the process of the development of ME / that goes through the /aei/ stage originates in Manchester and


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Sheffield; the development that goes through the /oi/ stage originates in the central part of northern England; and the development that goes through the /si/ stage originates in the area farthest north. Today, the wave northward from Manchester and Sheffield merges with the waves that radiate from the central part of northern England and from the area farthest north. And in the central area of northern England, these three processes of the development are competing. Moreover, the road that goes through York, Durham, and Newcastle was one of the important channels in the sixteenth-century communications (Darby, 1936, p. 342). Thus this road may have facilitated the network of oral communication and accelerated the change in the eastern part of northern England. The early spellings do not make it clear in which area the diphthongization started. Samuels (1963) gives examples dating from the early fifteenth century from East Anglia. From our maps we may assume that the diphthongization probably started around this area. We are not sure whether the change was monogenetic or polygenetic. One possibility is that the change began around Essex and then spread to Oxford, to Birmingham, and to several other places in the north of England, hopping among urban centers. The other possibility is that the change occurred independently around Essex, Oxford, Birmingham, and several places in the north of England at different times. Our maps show that the rates of the propagation of the waves are different. The waves from Essex and Oxford show more advanced stages than those from Manchester, Sheffield, and Birmingham.10 In the southwestern part of England most of the boundaries between stages run vertically, as we noted earlier for both primary and secondary vowels. It seems in many cases rivers have acted as barriers to the spread of the pronunciation, as can be seen in Figure 10. These rivers do not always coincide with county boundaries; thus, it is mostly the rivers that are responsible for the linguistic separation.11 They are the opposite cases of the modern form spread southward along the Rhone in France (Bloomfield, 1993, Section 19.5), and the penetration of the Standard German ims-form into the Mi-territory in Low German along the natural road formed by the Rhine valley (Palmer, 1972, Ch. 12). In these cases rivers have facilitated rather than reduced the spread of the changes.

The development of Middle English ϊ in England

Figure 10. Rivers as linguistic boundaries in the southwestern countries



Mieko Ogura - William S-Y. Wang - L Luca Cavalli-Sforza

4. Actuation of the diffusion of Middle English fin England Individual words do not have the same rate of change. This has already been pointed out for historical change and is likely to be true for geographic variation also, because variations in time and space are closely related phenomena. Then, in the process of diffusion, which words are the leaders and which ones are the laggers? Appendix 3 shows the patterns of diffusion in the development of ME Γ for each word. The phonetic environments where the vowel occurred in ME and the frequency counts of the modern words are also given. The frequency figures are based on the American Heritage Word Frequency Book (Carroll et al., 1971).12 In Appendix 3, each word is shown together with the frequencies of the reflexes at each of the 17 stages. For example, DRY is pronounced with the reflex d at 33 sites, with e at 3 sites, with g at 7 sites, and so on. Thus each row should sum up to 311, the total number of sites. We may compute a score for each word by simply summing the products of stage number and frequency. So the score for DRY would be (33 χ 4) + (3 χ 5) + (7 χ 7) + .... As the assumption is that the change progressed approximately in the manner shown in Figure 1, that is, a is the earliest reflex and q is the latest reflex, is follows that the larger score a word has the later it is in the chain of phonetic changes. This is to say, the words with the larger scores are the leaders and those with lesser scores are the laggers. On the basis of the scores computed in this manner, the following conclusions may be drawn. Words with final Γ are leaders; words in which the Γ precedes velars are laggers; the remaining words are intermediate. For words in this last class, there is also a clear tendency that words in which the post-i consonant is voiced have a lead over those in which the consonant is unvoiced. We have also tried to correlate the scores of the individual words with the frequency figures shown in the third column of Appendix 3. As Phillips (1984) has demonstrated, there may be a correlation within each narrow phonetic class between word frequency and schedule of change. No such tendency can be discerned, however, in the present data.

5. Correlation analysis In Table 3 we present a table of correlations between all possible pairs of the 17 reflexes. The correlations were computed after transforming the

The development of Middle English ϊ in England


observed occurrences into binary variables, with value 0 if a given reflex is absent at a given site and value 1 if it is present. This transformation minimizes the effect of the absolute number of occurrences, which otherwise would tend to create negative correlations, given that all occurrences at any one site sum to the total of 39. The majority of correlations in Table 3 are near zero, positive or negative. A minority, however, are clearly significantly different from zero. Of the total 136 correlations, there are 27 positive and 26 negative values outside the limits of significance for ρ = 0.05 (r = ±0.11). One might expect that pairs of reflexes that are neighbors to each other in development (i.e., joined by single arrows in Figure 1) should have positive correlations. In fact, if they truly developed from each other, and the change is still operating in some areas, then the tendency of both reflexes to present at the same sites should generate a positive correlation. Table 3. Correlation Table of the 17 Reflexes of Middle English ϊ a





0.25 1.00









0.02 0.04 1.00


-0.15 0.15 0.09 1.00


0.65 0.26 0.11 -014 1.00


0.28 0.21 0.20 -0.01 0.28 1.00


0.10 0.02 0.05 0.06 0.04 0.07 1.00


-0.13 -0.08 -0.04 -0.11 -0.16 -0.05 0.21 1.00


0.46 0.20 0.04 -0.13 0.59 0.21 0.14 0.16 1.00









0.30 0.00 -0.06 -0.09 0.20 0.01 0.25 0.32 0.30 1.00


-0.01 -0.11 -0.11 -0.26 -0.05 -0.17 -0.21 -0.09 -0.11 -0.08 1.00


-0.02 -0.12 -0.07 -0.23 -0.06 -0.07 -0.13 0.04 0.04 0.05 0.40 1.00


-0.30 -0.19 -0.10 0.04 -0.27 -0.15 -0.16 0.10 -0.07 -0.09 0.02 0.07 1.00


-0.14 -0.11 -0.06 -0.11 -0.27 -0.09 -0.10 -0.07 -0.36 -0.11 0.10 0.05 -0.05 1.00


-0.21 -0.08 -0.09 -0.18 -0.37 -0.13 -0.11 -0.10 -0.37 -0.16 0.20 0.13 -0.07 0.03 1.00



-0.01 -0.05 -0.02 -0.06 -0.09 -0.03 -0.04 -0.02 -0.12 -0.04 0.07 0.19 0.06 0.23 0.02 1.00


-0.14 -0.10 -0.06 -0.12 -0.25 -0.08 -0.09 -0.07 -0.34 -0.11 0.03 0.01 -0.10 0.29 0.46 0.15 1.00 a

















Testing this expectation in Table 3, we find that of the 19 pairs of neighboring reflexes (determined from Figure 1), 10 pairs show a


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

significant positive correlation. This is encouraging, but two observations remain that need to be discussed: 1. Most of the highly significant correlations involve pairs of reflexes that are nonneighbors, that is, not connected by a single arrow. For instance, the following pairs of reflexes have highly significant positive correlations (at the ρ = 0.001 level, r = ±0. 19): a,e=.65 a,f=.28 a,i=.46 a,j=.3Q

b,e=.26 b,f=.21 c,f=.20 e,j=.59

e,j=.20 g,j=.25 h,j=.32 l,p=.19

n,p=.23 o,q=.46

The preceeding correlations between nonneighbors are too numerous to be discarded as unimportant. There could be two explanations for them. The first is that some of the intermediate reflexes are phonetically less stable, in either production or perception. This instability makes such reflexes shortlived and therefore less attested in the data. This explanation would suggest that the unstable reflexes should be rarer across the database. Alternatively, we should also consider the possibility of multistep changes in the vowel space alongside the singlestep changes diagrammed in Figure 1. This hypothesis would complicate considerably theories of sound change, and should require strong independent support before it can be accepted. Nevertheless, some of the correlations between nonneighbors in Table 3 strongly indicate this possibility. 2. The other observation from Table 3 has to do with the fact that some correlations are significantly negative. The correlation between i and k, for example, is -.11 and close to ρ = 0.05, even though the two reflexes are neighbors. In principle, a negative correlation may arise if two reflexes had originated in different regions and propagated as independent waves that have as yet not reached the same sites, because of the distance between the origins or for some other reason. Whether this is indeed what happened can only be verified by more detailed historical investigations.13

Addendum (1996) After the above discussion, there is an exchange between Labov (1992, 1994, Chap. 17) and Ogura (1995) on the neogrammarian regularity an lexical diffusion in English dialect geography. Based on the data in

The development of Middle English Γ in England


Appendix 3 in the present study and Ogura (1987, Chap. 3, Apps. C and E), Labov (1992, 1994, Chap. 17) reanalysed the distribution of ME ϊ and ü words at 311 sites in England by chi-square test, multiple regression and multi-dimentional scaling, and maintains that the mathematical analysis of dialect distributions supports the regularity hypothesis as well as the claim of phonetic conditioning of sound change. Ogura (1995) examines the validity of Labov's claims. Since Appendix 3 in the present study and Appendices C and Ε in Ogura (1987) do not retain spatial information, there are many instances of lexical diffusion which do not show up in Labov's chi-square test. Ogura compares a given pair of ME Γ and ME ü words by counting the number of sites where the pair of words is pronounced differently. Ogura shows, contrary to Labov's conclusions, that the spatial distribution of the words through sites, which reflects the diachronic development of the words, is strongly suggestive of lexical diffusion and that phonetically conditioned change is not necessarily considered as evidence of the regularity of sound change. Ogura (1995) and Ogura & Wang (1996a, b) have further developed a 2dimensional diffusion model: diffusion from word to word in a single speaker or at a single site, which we call W-diffusion, and diffusion from speaker to speaker, or from site to site of a single word, which we call Sdiffusion. When W-diffusion is slower than S-diffusion, the difference is greater between words. When W-diffusion is faster than S-diffusion, the difference is greater between speakers or sites. W-diffusion may proceed so fast that it is difficult to observe it. This shows what is called the neogrammarian regularity. Our model synthesizes the neogrammarian conception of the regularity of sound change and the dialectologist's conception that each word has its own history. We would like to make these processes more precise based on historical and on-going changes in English in the future study.

Notes *

This is ail updated version of the paper originally published in New ways of analyzing sound change, ed. by Penelope Eckert, 1991, pp. 63-106, New York: Academic Press.

1. Two notable exceptions are the linguistic atlas of Gascogne, compiled by S6guy (19541973) and the atlas of English sounds, compiled by Kolb et al. (1979). In Vol. 6 of the Gascogne atlas, lexical frequencies is utilized to show gradients of change in a manner similar to that used in the present study. We thank Jerry Craddock of the University of California at Berkeley for calling our attention to S6guy's work. The atlas of English


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

sounds is based on data reported in the Survey of English dialects. In addition to the maps of isoglosses of individual words, it contains "collective maps," which show the average representation of the pronunciation of small groups of selected words. It would be interesting and useful to contrast the various methods used for compiling such atlases with respect to their visual impact and degree of informativeness on historical processes. Such work, however, falls beyond the scope of the present study (cf. Macaulay 1985). 2. Barrack (1975) supposes, on the basis of the degree of lexical diffusion, that back umlaut of *e in Old English occurred first outside West Saxon and then hit he West Saxon regions as successive waves, eo regularly appears as the reflex of *e before dentals and nasals in all dialects but at a very much reduced degree η West Saxon. Barrack (1976) also computes the degree of diffusion of the devoicing phase of the High German Consonant Shift as manifested in representative documents from both the north and the south. The data exhibit a higher degree of lexical diffusion in the south than in the north. This suggests that he change occurred first in the south and spread to the north. This interpretation supports the traditional view that the High German Consonant Shift spread northward from the south. 2a. Ogura (1990, Chapter 2) and Ogura & Wang (1990b) examine how exact the transcriptions in the Survey of English dialects are and develop a method by which we can judge whether the transcriptions by the fieldworkers are fieldworker isoglosses or not, especially when we are not familiar with areas. 3. The exact process whereby ME ϊ developed into /ai/ is controversial. Orton (1933) suggests a development of ME Γ for Northern English on the basis of its correspondences in the living dialects. He supposes that centralization did not take place until lowering had been completed. He suggests that current recorded variant /ai/ may be an imitation of Standard English or may have branched off from /ei/ or /ei/. Similarly, Ellis (1869-1871, Part V, pp. 926-827) supposes a development in which there is no intermediate centralization. He also notices the existence of /si/, but he is unsure how to fit it into the line of development. Wyld (1936, p. 225) proposes that the /ai/ pronunciation descended either from the line of the development suggested by Orton or from the process that branched off at the /ei/ stage. Luick (1914-1940, pp. 559-581) and Jesperson (1909, Section 8.21) suggest that ME Γ must through hi/ have lowered to /ei/, and from /ei/ or /ei/ the diphthong probably centralized to some kind of hi/ and then lowered again. Dobson (1968, Section 137) argues against the possibility of lowering before centralization, saying that if the development had been that suggested, ME Γ would have crossed the path of ME ai developing to /aei/ and /ei/. He assumes that all sound changes must occur gradually, and ME Γ and ai must have followed the same course after the merger. Thus, he proposes that ME Γ first became Ν and then became hi! or some closely related diphthong, and lowered. Kökeritz (1953, p. 216) adopts the same view as Dobson. Stockwell (1972) also presents a similar view. He agrees with Dobson in

The development of Middle English ϊ in England


rejecting the possibility of lowering before centralization on the basis of the merger of ME Γ and ai. He claims that ME / must have gone through the stages /iy > iy > ay > ay/, in which centralization occurred before any lowering took place. Chomsky and Halle (1968) adopt the same view as Luick and Jespersen. In generative phonology, however, even if the surface phonetic forms of ME Γ and ai had merged, the underlying forms must still have been distinct, so that it would be possible to explain the change by the addition of the rule to the grammar. They claim that the rule, which makes the high vowels mid and the mid vowels high, was added to the grammar. Stockwell objects to the use of the exchange rule on the grounds that the addition of such a rule to a grammar would result in an impairment of intelligibility among speakers who adopted the change and those who had not, or that the sounds involved in the exchange would merge. There is the evidence from orthoepists, rhymes, and occasional spellings. Hart (1569), Baret (1573), Minsheu (1599), Robinson (1617), Gil (1619), Sherwood (1632), Daines (1640), Gataker (1641), The English Schole-master (1646), Hexham (1647), J. Smith (1674), and Richardson (1677) show diphthongization by the transcription ei, or by statements that ME / was like Latin ei or Greek ei or foreign ei (Dobson, 1968, Section 137). Dobson argues that their view is impossible, because, if the development had been that suggested, ME Γ would have crossed the path of ME ai. But a straight reading of the orthoepists would favor the view that the vowel was /ei/, /ei/, or /aei/, and in terms of lexical diffusion, even if ME Γ merged partially with ME ai, the class of words that originally had ME Γ could still change to /ai/. Thus we may suppose that the orthoepists show the diphthong /ei/, /ei/, or /aei/, which are the stages of the three types of the development. Wallis (1653) describes ME ϊ as /ai/, Wilkins (1668) and Cooper (1685) as /ΛΪ/, and T. Jones (1688) as /ai/ or /ΛΪ/ (Dobson, 1968, Section 137). We can say that their hi/ or /ΛΪ/ indicates the /ai/ or /ei/ stage of the third or the second type of the development. [The rhymes between ME / and oi such as die:joy:annoy, joys:price, joy:eye, poise:advice, etc. (Shakespeare) (Kökeritz, 1953, p. 217), point to the /ai/ or /ei/ stage, and the rhymes between ME (and ei such as side:leyde, by:wey (Ed. Eth. about 1420), paradise:peyse 'weight' (Lydg.), etc. (Jordan, 1934, Section 279), show the /ei/ stage. And such rhymes as fly:he, me, eye:knee, die:me, thee:three (Shakespeare) (Kökertitz, 1953, p. 220) indicate that the diphthongization had not yet taken place before velars. The occasional spellings with ey, ei as in geiding, abeid, deifyrs 'divers', ei Τ (The hymn to the Virgin, ca. 1500), Eylle of Wyght, trey 'try' (Sir Thomas Seymour, 1544), feyre (Machyn, 1550-1553), etc. (Wyld, 1936, p. 223), point to the /ei/ or /ei/ stage, whereas the occasional spelling in smoile 'smile' (Shakespeare) (Kökertitz, 1953, p. 217) shows the /ai/ or /ei/ stage. 4. Bind, blind, climb, and find are not included here, because in the North they show a short vowel in which points back to ME i.


Mieko Ogura - William S-Y. Wang - L. Luca


4a. We called the primary, secondary, and tertiary vowels the first mode, second mode, and third mode reflexes respectively in Ogura (1987, 1990) and Ogura & Wang 1990a, 1990b). 5. The frequency of occurrence of the primary, secondary, and tertiary vowels in the whole data (i.e., 39 words χ 311 sites = 12,129) is 9,362, 1,829 and 608 times, respectively. 6. We include the /i:/ pronunciation in mice and lice at some sites in Kent, Essex, and Suffolk in our data on ME ϊ words in Kent, Essex, and Suffolk in our data on ME ϊ words for convenience sake, but in fact the /i:/ pronunciation at these sites comes from ME e. 7. The raising of ME ς in London was accelerated by the great influx of population into London from the eastern regions where the advanced pronunciation was more common. 8. Though the most populous area in the Central Midlands was not around the focal point of the wave but on the north of it, we may assume that there were a great number of migrants from around the origin of the wave after the midfourteenth century. 9.

Hashimoto (1983) examines the direction of linguistic developments in the East Asian continent, based on the lexical items of the Chinese languages related to eating and drinking, 'cooking vessles', '(cooked) dishes', and the verb 'to drink'. The diachronic variants or these words appear mainly in the order of the synchronic distribution from the southeastern to the northwestern corner of the continent. He assumes that this direction of change may have something to do with the successive invasions or migrations of the northerners. However, the direction of the High German Consonant Shift mentioned in Note 2 has been in the opposite direction from that of migration. We do not yet know how general is the correlation between the direction of the linguistic change and the migration routes.

10. Dobson (1968, Section 137) supposes that in the north the development was probably more rapid, and the /ai/ stage was reached by 1600, and in careful Southern speech the /ai/ was not heard before 1700. But these remarks are questionable. 11. Palmer (1972, p. 282, fn. 1) quoting Dauzat's remark, mentions that a river, if it is crossed by bridges and easily navigable, is not a hindrance to communication, and consequently it does not form a linguistic boundary. If, however, it is broad and bridges are rare, it becomes a linguistic boundary. Thus, the lower Loire in France forms a clearcut dialect boundary, and the Allier, which was not bridged anywhere in the department of Puy-de Dome until 1830, separates two widely diverging linguistic areas. He also cites the river Whorfe as the dividing line between the northern and Midland development of OE a. 12. The American Heritage wordfrequency book gives estimated frequencies per million tokens. 13. We have done parallel analyses of the reflexes for ME f, e, ä, u, ö and q and have found that these sound changes conform to each other to a high degree spatially. We have developed a three-dimensional maps to display dynamic patterns that reveal the waves of propagation. For details on the spatial distribution of the Great Vowel Shift in England, see Ogura (1990, Ch. 2).

The development of Middle English ϊ in England


We have also applied the methods of dynamic dialectology to the development of the Indo-European languages, consonantal and semantic changes in the dialects of Chinese and language contacts in the history of English. These developments appear to conform well

with diverse forms of evidence: social, cultural, genetic,


demographic, and so forth. For details, see Ogura (1990, chapters 3, 4, and 5) and Ogura - Wang (1990a).

References Asher, R.E. - Y.M.Y. Simpson (eds.) 1993

Encyclopaedia of language and linguistics. Oxford: Pergamon Press.

Bailey, Charles-James N. 1972

"The integration of linguistic theory: Internal




comparative method in descriptive analysis", in: Robert Stockwell - Ronald K. S. Macaulay (eds.), 22-31. 1973

Variation and linguistic

theory. Arlington, Virginia: Center for Applied

Linguistics. Barrack, Charles M. 1975

A diachronic phonology from Proto-Germanic to Old English stressing WestSaxon conditions. The Hague: Mouton.


"Lexical diffusion and the High German consonant shift", Lingua 40: 151175.

Bloomfield, Leonard 1933

Language. London: George Allen & Unwin.

Board, C. et al. (eds.) 1975

Progress in geography. Vol. 7. London: Edward Arnold.

Britton, Derek (ed.) 1996

English historical linguistics 1994. Amsterdam: John Benjamins.

Brogyanyi, Bela (ed.) 1990

Essays in linguistics offered in honour of Oswald Szemerenyi on the occasion of his seventy-fifth birthday. Amsterdam: John Benjamins.

Carroll, John B. et al. (eds.) 1971

American Heritage word frequency book. New York: Houghton and Mifflin.

Cavalli-Sforza, L. Luca - Mark W. Feldman 1981

Cultural transmission

and evolution. Princeton, N.J.: Princeton University

Press. Cavalli-Sforza, L. Luca - William S-Y. Wang 1986

"Spatial distance and lexical replacement", Language 62: 38-55.


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Cavalli-Sforza, L. Luca - Paolo Menozzi - Alberto Piazza 1994 The history and geography of human genes. Princeton, N.J.: Princeton University Press. Chambers, Jack K. - Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. Chen, Matthew Y. 1972 "The time dimension: Contribution toward a theory of sound change", Foundations of Language 8: 457-498. Chen Matthew Y. - William S-Y. Wang 1975 "Sound change: Actuation and implementation", Language 51: 255-281. Chomsky, Noam - Morris Halle 1968 The sound pattern of English. New York: Harper & Row. Darby, H. C. (ed.) 1939

An historical geography of England Cambridge University Press. Dobson, Eric J.

before A.D.



1968 English pronunciation 1500-1700. (2nd edition.) 2 vols. Oxford: Clarendon. Ekwall, Ellert 1951 Two early London subsidy rolls. Lund: Gleerup. 1956 Studies on the population of Medieval London. Stockholm: Almqvist & Wiksell. Ellis, Alexander J. 1869-1871 On early English pronunciation. [ 1969] [Reprinted New York: Haskell House.] Gabriel, K. R. 1971

"The biplot-graphic display of matrics with applications to principal component analysis", Biometrica 58: 453-467. Gerritsen, Mariel - Frank Jansen 1982 "The interplay between diachronic linguistics and dialectology: Some refinements of Trudgill's formula", in: J. Peter Mäher et al. (eds.), 11-37. Hägerstrand, Τ. 1952

"The propagation of innovation waves". Lund Studies in Geography. Series B. Human Geography 4. Lund: Gleerup. Hashimoto, Mantaro J. 1981 1983

Review of The lexicon in Phonological Change, Language 57: 188-192. "'Pan', 'dish', 'drink' in Chinese: A case study of longitudinal and latitudinal developments of languages", Journal of Chinese Linguistics 11:1 -35. Hickey, Raymond - Stanislaw Puppel (eds.) 1996 Language history and linguistic modelling: A Festschrift for Jacek Fisiak on his 60th Birthday. Berlin: Mouton de Gruyter.

The development of Middle English I in England


Hoenigswald, Henry M. - Linda F. Wiener (eds.) 1987 Biological metaphor and cladistic classification. Philadelphia: University of Pennsylvania Press. Hotelling, H. 1933 "Analysis of a complex of statistical variables into principal components", Journal of Educational Psychology 24: 417-441,498-520. Hsieh, Hsin-I 1973 "A new method of dialect subgrouping", Journal of Chinese Linguistics 1: 6492. Jespersen Otto 1909 A modern English grammar, Part I: Sounds and spellings. London: George Allen & Unwin. Jordan, Richard 1934 Handbuch der mittelenglischen Grammatik. 1. (Revised by Η. Ch. Matthes.) Heidelberg: Winter. Kirk, John M. et al. (eds.) 1985 Studies in linguistic geography. London: Croom Helm. Kökeritz, Helge 1953 Shakespeare's pronunciation. New Haven, Conn.: Yale University Press. Kolb, Edward et al. 1979 Atlas of English sounds. Bern: Franke Verlag. Krishnamurti, Bh. - Lincoln Moses - Douglas Danforth 1983 "Unchanged cognates as a criterion in linguistic subgrouping", Language 59: 541-582. Labov, William 1963 "The social motivation of a sound change", Word 19: 273-309. 1981 "Resolving the neogrammarian controversy", Language 57: 267-308. 1992 "Evidence for regular sound change in English dialect geography", in: Matti Rissanen et al., 42-71. 1994 Principles of linguistic change. Volume I: Internal factors. Oxford: Blackwell. Β. Lakshmi Bai - Β. Ramakrishna Reddy (eds.) 1990 Studies in Dravidian and general linguistics: A Festschrift for Bh. Krishnamurti. Bombay: Department of Linguistics, Osmania University. Lehman, W.P. - Yakov Malkiel (eds.) 1968 Directions for historical linguistics. Auston: University of Texas Press. Luick, Karl 1914-1940 Historische Grammatik der englischen Tauchnitz. [ 1964] [Reprinted Oxford: Basil Blackwell]


2 vols. Leipzig: Β.


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

Macaulay, Ronald, K.S. 1985 "Linguistic maps: Visual aid or abstract art?" Studies in Linguistic Geography, in: John M. Kirk, Stewart Sanderson & John Widdowson (eds.), 172-86. Mäher, J. Peter et al. (eds.) 1982 Papers from the 3rd International Conference on Linguistics. Amsterdam: John Benjamins. Malkiel, Yakov 1967 "Each word has a history of its own", Glossa 1: 137-149. Menozzi, Paolo - Alberto Piazza - L. Luca Cavalli-Sforza 1978 "Synthetic maps of human gene frequencies in Europeans", Science 201: 786792. Meyers, L.F. - William S-Y. Wang 1963 'Tree representations in linguistics", Project on Linguistic Analysis, Report No. 3. The Ohio State University Research Foundation. Columbus, Ohio. Ogura, Mieko 1987 Historical English Phonology: A Lexical Perspective. Tokyo: Kenkyusha. 1990 Dynamic Dialectology: A Study of Language in Time and Space. Tokyo: Kenkyusha. 1995

"The development of Middle English Γ and ü: A reply to Labov (1992, 1994)", Diachronica XII: 1.31-53.

Ogura, Mieko - William S-Y. Wang 1990a. "The development of the Indo-European languages: From a persepective of dynamic dialectology", in: B. Lakshmi Bai & Β. Ramakrishna Reddy (eds.), 280-337. 1990b 1996a

"Isoglosses: artifactual or real? in: Bela Brogyanyi (ed.), 153-181. "Snowball effect in lexical diffusion: The development of -s in the third person singular present indicative in English", in Derek Britton (ed.), 119141.


"Lexical diffusion and evolution theory", in: Raymond Hickey & Stanislaw Puppel (eds.), 1083-1099.

Orton, Harold 1933 The phonology of a South Durham dialect. London: Kegan Paul, Trench, Trübner. 1962 Survey of English dialects. A. Introduction. Leeds: E. J. Arnold. Orton, Harold - Michael V. Barry (eds.) 1969-1971 Survey of English dialects. B. The basic material, Vol. 2, Part I (1969), Part II (1970), Part III (1971): The West Midland counties. Leeds: E. J. Arnold.

The development of Middle English ϊ in England


Orton, Harold - Wilfrid J. Halliday (eds.) 1962-1963 Survey of English dialects. B. The Basic Material, Vol. 1, Part I (1962), Part II (1963), Part III (1963): The six Northern counties and the Isle of Man. Leeds: E. J. Arnold. Orton, Harold - Stewart Sanderson - John Widdowson. 1978 The linguistic atlas of England. London: Croom Helm. Orton, Harold - Philip M. Tilling (eds.) 1969-1971 Survey of English dialects. B. The Basic Material. Vol. 3, Part I (1969), Part II (1970), Part III (1971): The East Midland counties and East Anglia. Leeds: E. J. Arnold. Orton Harold - Martyn F. Wakelin (eds.) 1967-1968 Survey of English Dialects. B. The Basic Material, Vol. 4, Part I (1967), Part II (1967), Part III (1968): The Southern counties. Leeds: E. J. Arnold. Palmer, Leonard R. 1972

Descriptive and comparative linguistics: A critical introduction. New York: Crane Russak & Company.

Phillips, Betty S. 1984 "Word frequency and the actuation of sound change", Language 60: 320-342. Prins, A. A. 1972 A history of English phonemes. Leiden: Leiden University Press. Rissanen, Matti et al. (eds.) 1992 History of Englishes: New methods and interpretations in historical linguistics. Berlin: Mouton de Gruyter. Samuels, Michael L. 1963

"Some applications of Middle English dialectology", English Studies 44: 8194.

1972 Linguistic evolution. Cambridge: Cambridge University Press. Schleicher, August 1861 Compendium der vergleichenden Grammatik der indogermanischen Sprachen. Weimar: Hermann Bohlau. Schmidt, Johannes 1872 Verwandtschaftsverhältnisse der indogermanischen Sprachen. Weimar: Hermann Bohlau. Schuchardt, Hugo 1885 "On sound laws: Against the neogrammarians", in: [1972] Theo Vennemann - Terence Η. Wilbur (eds.) S6guy,Jean 1954-1973 Atlas Linguistique de la Gascogne. 6 vols. Paris: Editions du Centre Nationale de la Recherche Scientifique.


Mieko Ogura - William S- Y. Wang - L Luca Cavalli-Sforza

Smith, Michael E. (ed.) 1972 Studies in linguistics in honour of George L Trager. The Hague: Mouton. Stockwell, Robert P. 1972 "Problems in the interpretation of the Great English vowel shift", in: Michael E. Smith (ed.), 344-362. Stockwell, Robert P. - Ronald K. Macaulay (eds.) 1972 Linguistic change and generative theory. Bloomington, Ind.: Indiana University Press. The Times Atlas of the World, Comprehensive Edition 1980 New York: Times Books. Trudgill, Peter 1974 "Linguistic change and diffusion: Description and explanation sociolinguistic dialect geography", Language in Society 2: 215-246.



"Linguistic geography and geographical linguistics", in: C. Board et al. (eds.), 227-252. U. S. Board on Geographic Names. Gazatteer 1955-1979 Official Standard Names Approved by the United States Board on Geographic Names. Washington. Vennemann, Theo - Terence Η. Wilbur (eds.) 1972 Schuchardt, the Neogrammarians, and the transformational phonological change. Frankfurt/M.: Athenäum Verlag.

theory of

Wang, William S-Y. 1969 "Competing changes as a cause of residue", Language 45: 9-25. 1979 "Language change - A lexical perspective", Annual Review of Anthropology 8: 353-371. 1983

"Variation and selection of language change", Bulletin of the Institute of History and Philology 53: 467-491. 1987 "Representing language relationships", in: Henry M. Hoenigswald & Linda F. Wiener (eds.), 243-256. 1993 "Glottochronology, lexicostatistics, and other numerical methods", in: R. E. Asher - Υ. Μ. Y. Simpson (eds.), 1445-1550. Wang, William S.-Y. (ed.) 1977 The lexicon in phonological change. The Hague: Mouton. Weinreich, Uriel - William Labov - Marvin I. Herzog 1968 "Empirical foundations for a theory of language change", in: W. P. Lehman Yakov Mal kiel (eds.), 95-195. Wyld, Henry Cecil 1936 A history of modern colloquial English. Oxford: Basil Blackwell.

The development of Middle English J in England


Appendix 1: The 311 sites listed by counties 34 Read 35 Marshside 36 Eccleston 37 Harwood 38 Bickerstaffe 39 Halewood

Nb 1 Lowick 2 Embleton 3 Thropton 4 Ellington 5 Wark 6 Earsdon 7 Haltwhistle 8 Heddon-on-the-Wall 9 Allendale

Cu 10 Longtown 11 Abbeytown 12 Brigham 13 Threlkeld 14 Hunsonby 15 Gosforth

Du 16 Washington 17 Ebchester 18 Wearhead 19 Witton-le-Wear 20 Bishop Middleham 21 Eggleston

We 22 Great Strickland 23 Patterdale 24 Soulby 25 Staveley-in-Kendal

La 26 Coniston 27 Cartmel 28 Yealand 29 Dolphinholme 30 Fleetwood 31 Pilling 32 Thistleton 33 Ribchester

Y 40 Melsonby 41 Stokesley 42 Skelton 43 Egton 44 Dent 45 Muker 46 Askrigg 47 Bedale 48 Borrowby 49 Helmsley 50 Rillington 51 Burton-in-Lonsdale 52 Horton-in-Ribblesdale 53 Grassington 54 Pateley Bridge 55 Easingwold 56 Gargrave 57 Sppofforth 58 York 59 Nafferton 60 Heptonstall 61 Wibsey 62 Leeds 63 Cawood 64 Newbald 65 Thornhill 66 Carleton 67 Welwick 68 Golcar 69 Holmbridge 70 Skelmanthorpe


Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

71 Ecclesfiled 72 Tickhill 73 Sheffield

106 Mavesyn Ridware 107 Lapley 108 Edingale 109 Wigginton 110 Himley

Man 74 Andreas 75 Ronague


Ch 76 Kingsley 77 Rainow 78 Swettenham 79 Farndon 80 Audlem 81 Hanmer (Flintshire)

111 Brimfield 112 Weobley 113 Cradley 114 Checkley 115 Longtown 116 Whitchurch Wo

Db 82 Charlesworth 83 Bamford 84 Burbage 85 Youlgreave 86 Stonebrrom 87 Kniveton 88 Sutton-on-the-Hill

117 Romsley 118 Hartlebury 119 Hanbury 120 Clifton on Teme 121 Earls Croome 122 Offenham 123 Bretforton Wa 124 Nether Whitaere 125 Hockley Heath 126 Stoneleigh 127 Napton-on-the-Hill 128 Aston Cantlow 129 Lighthorne 130 Shipston-on-Stour

Sa 89 Weston Rhyn 90 Prees 91 Llanymenech 92 Montford 93 Kinnersley 94 Chirbury 95 All Stretton 96 Hilton 97 Clun 98 Diddlebury 99 Kinlet

Mon 131 Skenfrith 132 Llanellen 133 Raglan 134 Cross Keys 135 Llanfrechfa 136 Shirenewton

St 100 Warslow 101 Mow Cop 102 Alton 103 Β arlaston 104 Ellenhall 105 Hoar Cross

Gl 137 Deerhurst 138 Gretton 139 Bream 140 Whiteshill

The development of Middle English Ί in England

141 Sherborne 142 Slimbridge 143 Latteridge 144 Kingham 145 Steeple Aston 146 Islip 147 Eynsham 148 Cuxham 149 Β infield Heath 150 North Wheatley 151Cuckney 152 South Clifton 153 Oxton 154 Eastoft 155 Saxby 156 Keelby 157 Willoughton 158 Tealby 159 Wragby 160 Swaby 161 Old Bolungbroke 162 Scopwick 163 Beckingham 164 Fulbeck 165 Sutterton 166 Swinstead 167 Lutton 168 Crowland 169 Harby 170 Hathern 171 Seagrave 172 Packington 173 Markfield 174 Great Dalby 175 Sheepy Magna 176 Goadby

177 Carlton Curlieu 178 Ullesthorpe R 179 Empingham 180 Lyddington Nth 181 Warmington 182 Welford 183 Little Harrowden 184 Kislingbury 185 Sulgrave Hu 186 Warboys 187 Kimbolton C 188 Little Downham 189 Elsworth Nf 190 Docking 191 Great Snoring 192 Bückling 193 Grimston 194 North Elmham 195 Ludham 196 Outwell 197 Gooderstone 198 Shipdham 199 Ashwelthorpe 200 Reedham 201 Pulham St. Mary 202 Garboldisham cr Μ 203 Tuddenham 204 Mendlesham 205 Yoxford 206 Kedington 207 Kersey Bk 208 Tingewick 209 Stewkley



Mieko Ogura - William S-Y. Wang - L. Luca Cavalli-Sforza

210 Long Crendon 211 Buckland 212 Coleshill 213 Horton

245 Brompston Regis 246 Stoke St. Bregory 247 Horsington 248 Pitminster 249 Merriott

Beds 214 Turvey 215 Great Barford 216 Harlington

W 250 Ashton Keynes 251 Sutton Benger 252 Avebury 253 Burbage 254 Steeple Ashton 255 Netheravon 256 Sutton Veny 257 Fovant 258 Whiteparish

Herts 217 Therfield 218 Codicote 219 Wheathampstead

Ess 220 Great Chesterford 221 Belchamp Walter 222 Cornish Hall End 223 Henham 224 Stisted 225 West Bergholt 226 Little Bentley 227 High Easter 228 Tiptree 229 East Mersea 230 Netteswell 231 Little Baddow 232 Tillingham 233 Doddinghurst 234 Canewdon

Berks 259 Buckland 260 Uffington 261 West Ilsley 262 Inkpen 263 Swallowfield

Sr 264 Walton-on-the-Hill 265 East Clandon 266 Coldharbour 267 Outwood 268 Thursley Κ

Mx and London

269 Stoke 270 Farningham 271 Staple 272 Warren Street 273 Denton 274 Goudhurst 275 Appledore

235 Harmondsworth 236 Hackney

So 237 Weston 238 Blagdon 239 Wedmore 240 Coleford 241 Wootton Courtenay 242 Stogursey 243 Stogumber 244 Withypool

Co 276 277 278 279

Kikhampton Altarnun Egloshayle St Ewe

The development of Middle English ϊ in England

280 Gwinear 281 St. Buryan 282 Mullion 283 Parracombe 284 Swimbridge 285 Weare Giffard 286 Chawleigh 287 Gittisham 288 South Zeal 289 Kennford 290 Peter Tavy 291 Widdicombe 292 Cornwood 293 Blackawton 294 Handley 295 Ansty 296 Whitchurch Canonicorum


297 Portesham 298 Kingston 299 Hatherden 300 Oakley 301 King's Somborne 302 Alresford 303 Hambledon 304 Burley 305 Whitewell (I.o.W.) 306 Billingshurst 307 Harting 308 Sutton 309 Fletching 310Horam 311 Firle


Mieko Ogura - William S-Y. Wang - L Luca Cavalli-Sforza


r-~ po Μ — CS

ro es

es (N








ov Ν

η to

ts ro



ο CS

ο CS

a. O

















τ* po

CS po





CS —


a U)

( S f S P O P O f O T t T t i o v o v o r - r - r - t —




Mieko Ogura - William S-Y. Wang - L Luca Cavalli-Sforza




an ofslcegen scealt" are found in his corpus although the examples of the other five orderings of these three elements are observed. This also confirms (3) for the nonexistent pattern is the one in which the main verb beon/weor{)an is separated from the main verb to its right.

3. Theoretical considerations When a verb governs a VP-complement clause which is tenseless, the relative order between the matrix verb and the subordinate verb is free in Old English. Some other material can occur freely between them if the matrix verb precedes the subordinate verb. But when the order is reversed, no other material can come between. This can be generalized to include the cases of a matrix verb and the adjectival head of a small clause and the case of a copula and its adjectival complement. I think the point has been convincingly demonstrated. In this section, I will consider how the finding can be incorporated into grammars of the Government-Binding style and point out some problems that will arise, assuming that all the subcases considered can be subsumed under a single generalization (3).

On the inseparable nature of verb-auxiliary combinations


Within the GB framework, two different approaches have been taken to deal with word order. One involves the head parameter and movement rules, especially head-movement. In the other, word ordering is introduced into syntactic structures at a later stage in the derivation in terms of X-bar theoretic relations. I will show that (3) will pose problems to both approaches. Travis (1984), among others, developed a word order theory based on the head-movement rule and the head parameter, which dictates the position of the head in a phrase to be either at the left periphery or the right periphery. Suppose the parameter is set so that the verb is placed at the VPfinal position in Old English. (9)

[ s ...[vp...[ s [vp...verb]] a u x ] ]

The sequence verb-aux is a direct realization of the head-final parameter as is shown in (9). The matrix verb (aux) occupy the final position of the matrix VP and the subordinate verb (verb) is placed at the final position of the VP of the subordinate clause. Consequently, the main verb and the auxiliary are put side by side in this order unless they are moved or some other material is introduced between them along the derivation. The other orders are derived by (optionally) moving one or both of the verbs to some points to the left. But if we are to adopt this approach, we have to have some device to exclude the generation of *verb-X-aux. Some processes that quite freely place adverbial elements at the phrase boundaries can place such material between the main verb and the matrix verb. Furthermore, a head movement can easily detach the main verb from the matrix verb by moving the former to some other position to the left. Though I cannot go into detail here, this will pose a considerable challenge if we take the approach outlined above. Now I will turn to the other approach. Various relations, such as "command", "government", and "binding", are defined over syntactic structures consisting of D-structure, S-structure, and LF. These relations are based on primitive relations supplied by the X-bar theory.4 But in the recent developments of grammatical theories, X-bar theory has linearity relations removed from its primitive relations. Therefore, precede-follow (i.e., linearity) relations are not defined over syntactic structures in a narrow sense. The linearity will be introduced into syntactic structures later in the phonological component. This is done by saying that if some asymmetric relation holds of two elements, then the two elements must be


Katsuhide Sonoda

ordered in a specified way.5 O-role assignment and Case-marking, for example, have been used to impose linearity on syntactic sructures. In order to exclude the grammatical order *verb-X-auxiliary, we have to define or discover some new relation that holds between verb and auxiliary. Then we can say that any two elements in the relation must occur consecutively if the auxiliary verb precedes the other. To my knowledge, however, no one has ever proposed a syntactic relation that holds just between verb and auxiliary in our sense. But it is a relation that is conceptually very natural, hence it is indispensable for any descriptively adequate grammar of Old English.

Notes 1. Throughout the paper, the auxiliary verb and the governed main verb in question are italicised. 2. The examples are from Sweet's Anglo-Saxon Primer, 9th edition, revised by N. Davis (Oxford, 1953). 3. In his book, he gives only one of these examples. (a)

aes e heo gesittan swij>e werig on treodes telgum torhtum moste. [Gen A 1469-70]

Furthermore, we find in Jacobson (1981) the following example. (b)

a nolde se mildherta drihten ge afigen Jjaet heo (=seo rod) behyld alaenc wcere 'then the merciful Lord would not permit that it (=the cross) should any longer be concealed' [Legends of the Holy Rood 17.21]

Those are all the counterexamples to (3) that I have found so far. 4. For details of the network of syntactic relations see Chomsky (1986). 5. For a recent proposal along this line, see Chomsky (1995: §4.3).

References Allen, Cynthia L. 1990 Review of Ans van Kemenade 1987, Language 66: 146-152.

On the inseparable nature of verb-auxiliary combinations


Callaway, Jr., Morgan 1913 The infinitive in Anglo-Saxon. Washington, D.C.: Carnegie Institution of Washington. Chomsky, Noam 1986 Barriers. Cambridge, MA: The MIT Press. 1995 The Minimalist Program. Cambridge, MA: The MIT Press. Donoghue, Daniel 1987 Style in Old English poetry: The test of the auxiliary. (Yale Studies in English 196.) New Haven, CT: Yale University Press. Harris, David Payne 1964 "The development of word-order patterns in twelfth-century English", in: Albert H. Marckwardt (ed.), 187-198. Jacobson, Sven 1981 Preverbal adverbs and auxiliaries: A study of word order change. (Stockholm Studies in English 55.) Stockholm: Almqvist and Wiksell. Kemenade, Ans van 1987 Syntactic case and morphological case in the history of English. Dordrecht: Foris. Lightfoot, David W. 1979 Principles of diachronic syntax. Cambridge: Cambridge University Press. 1991

How to set parameters: Arguments from language change. Cambridge, MA: The MIT Press. Marckwardt, Albert H. (ed.) 1964 Studies in languages and linguistics in honor of Charles C. Fries. Ann Arbor, MI: The English Language Institute, University of Michigan. Mitchell, Bruce 1985 Old English syntax. 2 vols. Oxford: Clarendon Press. Travis, Lisa 1984 Parameters and effects of word order variation. [Unpublished Ph.D. dissertation, MIT]

The gerund in Chaucer, with special reference to its verbal character* Matsuji Tajima

0. Introductory remarks Originally the gerund was simply an abstract noun of action derived from a verb and was treated exclusively as a noun, syntactically as well as inflectionally. Although it has retained all of its nominal qualities, the gerund has subsequently, albeit gradually, also acquired most of the properties and, hence, syntactic characteristics of a verb. This syntactic development is clearly reflected in the following verbal characteristics that the gerund can exhibit: 1. it can govern an "accusative" or direct object (e.g., He practices writing leading articles [OED]); 2. it can govern a predicative complement (e.g., Your being so sick forbids me to discuss the matter with you now [Curme]); 3. it can be modified by an adverbial adjunct (instead of an adjective) (e.g., He has hopes of coming back speedily [OED]); 4. it can show tense and voice by means of compound forms (e.g., of having done it; the necessity of loving and being loved [OED]; and 5. it can take a subject in the common case (instead of the genitive) (e.g., I insist upon Miss Sharp appearing [OED]). There remains much divergence of opinion among scholars as to the origin of the 'verbal' gerund, or the gerund with verbal properties.1 Nevertheless, there seems to be general agreement that the gerund began to take a direct object and to be modified by an adverbial within the ME period, more specifically in the 14th century and that its assumption of compound tense and voice forms developed still later.2 (The second and fifth verbal features mentioned above remain to be properly explored.) Notwithstanding this limited general consensus and the general importance of this subject, however, Fourteenth-Century English, in which the most notable development of the gerund was supposedly taking place, or Middle English generally, for that matter, has not been subjected to any detailed examination.3 Most earlier studies, if they have discussed the verbal


Matsuji Tajima

function at all, have been satisfied with simply adducing a few examples from Chaucer or Langland as if they were the first to use the gerund with syntactic verbal force. In view of this unsatisfactory state of affairs, I previously made a somewhat detailed study of the 'verbal' gerund in Fourteenth-Century English (Tajima 1977: 113-33). The present paper is intended both to supplement and to continue the previous study. More specifically, it attempts to ascertain the extent to which the gerund has acquired certain verbal properties in Chaucer and to put it in historical perspective, thereby redressing the erroneous yet long-held view that the modern, full-fledged gerund (having verbal and nominal characteristics) is first instanced in Chaucer. For this purpose I have examined all of Chaucer's works,4 excluding The Romaunt of the Rose and Equatories of the Planetis because of their doubtful authorship. Schematic summaries of Chaucer's language are found presented in most of the modern editions of Chaucer's work. In addition, some of the remarkable aspects of his language have been treated in a number of books and articles. Nevertheless, Chaucer's use of the gerund has never been fully investigated. Indeed, there are only two limited studies of the subject. Emonds (1973) has a corpus restricted to The Parson's Tale, with the result that its conclusions do not necessarily reflect Chaucer's overall usage. In his second edition of Studies in the Language of Geoffrey Chaucer (1982: 118-21), Kerkhof has incorporated a section entirely devoted to the gerund, which provides a useful overview, its potential value being curtailed, however, by its exclusively synchronic approach. This, then, is the "reason for being" of this paper. Since there are no instances of the gerund which indicates tense and voice by means of compound forms in Chaucer, not to mention those of the gerund governing a predicative complement or taking a common-case subject, I will focus my discussion on the following two aspects: 1. the gerund with its logical object and 2. the gerund with its adverbial adjunct.

1. Gerund with object Initially a pure noun, the gerund expressed its logical object, if any, in the genitive, and later, as a result of the tendency to use analytic patterns in place of synthetic ones, in the periphrastic genitive with of, while in late ME the gerund began to take a direct object without the preposition of. Thus, in Fourteenth-Century English constructions of the gerund with an

The gerund in Chaucer


object of its own may roughly be classified into the following six categories: Type I = objective genitive (possessive) + gerund (e.g., Complaynt D'Amours 70 In any word to your displesinge); Type Π = object + gerund (e.g., CT I 447 Of clooth-makyng she hadde swich an haunt); Type ΙΠ = determiner + gerund + o/-adjunct (e.g., CT VII 1202 In the examynynge thanne of youre conseillour ye shul considere manye thynges); Type IV = gerund + o/-adjunct (e.g., Bo. I pr. 6.71-72 thow art confunded with foryetynge ofthiself)\ Type V = gerund + object (e.g., CT VE 1597 in getynge richesses ye mosten flee ydelnesse); Type VI = determiner + gerund + object (e.g., York MGame 76 3e may know a greet hert by fre beryng J?e woode [Visser]). Of these six types, I, ΠΙ, and IV are obviously strongly nominal in character, the -ing form having no syntactic verbal force at all, whereas in V and VI the -ing form is syntactically verbal in that it governs an object without a preposition. In Type Π it may be sometimes nominal, sometimes verbal, depending upon the degree of closeness of the combination 'object + gerund'. The following table provides the number of occurrences of each type in the respective works of Chaucer; for the sake of convenience, the number of occurrences of the gerund with an adverbial adjunct will also be given here, although discussion will be postponed to Section 2 (A = gerundial compound 'adverbial element + gerund'; Β = gerund + adverb; and C = gerund + adverb phrase):


Matsuji Tajima

Table 1. Frequency of Types I-VI of gerund with object (and frequency of gerund with adverbial adjunct) in Chaucer Text: Date of composition BD (1369) Anel. ( c l 3 7 5 ) //F(cl380) />F(cl380) Bo. ( c l 3 8 0 ) TC(cl385) LGW ( c l 3 8 6 ) Astr. (1391) S.P. (1370-99) CT ( c l 3 7 5 - a l 4 0 0 ) Total



0 0 1

1 0 1 0 0 2 0

With object ΠΙ IV 2 0

With adverbial adjunct V





0 0

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 2 1 0 0 3

0 0 0 2 6 3 0 0 1 16

0 0 1 0 2 5 1 1 3 19





1 0 2 2


0 0 18

0 0 44 4 4 2 1 33


0 1 1 11






0 3 2 0 0 1

19 15 7 7 1

0 0 4 2

Table 1 shows that in Chaucer: (i) Types ΙΠ and IV, particularly the latter, are by the most frequent; (ii) Types I and Π are not very common; (iii) Type V, with which we are especially concerned, does appear, although only very sporadically; and (iv) Type VI is non-existent. Table 1 also indicates that there is some difference in frequency between verse and prose. Table 2 gives the number and percentage of occurrences of each type in verse and prose in Chaucer. Table 2. Frequency of Types I-VI of gerund with object in verse and prose in Chaucer.

verse prose Total








9 (9.5)

15 (15.8)


95 (100%)

7 (3.9)

50 (52.6) 83 (45.6)


4(2.2) 13 (4.7)

17(17.9) 73 (40.1)

15 (8.2)



22 (7.9)

90 (32.5)

133 (48.0)

19 (6.9)


277 (100%)

Table 2 indicates that: (i) in verse, Type VI is by far the most frequent, and Type ΙΠ considerably less frequent, but Types I and Π are comparatively high in frequency; (ii) in prose, Types ΙΠ and IV are most common, but Types I and Π are extremely rare. More noteworthy than these points, however, is the fact that the 'modern' Type V construction occurs 15 times in prose as against only four in verse, despite the very limited corpus of the former.

The gerund in Chaucer


Let us compare these results with the following statistics based upon my previous research on the use of the gerund by the 14th-century writers (Tajima 1977: 118): Table 3. Frequency of Types I-VI of gerund with object in 14th-century English I


1300-1325 14(16.3) 33 (38.4)




9 (10.4)


11 (12.8) 6 (3.5)

VI 0(0)

Total 86 (100%)


6 (3.5)

14 (8.3)

59 (34.7)

84 (49.4)




52 (15.6)

202 (60.7) 37(11.1) 2 (0.6) 333 (100%)


31(5.2) 76(12.9) 120(20.4)


54 (9.2)

1 (0.6) 170 (100%) 3 (0.5) 589 (100%)

On the whole, Chaucer shows much the same tendency as was current in the 14th century, except for the fact that Type ΙΠ is of relatively high frequency in Chaucer. On closer examination, however, Chaucer's usage in verse is rather conservative or archaic, namely, closer to that of the first quarter of the 14th century, in that he has retained to a considerable extent vestiges of OE synthetic expressions, such as Types I and Π. With this overview in mind, a separate discussion and exemplification of each of the six types of gerund with object is now in order. 1.1. Type I: objective genitive (possessive) + gerund This is the type of construction in which the gerund takes the genitive or possessive as its object and accordingly has nominal force. In OE genitive was commonly used in this function, as in: King Alfred's Orosius 17/35 "toeacan pees landes sceawunge 'in addition to observing the land'", but in ME it is rare, its function being supplanted by the periphrastic genitive with of. On the other hand, the possessive in this function was rare in OE, but in ME it is quite common (cf. Visser 1966: §1106). In Chaucer, apart from Type VI, this OE type of construction occurs least often, but the general practice mentioned above applies to Chaucer as well, in which the gerund is found four times with the genitive object and nine with the possessive object: (1)

a. b.

TC V 1499 Of Archymoris brennynge; HF 636 in his folkes furtherynges. (Other examples: Bo. Π pr. 3.55-57; CTWHL 766.) Complaynt D'Amours 68-70 I drede That I have seid here ... in any word to your displesinge', CT ΥΠ 459 Somtyme shewen


Matsuji Tajima

they thyn heriynge. (Other examples: Bo. ΠΙ pr. 6.10-12, ΙΠ pr. 7.8-9; TC1853; CT 12748, IV 1040, VII847, VII1301.) 1.2. Type II: object + gerund Before the modern word-order 'gerund + object' established itself, there existed another word-order 'object + gerund' in ME. According to van der Gaaf (1928: 33-41) the construction 'noun (i.e., object) + gerund' was developed from OE compounds consisting of 'acc. + gerund' or OE synthetic expressions such as Vless gen. sg. + gerund' and 'gen.pl. + gerund'. It is, therefore, often difficult to decide whether the noun in constructions of this type should be regarded as accusative or genitive; in other words, whether the -ing form in Type Π constructions has verbal force or nominal force. It is, however, certain that this type was comparatively freely used throughout the ME period,5 particularly around 1300, as is shown above in Table 3. In Chaucer I have noted 22 instances of this type, of which 15 are found in verse and seven in prose (only in CT.Mel.) First let us take a close look at examples of the combination 'noun (sg.) + gerund': (2)

HF 1241 in fight and blod-shedynge; TC ffl 48 to Venus heryinge; CT I 447 Of clooth-makyng; CT VI 587 Of wynyevyng to hem; CTVU 1345 in vengeance-takyng, in werre; CT VII 1582 withouten wrong or harm doynge to any oother persone. (Other examples: C7 VII772, 850, 1032, 1429,1491.)

It is hard to decide whether in some of the above quotations the preceding noun is accusative or s-less genitive, or whether the noun and the gerund form a compound or not.6 The combination 'noun (sg.) + gerund' is often preceded by such an adjunct as the, his, my, and swich, in which case the adjunct qualifies the whole combination (8 exs.) or the first element of it (1 ex.). The former use dates back to OE and was common in ME. In Chaucer the following illustrate it: (3)

BD 1312-13 al was doon, For that tyme, the hert-huntyng\ CT1 3678-79 To Alison now wol I teilen al My love-longynge', CT VII 1392 of the vengeance-takynge upon that wolde engendre another vengeance; CT VIE 922 Somme seyde it was long on

The gerund in Chaucer


the fir makyng. (Other examples: CT 1913-14, 3348-50, 3705; CTV Π 1431.) In these examples, the noun and the gerund seem to be held firmly together, possibly being realized as a kind of compound. In the following instance, however, the adjunct appears to modify only the first element of the combination in question: (4)

TCI 740-44 Man maketh ofte a yerde With which the maker is hymself ybeten In sondry manere ... And namelich in his counseil tellynge That toucheth love oughte ben secree. Cf. Mannyng HS 7887 Whedyr hyt be yn a womman handlyng; PPI.B XV 76 in-to hiegh clergye shewynge.

In these quotations, the noun and the gerund seem to be less firmly connected, the former appearing somewhat detached from the latter. There are, further, two instances of the combination 'noun (pi.) + gerund' in Chaucer, in one of which the adjunct qualifies the first element, not the whole combination: (5)

CT V m 770 What sholde I ... bisye me to telle yow ... of the care and wo That we hadde in oure matires sublymyngj CT V m 796-97 Nat nedeth it for to reherce hem alle - Watres rubifiyng. Cf. PPI.B VE 87 vsage ... of seyntes lyues redynge.

This usage, well attested in ME (cf. Tajima 1977: 121), might have come from those OE combinations whose first element is a plural noun in the genitive. In ME, however, those nouns end in ~(e)s, and can no longer be recognized as genitives, their presence being consequently realized as direct objects. As is clear from the foregoing discussion, some of the examples quoted may have been felt to be compounds consisting of 'acc. + gerund', or vestiges of OE synthetic expressions such as Vless gen sg. + gerund' and 'gen. pi. + gerund'. Since formal distinction between j-less genitive singular and accusative singular, or between genitive plural and accusative plural is virtually impossible in ME, however, it seems more practical to take the noun in the combination as accusative object. By the same token, the fact that the noun occasionally takes an adjunct by itself or occurs in


Matsuji Tajima

the plural suggests that it is no longer a part of a compound, but is detached from the -ing form, functioning syntactically as its pre-posed object. 1.3. Type III: determiner + gerund + of-adjunct Here, while being preceded by such determiners as articles, possessives, and demonstratives, the gerund is followed by an o/-adjunct that is syntactically equivalent to a direct object and, accordingly, strongly nominal in character. Of 277 examples of the gerund with its logical object, 90, or 32.5%, belong to this type in Chaucer. This comparatively high frequency can, however, be accounted for by the fact that the construction is very much favoured in prose, especially in Bo. (cf. Table 1). The most representative pattern of this type - the 'the + gerund + ofadjunct' construction - is the regular construction, along with Type V, in Present-day English. In Chaucer, of these 90 examples, 75 are preceded by the definite article the. Chaucer's frequent use of it deserves some attention, considering that the construction with the begins to make its appearance around 1300 (Visser 1966: §1124), and that, interestingly enough, in Shakespeare it still seems to have been regarded as colloquial (Abbott 1870 [I960]: §93). A few examples will suffice to illustrate this pattern: (6)

Bo. ΙΠ pr. 10. 138-39 by the getynge of blisfulnesse men ben makid blisful; LGW 2459-60 Ye han wel herd of Theseus devyse In the betraysynge of fayre Adryane·, CT X 332 that resoun - that is to seyn, Adam - consented to the etynge of the fruyt.

The following quotations are of some interest in that, despite its strongly nominal character, the gerund also has the verbal force marked by the addition of an adverbial adjunct: (7)

CT X 843 This sacrement bitokneth the knyttynge togidre of Crist and of hooly chirche; CT VII 1148 the biwreiyng of thy conseil to a persone.

The other determiners which are found preceding the gerund are: his (1 ex.), my (1), no (4), more (1), a maner (2), and adjectives (6). Some examples follow:

The gerund in Chaucer



TC V 1833 And thus bigan his lovyng of Criseyde·, CT X 1008 this is no departynge of shrifte·, CT V 243 They speken of sondry hardyng of metal·, CT VIII 94-96 Cecile... Is joyned, by a manere conjoynynge Of "hevene" and "Lia". Cf. Mannyng HS 8334 To fordo a getyng of a chylde.

1.4. Type IV: gerund + of-adjunct This is the construction in which, though not preceded by any determiner, the gerund takes the o/-adjunct as its logical object. It is generally considered that this type of construction makes its first appearance about the time of Chaucer, and that it is the regular construction in ME and early ModE.8 However, I have a number of examples from about 1300 on, as in: Havelok 2320-24 "J>er mouthe men se f e moste ioie J>at mouthe be: ... putting of stone"\ KAlex. 583 "wijjouten doyng of any harm" (Tajima 1977: 123).9 As is also clear from Table 3, in the first quarter of the 14th century Type Π ('Object+Gerund') was the most frequent, and it is not until the second quarter of the century that Type IV establishes itself as 'the regular construction in ME', as is usually claimed (cf. Jespersen 1938: §208). This linguistic fact is well reflected in Chaucer, particularly in his verse in which the construction under discussion is by far the most common pattern. That is, out of the 277 examples of the gerund with its object, 133, or 48.0% (to be more exact, 52.6% in verse and 45.6% in prose) come under this category. The following demonstrate this type: (9)

HF 305-306 of oon he wolde have fame In magnyfyinge of hys name·, LGW 188-90 ne wene nat that I make In preysing of the flour agayn the leef, No more than of the corn agayn the sheef; CT TV 162 Lat me allone in chesynge of my wyf; CT VII 1636 Afterward, in getynge ofyoure richesses and in usynge of hem, yow moste have greet bisynesse.

Some of the examples belonging here are found qualified by an adverbial adjunct, which shows that even the strongly nominal -ing form was in the process of acquiring a verbal character: (10)

TC IV 422-24 The newe love, ... Or elles selde seynge of a wight, Don olde affecciouns alle over-go; CT ΠΙ 1979 Of buyldynge up of chirches;


Matsuji Tajima

CT VI 875 In caringe of the gold out ofthat place-, CTX 567 in vevynge of wikked conseil by fraude. Cf. Rolle Prose 13/10-11 Consaile es doynge awaye of worldes reches. 1.5. Type V: gerund + object The verbal character of the gerund is most prominent when it takes a direct object without the preposition of. This is the most representative pattern of the gerundial construction in Present-day English. Some scholars10 have placed the origin of this construction in the 14th century, especially in its second half, citing their oldest instances of it from Chaucer or Langland. More recently, Visser (1966: §1123) records clear examples of it from cl303 (Mannyng HS) onwards. My previous study (1977: 124-25) also shows that the gerund followed by a simple object is still extremely rare except in a limited number of texts such as Mandev. (cl400) and Dest. Troy (?al400), but appears as early as in Mannyng HS (cl303) and Cursor (a1325), over half a century earlier than has generally been considered. In Chaucer I have noted 19 instances of this type, which accounts for only 6,9% of all the examples of the six types under discussion. Considering the size of Chaucer's works examined for this study, the construction may be said to be very rare in Chaucer, which seems to agree with the practice current in the 14th century. It may, however, be of some interest to observe that, while it is extremely rare or practically nonexistent in most of his poems (six examples in all), it occurs unusually often in his prose works (14 in all), namely, seven times in CT.Mel., three in CT.Pars., three in Bo., and once in Astr.: (11)

Bo. IV pr. 7.61 in folwynge the opynioun of the people; Fortune 53-54 The negardye in keping hir richesse Prenostik is thou wolt hir tour assayle; CT VE 1624 in getynge of youre richesses and in usynge hem ye shul alwey have thre thynges in youre herte; CT IX 67 In liftyng up his hevy dronken cors. (Other examples: Bo. I pr. 4.221-23, ΠΙ pr. 2. 124-25, V pr. 5.20-22; Astr. Pt. I 1.1-3; CT VII 1018-19, 1435, 1522, 1549, 1597, 1648; CT Χ 464-65, 575, 1055-56.)

The gerund in Chaucer


The examples given above are all those of the gerund followed by a simple object. The following is an instance in which the gerund takes a noun clause as its object: (12)

TC IV 1016-17 But now η'enforce I me nat in shewynge How the ordre of causes stant.

Another interesting examples is found in: (13)

TC V 835-37 Troilus was nevere ... in no negree secounde In durryng don that longeth to a knyght.

Here the gerund appears to govern the bare infinitive as its object, although another interpretation may be possible.11 1.6. Type VI: determiner + gerund + object This is the construction in which the gerund preceded by determiners such as the, this, and his governs an object without the help of the preposition of. Accordingly, this type of construction has a mixed (i.e., verbal and nominal) character by having a direct object even though it is preceded by a determiner. It was common in early ModE (cf. Visser 1966: §1124 and Onions 1925: 130-31), but it is by no means usual in Present-day English except after possessives and genitives. I have found no instances of this Type VI in Chaucer, although Visser (1966: §1124) states, by quoting wrong or doubtful examples, that the 'the + gerund + object' construction the representative pattern of Type VI - occurs from the beginning of the 14th century. The earlier clear example of this construction Visser has no record turns out to be: cl410 York MGame 76 "3ε may know a greet hert by {>e beryng t>e woode". As far as I have been able to ascertain, there are no examples of 'the + gerund + object' in the 14th century, though there are a few instances of Type VI found after this and pyne, as in: (14)

Ayenb. 270/16 be {>yne ofseruynge Jjet he Jjolede myd guod wyle* PP1.B V 385 This shewynge shrifte...shal be meryte to the.11

That this type of construction is practically non-existent not only in Chaucer but in the 14th century generally is not unexpected, since the


Matsuji Tajima

construction 'gerund+object' (Type V) was not fully developed in the century.

2. Gerund with adverbial adjunct The verbal nature of the gerund asserts itself further when it is modified by an adverbial adjunct that can only be used together with verbs. As regards the development of the gerund with an adverbial adjunct, OED (s.v. -ing1 2) gives the following account: The first traces of it as yet pointed out ... occur cl340 in the Ayenbite of Inwyt and in the writings of Richard Rolle of Hampole, in the separation of the adv. in downcoming, downfalling, ingoing, etc., and the placing of it after the vbl. sb., coming down, falling down, going in, as in the finite verb, come down, fall down, go in. This was soon extended to adverbs and adverbial phrases generally, so that it became established that any vbl. sb. could, like the vb. to which it belonged, take an adverbial qualification. 13

Most scholars seem to agree with the OED statement, quoting their first instance of it from 1340 Ayenb.: 262/32 "at uerste guoinge in". In earlier studies (Tajima 1977: 127-28 and 1980: 34), however, I have demonstrated that the gerund with a simple adverb occurs as early as about 1200, as in: ?cl200 Hali Meidenhad 42/452-53 "his lokynge on ageasteö"; Ib. 46/49192 "in his fostrunge ford, moni earm hwille", but that in the first half of the 14th century it is still extremely rare, old compounds of 'adverbial element + gerund being more common, and in the second half of the century it gains ground gradually. In Chaucer (cf. Table 1), the old compound 'adverbial element + gerund' occurs only seven times, whereas the new analytic type 'gerund + adverb' is found as often as 28 times. Interestingly enough, the former is always found in verse, but not at all in prose, while the latter is found more often in prose, particularly in Bo. and CT.Pars. This may suggest that Chaucer's usage is rather archaic in verse, but more 'modern' in prose, as was already mentioned in connection with his use of the gerund with its object. Examples of the gerundial compound 'adverbial element + gerund' may be given below:

The gerund in Chaucer



HF 1522 her tyme of out-fleynge] TC Π 1308 at his in-comynge; TC V 503 at myn hom-comynge ('/zom-comynge' also in LGW 2100; CT 1884, 905; CT Π 765.)

As is shown above, there are, in fact, three different compounds in Chaucer: out-fleynge, in-comynge, and hom-comynge, of which the last is repeatedly used. Needless to say, these are nominal in character. The modern construction, i.e., the gerund with a simple adverb can be illustrated by such examples as the following: (16)

PF 655-56 Thanne wolde I that these foules were aweye, ... for taryinge lengere heere\ Bo. Π pr. 3.91-92 sehe ... forleete the in fleynge awey\ CTTK. 67 In liftyng up his hevy dronken cors; CT X 952 Slepynge longe in greet quiete is eek a greet norice to Leccherie; TC IV 423 selde seynge of a wight.

On the other hand, the gerund modified by an adverb phrase appears as early as c 1175, as in: Bod.Hom. 92/28 "Tactus, repung, oder grapung on alle limcen and J)aeh 3ewunelycost on \>am hondceri (Curme 1912: 352), and seems to occur twice as often as the one with a simple adverb in the 14th century as often in combination with as adverb phrase as with as adverb. Some examples follow: (17)

TC ΠΙ 1452 Acorsed be thi comyng into Troye\ Truth 16 The wrastlingfor this world axeth a fal; CT 1467 She koude muchel of wandrynge by the weye; CTX 1055 Thanne is discipline eek ... in scourgynge withyerdes.

3. Concluding remarks The principal object of this study has been to investigate the stage of development of certain features of the 'verbal' gerund in Chaucer and to put this in historical perspective. The major findings, which have already been treated at length, can now be summarized, preparatory to the formulation of some general conclusions. Section 1 discussed the size types of the gerund with its logical object. The 'nominal' construction 'gerund + ^/-adjunct' (Type IV) is much more frequently used than the other types. The strongly nominal 'determiner +


Matsuji Tajima

gerund + o/-adjunct' (Type ΙΠ) is comparatively frequent. The constructions 'objective genitive (possessive) + gerund' (Type I) and 'object + gerund' (Type Π) - survivors of OE synthetic expressions - are still occasionally found, particularly in his verse. The 'verbal' construction 'gerund + object' (Type V), with which we were chiefly concerned, is gradually developing, especially in his prose, such as Bo., CT Mel., and CT.Pars.u However, the 'determiner + gerund + object' construction (Type VI) does not occur at all, being a later development. Section 2 dealt with the development of the gerund with an adverbial adjunct. It reveals that the gerund modified by a simple adverb or an adverb phrase is also slowly developing in Chaucer, mainly in prose works such as Bo. and CT.Pars. Old compounds consisting of 'adverbial element + gerund' are very rare, but are found only in his verse. As a result, it can be concluded that Chaucer's syntax of the gerund follows the general practice of the second half of the 14th century. In other words, most of the examples of the gerund found in Chaucer, like those found in the works of his contemporaries, are still strongly nominal in character. The only major difference as to usage between Chaucer and his contemporaries is that the construction 'determiner + gerund + o/-adjunct' appears to be far more frequent in Chaucer, the reason being that it occurs unusually often in his prose work Bo. Historically speaking, however, it remains a fact of considerable interest that instances of the 'modern' gerund, or the gerund with certain verbal properties, do occur in Chaucer, more specifically in Bo., TC, and CT (particularly in prose portions such as The Tale of Melibee and The Parson's Tale), although Chaucer was obviously not the initiator. However, the limited development of certain aspects of the verbal gerund in the works of this period indicates that it is still far from fully developed. It can, therefore, be safely said that the verbal character of the gerund is at a stage of formative development in Chaucer, as it is in the works of his contemporaries.

Notes *

This is a revised version of a paper originally published in Poetica (Tokyo) 21/22 (1985).

1. For a recent overview and summary of important views concerning this much-discussed problem of English verb syntax, see Mustanoja (1960: 567-73). 2.

See, for instance, OED, s.v. -ing1 2; Jespersen (1938; §208); B0gholm, (1936: 256).

3. This lacuna has now been partially filled by Tajima (1985).

The gerund in Chaucer


4. All citations of Chaucer's works are taken from The Riverside Chaucer (1987). The abbreviations of Chaucer's works and their dates of original composition are generally based upon MED (1954 [Plan and Bibliography]: 31-33): BD (1369) = The Book of the Duchess·, Anel. (cl375) = Anelida and Arcite; HF (cl380) = The House of Fame·, PF (cl380) = The Parliament of Fowles; Bo. (cl380) = Boece [prose]; TC (cl385) = Troilus and Criseyde\ LGW (cl386) = The Legend of Good Women; Astr. (1391) = A Treatise of the Astrolabe [prose]; and CT (cl375-al400) = The Canterbury Tales (in which only The Tale of Melibee and The Parson's Tale are written in prose, being referred to as CT.MeL and CT.Pars. respectively as occasion arises). Chaucer's short poems (c1370-99) carry the same titles as appear in The Riverside Chaucer. 5. Cf. van der Gaaf (1928: 339); Mustanoja (1960: 574). 6. For example, Kittredge (1891: 100) regards the noun (Venus) in TC III 48 as genitive singular without ending. 7. Skeat (1894: Glossary, s.v. Matere) takes matires as genitive plural. 8. Cf. Jespersen 1938 [I960]: §208; Jespersen 1940: §8.4.6; Poutsma 1923: §33, etc. 9. Visser (1966: §1120) quotes his earliest instance of it from cl250 Gen. & Ex., however. 10. See OED, s.v. ing1 2; Jespersen (1940: §9.3.1.); Tmka (1930: 92); van der Gaaf (1928: 41) Brunner (1962: 355). 11. Robinson (1957: 834) interprets "In durryng don that longeth to a knyght" as 'In daring to do what belong to a knight', but MED (s.v. durring don) regards this durring don as a compound, defining it as 'courage to engage in feats of arms, prowess in battle; ...'. Davis (1979: s.v. durring) has adopted Robinson's interpretation. 12. I take it that this phrase contextually means 'This showing of shrift.' It may, however, be possible to look upon shewynge as present participle, and to interpret the phrase as 'This frank confession' (cf. Dunn and Byrnes 1973: 305). 13. See, for example, Jespersen (1940: §9.1.1); Tmka (1930: 92); Poutsma (1923: 163). Visser (1966: §1035) briefly touches on the ground with adverbial adjuncts, quoting only one instance from the 14th century (cl382 Wyclif) and two from the 15th century. 14. Chaucer's comparatively frequent use of the 'modem' gerund in Bo., CT.Mel, and CT.Pars. may be ascribed to the fact that in translating the Latin and French originals he might perhaps have rendered literally the Latin gerund and the French g6rondif, as in F. en venant, L. in veniendo. This point remains, however, to be further explored.

References Abbott, Edwin A. 1870 A Shakespearian grammar. (3rd revised edition.) London: Macmillan. [1966] [Reprinted New York: Dover Publications.]


Matsuji Tajima

Benson, Larry D. (gen. ed.) 1987 The Riverside Chaucer. (3rd ed.) Boston: Houghton Mifflin. B0gholm, Niels 1939 English speech from an historical point of view. Copenhagen: Nyt Nordisk Forlag. Brunner, Karl 1962 Die englische Sprache. Vol. II: Ihre geschichtliche Entwicklung. (2nd edition.) Tübingen: Max Niemeyer. Davis, Norman, et al. (eds.) 1979 A Chaucer glossary. Oxford: Clarendon Press. Dunn, Charles W. - Edward T. Byrnes (eds.) 1973 Middle English literature. New York: Harcourt Brace Javanovich. Emonds, Joseph 1973

"The derived nominale, gerunds, and participles in Chaucer's English", in Braj. B. Kachru, et al. (eds.), 185-98. Gaaf, Willem van der 1928 "The gerund preceded by the common case", English Studies 10: 33-41. Jespersen, Otto 1960

Growth and structure of the English language. (9th edition.) Oxford: Basil Blackwell.

1940 A Modern English grammar, Part V: Syntax. Copenhagen: Munksgaard. Kachru, Braj B., et al. (eds.) 1973

Issues in linguistics: Papers in honor of Henry and Renee Kahane. Chicago: University of Illinois Press.

Kerhof, Jelle 1982 Studies in the language of Geoffrey Chaucer. (2nd edition.) Leiden: E. J. Brill/Leiden University Press. Kittredge, George L. 1891 Observations on the language of Chaucer's Troilus. London: Chaucer Society. Kurath, Hans - Sherman M. Kuhn - Robert E. Lewis (eds.) 1952Middle English Dictionary. Ann Arbor, Michigan: The University of Michigan Press. Murray, James A. H. et al. (eds.) 1933 The Oxford English Dictionary. Oxford: Clarendon Press. Mustanoja, Tauno F. 1960 A Middle English syntax. Part I. Helsinki: Socidtö N6ophilologique. Onions, Charles T. 1925 An advanced English syntax. (5th edition.) London: Kegan Paul.

The gerund in Chaucer


Poutsma, Hendrik 1923 The infinitive, the gerund, and the participles of the English verb. Groningen: P. Noordhoff. Skeat, Walter W. 1894 The complete works of Geoffrey Chaucer, Vol. V. Oxford: Clarendon Press. Robinson, Fred N. 1957 The works of Geoffrey Chaucer. (2nd edition.) London: Oxford University Press. Trnka, Bohumil 1930 On the syntax of the English verb from Caxton to Dryden. Prague: Travaux de Circle Linguistique de Prague 3. Tajima, Matsuji 1977 "The gerund in fourteenth-century English, with special reference to the development of its verbal character", Studies in English Literature (Tokyo) 54: 113-33. [In Japanese.] 1980 "The gerund in the {Catherine group", Linguistic Science (Kyushu University) 15:31-35. 1985

The syntactic development of the gerund in Middle English. Tokyo: Nan'undo. Visser, Frederikus T. 1966 An historical syntax of the English language. Part II. Leiden: E.J. Brill.

Archaism in the vocabulary of jElfric* Schin 'ichi Takeuchi

1. Introduction Ever since Gneuss (1972) called our attention to the Winchester vocabulary,1 iElfric as its central figure has been extensively studied from the point of view of his unique diction and its implications in the context of late Old English standardization. Indeed, it has been pointed out on many occasions that yElfric uses certain words and expressions uniquely in his own way in preference to other synonymous words that can be found quite commonly in other prose works of late Old English. In my own research, too, forgyfan, for instance, occurs in an unbelievably high frequency in iElfric when sellan is the most common verb of giving in Old English. iElfric also uses the verb andwyrdan almost to the total exclusion of the verb andswarian, which, of course, is most widely and commonly used in other Old English texts including even those of the so-called Winchester school. On the other hand, jElfric makes frequent use of the noun andswaru in phrase such as cwedan to andsware 'say in answer', which us rarely used in other Old English works (Takeuchi 1994). In this short paper, I would like to discuss some unique aspects of forgyfan and andwyrdan as used in JElfric with a view to demonstrating that analyzing Old English vocabulary by focusing on its semantic structure and the process of its change could be one of the effective means of uncovering dialectal features as well as diachronic developments in Old English.

2. General distribution Table 1 below shows the general distribution of the chief verbs of giving in Old English. As I mentioned at the outset, forgyfan occurs most frequently in itilfric while sellan shows a proportionate and well-balanced distribution among the texts examined here. This shows that sellan was the most general and universal verb that was employed to denote the act of 'giving'




in Old English. Gesellan and gyfan, on the other hand, seem rather specialized semantically, hence occurring far less frequently than sellan. Especially, gyfan shows a curious distribution in that it occurs rather substantially in Bede2 and Chronicle E, which are said to be of Anglian origin, whereas its occurrence in the other texts are relatively meager. Table 1. CP





































































JE=M lfric, WH=WHom, WS=WSCp, ChA=Chron A, ChE=Chron Ε

Table 2 below shows the general distribution of andswarian and andwyrdan. What is really striking here is the fact that iElfric exclusively uses the verb andwyrdan in contexts where the most preferred and common verb in Old English is andswarian. Of the total of some 400 occurrences in jElfric, andwyrdan occurs 374 times, amounting to more than 90 per cent, which is really striking, considering the fact that andswarian is undoubtedly the major verb in the rest of the texts in Old English. Even in Gregory's Dialogues, which belongs to the Winchester group of texts, according to Gneuss (1972: 81), andswarian is the most prevalent. Table 2. Or Bede GD(C) GD(H) iElfric WHom WSCp Chron A Chron Ε



























These facts suggest that both forgyfan and andwyrdan can be key words that will give us a clear picture of and an insight into what lies behind jElfric's style of prose writing. What are the general syntactical and collocational features of each one of these two sets of verbs and what significant difference(s) can we observe between vElfric and the rest of the texts examined here?

Archaism in the vocabulary of AZlfric


3. Syntactico-semantic features 3.1. Sellan As the following examples in 1 to 20 show, sellan displays a wide range of collocational capacity, occurring with almost any kind of accusative object, material or non-material, animate or inanimate. Sellan often occurs with material objects such as hlaf as in (6), drincan as in (14), and gisl as in (17), but it equally frequently takes as its accusative object many different kinds of abstract nouns such as ap, lif\ gast, gewit, bietsung, heordnes or gyfu. It also occurs in the sense of 'to sell' when it is accompanied by 'w/a ilcan tid he oncneow, {)aet him waes eft lif seald ChonA 946.2 Scottas him aJ) as sealdon, J>aet hie woldan eal J)aet he wolde Beo 1269 gimfaeste gife öe him god sealde jECHoml 9.144.16 And se symeon him öa sealde bletsunge jECHoml 10.164.16 Crist Jjrowade for us and sealde us bisne .ECHomll 29.233.101 Se haelend da tobraec da hlafas and sealde his leorerum






(11) (12) (13)

(14) (15)





Schin 'ichi Takeuchi

JELS (Maccabees) Ne fegde is eowre lima, ne ic eow life ne forgeaf, ac middaneardes schippend eow ealde gast and lif he eft forgif J>aet ece life mid him swa swa ge nu sylla eow sylfe JElet4 (SigeweardB) 213 ac God sylf com Jjerto & sceawsede heora weorc & sealde heora aelcum synderlice spaece, |>aet heora aelc waes uncuö hwaet oöer saede GD2(C) 24.154.26 J)a hraöe urnon hi to Benedictus fotum & hine baedon mid mycclum wope, Jjaet he waere geeadmodod, J)aet he his forgyfnesse sealde J>am unlifigendan cnihte Bede & ne waes aenig se öe bebyrignysse sealde J)am öe swa hreowlice acwealde waeron ChronE 604.5 And Iusto he sealde Hrofesceaster ChronE 626.13 J>aer se cining sealde Pauline biscopsetl Jn(WSCp) 17.2 & swa Jdu him sealdest anweald aelces mannes, Jsaet he sylle ece life eallum Jjam J)e |ju him sealdest Bede & in waeter sendon & untrumum drincan sealdon GD2(H) 16.134.29 Ac t»a halgan Godes martiras noldon him syllan his haele, J>aet hi mid WHom 20.3.79 & syööan wiö weoröe syllaö of lande feondum to gewealde Godes gesceafte & his agenne ceap J>e he deore gebohte ChronE 876.2 Siööan wiö J)one here se cyning friö nam & him {)a gislas sealdon GD Pref/3(c)20.222.9 gif we us selfum ne sellaö ure heordnesse, ac J>aere upplican Godes gife jECHoml 1.26.8 he forgeaf blindum mannum gesihöe...dumbum he forgeaf getincnysse... deofulseocum and wodum he sealde gewit

Archaism in the vocabulary ofALlfric

(19b) (20)


jECHoml 32.480.11 and wodum mannum gewit forgeaf and blindum gesih^e Bo 41.142.11 He sealde swiöe faeste gife & swiöe faeste ae mid öaere gife aelcum men οδ his ende

3.2. Gesellan Gesellan pn the other hand mostly occurs with concrete objects, material and animate, as shown in the instances from (21) to (30) below. In (22) 'Iosep' is desribed as something that can be traded with money. In (23) Wulfstan uses Gesellan in the sense of 'to sell human beings' although it is accompanied by wid weoröe 'in exchange for worth'. In (24) and (29), too, gesellan occurs with the phrase wid weoröe, suggesting that it is used in the sense of 'to sell'. In (30), it occurs with ceht which, again, is a material object. In vElfric, too, the situation is the same. He uses gesellan in the sense of 'to sell' as in (21), (22) and (27), or it also occurs with a material object feo as in (28). jElfric actually applies gesellan to Latin vendere 'to sell' in (21).

These observations convince us that gesellan is semantically more delimited, hence collocationally more restricted, than sellan. (21) (22)


(24) (25)


vEGram 201.6 Ueneo ic gange to ceape oööe ic beo geseald jEGenPref 74 Eft Iosep, öe waes geseald to Egypta lande & he ahredde δ set folc wiö öane miclan hunger WHom 20.1.75 Eac we witan ful georne hwaer seo yrmöe gewearö Jjaet faeder gesealde beam wiö weoröe, & beam his moder & broöer oöerne fremdum to gewealde Or & siJ)J)an t>aet folc eall on ellJ)eodge him wiö feo gesealde Bede Ond f)a ilcan his dohtor to gehalgienne Criste J)am biscope to wedde gesealde ChronA 728.1 Her Ine ferde to Rome & J>aer his feorh gesealde & feng aej>elheard to Wesseaxa rice & heold [xiiii] gear


Schin 'ichi Takeuchi





jEHom 22 202 Smiöas hi worhtan smealice/mid craefte, and oft gesealdon J>a sylfrenan godas, sumne to maran wurde, be J)am J>e he gemacod waes, sumne eac waclicor, be Jjam J)e his wurö waes ^ C H o m n 24.205.171 Se frysa öa. J>a öa he hine gehaeftan ne mihte. Let hine faran on his truwan aefter 9am feo öe he him fore gesealde. and he swa dyde ChronE 1100.17 Ealle ne hi oööe wiö feo gesealde. oööe on his agenre hand heold. & to gafle gesette GD 2(C) 17.141.5 ...j)a he geheolde {>a feorh J>ara brodra, J>eah J>e he gesealde {)a achte haej>enum mannum

3.3. Gyfan Gyfan does not occur substantially in Old English as Table 1 shows, and it mostly occurs in a rather specialized sense of 'freely and gratuitously conferring on a person the ownership of a thing, as an act of bounty', accroding to the Oxford Enlish Dictionary. This is well documented by the examples in (31) to (40) below. But gyfan comes to occur more frequently, towards the end of the Old English period, in that sense which sellan previously covered. Gyfan in (32) and (34) is apparently used in the sense of 'to grant something' or 'to confer gratuitously', but in (31), which is from a twelth century manuscript, is used in the same sense as sellan. The passage dcet ofercom Rome pet ofercumed eall weoruld facet is gold & seolure clearly attests the fact that gyfan had replaced sellan by this time. In vElfric, gyfan occurs in contexts where some supernatural being, like God or Christ, for instance, bestows something on someone as in (36), (37), (38), (39) and (40). However, it does not occur very frequently even in iEfric, as Table 1 shows. iElfric seems to prefer forgyfan to gyfan in such contexts. (31)

ChronE 1123.61 Ac J>aet ofercom Rome J>et ofercumeö eall weoruld Jjaet is gold & seolure & se pape sweöolode & gaf him his palium

Archaism in the vocabulary ofALlfric











ChronA 959.1 Her he saente efter Sancte Dunstane & gaef him biscoprice on Wigracaestre & J>aer aefter J)aet biscoprice on Lundene Apt 48.21 öa underfengc me Arcestrates se cynge mid swa micelre lufe J>aet ic aet nyhstan geearnode J>aet he geaf me his ancaennedan dohtor to gemaeccan Bede Geaf he & sealde J>aet betste hors & {>aes faegerestan eondes Aidane J>am biscope ChronE 1103.9 & aer he heonne ferde he forgeaf {>a |)reo ^usend marc J)e him seo cyng Heanrig be foreweard aelce geare gifan sceolde iEHomM l(Bel 9) 152 God sylf gyfaej) alle monnum lif and gast [ipse Deus dabit omnibus uitam & spiritum] iEHomM 2(Bel 7) 145 His [apostoli] wrohten fela wyndrae & tancae aefter his upstige. & eac heom becom to öe Haiige gast of heofenum. & heom alle geaf alle Jjeodae spaece J)e on J)issere worulde beoö ^ECHomn 45.340.169 Salomon eac forgeaf jjaere cwene swa hwaes swa heo gyrnde aet him. toforan Jiaere cynelican lace öe he hire geaf jELet 4 (Sigeweard B) 107 öa on J>am sixten daege, syööan ()is idon waes, sceop J>e almihtigae God mannan of earöan, Adam mid his handum, & him sawle geaf iELet 4 (Sigeaweard B) 284 & {)e aelmihtiga God aefter Noes flode aellum moncynne geaf heom imaenelice fisccynn & fiigolcynn & öa fiöerfetan deor



Agyfan occurs as a full verb in the sense of 'to give back' as in (41), and in the sense of 'to restore' as in (42) below. It also occurs in the figurative sense of 'to die' in the phrase gast agyfan as in (43) and (46). But this is


Schin 'ichi Takeuchi

used basically in the same sense as 'to give back', because 'to die' means 'to give back one's soul to God'. Agyfan also occurs in the same function as sellan when it occurs in the idiomatic phrase to cwale agyfan 'put someone to death' as in (44) and (45).4 jElfric uses the phrase gast agyfan in 12 cases out of the total of 52 occurrences but in the rest of the cases he uses agyfan as a full lexical verb and does not use it in the sense in which sellan or forgyfan is used. In a context such as the one in (42) below, jElfric tends to use forgyfan. (41) (42) (43)




Or for J>am he hiere rice eft ageaf GD 1(C) 9.60.4 £>a [)a he ageaf leoht & gesihde twam blindum mannum iELS(Forty Soldiers)255 Jja cwaedon hi amen, and heora gastas ageafon and derdon swa gemartyrode to J>am aelmihtigan Drihtne J)e him aer gefultumode on dam frecednyssum Bede J>a sende he sone aerendwrecan to him & micel feoh, wid J)on öe he hine ofsloge oöJ>e him to cwaele agefe Bede ...jjaet he noht laöes me gedo ne J)ec {)inum feondum to cwale agife? vECHoml 26.382.22 ne beoö hi hyrdelease J)onne hi J>e habbaeö and he mid f>ysum wordum agef his gast

3.5. Forgyfan Forgyfan occurs either in the general sense of 'to give' or 'to bestow' or in the sense of 'to forgive', and it takes a variety of accusative objects, ranging from something material and concrete such as earldom and eall pet he cer ahte as in (48), or andleofen an in (53) to something non-material such as gesihp as in (47), where sellan could also occur, as we have seen. Forgyfan can also occur in the sense of 'to give someone in marriage' as in (49) where we find 'she was given in marriage to Aldfrith, king of Northumbria.' In (51), (54) and (55) forgyfan is used apparently in the

Archaism in the vocabulary of jElfric


sense of 'to forgive'. These instances indicate that forgyfan can be used for sellan almost interchangeably. JElfric uses forgyfan in the sense of 'to give' or 'to grant' as in (58), and in many cases he uses it in contexts where it is a supernatural being such as God or Christ that does the act of giving as in (59), (60) and (61). Forgyfan also occurs in the sense of 'to give up' or 'to cancel a debt', hence 'to forgive' as in (62) and (63). It is interesting to note, however, that iElfric uses forgyfan mostly in the sense of 'to give', as Table 3 shows. (47)


(49) (50)






LK(WSCp) 7.12 soölice on {>aere tide he gehaelde manega of adlum ge of witum & of yfelum gastum & manegum blindum he gesihöe forgeaf ChronE 1052.68 & se cyng forgeaf {mm eorle & his bearunum his fulne freondscype & fulne eorldom & eall jret he aer ahte ChronA 718.1 & hio waes forgifen Norman hymbra cyninge alder{>e WHom 7.135 God is swyöe mildheart & wile swyöe gemildsian & mycel forgyfan j)am [>e mid inwerdre heortan yfeles geswica & geornlice beta LK(WSCp) 11.3 Syle us todaeg urne daeghwaemlican hlaf & forgyf us ure gyltas swa we forgyfaj) aelcum J>ara J>e wiö us agyltaö GDl(c) 9.69.15 forJ)on l>e he J)am cnihte swylce gife forgaef J)aet he swa hraedlice mihte onfon swa hwaet swa he baed Bede ond swa swa he geheht, him ondlifen forgeaf & weoruldt>earfe ond eac swylce leafnesse sealde, J>aet heo mosten Cristes geleafan bodian & laeran Or J>a baed his faeder, waes eac Fauius haten, {)aet |)a senatum forgeafen J>aem suna J)one gylt CP 52.411.9 Hire sint forgifena swiöe manega synna foröamöe hio swiöe hreowsade



(57) (58)






Schin'ichi Takeuchi

Bo 38.119.26 3a cwaed ic: Ic eom swide gedrefed mid öisse spraece & wundrie forhwy swa rihtwis dema aenige unrihte gife wille forgifan GDl(c) 9.69.15 fortan J>e he J>am cnihte swylce gife forgaef iECHomII6.52.10 and he (cyning) forgeaf him öa wununge on cantwarebyrig. seo waes ealles his rices heafodburh iECHoml 1.26.8 He awende waeter to wine and eode ofer sae. mid drium fotum. and he gestilde windas. mid his haese. and he forgeaf blindum mannum gesihöe. and healtum and lamum rihtne gang, and hreoflium smeönesse and deafum heorcnunge. deofulseocum and wodum he sealde gewyt .ECHomll 21.189.280 He abaed eft siddan. aet öam soöan gode. J>aet he renas forgeaf. and eorölice wastmas ^ C H o m n 33.253.118 Eft siddan him forgeaf se aelmihtiga wealdend his gewitt. and he cwaed iECHoml 3.54.33 Hwi neltdu forgyfan j>a lytlan gyltas anum men. Jjaet se aelmihtiga god J>e öa miclan synna forgyfe MHom 6.258 Be 5am ge magon witan J)aet he is eallwealdend God. J)onne he swa mihtig is J?aet he maeg forgyfan ealra synna J>e hiom soölice behreowsiaö heora misdaeda

Table 3.

vECHoml 'to give' 60 'to forgive' 40

jECHomll 81 18

/ELS yEHom iEHomM 51 47 23 6 10 10


6 1

jEGram 5 3

While yEfric uses forgyfan both in the sense of 'to give' and in the sense of 'to forgive', it is particularly in his early works, i.e., Catholic Homilies I and II that he uses forgyfan more in the sense of 'to forgive'. However, he comes to use it less often in the sense of 'to forgive' in his later works where he comes to use forgyfan in the sense of 'to grant' more frequently

Archaism in the vocabulary ofAilfric


and consistently as Table 3 clearly indicates. Here we can detect a touch of design on jElfric's part to standardize the sense in which he uses forgy fan. To recapitulate, sellan, which is a context-bound verb, occurs with any type of accusative object, which shows that it is the most common verb of giving in Old English. Gyfan, which occurs in the sense of 'to give something freely and gratuitously', was on its way to gaining franchise as an authentic verb of giving, gradually replacing sellan towards the end of the Old English period. Gyfan is a context-free verb in this sense. Forgyfan, on the other hand, occupies a place somewhere in between. It shares a considerable part of its semantic field with sellan, occuring in the sense of 'to give', and it also occurs in the sense that gyfan denotes. Forgyfan also uniquely denotes 'to forgive'. Forgyfan is context-free when it is used in the sense of 'forgiving'. When it occurs in the most general sense of 'giving', it is context-bound. Based on this analysis, we can place these three verbs on the semantic scale as shown below.

Context-bound Sellan

Context-free Forgyfan

• • • Gyfan

Since sellan is context-bound and more lexically empty5, we can expect, at least in theory, that it should be used in combination with deverbal nouns and abstract nouns more readily than the other verbs of giving. I shall next examine these verbs from the point of view of their collocation with gyfu, forgyfnes and andswaru, because, as Table 4 below shows, jElfric uses gyfu, forgyfnes and andswaru frequently, whereas in the other late Old English texts their occurrences are almost negligible. Particularly remarkable in this regard is the fact that while jElfric rarely uses the verb andswarian as we have seen above, he often uses the noun andswaru in combination with various verbs to denote the act of answering. Is there any principle or philosophy behind vElfric's use of these nouns? We will take a brief look at some collocational features of the nouns gyfu, forgyfnes and andswaru in the following paragraphs.

352 Schirl'ichiTakeuchi Table 4.

^lfric 193 69 44 0

gyfu forgyfnes andswaru andwyrde

WHom 21 2 0 0

WSCp 7 5 6 0

Chron A 1 0 0 0

Chron Ε 8 1 0 0

4. Collocational features 4.1.


In Old English, sellan, gyfan and forgyfan all occur with gyfu as shown in the following examples in (64), (65) and (66), but agyfan does not occur with gyfu, which suggests that agyfan is more semantically restricted and specialized than the other three verbs. All these three types of collocation occur in ^Elfric, but the type that most frequently occurs is the one in (64). The type gyfan + gyfe occurs only once in JElfric, but considering the fact that it comes from a twelfth century manuscript, this may be due to semantic change, not due to iElfric's personal preference or style. (64)



Sellan + gyfe a) Beo 1269 gimfaeste gife öe him god sealde b) .ECHomll 22.193.118 l>a cwaeö se engel he gelyfde öaet gehwilc öe him aenige gife sealde Gyfan + gyfe a) ChronE 1125.20 & ofer eall he w!s underfangen mid wuröscipe & eallehine iaefen micele gife & maere b) ,ΈΗοπιΜΙ (Bel 9) 88 aefre he bid gyfende his gyf J)am öe he wyle Forgyfan + gyfe a) jEHom 10.113 ... his gife eow forgifj} b) GDl(c) 9.69.15 ίοφοη J)e he l>am cnihte swylce gife forgaef J>aet he swa hraedlice mihte onfon swa hwaet swa he baed

Archaism in the vocabulary of jElfric




Agyfan + gyfe No occurrence V+forgyfnes

Although sellan, gyfan and forgyfan all occur with forgyfnes, it is sellan that occurs most generally among the three. Gyfan and forgyfan occur with forgyfnes only once as in (69) and (70) respectively. Agyfan does not occur with forgyfnes in our corpus. The fact that forgyfnes commonly occurs with don, another context-bound verb, as in (72), again suggests that both gyfan and forgyfan are more semantically specialized than sellan, but the type gyfan + forgyfnes in (69), which occurs in a Peterborough interpolation of the twelfth century, also suggests that gyfan was beginning to take over the work of sellan. Under theses circumstances, however, jElfric uses both sellan + forgyfnes and don + forgyfnes, but does not use gyfan + forgyfnes and forgyfan + forgyfnes. (68)




Sellan forgyfnes a) GD2(c) 24.154.26 t>a hrade urnon hi to Benedictus fotum & hine baedon mid mycclum wope, J)aet he waere geeadmodod, £>aet he his forgyfnesse sealde Jiam unlifegendan cnihte b) HomS 11.1 (Belf5) 141 & heo sylö synne forgifenesse c) jECHomll 31-32.245.138 We syngodon. We dydon unrihtlice, syle us forgifennysse d) .ECHomn 3.20.57 for 3i J>e he ne sealde nane synne forgifenysse Gyfan forgyfnes a) ChronE 963.54 And ic wille J?aet ealle J)a freodom & ealle J)a forguenesse t>e mine forgengles geafen Forgyfan forgyfnes a) HomS 6 (Ass 14) 20 J)onne forgifö us drihten ure synna forgifenesse Agyfan forgyfnesse No occurrence

354 (72)

Schin'ichi Takeuchi

Donforgyfnes a) vEHom 10.83 ...and he sylf gedej) ealra forgifennysse on eallum middanearde b) iEAdmon 2.41 ...he sceal don J>onne forgifenysse eallum dam mannum c) ChronE 1066.137 Se aelmihtiga God cyjje his saule mid mildheortnisse & do him his synna forgyfenesse d) WHom 14.39 & he wile miltsian & mycle forgyfnesse don forsyngedan mannum e) Conf 3.1.1 (Raith Y) 3.14 ac do man him forgifenesse for godes lufan

4.3. V + andsware In collocation with andswaru, the verb most commonly used is sellan as in the following examples in (73). Gyfan occurs with andsware only in 4 cases in Old English as shown in (74). This may be due to the fact that gyfan was still a highly lexical verb in Old English as Yamanouchi (1994: 48-49) points out. Forgyfan does not occur with andsware, as we can theoretically predict. Agyfan occurs with andsware only twice in prose while we have 26 instances in verse. This may be partly due to its alliterative function and partly due to the general archaism inherent in verse. (73)

Sellan + andsware a) jECHoml 31.460.34 J>aet he is gewriten and nane andsware syllan ne maeg b) CHomn 38.280.8 ...ne mihton him andsware syllan for J)am twam apostolum öe £>aere cumene c) LS 13 (Machutus) 24.V.15 Jja heora nan him andswara naes sellende d) Nie (A) 67 Hig suwedon 7 ne cuöon nane andsawre syllan e) BenR 66.1 J>e mid gesceade cunne andswara syllan underfon

Archaism in the vocabulary of Ailfric




(75) (76)

BenRW 31.21 heo nabbe aehta to syllenne, sylle gode andswara fortan J>e hit is gewriten {iaet seo gode andswaru sy ouer da selestan selene g) ByrMl (Crawford) 104.13 Ic wat J>aet ic J>e mot andsware syllan Gyfan + andsware a) LS14 (MargaretAssl5) 144 And seo faemne andswaro geaf: Hwaet J)u nu, earming, mid... b) LS 29 (Nicholas) 451 heo swiöe ofwundrode and ne geafaon him nan andswaere c) BenRW 2.99 underfangen to witienne, gearewige hi seolfe to gifonne andsware for heom d) BenRW 66.1 J>e wel cunne andsware gyfe and underfo Forgyfan + andsware No occurrence Agyfan + andsware a) LS 14 (MargaretAssl5) 276 And hi J)a andswera ageaf: Ne J?e ne {sinum godum ic ... b) Alex 808 da bad se socerd sunnan setlgongen for |)on sunnan trio agefed ondsware aet J>am upgonge 7 eft aet setlgonge

5. Cwedan to andsware As a paraphrase of andswarian, the frequent use in jElfric of the type cwedan to andsware 'say in answer' draws our special attention. Of the total of 34 instances that occur in Old English, 24 cases occur in ^Elfric alone as in (77) and (78) below (Takeuchi 1994: 24), as if to suggest that vElfric deliberately avoids the use of the verb andswarian. In other Old English texts, however, different types of paraphrase are used as in (79) and (80).






Schin 'ichi Takeuchi

Cweöan to andsware a) JELS (Book of Kings) 355 öa cyddon öa cnihtas £>am cynincge J>aet, and he cwaeö to andsware, öaet, hit waes aer swa gewitegod b) iEHomM2 125 Heo cwaeden öa to andsware, Nute we na to saecgenne hwanon Iohannis fulluht beo c) yECHoml 364.20 Petrus him andwyrde, öu eart Crist, öaes lifigendan Godes Sunu. Brihten him cwaeö to andsware, Eadig eart öu, Simon... d) iELet 4 (Sigeweard 2) 1107 {ja befran Iohannes faerlice, cwaeö: Hu ys he la dead oöö hwilum deaöe? He cwaeö him eft {ras to andsware: He is Gode dead. Seegan to andsware a) vEHom 17.179 and se Haelend {>a saede dona him to andsware... b) ΛΕΗοπιΜ 2(Bel 7) 222 Heo saedon him to andsware, Beo öu his leorningeniht. Cweöan andswariende a) Lk (WSCp) 7.40 {ja cwaeö se Haelend him andswariende, Simon ic ... b) LS 13 (Machutus) 20r. 16 Se halga machu him andswarigende |ras cwaeö: Cweöan on andsware a) GDPref and 4(c) 49.338.1 He hraöe gehyrde eft cweöan to him on andsware

6. Conclusion We have discussed the chief verbs of giving and those of answering, focusing on how vElfric uses them in the general context of Late Old English standardization. iElfric was certainly trying to standardize the use of certain words and phrases, but this does not immediately lead us to conclude that he was trying to set standards, as Gneuss and his followers say, for the dialect of West Saxon, or late Old English in general, because IElfric's use of

Archaism in the vocabulary of AZlfric


andwyrdan is clearly a case of idiolect, deviating from the main streams of both the Winchester vocabulary and of late Old English in general. The frequent use of forgyfan may be characteristic of jElfric, but it is mainly in the sense of 'to give', or 'to grant' that iElfric uses it most frequently, as we have seen above. Considering the fact that sellan, gyfan and forgyfan were undergoing what seems to be a structural semantic change, it may be said that ./Elfric was using the verb forgyfan deliberately in its old, well-established, or, more properly, in its archaic sense. The same observation holds with regard to iElfric's exclusive use of andwyrdan in preference to andswarian, which, as we have seen, was the verb most commonly used throughout Old English. Andwyrdan, however, eventually goes out of use in the course of Middle English and andswarian instead survives as answer in Presentday English. Here again, we must conclude that iElfric seems to deliberately use archaic vocabulary, a stylistic feature that is not necessarily shared with the group of texts that Gneuss classifies as the Winchester Group. jElfric was a careful and meticulous writer, and as he himself says in the preface to his First Series of Catholic Homilies, he was trying to be strict in his use of language in order to avoid errors in what he says and how he says it. In order to determine whether the Winchester vocabulary was really a collaborative effort, originating in ^thelwold's school at Winchester, to standardize the dialect of West Saxon or not, we need further extensive studies not just in syntax and phonology, but in semantics as well, of which the present study is only a tiny example.

Notes *

I am grateful to Professor Peter Clemoes, Emaniel College, Cambridge, for giving me his kind permission to take photocopies of his Ph.D. dissertation. I am also grateful to Professor Antoinette diPaolo Healey, Dictionary of Old English, University of Toronto, for giving me a chance to use the DOE library for the present research. My acknowledgement also goes to Dr. Hugh E. Wilkinson, Professor Emeritus Aoyama, Gakuin University, for reading the present paper in draft and making many useful and invaluable comments.

1. There were some scholars, even Gneuss, who pointed out some unique aspects about yElfric's vocabulary such as Dietrich, Jost and Schabram, just to mention a few (see Hofstetter (1988) for details), but it was Gneuss who systematically and in a convincing manner advanced his hypothesis that there must have been some strong current at Winchester to 'standardize' the use of vocabulary in late Old English.


Schin 'ichi Takeuchi

2. The abbreviated titles of the source materials used in the present paper are those of A microfiche concordance to Old English, University of Toronto. 3. The word 'contex-bound', in contrast to 'context-free', is used in the sense that a verb like sellan in Old English exclusively depended on its syntactical context for the actual 'surface' meaning. Sellan, in thissense, could be called 'functional' as with make in to make fun of, to make answer to in Presentday English. 4. See Fascicle A: Agyfan, Dictionary of Old English, University of Toronto (1994) 5. Sellan is more functional, in other words. See (6) above.

References Clemoes, Peter 1955-1956 Mine's Catholic Homilies First Series: The Text and Manuscript Tradition. [Unpublished Ph.D. dissertation, Univ. of Cambridge, King's College.] Clemoes, Peter - Kathleen Hughes (ed.) 1971 England before the Conquest". Studies in primary sources presented to Dorothy Whitelock. Cambridge: CUP. Gneuss, Helmut 1972 "The origin of Standard Old English and jEthel wold's school at Winchester", Anglo-Saxon England 1: 63-83. Healey, Antonette diPaolo - Venezky, Richard (eds.) 1980 A Microfiche Concordance to Old English. Toronto: University of Toronto. Hofstetten Walter 1988 "Winchester and the standardization of Old English vocabulary", AngloSaxon England 17: 139-161. Hogg, Richard M. (ed.) 1992 The Cambridge history of the English language. Vol. 1: The beginnings to 1066. Cambridge: Cambridge University Press. Takeuchi, Shin'ichi 1994 "Some notes on Heo eweed him to andsware in Old English", Studies in Humanities and Social Sciences 70: 27-36. Thorpe, Benjamin 1844-1846 The Sermones Catholici or Homilies of AZlfric, 2 vols., London: jElfric Society. Wormald, Francis 1971 "The Winchester School before St iEthelwold", in: Peter Clemoes - Kathleen Hughes: 305-313.

Archaism in the vocabulary of&lfric


Yamanouchi, Kazuyoshi 1994 "Eigono Korokeshonto Idiomuno Keisei 'Formation of collocations and idioms in English", Korokeshonto Idiomu 'Collocations and Idioms." Tokyo: Eichosha Publishing Company.

The syllable structure and phonological processes in the history of English Michiko Terajima

1. Introduction The syllable is hierarchically structured such that a number of levels are related to each other: the syllable node is linked to two immediate constituents, the onset and the rhyme, the latter of which, in turn, is linked to the nucleus and the coda. Phonological processes often depend on the weight of syllables. The syllable weight distinction is not universal, however. It sometimes varies across languages and even changes over time within one and the same language (Terajima 1994b). In Old English, for examples, a syllable is 'heavy' if it contains a long vowel/diphthong, or a short vowel/diphthong obligatorily followed by two or more consonants as in (lc-d). An open syllable which contains a short vowel/diphthong or a closed syllable which contains a short vowel/diphthong followed by one consonant is 'light' as in (la-b).



lufu 'love'

giefu 'gift'


bcec 'back'

lim 'limb'

c. d.

hüs 'house' prymm 'glory'

eage 'eye' eorl 'chief

It seems that phonological processes operate differently depending on whether the onset/coda position in which they are implemented is peripheral or internal to a word. In this paper we will be concerned with how phonological stimuli behave in word-initial/internal onset position and in word-internal/final coda position. The phonological processes we will discuss here with respect to two major constituents of syllable structure Eire voicing, devoicing, deletion, gemination, spirantization and palatalization. Our data is from English diachronic phonology.


Michiko Terajima

2. Analysis 2.1. Voicing 2.1.1. Onset We will begin with the voicing process in OE, ME and ModE where it is implemented in word-initial and word-internal onset position. The voicing process from the Middle English period seems to have taken place in the Southwest, Southeast, and Southwest Midland dialects (3c). The voicing from [Θ] to [δ] in ModE seems to have occurred only in a restricted group of words. (4c) illustrates the examples with the prefix plus the stem. The onset occupies the beginning of the stem. (2) a. b.

f> ν vader 'father' / vinden 'find' gie *fan 'give' / heo *fon 'heaven'

a. b. c.

θ>a panken 'thank'/pombe 'thumb' fe · per 'feather' / δ · per 'other' / weor · dan 'become' pis/that/these/thos 'those' / thens 'thence' / thider 'thither'

a. b. c.

s>ζ ζorje 'sorrow' / ζone 'son' po · sess / ver · sion / pan · sy re· solve I re· sign I de· serve I de· sire



2.1.2. Coda Let us now have a look at the voicing change that takes place in word-final and word-internal coda position. The processes in (5e) and (5f) were observable from the end of the 15th and the 16th centuries, especially in the northern parts of Suffolk. (5)


θ>δ with / with · out

The syllable structure and phonological processes b.


S> Ζ

is/was/his/hers/as/these/those/bryddez 'birds' / dayez 'days' / Godez 'God's' c.

f >ν actif 'active' / passif 'passive'


tj>d3 knowleche > knowledge / caboche > cabage 'cabbage' / spinach > spinage


ρ >b coppeweb > cobweb / cupboard > cubberd 'cupboard' / loppestre > lobster


k>g sack · but > sugbut

2.2. Devoicing Let us next go to the devoicing process. As compared with the reverse process, this process is low in frequency because it is restricted to coda position. All the examples we have appear in word-final position. The process in (6) began in OE and was succeeded by the Northwest Midland dialect in ME. The change in (7a) seems to have occurred particularly in the 13th century West Midland dialect. The examples illustrated in (7b) can be found in unstressed suffixes. They come from OE and ME. The g > k process in (8) is seen in OE and inherited by the Northwest Midland dialect in ME. The evidence in (10b) is seen in unstressed suffixes. (6)

b>ρ lomb > lomp 'lamb' / domb > domp 'dumb' / godsib > gossip


d>t a. elpend > elpent 'elephant' / püsend > püsent 'thousand' / freond > freont 'friend' b. cleopode > cleopede > clepde > clepte '(he) called' Ißlde > feite '(he) felt' / dremed > dremt 'dreamt'


Michiko Terajima


heardlice > heartlice 'hardly' / bredful > bretful 'brimfully / goldsmith > goltsmith


g>k ding > dine 'thing* / strong > stronc / strengp > strencp 'strength'


ν>f deav > deaf 'deaf / llv > IJf 'life' / luv > luf 'love'


ζ>s a. nos 'nose' / ris 'rise' b. shipes[s] / books / monthes / keeps / looks

2.3. Deletion It is possible to view the process as the extreme degree of consonant weakening. The deletion process is usually preceded by weakening or reduction processes in the syllable where the consonant to be deleted occurs. As is well-known, [t h η w] are the most, [d k g χ 1 r] are the second most, [p b d3 s ζ θ m] are the least vulnerable to deletion process (Nakao 1985: 2.3.1. Onset We will now look at the deletion phenomena which occur in word-initial /internal onset position in OE, ME and ModE. /t/ tends to be deleted when it is in the vicinity of its phonetically coarticulatory segments /s n/. The examples in (11c) show the [t]-dropping of /st/ cluster. /w/ is deleted after the word-initial /c s h t/ in (12b). This shift began in OE and continued into ME. In word-initial position it occurred in ModE as in (12a). The deletion process is also observed in onset position of the second elements of compound words, after ne or after a pronoun ich Τ as in (12c). We have the examples for word-internal deletion in (12d). [h] is ready to drop out in word-initial position. This deletion process is also observed in onset position of the second elements of compound words, after a negator ne, or the prefix ge- as in (13b). From early ME it is seen in

The syllable structure and phonological processes


onset position. They revived in the South-Humbrian areas in early ModE. As a consequence of Standard English having restored the lost [h], however, it has come to be branded as a vulgarism (13c). [x] occurs after back vowels and the diphthongs eo / eo, cea / äsa in Old English. This deletion process takes place in OE. It appears in word-initial position. [k]-dropping in initial position begins from the 17th century onward. Dd is rarely deleted in word-internal onset position in (15b). /d/ is dropped out in initial onset position after [n] in the EModE times as in (16). From the 13th century [g] begins to be lost in word-initial onset position if immediately followed by /i/ or /e/. [g] disappears in onset position from the 16th to the 17th century. /g/-deletion also occurred between /n/ and Irl or IV as in (18b). This takes place very sporadically as early as OE. fbl is deleted at onset position after fml in (19). [v]-dropping occurs before consonants such as Im r 1/ in ME. [j] is effaced at onset position. This is a typical process in ModE. (Π)


a. b. c.

be · te · sta > besta 'best' sof· ten 'soften' [sofn] / swif· ten 'swiften' [swifn] / o f · ten li · sten I fa· sten / gli · sten / ha· sten


[w] a. write / wrong / wlate 'loath' / wlatsum 'loathsome' / ME sword > ModE surd b. cwucu > cucu 'alive' / swuster > suster 'sister' / *hwu > hü 'how' / twä > two > to 'two' c. ne wät > nät 'know not' / ne wäBron > nseron 'were not' / ich wulle > ichulle Ί will' d. bet · wux > betuh 'between' / hläf· weard > hlaford 'lord' I full · wuht > fulluht 'baptism' OE Cant · wa · ra · by · rig > ME Canterbury / OE his · wif> ME hussif 'hussy'


[h] a. his > is / hcefde > cefde '(he) had' / OE hring > ME ring / OE hlaford > ME lord / OE hnutu > ME nute 'nut' b. ne habban > nabban I ge · hwelc > gewelc 'each'


Michiko Terajima


harm > arm / hard > ard / hundred > undred / house > ouse


[x] *rä · ha > OE rä 'roe' / *fo · hart > OE fön 'take' / *slea · han > OE slean 'slay'


[k] a. knave / knee / knife / knight / knit /fcnofc/ knot / b. OE hor · sc /ii'ce > horslice 'readily' / OE mu · scle > musle 'mussel' / ME ma· ke · de > made / ME α · sked > ast 'asked'


[d] EModE Lon · don > Lunnon / EModE thun · der > thuner / EModE hun · dred > hundred


[g] a. OE frador > eador 'together' / OE $ if> if / LOE φ icel > ikle 'icicle' b. OE win · geard > wlnard 'vineyward'


[g] a. gnaw > naw / gnat / gnarled / gnash / gneiss / gnostic b. /ryn · gran > hinran 'hunger' / cen· glum [dat pi] > cenlum 'to the angels'


[b] ME tim · ber > timer 'timber' / ME chaum · ber > chaumer 'chamber'


[v] _ OE as · fre > ME er 'ever' / OE ο · fer > ME or 'over' / mar · ve/ > marl / pave · ment > pament Μ year > ear / yeasi > east / look yee > look ee


The syllable structure and phonological processes


2.3.2. Coda Let us now take a closer look at the cases in which the deletion process is realized in word-final and word-internal coda position. /n/ tends to enhance the process before the word-boundary in OE as in (22d). In OE word-final /n/ is dropped across the board in unstressed syllables. This /n/-effacement process was already set in motion in the Northumbrian dialect before the middle of the 10th century, affecting /n/ that appears in nearly all of what once constituted the inflectional affixes as in (22a). This process also appears in word-final position in ME as illustrated (22b). From the 15th century word-final [n]-deletion started after Iml as can be seen in (22c) and continued into EModE. The [t]-deletion occurs at the coda which consists of three- or fourconsonant clusters, [t] usually disappears in the middle of the cluster. [x] and [9] are in complementary distribution with [h], as has already been mentioned in 2.3.1: [x] occurs after back vowels, while [9] after front vowels. In ModE post-vocalic [r] is deleted in coda position. This [r] is deleted, after an intervening stage of being vocalized to [a]. As is exemplified in (25b), we have instances in which the change takes place in word-internal position. IM drops in coda position. /l/-deletion is also occasioned at coda position where Ν is sandwiched between stressed vowels and labials or velars [f ν m k]. The process took place in the 14th to the 15th century as in (26b) and from the 15th to the early 16th century as in (26c). /d/ is subject to loss particularly when it is in the neighborhood of segments [n s 1] in coda position. Palatal fricative [g] is lost in coda position where it occurs between [back] vowels and consonants such as [d η 1 δ]. This deletion is a feature of the West Saxon and Kentish dialects of Old English. In ME [g] is deletable at word-final coda position where it is preceded by fo]. /b/ drops out at word-final coda position where it is preceded by Iml. This process begins c. 1300. ME [Θ] and [δ] are occasionally subject to loss in coda position. The latter is effected in a few words (31b). [v] and [f] are deleted in coda position in ME.


Michiko Terajima


[η] a. heortan > hearta 'hearts' / gödan > god 'good' / biddan > bidda 'ask' b. min > mi / pin > pi / abuten > abute 'about' / beforen > before / maiden > maide c. damn > dam / hymn > hym / autumn > autum d. wsBpn · man > wäepmann 'man' / ein · boga > elboga 'elbow' / emn · lice > emlice 'equally'




b. c.


myjt > my$ 'might' / brojt > broj 'brought' / seint > sein 'saint'

wcestam > wcesm 'fruit' / bintst > binst '(thou) bindest' / eetst > ehst '(thou) attackst' biet · sian > blessen 'bless' / blost · ma> blosma 'blossom' / nost · le > nosle 'fillet' / nest · le > nesle / cast • le > casle / host · /er > hosler / chest · nut > chesnut

[x] and [9]


*fux · sftj >fyst 'fist' / *eo/i · reed > eored 'band' / O E doh

•tor> (14c) doutor 'daughter' / OE poh *te> (14c) thout 'thought'


OE niht > nit 'night' / cniht > knit 'knight' I flyht > flyt 'flight'




scarce > scase / Worcester > Wocester / burst > bust / nurse > nuss /fc/rd/ storm I fear / fire b. par · fe · si > pallsie 'palsy' / mor · ce/ > mossel 'morsel' / smor · smotheren 'smother'


[1] a.

b. c. (27)


> M E ech 'each' / OE hwylc > ME hwich 'which' /

swylc > ME such aim / äm 'aim' / alle > äk 'auk' / calf > cäf 'calf chalk > chauk / talk > tauk / half> hauf


a. findst >finst '(thou) findest' I fundn >fun 'found' b.

wed · nes · day > wenesday 'Wednesday' / and · swe · rian

> answeren 'answer' / /umJ · some > hansome / hand · kerchief > hankerchief

The syllable structure and phonological processes


[g] a.

'thane' / sneg

pegn > pen

I > snel

'snail' / sceg




'said' b.

mceij · den > mäsden

'maiden' /

dig · nen >


'handmaid' (29)

[g] ferthing teching


~ ferthyn ~

'farthing' /


~ drynkin

'drinking' /


[b] lamb > lam / dumb > dum / tumb > tum / comb > com


[Θ] and [δ] a.

b. (32)

O E Norf) *folc > M E Norfolk / O E weord weorscipe 'worship' / months > mons odr > or 'or' / hwedr > hwer 'whether'

[v] and [f] a. evn > en 'even' / abuvn five pence



OE heaf·

des >


M E lady half · peni > halpeni 'hussy' / hcelf· tre >

> abün

'above' /

· scipe > M E

tar · div > tardy



ME hedes 'head's' / OE hlsef· 'halfpenny' / hüs

· wif>

di · ge > huzzy


2.4. Epenthesis The epenthetic processes occurred in the history of English fairly frequently. As compared with deletion mentioned just above, however, the incidence of this process is rather low. The epenthesized consonants tend to be stops [t d ρ b k]. It is only rarely that the nasals [q n], glides [j w], and fricative [h] are epenthesized. 2.4.1.


We will first consider the epenthetic processes in word-initial/internal onset position.


Michiko Terajima

First [h]-epenthesis. It takes place from EME in stressed initial syllables as in (33). Next [j w]-epenthesis. The front glide [j] is inserted between [h kl] and the stressed vowel [e] as in (34a), or between the initial word boundary and stressed front vowel [e ε] as in (34b). The back glide [w] is, on the other hand, epenthesized in initial onset position before stressed back vowels [δ, 5] as in (35a). In the South including Kentish and West Midland the [w]-insertion is also implemented between [b ρ g] and [δ, 5] as in (35b). The examples in (36) are those in which [k] is inserted at word-initial onset position after [s]. (33) (34)



0 -> [h] herthe 'earth' / helde 'age' / harme 'arm' / halle 'all' 0 [j] a. hier 'here' / hieren 'hear' / clier 'clear' b. (15c) j ende 'end' / (15c) j euery 'every' / yese 'ease' / yeere 'ear' 0 [w] a. wöther 'other' / won 'one' / wöth 'oath' / wöld 'old' b. guöd 'good' / budn 'bone' / pwöpe 'pope' 0 [k] OE slät > sclät '(he) slit' / OE smlgende > scmegende 'meditating' / OE snecendce > scnicende 'sneaking' / ME slak > sclak 'slack'

Next let us look at the epenthetic processes which take place in wordinternal onset position. Those epenthesized here are limited to stops [d ρ b]. [b] and [p] are epenthesized between [m] and [1 r n] as in (37) and (38). [d] is inserted between [n] and [1 r] as in (39). (37)



0 [b] OE brae · mel ~ brcem · bei 'bramble' / OE rise · mel > ME nim • ble · ble 'nimble' / ME slu · me· ren > slum · bren 'slumber' 0 - » [ft OE nem · nan > ME nem · pnen 'name' / OE hier · sum · nes > ME her · sum · pnesse 'obedience' 0 -» [d] OE gan · ra~ gan · dra 'grander' / ME kyn · red > ModE kin · dred 'kindred' / ME spin · die 'spindle'

The syllable structure and phonological processes


2.4.2. Coda Let us proceed to see how the processes in question operate in wordinternal/final coda position. We will begin with /t/-epenthesis, which is induced both in wordinternal and word-final coda position. In the former position it is seen only after [s] as in (40a). In the latter position it can be observed in word-final coda position after /n s f x/ as in (40b). [d] is inserted in word-final coda position after [n 1] as in (41a). In word-internal coda position the insertion also takes place as in (41b). [p] is epenthesized between [m] and [t s] as in (42a, b). Finally nasal insertion. Historically this insertion began in the early 13th century and continued up to the ModE period, [n] is epenthesized before [d3, t], and [rj] is inserted before [g] in word-internal coda position as in (43a, b). (40)




0 a.

[t] OE mae —> sling ~ maest · ling 'brass' / OE mi · slic ~ mist • lie 'various' / NrthB *hly · sna > ME list · nen 'listen' b. against / middest 'mids' / betwixt // OF auncien > ME auncient/ME orphane > orphant//draff > draft/thrught 'through' 0 —> [d] a. soun > sound/ expoun > expound/ col > colde 'coal' b. OE en · le · fan ~ end · leo ·fan 'eleven' / hoi · some > hold · some 'wholesome' / ME sud · den · ly > ModE sud · dand · ly 0 -> [p] a. ModE dream(p)t / mushrum(p)s / ME glim · sen > ModE slimpse b. OE läe · tig ~ xmp · tig 'empty' / ME dem · ster > demp · ster 'judge' / ModE Ham(p) · ton 0 [n, g] a. mes · sa· ger ~ mes · san · ger/stal · la · ge > stal · lan · ger/her •be* ger > (15c) har · bin · ger/por · ra· ger > por · rin · ger b. OE nih · te · ga · le > ME nih · ting · a· le > OE pa· re· ga· le> ME pe · ring ·α· le 'equal' I far · thing ·α· le/ whar · fing · er


Michiko Terajima

The evidence clearly shows that the inserted consonant is usually one that is homorganic with what precedes or follows it: [p-m] / [b-m] / [d-n] / [d-1] / [n-1] / [r-s] / [r-1] / [z-1] / [z-r]. It has also been demonstrated that the epenthetic processes do not particularly conspire to yield the unmarked CV or the CVC syllable structure that is characteristic of English. The distribution of epenthesized consonants in onset/coda position is given in Table 1. Table 1. The distribution of epenthesized consonants at onset/coda position.

word-initial onset word-internal onset word-final coda word-internal coda

(##)_V 32 7 CV 0 23

(##)C_V 41 19 CVC / CV 47 10


2.5. Gemination There has been much controversy about whether Gemination produces simple consonants or complex consonants. In the history of English Old English was the only period when Gemination was 'productive'. In what follows we will discuss whether OE geminates were complex or simple in terms of diachronic evidence. Let us first consider whether the output of Gemination should be syllabified as complex or simple. We take apple as an example. If were a simple consonant, the Maximal Onset Principle would require that cepple be syllabified as ce · ppel not as cepp · el because the latter is against the Principle. The evidence from Old English denies the existence of words where two orthographically identical consonants are allowed in wordinitial position. This rejects the possibility that the output of Gemination is phonetically simple. Thus ceppel should be syllabified as cep · pel, that is, as a complex consonant. (44)

cep · pel 'apple' / cup · pe 'cup' / lib · ban 'live' / rec · can 'stretch' / docga [do r · ra] 'dog' / frocga [fror · ra] 'frog' / pyf'fan 'puff / seed · dan 'scathe' / cys · san 'kiss' / croh · ha 'crock' / swim ·τηαη 'swim' / feal · lan 'fall' / steor · ra 'star'

The syllable structure and phonological processes


We will next proceed to phonological evidence in favor of the view that geminates act as double consonants. It is possible to find cases where in Old English geminates contrast with simple congeners word/stem-finally as in (45). See also the examples in (47). (45)

hyppe 'gain' [dat sg] Φ hype 'covenant'

Let us have a look at the examples in (46). In (46a) the coda of the first component [p] has joined to form a geminate [bb] in the wake of its having been assimilated to the onset of the second component. The examples in (46b) may also belong to the same category as (46a). (46) a. b.

cup · board > cubberd 'cupboard' cop · pe · web > cobweb 'cobweb' / Camp · bell > Cambbell > Cambell 'Campbell' / loppestre > lobbester > lobster

In so far as Old English phonology is concerned geminates behave like complex consonants and are allowed to appear as such in word-final coda position only, but not elsewhere. 2.6. Spirantization For spirantization in Germanic languages Foley (1977) establishes the following implicational relations: b>ß implies g>v, d>3 implies g>Y and g>ß. These implicational relations enable one to set up a strength scale for spirantization shown in (47). (47)

velars labials dentals g b d 1

The historical data from English, however, demonstrate that dentals and velars are more likely to spirantize than labials. The same strength scale can be established for the opposite process, despirantization, as well. (48)

dentals velars

labials •




Michiko Terajima

Spirantization is scarcely documented in historical phonology. Despirantization, on the other hand, occurs a little more frequently. Historical Spirantization can be summarized as follows: (50)


d ö (OE, ME, ModE) t s (ME) t -> θ (ModE) tf —»J" (ModE) d3 3 (ModE)

p - > f (OE) b —» ν (OE) k —> χ (OE)


We will begin with the spirantization processes in OE, ME and ModE which are implemented in word-initial and word-internal onset position. The process from [b] to [v] in (50) occurs between resonants after the early OE period. This, however, seems to have neared completion in the late 8th century. It has to be noted that [b] is probably [ß], an allophone of /b/. In OE alternations between [d] and [ö] can be observed in the context V_V as in (52a). After 1400 [d] is also weakened to [δ] between stressed vowels and syllabic [r]. Some items such as 'mother' and 'father', however, exhibit vacillation between [d] and [Ö] up to ModE in (51b). The spirantization of [t] > [Θ] operates at word-initial onset position as in (52a) and word-internal onset position as in (52b). This process began in French loanwords in late ME. It has not affected a few items in British English, such as Thames, Anthony, thyme. (50)

b>f ο · ber > ο · fer 'over' / gi · baen bae > hal *fae 'half [Dat sg]


'give' / hal ·

d>ö a. sm · dan > snipan 'cut' / weor · dan > weo · pan 'become' / gi · do· pta > ge· do· fta 'comrade' b.


> gi · faen

fa · der >fa · ther/mo · der > mo· ther/ga ther > ge· ther /whi · der > whi · ther

· der >


t> θ a.

teme > theme / trone > throne / te· a· tre > the · a· te · chying > the · chyng 'teaching'


The syllable structure and phonological processes



au · tour > au· thour / Ca· to· lie > Ca· tho · lie /or · to • gra · phy > or · tho · gra · phy / diph · tong > diph · thong /me · tod > me· thod

2.6.2. Coda Let us now have a look at the spirantization process in word-final and word-internal coda position. The change from [p] to [fj as in (53) appears before [t] in word-final / internal position. In OE [k] spirantizes to [x] in word-final position in function words. This process occurred before the 9th century especially in Northumbrian. We also have the [t] > [Θ] alternation in word-final coda position, [tj] and [d3] were also spirantized particularly after /n/ as in (56a, c) and l\l as in (56b, d) in early ModE. This spirantization is restricted to monosyllabic words. (53)

ρ>f sceapt > sceaft 'shaft' / gi · dop ·ta> ge · dofta 'comrade'


k>χ ac > ah 'but' / ic > ih Τ / dec > deh 'though'


t θ dout > douth 'doubt' / debt > debth 'debt'


t/>Jandd3>J a. branch / inch / clensh 'clinch' / hansh 'haunch' / bench / quench / wench / French > frensh / lunch b. belch /filch c. singe / cringe / hinge / gin · ger / change / strange / dan • ger/an · gel/ο · range d. bilge / bulge / in · dulge / di · vulge



2.7.1. [kkkgggsc]


Palatalization has taken place rather often in the history of English. In OE it affected [k kk g gg sc] and changed them into [tj d3 J], [k kk g gg sc]


Michiko Terajima

became first velar-softened to [c cc ae J J], and then palatalized to [tj d3 J]. Two fricatives [yx] also underwent the Velar softening process and were converted into palatal fricatives [g ς], which are responsible for Palatal Diphthongization. The shift to [tj d3] is believed to have begun about the middle of the 9th century and was completed in the later half of the 12th century. The shift to [f], on the other hand, seems to have begun in word-initial position around the same time as the shift to [tf d3] and then spread to word-internal/final position. Palatalization processes revive in ME, having effects on three dentals /s ζ d/ this time in such a way that they generate LT d3 3]. Onset Let us now begin with [k kk] > [tf] shift where Palatalization operates in word-initial (57a) and word-internal (57b) onset position. Two geminates [kk] and [J ] have the second elements palatalized first, and then have the first elements assimilated to what follows, which results in the surface forms [tf t j d3 d3]. The outputs are to be simplified later in terms of Degemination (Nakao 1985: 390-391). In (59) we will discuss the [sk] > [J-] shift. Palatalization also affects [c] after /s/ differently from unprotected [c] and produces [J]. The stage of [J] must have been arrived at before 950 (Nakao 1985: 390). The immediate output [s J] has the first constituent [s] assimilated to the following [J], and reaches the geminate sequence [J J"]. Later it triggers Degemination and surfaces as [f]. (57)

[k kk] > [tf] a. ci · ή · ce 'church' / cinn 'chin' / cl· dan 'chide' / ceald 'cold' / ceorl 'churl' / ceo · san 'choose' b. mi · eel 'much' / se · can 'seek' / pen · can 'think' / bir · ce 'birch' c. wrec · ca 'wretch' / strec · can /stretch'


[g gg] > [d 3 ] sen · gan 'singe' I fin · ger 'finger'


[sk] > [J] a. scip 'ship' / scoh 'shoe' / scot 'shot' / sceal '(he) shall' / sceap 'sheep' / serüd 'dress' I sei· fan 'decree' b. wa · scan 'wash' / wy · scan 'wish'

The syllable structure and phonological processes

377 Coda We will now look at the palatalization phenomena which appear in wordinternal/final coda position. The processes [k kk] > [tj], [g gg] > [d3] and [sk] > LT] followed the same course as in onset. (60)

[k kk] > [tj] a. pic 'pitch' / ic Τ / bene 'bench' / drinc 'drink' / swile 'such' b. crycc crutch'


[ggg]>[d3] a. ping 'thing' / streng 'string' /hring 'ring' b. secg 'man' / ecg 'edge' / brycg 'bridge' / hryct 'ridge' c. i · cgan 'lie' / le · cgan 'lay' / se · cgan 'say'


[sk] > U] cesc 'ash' / fisc 'fish / flsesc 'flesh' / men · nisc 'human'

2.7.2. [s ztd]



If ModE [s ζ t d ] are palatalized they are converted to two fricatives [J 3] and two africates [tj d 3 ] in the context of (a) V J and sometimes (b) {##, + } _V. [j] in (a) comes from unstressed vowels: ME i > j, ModE i > j (-ia / -ion / -ial / -ie-, etc.) and ModE iu (< ME iu, eu, y) > ju: (sugar/nature/gradual). This process began in the early 15th century and continued up to the later 17th century. The output of this shift was analogical with nonpalatalized [sj zj tj dj] for a long time. It is important to note that palatalization appeared largely in wordinternal onset position. In contrast, a very small number of examples take place in word-initial onset position. [sj] > [/] The [sj] to [f] shift was generalized in the 18th century. In word-initial position, however, it reverted to [s] except in sure and sugar. (63)

pro · fes · sion [prefesjan] > [prefejan] / gra · cious I an · xious I an · cient / as · so · ci · ate / mis · sion / pressure / pursue / sexual / luxury II sure / sugar II sue / suicide


Michiko Terajima [tj] > [/] [tj] that is not preceded by [s] is converted to [f]. [tj] that is preceded by [s], on the other hand, changes to [tf]. (64)

nation [neitjon] > [neijan] / par · tial I es· sen · tial / po · si · tion I mo· tion / am · bi · tious I pa· tient / ne· go· ti · ate [zj] > [j] [zj] changed to [3] and became general by the end of the 18th century. (65)

vi · sion [vizjan] > [vi33n] / con · clu · sion I de· ci · sion / gla • zi · er / bra · si · er / plea · sure /a· zure / u · su · al / vi · su •al [tj] > [tj] This shift was generalized during the 18th century, [tj] as in (66) coexisted with [tf] during that century and [tj] due to the spelling appeared again in the 19th century as in (67). These two [tj]* s have been succeeded by [tf] in PE. The examples in (68) exhibit the shift [tj] > [tf]. (66) (67) (68)

du· te · ous / pi · te · ous / plen · te · ous / boun · te · ous be · sti · al / fron · ti · er / cour · te *ous que · stion / Chri · stian / na· ture / pic · ture I for · tune / ac · tu· al! vir · tu· ous / cen ·tu· ry [dj] > [d3] This process was widespread in the 18th century. However, there are some examples where [dj] was retained due to the spelling, [dj] sometimes alternates with [03] in PE. (69)

sol · dier / o· di · ous / i · di · ot/ cor · di · all guar · di · an I due I du· el I gra · du· al I in · di · vi · du · al I e · du · ca · tion / pro · ce · dure

The syllable structure and phonological processes


2.8. Assimilation Assimilation can be defined as a phonological process in which a segment shares a feature or features with the segment that triggers it. In a majority of cases a conditioner is immediately adjacent to that which is to be assimilated, but this is not invariably the case. Diachronic evidence from OE and ME reveals that the outputs share the features [voice] and [place] with conditioning segments as in bm>mm / Of > ff/lr > rr. In a few exceptional cases there may be discrepancy in either of them, but not both simultaneously, as in pn>mn / fn>mn / fn>sn / mt>nt. 2.8.1. Onset Now we will see how the assimilative processes behave in wordinitial/internal onset position. We have only one example where the process appears in word-initial onset position: /f/ is assimilated to /s/ by a feature of /n/ as in (70). On the other hand, the examples in word-internal onset position occur much more frequently. The examples in (71a) exhibit Progressive Assimilation process. The example in (71b) shows Regressive Assimilation process. The instances exemplified in (71c) show Mutual Assimilation process. (70) (71)

OE fnesan > ME snese 'sneeze' a.

b. c.

OE ä · cum · ba > ä · cum · ma 'oakum' / OE lär · peow > lar · reow > lareow 'teacher' / OE neah · gea · bür 'neighbor' / OE sam · nian > ME sam · men 'join' / OE Ises · ra > läes · sa 'less' / OE wier · sa> wier · ra 'worse' / ME up · ward > up · pard / OE gödwlf > good · dive > goodive, goody / ME ham · ward > ham · mard 'homeward' OE Γ · sern > ϊ · r(e)rn >ϊ· ren 'iron' OE nem · nan > nem · pan > ME nempe 'name' / ME Ma · me· che · stre > Man · che · ster

2.8.2. Coda Let us now consider the examples in word-internal/final coda position. In word-internal coda position almost all the examples exhibit Regressive Assimilation as in (72a). In word-final coda position some show Mutual


Michiko Terajima

Assimilation, some Progressive Assimilation and some Regressive Assimilation as in (73a, b, c). (72)





b. c.

OE ceap · faru > ME chaf · fare 'chaffer' / OE wif · mann > ME wim · man 'woman' / OE süö · folk > ME Suf· folk / OE heah · ra> heah · ra 'higher' / OE en(d) · leo · fan > ME el · le · uen 'eleven' / ME wor · sted > wos · sted > wo · sted 'worsted' ein · bo · ga> elm · fcoga 'elbow' / ME cun · fort > com · fort / ME in · /?os · sible > im · pos · sible / Bran · bury > Bram · bery OE vv^pn > waemn 'weapon' / OE stefn > stemn 'voice' / OE nemne 'unless' / ModE cheapen > cheapm / OF randon > ME random / OF ransoun > ME ransom / ME amie > ante 'ant' / ON skammt > scamt > scant OE hremn > hremm > hrem 'raven' / ME dampne > damppe > dam 'dam' OE myln > ME myll 'mill' OE blips > bliss / lips > liss 'grace' / OE hors > ME hoss 'horse'

To summarize the assimilative processes operate in word-internal coda/onset position. In addition, in word-internal onset position Progressive Assimilation occurs more frequently than the other two types. In word-internal coda position, on the other hand, Regressive Assimilation appears more often.

3. Conclusions In the paper we have discussed how consonant changes behave in wordinitial vs. word-internal onset or in word-final vs. word-internal coda position on the bases of the evidence from the history of English. The main findings of our discussion can be summarized as follows: 1)

Onset position is more resistant against phonological processes than coda position (cf. Mohanan 1993: 90). Word-internal onset as well as word-internal coda position, however, triggers them comparatively readily.

The syllable structure and phonological processes




Phonological processes do not necessarily behave identically in word-initial (w-ini.) and -internal (w-int.) onset positions and word-internal and -final (w-fin.) coda positions, as is shown in the table below. a. Voicing

b. Devoicing

Onset (i) w-ini. f —> ν + θ —»δ + s -»ζ + Coda (i) w-fin. θ 3 + s —> ζ + f —> V + tj d3 + ρ —»b k —> g

(ii) w-int. + + + (ii) w-int. +

Onset b ρ d -»t

(i) w-ini.

(ii) w-int.

(i) w-fin. + + + + +

(ii) w-int. +

(i) w-ini.

(ii) w-int. + + + + + +

ν —> f ζ —> s Coda b —> ρ d —»t g —> k ν —> f ζ —> s

c. Deletion

Onset t w h χ k d

+ + +

+ +


Michiko Terajima

g g b V i Coda η t X 9 r 1 d g g b θ δ ν f

d. Epenthesis

+ +

+ + + +


(i) w-fin.

(ii) w-int.

+ +

+ + +

+ + + + + + + + + + +

Onset 0 —» h 0->j

0 —> w 0 —> k 0 —> b

+ + +

+ + +

(i) w-ini. + + + +

+ + +

0 ρ 0 —> d

Coda 0 —> t 0 —> d 0 -> ρ 0 —> η, η

(ii) w-int.

(i) w-fin. + + +

(ii) w-int. + + + +

The syllable structure and phonological processes

e. Spirantization

Onset b —> f d δ t —> θ Coda k —» χ t-> θ

f. Palatalization

g. Assimilation


Onset k, kk -> t; g. gg -> J zj - > 3 tj —> t j dj -> d} Coda k, kk -> t j g. gg -> d3 sk-»J

Onset Regressive Progressive Mutual A. Coda Regressive Progressive Mutual


(i) w-ini. (ii) w-int. + + +


(i) w-fin. (ii) w-int. + + + + +


(i) w-ini. + +

(ii) w-int. + + + + + + + +

(i) w-fin. (ii) w-int. + + +


(i) w-ini.

(ii) w-int. + + + + (i) w-fin. (ii) w-int. + + + + +

If we try to capture sound changes in terms of the syllable structure alone, we will have to say the voicing process, for example, is realized in coda position in the context of $_V. Or


Michiko Terajima

we will have to say it is realized in foot-internal position. In the voicing process, however, the vowel immediately to the left of the syllable boundary plays an important role. In fact, in some languages such as Modern Dutch and Cuban Spanish voicing processes are not implemented unless the first V is either a tense vowel or a diphthong (Nakao in press). That is, without reference to the first vowel as part of the voicing context we will suffer a loss of the important generalisation as to voicing processes. This suggests the important fact that some phonological processes can more satisfactorily be accounted for in linear perspective than in non-linear perspective.

References Chiba, Shüji et al. (eds.) 1994

Synchronic and diachronic approaches to language: A Festschrift for Toshio Nakao on the occasion of his sixtieth birthday. Tokyo: Liber Press.

Dressier, Wolfgang U, et al. (eds.) 1992

Phonologica 1992. Cambridge: CUP.

Foley, James 1977

Foundations of theoretical phonology. (Cambridge Studies in Linguistics 20.) Cambridge: Cambridge University Press.

Goldsmith, John (ed.) 1993

The last phonological rule. Chicago: University of Chicago Press.

Giegerich, Heinz J. 1985

Metrical phonology

and phonological


Cambridge: Cambridge

University Press. Hayes, Bruce 1989

"Compensatory lengthening in moraic phonology", Linguistic Inquiry 20: 253-306.

Hulst, Harry van der 1984

Syllable structure and stress in Dutch. Dordrecht: Foris.

Hyman, Larry M. 1985

A theory of phonological weight. Dordrecht & Cinnaminson: Foris.

Katamba, Francis 1989

An introduction to phonology. London: Longman.

Lutz, Angelika 1986

"The syllabic basis of word division in Old English manuscripts", English Studies 67: 193-210.

The syllable structure and phonological processes


Mohanan, K.P 1993 "Fields of attraction in phonology", in: John Goldsmih (ed.), 66-116. Nakao, Toshio 1985 English historical phonology. Tokyo: Taishukan. in press Diachronic universals in phonology - minimality principle of change Tokyo: Liber Press. Sisam, Kenneth


1953 Studies in the history of Old English literature. Oxford: Clarendon. Terajima, Michiko 1993-4a The syllable structure in the history of English. 1994b

"On VC syllable structure in the history of English", in: Shüji Chiba et al. (eds.), 183-189. Treiman, Rebecca 1992 "Experimental studies of English syllabification", in: Wolfgang U. Dressler et al. (eds.), 273-281.

The helle sequence in Old English poetry Jun Terasawa The distinction between a compound and a syntactic phrase is one of the major morphological problems that have troubled Old English lexicographers and editors. In Judith (317a), for instance, there is some disagreement as to whether the sequence bräd swyrd 'broad sword' is a compound or a syntactic phrase: bord ond brad swyrd Dobbie's edition takes the sequence as a compound, with a brief note that "in view of the later form broadsword, this word is probably a compound".1 Clark Hall-Meritt (A Concise Anglo-Saxon Dictionary, Supplement) and Healey-Venezky (A Microfiche Concordance to Old English) also list it as a compound. On the other hand, editors of Judith such as Cook and Timmer treat bräd swyrd as a syntactic phrase, and Bosworth-Toller (An Anglo-Saxon Dictionary) do not include it as a compound. It is important to add here that, although the sequence appears as two words in the Nowell Codex (MS Cotton Vitellius A XV), the manuscript form does not support the syntactic phrase analysis: Old English scribes usually wrote separately forms which were clearly elements of a single compound. Neither will the morphological evidence enable us to decide between the compound and syntactic phrase analyses of the sequence because bräd is a strong long-stemmed adjective modifying the accusative plural neuter noun swyrd and would have no inflectional ending in a syntactic phrase. Above all, the treatment of helle 'hell' sequences (e.g., helle jyr 'fire of hell') has presented serious difficulties to Old English lexicographers and editors because they are ambiguous between a compound and a syntactic phrase: in the former, helle is regarded as the stem form where the ending e is a relic of the earlier stem vowel; in the latter, it is regarded as the genitive of the later stem form hel(l) where the ending -e is a genitive inflection. Before listing helle compounds in An Anglo-Saxon Dictionary, Bosworth-Toller admit that "[i]n the case of at least some of the following words which are given as compounds, they might be taken as independent words, the first of which is the genitive of heP\ The uncertainty concerning


Jun Terasawa

the interpretation of helle sequences results in a great deal of disagreement among lexicographers and editors as to which instance of such a sequence should be regarded as a compound and which as a syntactic phrase. To the 33 helle compounds listed in Bosworth-Toller, Toller (An Anglo-Saxon Dictionary, Supplement) adds 8 new ones such as helle-cäsge 'hell-key', helle-loc 'hell-prison', helle-maegen 'hell-troop', etc.2 Clark Hall-Meritt add to Bosworth-Toller and Toller helle-müö 'hell-mouth' and helle-ρϊη 'hell-torment', while excluding helle-god 'a god of the infernal regions' and helle-smip 'hell-smith', the latter of which is added by Toller. HealeyVenezky (1980) seem to be more hesitant about taking a helle sequence as a compound,3 and delete 11 compounds listed in Bosworth-Toller, Toller, or Clark Hall-Meritt, like helle-bealu 'hell-bale', helle-ceafl 'hell-mouth', helle-cinn 'hell-race', etc. Although they list helle Jyr 'fire of hell' as a compound, Healey-Venezky only give prose examples and exclude helle jyr in Christ (1269) and The Meters of Boethius (viii.51). The two verse instances of helle jyr are also analyzed as syntactic phrases by Krapp and Dobbie. Bosworth-Toller, however, represent the sequence as a compound, giving references to the two verse passages. In the treatment of helle grund 'abyss of hell', the Anglo-Saxon Poetic Records reveal some inconsistency. There are 9 verse instances of helle grund, two of which are regarded as compounds: Het hine j)2ere sweartan helle / grundes gyman ah in helle grund hat ne helle grund in hellegrund in helle grund in helle grund sec an hellegrund sec an helle grund ac J>£r is helle grund

(Genesis Β 345b-46a) (Christ & Satan 448b) (Christ & Satan 454a) (Elene 1305b) (Christ 265a) (Christ 562b) (Soul & Body 1104a) (Soul & Body II 98a) (Judgment Day 124b)

Notice first that in the Anglo-Saxon Poetic Records, helle grund in Soul & Body I (104) is taken as a compound whereas the same sequence in Soul & Body II (98), which appears in the same formula, is taken as a syntactic phrase. Since these two instances of helle grund occur in corresponding passages of the same poem, it would be hard to find any valid reason why they should be treated in a different way. In his edition of the Old English Soul & Body, Moffat considers both instances of helle grund to be compounds. However, we shall later demonstrate that the syntactic phrase

The helle sequence in Old English poetry


analysis of helle grund is preferable. There are 4 instances of helle grund preceded by the preposition in. The editors of the Anglo-Saxon Poetic Records seem to be inconsistent again in dealing with these instances, which occur in similar grammatical constructions: the sequence in Elene (1305b) is regarded as a compound whereas those in Christ and Satan (448b) and Christ (265a, 562b) are regarded as syntactic phrases. In the present paper I examine the verse occurrences of the helle sequence and provide metrical (both alliterative and rhythmical) evidence for the syntactic phrase analysis of the helle sequence in verse. Except for those in Beowulf, verse instances are taken from the Anglo-Saxon Poetic Records. The Beowulf text used here is the third edition of Klaeber's Beowulf and the Fight at Finnsburg. To begin with, let us look at an instance of helle fyr in Christ (1269) again. (The sounds in alliteration are in italics): grim helle fyr, gearo tö wlte It is important to note here that alliteration does not fall on the first element of helle fyr but on the preceding adjective grim "fierce', which alliterates with gearo 'ready'. In Old English poetry, it is extremely rare for the first element of a compound to fail to alliterate. In Beowulf, güö 'battle', one of the most productive first elements, forms 33 different compounds, such as güö-beorn 'war-hero', gud-bil(l) 'war-sword', güd-cyning 'war-king', and so on. Each of these compounds invariably alliterates on the sound [g].4 Consider, for instance, line 803 from Beowulf where the first element of güö-bill 'war-sword' alliterates with gretan 'to touch': güöbilla nän

gretan nolde

If the sequence helle fyr in Christ (1269) is taken as a compound, then, the alliterative pattern is peculiar in that the first element does not alliterate. Alliteration does not have to fall on the first element of a syntactic phrase like llfes bläed 'glory of life', as in the Seafarer (79): Äwa tö ealdre, ecan llfes bläed In view of these points I should like to read helle fyr in Christ (1269) as a syntactic phrase. We shall now turn to another helle sequence, i.e., helle grund. Consider first helle grund in Genesis Β (345b-46a):


Jun Terasawa .. .het hine J)Sre sweartan helle grundes gyman...

The sequence here has to be analyzed as a syntactic phrase because both the article päere and the adjective sweartan 'black' are in the genitive singular feminine and agree with helle, i.e., the genitive singular of the feminine hel(l). On the other hand, Kastovsky (1992) regards helle grund in Christ & Satan (454a) as a compound since the adjective hätne 'hot' agrees with the second element grund 'abyss': hä.tne helle grund The agreement between the adjective and the second element, however, does not necessarily point to the compound analysis of helle grund. In the following instances, the adjectives (e.g., leofne 'beloved' and mycel 'great') agree with the second elements of the sequences that are clearly syntactic phrases, i.e., genitive + noun: /eofne fifes weard (Christ 1642a) mycel modes hiht (Andreas 287a) mycel modes sorg (Andreas 1690a) We may note here that unlike helle Jyr, alliteration does not provide a crucial criterion to decide which analysis is preferable in the case of helle grund: except in Genesis Β (345f.) alliteration always falls on the first element of helle grund. In what follows, I shall provide a piece of rhythmical evidence for the syntactic phrase analysis of the helle sequence. Before going on, let us pause and look at the use of the compound hildedeor 'brave in battle.' It is generally agreed that unlike helle, hilde is an earlier stem form with the ending -e and serves as the first element of archaic poetic compounds (Bammesberger 1980: 7f). It is to be noted that hilde-deor is metrically comparable to helle grund: each first element consists of a sequence of a long stressed syllable and an unstressed one whereas each second element consists of a long stressed syllable.5 The Old English verse texts provide 9 instances of the compound hilde-deor in its uninfected form:

The helle sequence in Old English poetry


Ziaefe /nldedlor (Andreas 1002a) Type D4 /iaefeö /nldedeor (Elene 935a) Type D4 Him t>a /lildedeor (Beowulf 3\2ά) Type Β X



X /

syjjöan /lildedeor (Beowulf 834a) Type Β Aaefe /lildedeor (Beowulf 1646a) Type D4 Aaefe /lildedeor (Beowulf 1816a) Type D4 hwflum /lildedeor (Beowulf 2107a) Type Β heold /lildedeor (Beowulf 2183a) Type Β Ziaefe /iildedlor (Beowulf 311 la) Type D4 According to Sievers' system of versification, 5 instances of the half-lines with hilde-deor are analyzed as Type D4 and 4 instances as Type B.6 In half-lines 2107a and 2183a of Beowulf, I assume that the adverb hwilum 'sometimes' and the finite verb heold 'held' are weakly stressed and accidentally take alliteration.7 Old English verse texts provide 13 other hilde compounds like hilde-deor:8 among 19 instances of these compounds, 9 occur in Type D4 verses and 10 in Type B.9 This would lead us to say that the use of hilde compounds like hilde-deor are preferred both in Type D4 and in Type B. On the other hand, syntactic phrases like wuldres god 'God of glory', which is comparable to the sequence helle grund, is predominantly found in Type Β verses: among the 13 instances of half-lines with wuldres god, 11 are scanned as Type Β and only 2 as Type D4: γ / / X \ Him δη wuldres god (Genesis A 2916b) Type D4 X X / X / E>^t waes wuldres god (Daniel 277b) Type Β / χ / χ \ weoröan wuldres god (Andreas 758a) Type D4 χ / χ / wrät, wuldres god (Andreas 1510a) Type Β


Jun Terasawa

XX / X / J)ä him wuldres god (Andreas 1661b) Type Β XX / X / J)aet ic wuldres god (Guthlac 1081b) Type Β X / X / ond wuldres god (Juliana 180b) Type Β X X / X / Hwast, J)e, wuldres god (Paris Psalter 76.13.1a) Type Β XX X / X / Hü wilt J)ü, wuldres god (Paris Psalter 78.5.1a) Type Β X X X / χ ^ ic me tö wuldres gode (Paris Psalter 87.9 2b) Type Β XX / X / |dü eart, wuldres god (Paris Psalter 89.2 3b) Type Β X X / X / Hwaet, J)ü, wuldres god (Meters ofBoethius xx. 57b) Type Β X X / X / ac J)ü wuldres god (Exhortation to Christian Living 28b) Type Β We regard half-line 1510a in Andreas as Type B, since the finite verb wrät 'wrote' could be weakly stressed and accidentally participate in alliteration. A marked preference is also shown for Type Β over Type D4 when wuldres forms syntactic phrases with other words of the god type:10 among 79 instances of such sequences with wuldres, 75 are attested in Type Β verses and 4 in Type D4.11 A question which arises is why, in contrast to compounds like hilde-deor, the use of syntactic phrases like wuldres god is much less common in Type D4. The explanation could be that the second element of syntactic phrases tend to retain sufficient stress to resist being with a secondary arsis or rhythmic stress. With these facts in mind, we may return to the sequence helle grund. If helle grund were to be regarded as a compound, we would expect that like hilde-deor the sequence should occur in Type D4 as frequently as in Type B. This expectation is, however, not borne out. Except for verse 454a in Christ & Satan, each of the half-lines with helle grund is used in Type B. (The half-lines with helle grund are repeated below for convenience. The sequence in Genesis Β (345b-46a) is excluded since it does not occur in a single half-line):

The helle sequence in Old English poetry


XX / X / ah in Aelle grund (Christ & Satan 448b) Type Β / X / \ hätne helle grund (Christ & Satan 454a) Type D4 χ

/ χ


in Aellegrund (Elene 1305b) Type Β χ


in Aelle grund (Christ 265a) Type Β χin Aelle / χ grund / 562b) Type Β XX Aellegrund / X / (Christ secan (Soul & Body 1104a) Type Β XX

/ X


secan Aelle grund (Soul & Body II 98a) Type Β X


/ X


ac J>sr is Aelle grund (Judgment Day 124b) Type Β

When they take second elements like grund, other helle sequences also occur much more commonly in Type Β than Type D4: among 21 instances of such helle sequences 16 are used in Type Β and 5 in Type D4. A marked preference for Type Β in the use of helle grund and other helle sequences of the same metrical pattern would thus associate the sequences more closely with syntactic phrases like wuldres god (rather than compounds like hilde-deor) and suggest that the sequences should be analyzed as syntactic phrases.13 As in the case of helle fyr and helle grund, some disagreement is found as to whether the sequence helle deofol 'devil of hell' should be analyzed as a compound or a syntactic phrase. Among three verse instances of helle deofol, Krapp and Dobbie regard two as syntactic phrases and one as a compound, the latter of which Gradon takes as a syntactic phrase:

helle dioful (Andreas 1298b) Type A / χ / χ helle deofol (Juliana 629b) Type A helledeofol (Elene 900b) Type A Notice that each instances of helle deofol is used in Type A. Other helle sequences of the helle deofol type are also exclusively attested in Type A verses.14 Important to note here is that compounds like hilde-mece 'battle-


Jun Terasawa

sword', which is metrically equivalent to helle deofol, are also exclusively used in Type A: 15



It follows that the verse rhythm does not help distinguish sequences like helle deofol from compounds like hilde-mece. However, a careful comparison between helle sequences and hilde compounds clarifies a sharp distinction: helle is often accompanied by a second element which begins with the sound [h] and alliterates with the first element: e.g., helle-heafas 'hell-wailings', helle hatftas/hatfton 'prisoners) of hell', helle hatftling 'prisoner of hell', helle hlenfru 'humiliation of hell', and helle hinca 'hell limper'. Among compounds of the hilde-mece type, on the other hand, none of the second elements alliterates with the first element except for hilde-hlaemmum/-hlemma 'battle-crashes' (Beowulf2201, 2351, 2544). It is to be noted that the use of self-alliterating compounds is quite restricted in verse texts. The tendency to avoid double alliteration with compounds could be due to the fact that the stress on the second element is subordinate to that on the first element, so that the former is less likely to be associated with alliteration than the latter. Under the syntactic phrase analysis, alliteration could fall on both the first and the second element of sequences like helle deofol because the non-subordinate second element retains as much metrical prominence as the first. Before concluding, I should like to consider briefly the helle sequence in Old English prose texts. As mentioned earlier in this paper, while HealeyVenezky are somewhat hesitant about the compound analysis of helle sequences in verse, they frequently treat prose occurrences of the sequence as compounds, e.g., helle-bröga 'hell-terror', helle-bryne 'hell-fire', helleduru 'hell-gate', helle-fyr 'hell-fire', helle-werod 'helle-host', etc. Our metrical (i.e., alliterative and rhythmical) criteria cannot tell us whether theses prose instances of the helle sequence are compounds. It seems unlikely, however, that the earlier stem form helle is preserved in Old English prose, the language of which generally retains fewer archaic features than poetry. It is worth noting here that the earlier stem form hilde is attested almost exclusively in verse texts to form archaic poetic compounds. These considerations would lead me to treat prose instances of the helle sequence as syntactic phrases.

The helle sequence in Old English poetry


Notes *


2. 3. 4.

This is a revised version of Terasawa 1995. A shortened version of the present paper was read at the Kyoto Conference on English Historical Linguistics and Philology, 3 November 1995. I should like to thank those who commented on my paper on that occasion. Thanks are also due to Professor Paul E. Davenport, who read the draft and suggested stylistic improvements. Dobbie also takes as a compound bräd swurd in Maldon (15a), which occurs in the same formula as in Judith (317a). So do Healey-Venezky. The Oxford English Dictionary (2nd ed.) gives this instance as the first citation of the compound broadsword. On the other hand, editors of Maldon such as Gordon and Scragg print bräd swurd as two words. Alistair Campbell (An Anglo-Saxon Dictionary, Enlarged Addenda and Corrigenda) neither adds nor deletes any of the helle compounds in Bosworth-Toller or Toller. However, some helle compounds such as helle-carcern 'hell-prison', helle-waran 'hellinhabitants', helle-werod 'hell-host' are newly added by Healey-Venezky. The obligatory alliteration on the first compound element is due to the fact that Old English poets used a compound as an important means to provide alliteration. The poets might substitute hilde- 'war' for güö- if they wanted alliteration on h, as in Beowulf (557): /lildebille; /jeaJ>or£s fornam

5. A syllable is long if it has a long vowel or diphthong optionally followed by a consonant (e.g., sm 'sea', säel 'time', frea 'lord', sceat 'comer, region') or if it has a short vowel or diphthong followed by one or more consonants (e.g., seep 'juice', sceap 'shape', web-ba 'weaver', neal-les 'not at all', niht 'night', weord 'worthy'). On the other hand, a syllable is short if it ends with a short vowel or diphthong (e.g., fe-la 'many, much', scea-fra 'fiend, devil'). Note that a single medial consonant belongs to the following syllable, while at least one member of a medial consonant cluster belongs to the preceding syllable. 6. According to Sievers, the basic structures of Types Β and D4 are: Type Β (X IX _J and Type D4 I X An acute ( ' ) , a grave ( ), a cross (X) and a vertical stroke ( I ) stand for a primary arsis, a secondary arsis, an unstressed syllable and a foot boundary respectively. In Type Β the number of unstressed syllables may vary. A verse of Type D4 may be expanded by an unstressed syllable after the first arsis ( X I X ^J. 7. Bliss regards both half-lines as Type D4 (1D5 in his notation) while Kendal analyzes them as Type B. Pope, on the other hand, takes half-line 2107a as Type Β (B2 in his notation) and half-line 2183a as Type D4 (D16 in his notation).


Jun Terasawa

8. Hilde-bil 'war-sword', hilde-bord 'battle-shield', Hilde-burh (proper name), hilde-grap 'hostile grasp', hilde-heard 'bold in battle', hilde-leod 'war-song', hilde-rses 'battlestorm', hilde-rinc 'war-hero', hilde-sceorp 'war-dress', hilde-setl 'war-seat, saddle', hilde-swät 'battle-sweat, hostile vapor', and hilde-sweg 'battle-sound'. We might add hilde-bläc 'battle-pale' (Beowulf 2488a), which has been proposed for MS. blac. Klaeber, however, reads heoro-bläc 'battle-pale'. 9. I regard half-lien 1114a in Beowulf (Hit ö l Hfldeburh) as Type Β while Pope (1966) takes it as Type D4 (D49 in his notation). 10. Wuldres beam 'tree of glory', wuldres beam 'child of glory', wuldres blSd 'riches of glory', wuldres boda 'messenger of heaven', wuldres byrig 'city of glory', wuldres crxft 'strength of glory', wuldres cyning 'king of glory, wuldres däel 'heavenly part', wuldres dream 'joy of heaven', wuldres eard 'abode of heaven', wuldres ful 'full of glory', wuldres gäst 'spirit of glory', wuldres gim 'jewel of heaven', wuldres helm 'protector of heaven', wuldres hyht 'joy of glory', wuldres lean 'reward of glory', wuldres leoht 'light of glory', wuldres lof 'praise of glory', wuldres mseg 'glorious maiden', wuldres miht 'power of glory', wuldres sang 'song of glory', wuldres setl 'seat of glory', wuldres sweg 'sound of glory', wuldres treow 'tree of glory', wuldres fregn 'thane of glory', wuldres fjreat 'troop of glory', wuldres prym 'glory of heaven', wuldres wlite 'brightness of glory', wuldres word 'word of glory', and wuldres wynn 'joy of heaven'. 11. We regard the following half-lines as Type B, sine the finite verbs onwräh 'revealed' and onwrige 'showed' could be weakly stressed and accidentally take alliteration: Χ X / X / X Χ X / X ^^ "onwräh, wuldres helm" (Christ 463a) and "onwrige, wuldres cyning" (Juliana 516a). 12. Helle bealu 'bale of hell', helle-bryne 'hell-fire', helle ceafl 'mouth of hell', helle cinn 'hellish race', helle duru (dorum) 'gate(s) of hell', helle fyr 'fire of hell', helle giest 'spirit of hell', helle heap 'troop of hell', helle hits 'house of hell', helle locum 'prisons of hell', helle nid 'affliction of hell', helle-sceafran 'hell-foe, devil', and helle sead 'pit of hell'. 13. For another piece of rhythmical evidence for the syntactic phrase analysis of the helle sequence, particularly those of the helle duru type, see Terasawa (1994: 22f.; 1995: 619ff.) 14. Helle clommas 'fetters of hell', helle-flöras 'hell-floors', helle-heafas 'hell-wailings', helle haeftas/hatfton 'prisoner(s) of hell', helle hacftling 'prisoner of hell', helle hlenpu 'humiliation of hell', helle hinca 'hell limper', helle-scealcas 'hell-subjects', helle pegna 'thanes of hell', helle weallas 'walls of hell', and helle wita/wUes/witu/wUum 'torments of hell'. 15. Other hilde compounds like hilde-mece that occur in Type A include: hilde-bille 'battlesword', hilde-bordum 'battle-shields', hilde-calla 'war-herald', hilde-cordre 'warlike band', hilde-cystum 'battle-virtue', hilde-diore 'brave in battle', hilde-fröfre 'battlecomfort', hilde-geatwa/geatwe 'war-equipment', hilde-gicelum 'battle-icicles', hilde-

The helle sequence in Old English poetry


gieste 'war-enemy', hilde-hlxmmum/hlemma 'battle-crashes', hilde-leoman 'battle-light, sword', hilde-mecgas 'battle-warriors', hilde-nxdran 'war-snakes, arrows', hilde-pilum 'battle-darts', hilde-randas 'battle-shields', hilde-rinca/rincas/rince/rinces 'war-heroes', hilde-scürum 'showers of darts', hilde-sercum 'coats of mail', hilde-spelle 'warlike speech', hilde-strengo 'battle-strength', hilde-torhtum 'shining in battle', hilde-tüxum 'battle-tusks', hilde-jjremman warriors', hilde-prymme/f>ryf>e 'warlike strength', hildewsepen/wspnum 'war-weapon(s)', hilde-wisan 'battle-leader', hilde-wöman 'battlecrash', hilde-wrssne 'fetter for captives', and hilde-wulfas 'war-wolves, warriors'.

References Primary Sources

Cook, Albert S. (ed.) 1889 Judith: An Old English epic fragment. (2nd edition, revised and enlarged.) Boston: Heath. Gordon, Eric V. (ed.) 1937 The battle of Maldon. London: Methuen. Gradon, Pamela O.E. (ed.) 1977 Cynewulfs 'Elene'. (Revised edition.) Exeter: University of Exeter Press. Klaeber, Frederick (ed.) 1950 Beowulf and the fight at Finnsburg. (3rd edition with 1st and 2nd Supplements.) Lexington, Mass: Heath. Krapp, George Philip - Elliott van Kirk Dobbie (eds.) 1931-53 The Anglo-Saxon poetic records. 6 vols. New York: Columbia University Press. Moffat, Douglas (ed. and trans.) 1990 The Old English Soul and Body. Woodbridge: D.S. Brewer. Timmer, Benno J. (ed.) 1978 Judith. (Revised edition.) Exeter: University of Exeter. Scragg, Donald G. (ed.) 1981 The battle of Maldon. Manchester: Manchester University Press.

Secondary Sources Baba, Akira, et al. (eds.) 1995 Essays in linguistic and philology presented to Professor Kinsuke Hasegawa on his Sixtieth Birthday. Tokyo: Kenkyusha.


Jun Terasawa

Bammesberger, Alfred 1980

"Altenglische Komposita mit Sprachwissenschaft 39: 5-10.





Bammesberger, Alfred (ed.) 1985

Problems of Old English lexicography: Studies in memory of Angus Cameron. Regensburg: Pustet.

Bliss, Alan J. 1967 The metre of Beowulf. (Revised edition.) Oxford: Basil Blackwell. Bosworth, Joseph - T. Northcote Toller (eds.) 1882-98 An Anglo-Saxon dictionary based on the manuscript collections of the late Joseph Bosworth, edited and enlarged by T. Northcote Toller. [ 1972]

[Reprinted London: Oxford University Press.] [Cited as Bosworth-Toller]

Campbell, Alistair (ed.) 1972 An Anglo-Saxon dictionary based on the manuscript collections of Joseph Bosworth. Supplement by T. Northcote Toller. Enlarged addenda of corrigenda. Oxford: Clarendon Press. Carr, Charles T. 1939 Nominal compounds in Germanic. (St. Andrews University Publications 41.) London: Oxford University Press. Hall. J.R. Clark (ed.) 1960 A concise Anglo-Saxon dictionary. (4th edition, with a Supplement by Herbert D. Meritt.) Cambridge: Cambridge University Press. [Cited as Clark HallMeritt] Healey, Antonette diPaolo - Richard L. Venezky (eds.) 1980 A microfiche concordance to Old English. Toronto: Centre for Medieval Studies, University of Toronto. [Cited as Healey-Venezky] Hogg, Richard M. (ed.) 1992 The Cambridge history of the English language. Vol. I: The Beginnings to 1066. Cambridge: Cambridge University Press. Kastovsky, Dieter 1992 "Semantics and Vocabulary", in: Richard M. Hogg (ed.), 290-408. Kendall, Calvin B. 1991 The metrical Grammar of'Beowulf. Cambridge: Cambridge University Press. Pope, John C. 1966 The rhythm of Beowulf (Revised edition.) New Haven: Yale University Press. Sauer, Hans 1985 "Die Darstellung von Komposita in altenglischen Wörterbüchern", in: Alfred Bammesberger (ed.), 267-315.

The helle sequence in Old English poetry Sievers, Eduard 1885 "Zur Rhythmik des germanischen Alliterationsverses", Beiträge Geschichte der deutschen Sprache und Literatur 10: 209-314, 451-545. 1893 Altgermanische Metrik. Halle: Max Niemeyer. Simpson, John Andrew and Admund S.C. Weiner (eds.) 1989



The Oxford English dictionary. 20 vols. (2nd edition.) Oxford: Clarendon Press.

Terasawa, Jun 1994 Nominal compounds in Old English: A metrical approach. (Anglistica 27.) Copenhagen: Rosenkilde and Bagger. 1995

"The Helle sequence in Old English poetry: A compound or a syntactic phrase?" in: Akira Baba et al. (eds.), 615-627. Toller, T. Northcote (ed.) 1921 An Anglo-Saxon dictionary based on the manuscript collections of the late Joseph Bosworth. Supplement. Oxford: Clarendon Press.

Some etymological and semasiological notes on girl* Yoshio Terasawa Etymological cruces sometimes find their way even into such common words as boy, girl, lad, lass. Numerous etymological theories have been proposed for these words, but none seems firmly established. They are usually labelled 'of uncertain etymology'. In this article we review and examine some theories on the origin and early development of the word girl as it appears in early Middle English citations1 with the attested meaning of 'a child or young person of either sex'. The earliest attested forerunners of ModE girl appear in about 1300 and after. Men mi3tten seen j>ere hondes wrynge ... Wymmen shrikyng, gyrles gradyng. cl300 Kyng Alisaunder (Laud Misc. 622), 2795-982 In daunger hadde he at his owene gise The younge girles of the diocise, And knew hir conseil, and was al hir reed. c 1387-95 Chaucer, Canterbury Tales, Prologue 663-653 Thoru3 wyn and t>oru3 wommen J>er was loth acombred, And t>ere gat in glotonie gerles J)at were cherles. cl378 Piers the Plowman B. i. 32-34 Here knaue gerlys I xal steke. α 1475 Ludus Coventriae 171.59s In these examples there are several points which will be relevant to our discussion. In the Chaucer quote there are variant MS readings geerles, gerles, gurles, etc. for girles, and in the first two examples of Alisaunder and Canterbury Tales the word girle is used in the sense of 'a young person of either sex', 6 in the third it evidently refers to Lot's two sons, Moab and Ammon, and in the last knaue gerlys means 'boys'. On the other hand, the current use of girl will be traced back at least to the latter half of the 14th century, though OED citations only begin with


Yoshio Terasawa

Palsgrave (1530). The OED cites the following instance as an example of the sex-neutral use of girl, but it is contextually apparent that gayue gerles there refers to the young princess Meilor and her favourite maiden Alexandrine. And whan J>e gaye gerles were in-to J>e gardin come, Faire floures J>ei founde of fele maner hewes. α1375 William ofPalerne, 8167 The MED can add further examples of this use before the 15th century. A py3t coroune 3et wer {>at gyrle. cl380 Pearl 2058 What eyleth yow? Som gay gerl, God it woot, Hath broght yow thus upon the viritoot. cl390 Chaucer, Canterbury Tales, Miller's Tale A 3769-70 Thanne mayst thou bultyn in this boure, And serdyn gay gerlys. α 1450 Castle of Perseverance 1159f.9 In the first quote above, gyrle refers to the poet's beloved daughter, an only child who died before she was two,10 and in the following two examples, it will be noticed, gerl(ys) collocates with an alliterative adjective gay as in the William ofPalerne example (gay girl always seems to represent the feminine sex). Perhaps it is not without significance, as well be seen later, that Kyng Alisaunder, Piers the Plowman and William of Palerne are alliterative poems which follow directly or indirectly the Old English tradition, while Castle of Perseverance and Ludus Coventriae are medieval plays. That girl denoting the female sex is found in the medieval plays makes us suspect it was used on the spoken level or in the speech of the lower classes, and this is further supported by the fact that the 16th early 17th century examples in the OED appear to cluster round the plays written by Redford, Heywood, Shakespeare and Brome. If we examine the use of girl in Shakespearean works and in the Authorised Version, we find opposite tendencies in their use of girl. In Shakespeare the commonest word for girl is maid (27 lx), followed by wench (98x), n girl (68x), maiden (61x), lass (llx) and one instance of maid-child. In the Authorised Version, on the other hand, most commonly used is damsel or damosel

Some etymological and semasiological notes on girl


(51x), next come maid (46x) and maiden (26x), and there are only two instances of girl and a single instance each of wench and maid-child. And they haue cast lots for my people, and haue giuen a boy for a harlot, and colde a girle for wine, that they might drinke. Joel 3:3 12 And the streets of the citie shall be full of boyes and girles playing in the streets thereof. Zechariah 8:5 13 The literary, conservative Authorised Version and the literary-colloquial, innovative Shakespearean works thus show a quite different distribution of girl. As the semantic change of girl from 'a young person of either sex' to 'a young female' is easily explained as a specialisation of the semantic dimension of sex, and the phonological development from ME girle to ModE girl does not seem to pose any serious problem, it will be admitted that the continuity of girl from ME girle is established. The crucial point about the etymology of girl, then, is what the origin of ME girle is: does it reflect any OE word or does it have to be regarded as a loanword in the early Middle English period? The following list14 epitomises the numerous etymologies so far proposed. (source) (date) John Minsheu 1625 Meric Casaubon 1650 1671 Stephen Skinner 1689 George Hickes (pub. 1703-5) 1721-30 Nathan Bailey Francis Junius 1677 (pub. 1743) 1755 Samuel Johnson John Ash 1775 G.W. Lemon 1783 1828 Noah Webster

(proposed etymology) L garrula Gk kore OE *ceorla Icel. Karlinna OE *ceorla Welsh herlodes (cf. harlot) (various etymologies cited) of uncertain derivation =1650 Casaubon LowL gerula 'a young woman employed to tend children'


Yoshio Terasawa

1836-37 1860 1872 1878-79 1879-82 1889-91 1898 1898 1921 1928 1932 1947 1949 1951 1953 1958 1961 1966 1966

1966 1971 1973 1979 1986 1987

Charles Richardson J.E. Worcester H.A. Wedgwood15 (2nd ed.) Eduard Müller16 (2nd ed.) W.W. Skeat W.D. Whitney Friedrich Kluge & Frederick Lutz Karl Luick Ernest Weekley (Etym. Die.) OED H.C. Wyld (UED) Kemp Malone CACD) Ferdinand Holthausen Rolf Berndt Harold Whitehall ('WNWD) Eric Partridge (Origins) Webster3 C.T. Onions (ODEE) Ernest Klein (Comprehensive Etym.) RHD MED G.W.S Friedrichsen (SOED 3rd ed.) T.H. Long (CED) T.F. Hoad (Concise ODEE) RHD (2nd ed.)

=1828 Webster (various etymologies cited) LG gör, göre, görr 'kleines kind, mädchen' OE ceorl... OE gyrla 'kleidung' OE *gyr-el-... Gmc *gurLG gör OE *gyrel OE *gyrele ? cf. LG geere of obscure etymology OE gyrl(e): cf. LG gör OE gyrl- (cf. gyrlgyden) cf. OE gor 'gore' OE *gyr(w)ela 'der, die Reifende, Heranwachsende' OE *gyrele of obscure origin (cf. 1953 Whitehall) ME girle, gerle, gurle OE *gyrela, *gyrele ?cf. LG göre

=1947 Kemp Malone ?OE *gyrela (cf. OE gierela 'garment') OE *gyrela (?cf. LG gör(e)) ?LG göre of uncertain origin cf. OE gyrela 'dress, apparel

Some etymological and semasiological notes on girl

1988 1990

R.K. Barnhart John Ayto

1993 1993 1993 1995

RHWCD Chambers AHD (3rd ed.) New SOED COD (9th ed.)


?cf. OE gierela 'garment' ?cf. LG göre 'child, kid'& Norw. (dial.) gurre 'lamb' =RHD (2nd ed.) ?cf. LG gör(e) 'child' (of uncertain origin) ?cf. LG gör 'boy, girl' ?cf. LG gör 'child'

The lexicographers and etymologists preceding Worcester, who were not versed in historical comparative linguistics, do not go beyond folketymological speculation; most of them arbitrarily strain the phonological correspondence of the word, on the basis of some semantic association. It would be outside the scope of this paper to enter into a detailed discussion of each etymology in the above list, so we shall merely mention some points of special interest and importance. Among the suggestions made after Wedgwood, it is possible to detect two main types of etymological interpretation. One type of etymology postulates OE *gyrela from the ME dialectal forms girle, gerle, gurle, and the other ascribes girl to a loanword, preeminently to LG gör(e) 'child'. The hypothetical OE etymon *gyrela for girl was briefly but authoritatively proposed by Luick17 and has been supported by many scholars, but we are left with the problem that no instance of gyrela 'a young person' is attested in Old English. The LG etymon gör(e) 'child, boy, girl' for girl has also bee advocated by many scholars since Wedgwood, but the quandary is that this Low German word has not been found in literature before the 16th century. For the latter etymology, some cognate words such as G (dial.) Gör(e) 'wanton girl', Gören 'children', Norw. (dial.) gorre 'child, boy', Swed. (dial.) gärre, gurre 'child' and so on were cited with their ultimate etymon IE *gher 'small'. 18 However, the absence of earlier instances of LG gör(e) and the phonological relation of ME i, e, u, to LG ö lay a heavy burden of proof on scholars. As to the first postulation, OE *gyrela as the etymon of ME girle, etc., have we really no indication of its previous existence? Wyld (1932) and Malone (1947) proposed gyrel(e) 'virgin' on the basis of gyrelgyden, an OE gloss to L Vesta 'virgin, goddess'. This would not be improbable, were it not for two reasons. Firstly, the semantic relation with early ME girle 'a young person of either sex' and the later semantic development is not clear enough; and secondly, OE gyrelgyden actually means 'clothing goddess', a description of the Roman goddess in which the association of her name


Yoshio Terasawa

with L vestis 'garment' is possibly due to Isidore's explanation quod herbis vel variis vesti (!) sit rebus (Etym. VHI.xi.61).19 Another more plausible proposal on the etymology of girl is found in the RHD,20 Barahart and most recently in the RHWCD, which follows the RHD. [1250-1300; ME qurle, girle, gerle child, young person; cf. OE gyrela, gi(e)rela, item of dress, apparel (presumably worn by the young in the late OE period, and hence used as a metonym)] This etymology, though not clearly expressed, must be based on Fred Robinson (1967). Robinson assumes the metonymical use of OE gyrela 'dress, apparel' as the precursor of ME girle, gerle, gurle, referring to the semantic pattern that words denoting an article of clothing tend to denote the wearer of it.21 He argues: For it is quite conceivable that OE gyrela had already developed the transferred meaning 'young person of either sex' in pre-Conquest times, but that the usage was limited to the domestic sphere and never got into literary record, where an ample series of non-colloquial terms such as beam, byre, cnafa, cnapa, cncepling, eafora, fcemne, geonga, geongling, hyse, lytling, mcegep, meowle, umbor, etc. adequately served this expressive need. But after the Conquest, all these native literary words began to fall out of common use, thus depleting the wordstock for this semantic category. Meanwhile, gyrela in its primary sense of 'dress, apparel' was being displaced by a rush of French loanwords such as array, attire, cloak, habit, mantle, robe, roket, vestment (all first recorded in thirteenth-century writings), apparel, coat, frock, garment, gown, livery, ray, vesture (all first recorded in early fourteenth-century writings). Having been rendered superfluous as a clothing term and being in high demand as a term for 'a young person', gyrela would seem, then, to have passed, during the centuries when it is absent from written record, into exclusive use in this latter, originally secondary sense (Robinson 1967: 236ff).

Some etymological and semasiological notes on girl


I would like to subscribe to Robinson's ingenious and persuasive proposal of OE gyrela 'apparel' as an etymon of ME girle. There is, however, a phonological problem of Anlaut as well as some semantic problems left to be explored: when and why OE gyrela 'dress, apparel in general' came to be applied to a person of a particular age, i.e., a child or young person; and why ME girle, etc., originally indeterminate with respect to gender, came to be limited to the female sex. As to the phonological problem, Late WS gyr(e)la, gir(e)la, corresponding to early WS gierela (e.g., Alfred's Cura Pastoralis, ed. Sweet, p. 135, 12) and Anglian gerela (e.g., Lindisfarne Gospels, Mt. 11.8) can hardly be derived from Gmc *jur-, but seems to reflect 'unstable /', i.e., probably Gmc *jar-, the further development depending on the etymology assumed: probably breaking followed by i-mutation.22 If, as Holthausen believes, the Gmc basis was *gar-w-il-, it would be derived from IE *ghor- with the labial formant seen in OE gearu 'ready', gienvan 'to prepare', plus a labial formant *-il-.23 The of ME girle (instead of ) does not reflect the OE initial palatal sound one should expect if OE gierela or one of its variants was the etymon. Gmc *gurwil-, on the other hand, if the basis of girl, would separate it by vowel gradation (with unknown semantic consequences) from gierela. Robinson suggested that Northern, i.e., Scandinavian, phonological influence may be responsible for the initial plosive. This forces one to believe that a Northern initial consonant should be assumed in ME in spite of the SW and SWM vowel found, for example, in the MED quote form the Becket Legend in the South English Legendary (cl300, Laud MS: South-Western, which reads, e.g., instead of the ON loan give). Before discussing the semantic problems, let us consider the use of gyrela in Old English. According to Healey and Venezky (1980), there are no less than 70 examples of gyrela with its variants and twelve kinds of compound. The word does not seem to be used in poetry except for Guthlac (2x) and The Meters ofBoethius (lx). oööe hwig eode ge ut geseon man hnescum gyrlum gescrydne [: 'mollibus vestitum']. Nu t>a öe synd hnescum gyrlum gescrydde [: 'mollibus vestiuntur'] synd on cynynga husum. West-Saxon Gospels (MS A), Matthew 11.824 (Lindifarne similarly renders the second 'mollibus' freely by mid hnescum gerelum', Rushworth renders the first as


Yoshio Terasawa

ncescum hrceglum; the late Hatton version of the WS translation replaces gyrlum twice by certlan). Hy hine J>a hofun on J>a hean lyft, sealdon him meahte ofer monna cynn Jjaet he fore eagum eall sceawode under haligra hyrda gewealdum in mynsterum monna gebaeru Jjara J>e hyra lifes t>urh lust brucan idlum aehtum ond oferwlencum, gierelum gielplicum, swa bid geoguöe J)eaw J)aer Jjaes ealdres egsa ne styreö. Guthlac 412-2025 And J>a, J>e he ne maeg to him gebiean mid golde ne mid seolfre ne mid godwebgyrlum. Ofi De temporibus Anticristi As in the above examples, OE gyrela is fairly frequently found referring to 'showy, luxurious apparel', which seems to be associated with worldly rank. In the parable of "the Prodigal Son" se selesta gegyrla refers in its context to the best robe, with the younger son is formally honoured on his return home: Bryngaö raöe f)one selestan gegyrlan. West-Saxon Gospels, Luke 15:2227 (Lindisfarne: J) stol aeriste; Rushworth: [leaf wanting in the MS]) In other contexts, OE gyrela often refers to clothing as a symbol of social standing, representation, ostentation or worldly values generally. In the Guthlac passage quoted above it is specifically associated with youthful vanity. It should be noted in this connexion that some (ge-)gyrela compounds (cynegyrela, kynegyryla, cyning-gierela, godweb-gyrla, hropgirela, weoruld-girela30) express connotations of worldly splendour, while others (bisceopgegyrela, diacongegyrela, mcesse-gierela, munucgegyrela) denote formal or vocational robes,31 and earmgegirela 'dextrale' may denote an ornament. As for the second semantic problem above mentioned - how and why the meaning of girl shifted from 'child' or boy' to 'girl', we should

Some etymological

and semasiological

notes on girl


consider the matter from the standpoint of the semantic or lexical field, which consists of synonymous words maid, maiden, wench, damsel, lass, etc. and, in the larger field, includes its word-pair boy and items related to this word. Generally speaking the commonest term for 'girl' in Old English was mcegden·, in poetry, on the other hand, it is ides, followed by fcemne, while mcegden occurs very rarely. Other terms are mceged, meowle, femne, mcegdencild, mcegdenmann (cf. wifmann), mceged-mann, fcemnhadesmann, geong wifmann, etc. In Middle English the central or dominating position of maiden, the ME reflex of OE mcegden, with its apocopated variant maide, was in due course replaced by wneche derived from OE wencel, and the OF loanwords damsel. The replacement in Middle English of maiden and maid(e) in the sense of 'a young woman' may be due to the fact that maiden increasingly occurred in the sense of 'virgin' religious contexts and especially the B.V.M., and maid(e) commonly came to denote 'virgin' on the one hand and 'maid-servant' on the other. In worschepe of god & of his moder maiden marie, & seint katerine f>e gloriouse virgine & martyr 1389 Gild of St. Katherine, Aldersgate33 Hie is ... aire maidene maide. cl200 Trinity Homilies 16134 Heil beo thow, Marie, Moodur and May. cl390 Heil ne thow Marie Moodur A similar fate later befell the word wench. During the 14th century wench acquired the pejorative overtone of 'mistress' as well as the sense 'maid-servant', and its function as a neutral term became less definite until it was finally ousted from its central position in the semantic field. I am a gentil womman and no wenche. c1395 Chaucer Canterbury Tales, Merchant's Tale Ε 2202 Thy ney3boris hows, wenche [v.l. T: maiden] ne knaue, coueyte hem nojt. cl380 The Lay Folk's Catechism L 86136


Yoshio Terasawa

In the mid-14th century boy is first attested in the sense of 'a young male' and the 15th century the word becomes prevalent in this sense. Besides, boy had definitely denoted the male sex since its supposed first appearance, without a single instance of sex-neutral or female reference. About the same time, child was also becoming a central term in the semantic field of 'child'. It seems reasonable to assume, therefore, that in the late 14th and the 15th century the tension which had existed among boy-child-girl broke down and the semantic field began to be restructured. Of the three allosemes of girl, 'child, youth', 'boy' and 'girl', the first two were gradually dying out except in some dialects, as boy and child were established in these functions. This might have been the situation in which girl 'a female youth' entered the centre of the semantic field of 'girl' - first perhaps in colloquial texts.38 In this paper we have contented ourselves with a brief attempt to clarify some relevant problems about the etymology and the semantic development of girl. Opinions, as cited above, are much divided as to the origin of the word. Except for the Anlaut problem involved, Robinson's proposal of OE gyrela 'apparel' being the etymon seems to us the most convincing of those so far offered, although here we can do little more than speculate for lack of direct evidence. However, concerning the later semantic development of girl, a fairly adequate rationale seems to emerge from the foregoing discussion. It would be further desirable to examine in detail the 15th-16th century literature with respect to the use of the word in colloquial as well as literary styles,39 since these two centuries were also a crucial and formative period in the making of ModE girl·, but this will need another paper.

Notes *

The original version of this paper was published in Anglo-Saxonica: Festschrift für Hans Schabram zum 65. Geburtstag, ed. by Klaus R. Grinda and Claus-Dieter Wetzel (München: W. Fink), 1993, pp. 335-45. Several minor additions and corrections have been made for this reprinting. I am indebted to Dr. Klaus R. Grinda of the University of Göttingen and Prof. Paul E. Davenport of Hitotsubashi University for their valuable comments and suggestions in preparing the original paper published in the Schabram Festschrift.

1. The ME quotations in this paper as well as their datings are based on those in the MED.

Some etymological and semasiological notes on girl


2. Smithers (1952-57). R.M. Lumiansky dates the text "Beginning of fourteenth century" and the Laud MS "ca. 1400" in Severs and Härtung (eds.), 1,105, 270. 3. All Chaucer citations are from Benson et al. (1987). 4. Kane and Donaldson (1975 I: 320; Alford (1988: 184-86). 5. Block (1922). 6. This was already recognised by Thomas Tyrwhitt in the glossary to his edition of the Canterbury Tales ((London, 1775-78) V, 85: "Girles, ... Young persons, either male or female". In The Riverside Chaucer, the word is glossed as 'young woman or young people'. 7. Skeat (1867). C.W. Dunn dates the translation "ca. 1350-61" and the MS "ca. 1360-75" in Severs and Härtung (eds.), I, 34, 223. 8. Gordon (1953). M.P. Hamilton dates the text "sometime within the period 1360-95" and the MS "1375-1400" in Severs and Härtung (eds.), II, 339, 503. 9. Eccles (1969). S. Lindenbaum dates the text "ca. 1400-25" and the MS "ca. 1440" in Severs and Härtung (eds.), V, 1366, 1606. 10. In the same poem the early use of boy appears, but the meaning may be 'wretch' or 'ruffian'. Cf. the notes to Pearl 806 in Gordon (ed.) as well as in Vantuono (ed.), II, 263, 319 (note to Cleanness 878). 11. In Shakespeare, wench is commonly used as a term of affectionate or familiar address or with a pejorative implication. 12. Early printed Bibles are quoted respectively from the facsimile reprint of the first edition: Coverdale (Folkestone, 1975), Great Bible (Tokyo, 1991), Geneva (Madison, 1969), Reims-Douai (Ilkley and London, 1975), A.V. (Tokyo, 1982). Coverdale (1535) & Great Bible (1539): the yonge men ... the Damsels; Geneva (1560): the child ... the girle; Rheims-Douai (1609-10): boy ... wench. 13. Coverdale & Great Bible: yonge boyes and damselles; Geneva: boys & girles; RheimsDouai: infants ... girles. 14. The list is not exhaustive. Source information is abridged as expected to be reasonably self-evident. 15 .Dictionary of English Etymology (London, 1859-65; 2nd ed.: 1872). This was not available to me. 16. Müller (ed.), s.v. girl: "vom ags. gyrla kleidung". So far Müller seems to have invented the etymology proposed independently by Robinson almost a century later. But Müller seems to equate gyrla erroneously with gyrdel ("Zusammenhang zwischen benennung des kleidungsstückes ags. gyrla, gyrdel, und des kleinen kindes ist nicht undenkbar"). He does not consider Wedwood's gör as a competing etymology, but as a parallel, being itself related to gehre 'rockschoss, zwickel'. 17. Luick (1897: 235f); cf. Berndt (1960: 339f). 18. Gör received an entry in Kluge, only in its 7th ed., when Kluge called its etymology "dunkel" and its relation with girl "ganz fragwürdig". Subsequently, Alfred Götze (in


Yoshio Terasawa

the 1 Ith ed., 1934) began to etymologise göre, but did not mention girl any more. Most recently, Elmar Seebold (in the 22nd ed., 1989) brings in girl again, but is vague about etymology and relationship of either word. 19. Cf. the interpretation by Meritt, 69. 20. RHD=77je Random House Dictionary of the English Language (1st ed. by Jess Stein; 2nd ed. by S.B. Rexner; etymological editor: J. Radar) (New York, 1966, 1987); Bamhart=77ie Barnhart Dictionary of Etymology, ed. R.K. Bamhart (New York, 1988); RHWCD=Random House Webster's College Dictionary, ed. R.B. Costello (New York, 1991). 21. Cf. Pauli (1919); Flom (1913). 22. Cf. Cosijn (1883-86,1: 32); Bülbring (1902: § 306); Holthausen (1934: ί.ν. gierela). 23. For the IE root, cf. Pokorny (1959,1: 494); for the suffixes, cf. Meid (1969: §§ 77.1, 87.3). 24. Grünberg (1967:11) 568-70. The other gospel versions are quoted from Skeat (1871-87). 25. Roberts (1979) 26. Napier (1967: XLII, 197-1). 27. Griinberg (1967:11) 993f. 28. Typical adjectives used with gyrela are cynelic, deorwurp, gilplic, godwebben, ofergyld(d), purpuren, scinende, wuldotfull·, such attributes characterise reaf or hrcegl far less frequently. It is a sign of 'vana gloria' to be on ... gyrlum ranc (cf. jElfiric's Pastoral Letters to Wulfstan, ed. Bernhard Fehr, II, § 187; III, § 173). 29. The only other example with this connotation seems to occur in Wulfstan, Horn. XI, 129 (ed. Dorothy Bethurum), abbreviating Is 3.16ff. Youth or youthfulness are otherwise not markedly associated with gyrela in the OE texts. 30. Such compounds seem to be lacking with reaf or hrcegl. 31. 32. 33. 34. 35

Gyrela shares this type of compound with reaf Cf. Bäck (1967); Diensberg (1985). Chambers and Daunt (1931: 49, 66f). Morris (1873: 161,2). Horstmann (1892: XXVIII, first line). The poem predominantly uses mayden (19, 37, 61,81) and once mayde (75). 36. Simmons and Nolloth (1901). L was issued in 1357 and Τ is dated "early fifteenth century" (MS "1400-1425") according to R.R. Raymo in Severs and Härtung (eds.), VII, 2270f„ 2492-94. 37. Cf. Dobson (1940); Dobson (1934); Diensberg (1981). 38. Weekley (ed.) suggests that limitation of sex may be due to association of girl with gil, gille 'name of a mare, girl, woman', which is originally a feminine name Gille, shortened form of Gillian (=Julian(e)). Yet this raises an Anlaut problem, as late ME/Early ModE gille is merely a graphic variant of Jille, both forms beginning with an

Some etymological and semasiological notes on girl


affricate: cf. OED s.v. Gill, sb. 4. Gil, gille seems mainly to appear in medieval plays and such informal texts as Paston Letters. 39. The late ME form garlement for garment in Promptorium Parvulorum (1440), Paston Letters (1422-1509) (MED s.v.) and Digby Plays (c 1480-90) (OED s.v.) may somewhat corroborate the colloquial or low status of ME girle, since garlement is nothing but a blend of girl 'apparel' with garment and this blending presupposes that the reflex of OE gyrela 'apparel' had continued to exist at least on the spoken or dialectal level in Middle English.

References Primary Sources Benson, Larry D. et al. (eds.) 1987 The Riverside Chaucer. (3rd ed.) Boston: H. Mifflin. Bethurum, Dorothy (ed.) 1957 The homilies ofWulfstan. Oxford: Oxford University Press. Block, K.S. (ed.) 1922 Ludus Conventriae or the plaie called Corpus Christi. (EETS, E.S., 120.) London: Oxford University Press. Chambers, Raymond W. - Maijorie Daunt (eds.) 1931 A book of London English 1384-1425. London: Oxford University Press. Eccles, Mark (ed.) 1969 The macro plays. (EETS, O.S., 262.) London: Oxford University Press. Gordon, E.V. (ed.) 1953 Pearl. London: Oxford University Press. Griinberg, Madeleine (ed.) 1967 The West-Saxon Gospels. A study of the Gospel of St. Matthew with the text of the four Gospels. Amsterdam: Scheltema & Holkema. Horstmann, C. (ed.) 1892

The minor poems of the Vernon Manuscript. Part I. (EETS, O.S., 98) London: Oxford University Press. Kane, George - E.T. Donaldson (eds.) 1975 Piers Plowman: The Β Version. Will's Visions of Piers Lowman, Do-Well, Do-Better and Do-Best. London: Athlone Press. Morris, Richard (ed.) 1873 Old English Homilies of the Twelfth Century, 2nd series, EETS, O.S., 53. London: Oxford University Press.


Yoshio Terasawa

Napier, Arthur Sampson (ed.) 1967 Wulfstan: Sammlung der ihm zugeschriebenen Homilien nebst Untersuchungen über ihre Echtheit, 2. unveränderte Aufl. mit einem bibliographischen Anhang von Klaus Ostheeren. Dublin/Zürich: Weidmann/M. Niemeyer. Roberts, Jane (ed.) 1979 The Guthlac poems of the Exeter Book. London: Oxford University Press. Simmons, T.F. - H.E. Nolloth (eds.) 1901

The Lay Folk's Catechism. EETS, O.S., 118. London: K. Paul, Trench, Trübner. Skeat, Walter Williams (ed.) 1867 William ofPalerne. EETS, E.S., 1. London: K. Paul, Trench, Trübner. 1871-87 The Gospel According to Saint Matthew (...Mark, ...Luke, ... John). Cambridge: Cambridge University Press. Smithers, G.V. (ed.) 1952-57 Kyng Alisaunder EETS, 227, 237 London: Oxford University Press. Vantuono, William (ed.) 1984 The Pearl poems: An Omnibus edition. New York: Garland. Secondary Sources Alford, J.A. (ed.) 1988 A Companion to Piers Plowman. Berkeley: University of California Press. Arndt, W.W. et al. (eds.) 1967 Studies in historical linguistics in honor ofG.S. Lane. Chapel Hill: University of North Carolina Press. Bäck, Hilding 1967 The synonyms for "child", "boy", "girl" in Old English: An etymologicalsemasiological investigation. Lund Studies in English, II. Lund: C.W.K. Gleerup. Bemdt, Rolf 1960 Einfuhrung in das Studium des Mittelenglischen. Halle: Μ. Niemeyer. Bülbring, Karl D. 1902 Altenglisches Elementarbuch, I: Lautlehre. Heiidelberg. Cosijn, Peter J. 1883-86 Altwestsächsische Grammatik. Haag. Dienesberg, Bernhard 1981

"The etymology of Modern English boy. A new hypothesis", Medium Aivum 50: 79-87.

Some etymological and semasiological notes on girl 1985

"The lexical fields boy/girl-servant-child logische Mitteilungen, 86: 328-36.


in Middle English", Neuphilo-

Dobson, Eric John 1940 "The etymology and meaning of boy", Medium JEvum 9: 121 -154. 1943 "Middle English and Middle Dutch boye", Medium Mvum 12: 71 -76. Flom, George T. 1913 "Semological notes on Old Scand. flik and derived forms in the Modern Scandinavian dialects", JEGP 12: 78-92. Grinda, Klaus R. - Claus-Dieter Wetzel (eds.) 1993 Anglo-Saxonica: Festschrift für Hans Schabram zum 65. Geburtstag. München: W. Fink. Healey, A.diP - R.L. Venezky (eds.) 1980 A microfiche concordance to Old English. Toronto: Centre for Medieval Studies, University of Toronto. Holthausen, Ferdinand (ed.) 1934 Altenglisches etymologisches Wörterbuch. Heidelberg: C. Winter. Kluge, Friedrich (ed.) 1910 Etymologisches Wörterbuch der deutschen Sprache. (7th edition) Straßburg: K.J. Trübner. Kurath, Hans et al. (eds.) 1952Middle English dictionary. [MED] Ann Arbor: University of Michigan Press. Luick, Karl 1897 "Die herkunft des ne. girF', Anglia Beiblatt 8: 235-236. Meid, Wolfgang 1969 Germanische Sprachwissenschaft, III: Wortbildungslehre. Sammlung Göschen. Berlin: Walter de Gruyter. Meritt, Herbert Dean 1959 The Old English Prudentius Glosses at Boulogne-sur-Mer. Stanford Studies in Language and Literature, 16. Stanford: Stanford University Press. Müller, Edward (ed.) 1865-67 Etymologisches Wörterbuch der englischen Sprache. Cöthen: P. Schettler; 2. Aufl.: 1878-79. Murray, James Augustan Henry, et al. (eds.) 1884-1928 The Oxford English dictionary. 10 vols. Oxford: Oxford University Press. Pauli, Ivan 1919 'Enfant', 'Garfon', Fille' dans les langues romances. Lund: Gleerup. Pokorny, Julius (ed.) 1959-69 Indogermanisches etymologisches Wörterbuch. 2 Bde. Bern & München: Francke.


Yoshio Terasawa

Robinson, Fred Colson 1967

"European clothing names and the etymology of girl", in: W.W. Arndt et al. (eds.), 233-239. Severs, J. Burke - A.E. Härtung (eds.) 1967-

A manual of the writings in Middle English 1050-1500. New Haven: The Connecticut Academy of Arts and Sciences. Weekley, Ernst (ed.) 1921

An etymological dictionary of Modern English. London: J. Murray.

On the functional motivation of phonological changes in English* Norio Yamada

1. Introduction In view of the fact that phonological changes do take place in all languages, historical phonologists obviously have a responsibility to uncover causes of phonological change, or to identify its raison d'etre. In order to tackle this conundrum it may be advantageus first of all to break it down into the following three questions: (1)




The repetition riddle Why do phonological changes occur repeatedly throughout the temporal continuum? The causation riddle Do these changes occur independent of each other? Or do they have some causal relations? The opposition riddle Why do mutually opposing changes, such as lengthening and shortening of vowels, sometimes occur?

The present paper aims to answer these questions from a functional point of view. It will be shown as a conclusion that the majority of phonological changes occur so as to "conspire" to add to the perceptual salience of stressed syllables. On the other hand, once a change that serves to facilitate perceptual processing either goes beyond a certain limit or reaches its utmost end, a diametrically opposite change may well happen as a functional reaction that relaxes the heightened muscular tension and thereby renders segments or segment sequences easier to pronounce.


Norio Yamada

2. The basic assumptions In this section, we provide some basic assumptions for our exposition of phonological change. Since changes in stressed syllables seem to be intrinsically different from those in unstressed syllables, we should be able to distinguish them in some way or other. For this, we may well take the following as a basic working hypothesis. (2)

Diachronic Perceptual Principles a. Stressed syllables tend to be strengthened. b. Unstressed syllables tend to be weakened, especially in the posttonic environment.

Let us first consider strengthening processes in stressed syllables. These can be classified into peak strengthening and marginal strengthening, of which the latter can be further divided into initial strengthening and final strengthening, as in (3). (3)

Strengthening processes a. Peak strengthening vowel lengthening/diphthongization/long vowel raising b. Marginal strengthening i) Initial strengthening fortition processes (such as affrication) / obstruent devoicing / initial C addition / noninitial C deletion / noninitial C postposing (or metathesis) ii) Final strengthening fortition processes (such as despirantization) / obstruent devoicing / final C addition (especially by gemination)

The peaks of stressed syllables are likely to undergo lengthening, diphthongization, or raising especially when they are long vowels. Such changes are lumped together here under the general name of peak strengthening. Every phonological change belonging to the category of marginal strengthening seems to have the effect of enhancing the salience of a peak vowel by increasing the contrast in sonority between it and the surrounding consonants within a stressed syllable. In other words, both the onset and the coda of a stressed syllable serve as foils for the peak.

On the functional motivation of phonological changes in English


Similarly, weakening processes in unstressed syllables can be classified into peak peak weakening and marginal weakening, the latter of which can in turn be subcategorized into initial weakening and final weakening, as in (4). (4)

Weakening processes a. Peak weakening monophthongization / vowel shortening / vowel reduction / vowel loss / sonorant syllabification b. Marginal weakening i) Initial weakening lenition processes (such as spirantization) / obstruent voicing / initial C deletion ii) Final weakening obstruent voicing / final C deletion (especially by degemination)

The peak of an unstressed syllable can be weakened by dint of processes such as monophthongization, vowel shortening, vowel reduction, vowel loss, and sonorant syllabification. When compared with marginal strengthening, every instance of marginal weakening turns out to have exactly the opposite effect of making the peak vowel of an unstressed syllable less prominent by diminishing the sonority contrast between it and the marginal consonants which surround it. While it can surely be considered an inherent property of unstressed syllables, this weakening seems to be particularly favored in the environment immediately following a stressed syllable, for it makes it possible to bring that stressed syllable into more prominence. Thus in those cases, the unstressed syllable at issue can be regarded as serving as a foil for a stressed syllable immediately preceding it. Thus, it follows that in the majority of cases phonological changes which strengthen stressed syllables and those which weaken unstressed syllables do have a functional unity in that both contribute to increasing perceptual salience in stressed syllables. However, if such a change pushes ahead to the very limit or somewhere near to this point, then exactly the opposite change may well happen so as to counteract that change. Thus, we can often find in stressed syllables such counteractive processes as monophthongization and vowel shortening, processes which undermine the tendency to add to salience in stressed syllables, and thus help preserve an equilibrium between perceptual and productive requirements.


Norio Yamada

3. Peak strengthening in stressed syllables The purpose of this section is to discuss some examples of peak strengthening in stressed syllables. In particular, two important changes that had a far-reaching influence on the establishment of the phonological word structure of Modern English, the Middle English Open Syllable Lengthening and the Great Vowel Shift, will be examined in search of their causes and motivations.

3.1. The Middle English Open Syllable Lengthening Let us first briefly consider, as a representative of vowel lengthening phenomena, what is generally called the Middle English Open Syllable Lengthening (MEOSL). We agree with Minkova (1982: 50-51, 1985: 170171, 1991: 116), Lass (1985: 250-252), Hock (1986), and Hayes (1989: 266-269) in assuming that this process of change can be qualified as an instance of compensatory lengthening triggered by the loss of final schwa. Thus, the change can be graphically represented as in (5). (5)


(tale 'tale' > täl)

Moreover, as Jones (1989: 64-66) and Ritt (1992: 208, 1994: 117) point out, the lengthening phenomenon itself should be seen as a perceptuallymotivated "foregrounding" (i.e., strengthening) process.

3.2. The English Great Vowel Shift Recently, not a few articles have been published that have all dealt with the basic question as to whether the individual changes that have been supposed to fall under the so-called "Great Vowel Shift (GVS)" can ever constitute a coherent whole. It is Stockwell - Minkova (1988) who have taken the initiative in the recent lively discussion on this question; in fact, they have taken issue with the traditional view of the GVS as a temporally localizable and internally coherent event. When we take into consideration the fact that many isolated changes of diphthongization and vowel raising that have no connection with chain shifting can be attested in both synchronic and diachronicaspects of a great variety of languages (see Wells 1982: 308, Labov 1994: 122, and Nakao

On the functional motivation of phonological changes in English


1985: 163-167 for examples and further discussion), it does not seem unnatural to suppose that we may possibly gain a more general and reasonable explanation if we look upon the individual changes of raising and diphthongization that have usually been lumped together as the GVS, not as a temporally delimited chain shift, but rather as sharing a common motivation with many analogous changes that can be attested in certain other periods in the history of English or in some other languages, as has been argued in Stockwell - Minkova (1988). Furthermore, as Samuels (1972: 41) and Labov (1972: 154, 1994: 261) correctly point out, the process of long vowel raising can be seen as an instance of strengthening that heightens the perceptual salience of stressed syllables, just as diphthongization can be so interpreted (cf. Sweet 1888: 21, Stockwell - Minkova 1988: 381, and Johnston 1992: 220).

4. Peak weakening in unstressed syllables As we have argued in section 2, weakening may be considered to be a phenomenon characteristic of unstressed syllables which serves to lower their degree of salience. Thus monophthongization, a typical weakening process, has the basic function of decreasing salience in unstressed syllables. Another more striking fact about this process - or any weakening process, for that matter - is that they tend to affect unstressed syllables most easily when they tend to affect unstressed syllables most easily when they are either immediately preceded or followed by stressed syllables in the same words, as is clear from examples such as the following: (6)

Peak weakening in unstressed syllables a. Monophthongization in EModE (Nakao 1985: 312) aug.ment,









Vowel shortening in ME (Nakao 1985: 312-314) wis.döm > wis.dom, right.wis 'lawful' > right.wis, sti.räp 'stirrup' > sti.rop, wed.läc 'wedlock' > wedlok Vowel reduction from late OE to ME (Nakao 1985: 314-315, Campbell 1959: §§369-380, Minkova 1991: 87-90) cnot.ta > knot.te 'knot', heo.fon > he.ven 'heaven', se.nu > se.ne 'son'. StÖ.ne > sto.ne 'stone' [dat. / sg. / masc.] Vowel syncope in OE (Nakao 1985: 318-319) cy.ning > cyng 'king', b6.cere > b6cre > writer


Norio Yamada


ώ/.terra > xf.tra nexi,fiil.tumian >ful.tmian 'assist' Sonorant syllabification in late ME and EModE (Dobson 1957: §§318-326) sdd.dle, me.tal, e.fr 'ever', d0uh.tr 'daughter', o.xen, benign ([gn], schi.sm, cdtechi.sm

The reason for this seems to be that an unstressed syllable appearing in this environment, if it undergoes weakening of some sort, can then serve as a foil for an immediately adjacent stressed syllable, indirectly enhancing its salience. Thus, weakening in this particular environment can be seen as functionally similar to changes which increase perceptual salience of the peaks of stressed syllables themselves, such as diphthongization and low vowel raising.

5. Counteractive processes Let us next turn our attention to what we call counteractive processes. Phonological changes such as monophthongization and vowel shortening can be interpreted essentially as weakening phenomena. It thus seems quite natural to suppose that they may usually occur in unstressed syllables. For all that, we are occasionally confronted with those cases where such changes actually occur in stressed syllables. This causes us no surprise, however. As we noted at the outset, once a phonological change which serves to increase perceptual clarity in stressed syllables reaches its utmost end, or at least advances beyond some fixed limit, the exact opposite of that change may very well happen as a counteractive process which aims at either relaxing muscular tension, and thereby making the act of production easier to perform, or remedying an unbalanced situation caused by the overexecution of a former, perceptually-based change. Processes of the former type can be illustrated by examples: (7)



(ME /i:/ >) [ai] > [a:] (Southern States of America, parts of England, South Africa (cf. Wells 1982: 208-209, Labov 1994: 170-171)) (ME Im:/ >) [au] / aeu] > [a:] / [χ:] (American Tidewater South, Pittsburgh dialect, popular London speech (cf. Stockwell Minkova 1988: 372, Labov 1994: 281, Gimson 1989: 138))

On the functional motivation of phonological changes in English


These monophthongization processes can be interpreted as assimilatory phenomena which may occur to counteract the effect of maximal differentiation brought about by diphthongization processes. 5.1. The Late ME-to-EmodE Monosyllabic Shortening As an example of the latter type of process, we consider a process of monosyllabic shortening in English. 5.1.1. The basic facts We know that from Late Middle English on to Early Modern English, long vowels were shortened in stressed monosyllabic words ending in single consonants. This shortening most commonly affected ME open ξ and close ö, though its influence can be discerned in words with other long vowels as well (see Dobson (1957: §§ 24-39), Araki - Ukaji (1984: 150-153), and Nakao (1985: 148-151)): (8)

The LME-to-EModE Monosyllabic Shortening a. ME /ε:/ > [ε] death, breath, deaf, bread, dead, thread, haed, dread, stead, spread, sweat, threat, wet, lead, shed, red, shred, tread b. (ME /o:/ > ) [u:] > [u] (> [λ]) i) Shortening earlier than the unrounding of ME /u/ blood, flood, glove, love, stood, foot, shoot, move, remove, prove, behove, struck, roof ii) Shortening later than the unrounding of ME /u/ good, stood, foot, sooth, tooth, book, hook, look, took, cook, brook, crook, shook, nook, rook c. (ME /e:/ > ) [i:] > [i] sick, heel, sheep, seem, creek, green, keen, keel, deed, seen, been, rick, wick, grit d.

e. f.


cloth, froth, wroth, loath, both, loaf wood, rode, broad, abroad, hot, wot, wrote, boat, whom, gone, shone, anon ME/u:/>[u] enough, rough, though (ME /ai/ > ) [ε:] > [ε] again, said, saith, says


Norio Yamada g.

ME /au/ > [a] laugh, calf, half

There is one further point that should not be overlooked. Thus the shortening at issue did not apply automatically to all relevant words satisfying the conditioning environment. There are in fact many monosyllabic words that have remained to be shortened: (9)

Monosyllables with long vowels a. ME /ε:/: sheath, heath, beneath, underneath, wheat, eat, beat, cheap, speak, leave, ease, stream, heat, meat, seat, leaf, sheaf, read, mead, bead b. ME /o:/: food, mood, boot, root, goose, loose, choose, lose, ooze, brood, rood, hood, moot, sooth, forsooth, soot, room, broom, spoon, hoof

5.1.2. The relevant period There are at least two questions that need to be asked about the chronological order of this shortening: (10)

a. b.

When did this shortening happen? At what period did it reach its climax?

Of direct relevance to these questions are the following remarks made by Dobson (1957) and Kuiytowicz (1966): ...it [shortening in monosyllables before single final consonants, NY] probably operated most fully in the early fifteenth and may have already occurred occasionally in the fourteenth century, i.e., contemporaneously with lengthening in open syllables ... (Dobson 1957: 497) This second wave of shortenings [mostly a shortening of e, ö before final consonants, NY], entailing e > ϊ, ö > ü, took place after the loss of final -e, i.e., in late M.E. from the fourteenth century on, though earlier in the North (Kurylowicz 1966: 184).

On the functional motivation of phonological

changes in English


As we noted earlier, MEOSL and schwa loss occurred simultaneously, both peaked around c. 1250 to c. 1400, and continued perhaps into the beginning of the fifteenth century. Now, if we build on Dobson's and Kuryiowicz's well-founded assumptions, we may say that the monosyllabic shortening took place nearly concurrently with, or perhaps directly after, the other two interdependent changes of MEOSL and schwa loss. In other words, it appears that the shortening in question occurred as if to run after the MEOSL-cum-schwa-loss change. 5.1.3. Monosyllabic

Shortening as a counteractive


As we discussed in section 3.1, MEOSL can be qualified as a compensatory lengthening process triggered by schwa loss, yielding long vowels in precisely the same environment as in the shortening process under consideration, namely in stressed monosyllables ending in single consonants, as in tdle 'tale' > ΐά. Furthermore, the lengthening process, which completely squares with the general tendency toward strengthening in stressed syllables, exercised a quite systematic influence upon a huge number of relevant words. We may, therefore, go on from these considerations to the conclusion that the monosyllabic shortening at issue occurred as a counteractive change in the same environment as the output structure of MEOSL in order to cope with the sweeping spread of this lengthening through the vocabulary, and to preserve the variety of words in such a way that not only monosyllables with long vowels but those with short vowels may also be admitted as possible lexical items in the language. On the other hand, there are also many monosyllabic words that failed to be shortened, as in (9) above. This seems to reflect nature of the change as a mere counteractive process against MEOSL.

6. Marginal strengthening in stressed syllables We will now proceed to the discussion of a wide variety of consonantal changes which can be observed in the history of English. First of all, we have to inquire into several types of phonological change falling under the category of marginal strengthening. It will be shown as a conclusion that these types of phonological change each serve to enhance salience in


Norio Yamada

stressed syllables by increasing the contrast in sonority between the peak and the surrounding consonants within a stressed syllable. 6.1. Initial strengthening Changes of initial strengthening can be classified into several types. First, let us consider fortition processes which make consonants stronger or less sonorous by making them to the right along the following scale: (11)

The universal scale of consonantal strength (Hooper 1976: 206)

voiced voiceless continuant glides liquids nasals continuant voiced stop 1





voiceless stop • 6

Thus in the history of English, some sporadic changes of glide strengthening can be observed in word-initial position. In the following examples the glide /w/ is strengthened to the stop [b] (Lutz 1992: 158): (12)

Glide strengthening OE wlott 'blot' > 14th c. blot / OE wlabffian 'stammer; speak indistinctly' > 14th c. wläffen ~ bläffen (cf. EModE bläff 'bark') / ME wräbben 'dispute, quarrel' > c. 1500 bräbble

Another type of initial strengthening is the addition of a consonant to the initial position of a stressed syllable, whose peak vowel will thereby be made perceptually more salient. Thus, from the period of Early English onward /h/ was sometimes added to word-initial stressed syllables beginning with vowels (Nakao 1985: 430): (13)

Initial /h/-addition in EME herthe 'earth', howyn 'own', helde 'age', houp 'out', hinde 'end', härme 'arm', here 'ear', hetten 'eaten', hoyth 'oath'

Special mention must be made of the third type of initial strengthening, namely the deletion of a noninitial consonant from the onset of a stressed syllable. According to our hypothesis presented in section 2, the basic function of the onset of a stressed syllable is to increase the perceptual

On the functional motivation of phonological changes in English


salience of the peak vowel by placing in initial position a consonant as low in sonority rank as possible. This makes it quite natural to expect that under normal conditions the initial consonant of a stressed syllable will not be deleted by any phonological change. In fact, there has hardly been any such change during the long history of English, though a few changes deleting certain extrasyllabic consonants have been attested. By contrast, when a nominal consonant is deleted from the onset of a stressed syllable, its first half (i.e., the initial demisyllable) comes to show a sonority profile that rises more sharply toward the peak, making the peak vowel the more conspicuous. It is probably for this reason that the following changes occurred in OE (Nakao 1985: 421-423, Campbell 1959: §475): (14)

Noninitial C deletion a. OE /w/-deletion cwucu 'alive' ~ cucu, cwüdu 'cud' ~ cüdu, *hwu > OE hu 'how', swtigian 'be silent' ~ sugian, swütol 'clear' ~ sutol, be.twux ~ be.tuh 'between', swuster 'sister' ~ suster, cw6mon ~ cömon '(we) came' b. Late OE /r/-deletion sprecan > specan 'speak', spr&c > spaec 'twig', pr£ttig ~ pdttig


The final type of initial strengthening is a process usually called metathesis, which shifts a noninitial consonant from the onset of a stressed syllable to a post vocalic position. This bears a striking resemblance to the process of noninitial C deletion just discussed in that it transforms a stressed syllable into one with a sharper rise in sonority at the beginning. What has been called the /r/-metathesis in OE seems to offer a case in point. Campbell (1959: §459) gives the following examples: (15)

OE /r/-metathesis brdstlian 'crackle' ~ b&rstlian, briinna 'stream' ~ burna, cresse 'cress' ~ cerse,frist'period' ~ first, dr&stan 'dregs' ~ daerstati, frosc 'frog' ~ forsc, frost 'frost' ~ forst, grass 'grass' ~ gc&rs, dridda 'third' ~ dirda, cyrps 'curly' / hors 'horse', birdas 'young birds', gescryd(ed) 'clothed' ~ gescyrd


Norio Yamada

6.2. Final


The coda, as well as the onset, of a stressed syllable can serve as a foil in enhancing the salience of the peak vowel. Changes of final strengthening can be distinguished into three main types: (i) fortition, (ii) obstruent devoicing, and (iii) final C addition. As an example of the first type, let us consider an OE-to-ME change of despirantization, about which Jordan - Crook (1974) observe that "in continuation of an already OE tendency the voiced d before an immediately following, not syllabic n, m, r goes over into the stop d" (p. 186). The fact that this fortition occurred before the nonsyllabic sonorant consonants implies that the consonants undergoing this process were located in the final position of stressed syllables. The following are some representative examples: (16)

OE-to ME final despirantization OE fsbd.me (inflectional) > fsdme = ME fedme, fadme (—» fdddom), OE byrd.ne (inflectional) > *byrdne (—> ME burden 'burden'), Cuj?.(wi)nes dün > *Cudnes- (—» Cuddesdon), röd.re- (inflectional) > *r6dre (—> ME röder 'rudder'), spfd.ra > ME spfdre (—» spider 'spider'), mord.rian 'to murder' > ME mürdren

The second strategy for final strengthening is devoicing of obstruent, which can be illustrated by the following change in OE (Campbell 1959: §446): (17)

OE/yAdevocing a. Early West Saxon wah 'wall', gen6h 'enough', bürh 'city', -sl6h 'struck', löh 'blamed', gef6h.stän Titted alone', bürh.ware 'citizens' b. Late West Saxon bSah 'ring', pl6h 'plough', stSi 'ascended', sorh 'sorrow', bealh 'was angry', läh.bryce 'breach of law', bSah.gifa 'ringgiver', 6eh.\>yrl 'eye-hole', flih.las 'birds', dh.nian 'possess'

We now turn to the remaining type of final strengthening, i.e., the addition of a consonant to the final position of a stressed syllable. Let us take as an example what has traditionally been called the West Germanic

On the functional motivation of phonological

changes in English


Gemination, of which Prins (1972) gives the following succinct description as in: All single consonants with the exception of r were doubled after a short vowel before a following j. This j was generally lost in OE, but mostly retained in Old Saxon. (Prins 1972: 195). He then illustrates the effects of this change by the following examples: (18)

West Germanic Gemination Gothic (Go) hlahjan 'to laugh' ~ OE hlieh.han, hleh.han / OE High German (OHG) hläh.han Go lagjan 'to lay' ~ OE lec.&m / Old Saxon (OS) leg.gian / OHG leg.gen, Dutch (Du) leg.gen Go satjan 'to set' ~ OE set.tan, OS set.tian/Du zet.ten Go skapjan 'to create' ~ OE sciep.pan, scep.pan / OS skep.pian Du schep.pen

Of particular interest here is his observation that in OS the glide /j/ was mostly retained even after the preceding consonants had been doubled. This means that in the OS outputs of the gemination the second consonant of a geminate belonged to a second unstressed syllable and thus formed its onset together with the following glide. In view of the fact that, unlike stressed syllables, unstressed syllables do not allow any consonants to be added to their initial position, it seems to be the case that, even before the gemination, a consonant that was to be geminated and a following glide /j/ were tautosyllabic, and that the gemination adjoined a consonant to the final position of the preceding stressed syllable so as to heighten the prominence of its peak vowel. In short, the gemination in question seems to have operated in such a way as to spread a consonant, not from left to right as in Figure la, but in the reverse direction as in Figure lb. Figure 1. Direction of spreading in West Germanic Gemination

a. *VC.jV > VC.DjV

b. v . c j v > v m . c j v


Norio Yamada

7. Marginal weakening in unstressed syllables The present section focuses on the problem of why changes of consonantal weakening characteristically occur in unstressed syllables. Specifically, it will be shown that every instance of marginal weakening in an unstressed syllable contributes to making its peak vowel less salient and thereby making the immediately preceding stressed syllable the more prominent. 7.7. Initial


We will begin by considering changes of initial weakening in unstressed syllables. Such changes fall into at least three types: (i) lenition, (ii) obstruent voicing, and (iii) initial C deletion. In the case of the first type of change, the initial consonant of an unstressed syllable will be lenited or weakened into something else which is more sonorous. Thus in OE the voiced stops Ibl and /d/ were spirantized into [v] and [δ] respectively at the beginning of posttonic unstressed syllables (Nakao 1985: 384): (19)

The OE spirantization of Ibl and /d/ o.ber > ofer 'over', gi.baen 'give' > [v], hal.bae 'half > [v], sni.fyan 'cut', weor.dan 'become', ewe.pan 'speak', Ιίφαη 'go'

Secondly, we often encounter cases in which obstruents become voiced in the initial position of posttonic unstressed syllable. Thus in OE the anterior fricatives I f , Θ, si underwent voicing in this very weakening environment (Nakao 1985: 377, Campbell 1959: §§ 444-445, Kim 1991): (20)

OE fricative voicing a. HI > [v]: gie.fan 'give' - geaf '(he) gave', heo.fon 'heaven', ofer 'over', si.fiin- 'seven', clo.fae 'buckle', ho.fr 'hump', girde.fa 'reeve', sco.fi 'shovel', fi.fe (plural) 'five', hr6.fe [dat./sg.] 'roof / wülfas 'wolves' b. IQI > [3]: weor.dan 'become' ~ weard '(he) became', fe.per 'feather', 6.per 'other', fsb.Öm 'fathom', eor.de 'earth', sni.fjan 'to cut', S.pel 'home', brö.por 'brother' c. Isl > [z]: rtsan 'rise' ~ rds '(he) rose', no.su 'nose', b6.sm 'bosom', 6.sle 'ouzel', hriop.san 'to fell', cio.san 'choose'

On the fimctional motivation of phonological changes in English


Thirdly, posttonic unstressed syllables can also be weakened by processes deleting their initial consonants, as is exemplified in (21). (21)

Some instances of initial C deletion a. OE /k, g/-deletion (Nakao 1985: 406-408, Campbell 1959: § 477) i) hors.cllce > horslice readily', mus.cle > musle 'mussel' ii) hyn.gran > htnran 'hunger', mor.gen > mornes 'morn' b. OE /x/-deletion (Nakao 1985: 413, Campbell 1959: §§ 461, 464-465, Prins 1972: 190-191) i) *or.hBtan > Örettan 'fight', *seol.hes > sSoles 'seal's' ~ seolh, *hweo.hI> hwSol 'wheel', scyld.hreoda 'phalanx' ~ scyldreda, dur.here 'folding door' ~ dürere, *eohräed >6ored 'band' ii) be.hindan 'behind', be.healdan 'hold', ge.helpan 'help', to.heald 'learning', to.hräosan 'fall', to.hweorfan 'separate' iii) feoh.tan 'fight', s6h.te 'he sought' c. EModE /h/-deletion (Nakao 1985: 417, Dobson 1957: §426)






annihilate, co.hesion,

nihilism, pro.hibit,

Gm.ham compre.hend,

EModE /b/-deletion (Nakao 1985: 404, Dobson 1957: §403) assem.ble,






EModE /t, dJ-deletion (Nakao 1985: 405-406, Dobson 1957: §§401,405-406,410, Prins 1972: 226) i)

nes.tle, apos.tle, mois.ten,



whis.tle, mus.tle, Wes.ton, kin.dle,

thros.tle, lis.ten, of.ten, won.der,

c äs .tie, epis.tle, jus.ten,


wres.tle, has.ten,

sof.ten thun.der,


In (21b-c) /xJ and /h/ regularly dropped in the weakening environment of case (i), but when located within stressed syllables as in cases (ii) and (iii), they never deleted, serving as foils for their peak vowels. 7.2. Final weakening Final weakening in unstressed syllables, like initial weakening just discussed, helps make stressed syllables abutting on the left more perceptible. There seem to be only two types of final weakening which can


Norio Yamada

be attested in the history of English, namely obstruent voicing and final C deletion. A typical example of obstruent voicing can be seen in an EModE change which voiced an affricate /tf/ in the final position of posttonic unstressed syllables (Prins 1972: 225, Jordan - Crook 1974: §180, Dobson 1957: §363): (22)

EModE ch-voicing ME know.leche > EModE knowledge, ME pdr.triche > EModE partridge, ME spinach > EModE spinach [spinidj], Ν or.wich [norid3], Green.wich [grinidj], Wool.wich [\vulid3], Hdr.wich [haerid3], Brom.wich [brXmid3], Ips.wich [ipsid3] (now obsolete), os.tridge 'ostrich', cdb.bage, sdu.sage

Let us now turn to another type of final weakening, i.e., final C deletion. This can be illustrated by a change of degemination in Old English, as a result of which a geminate consonant in the final position of an unstressed syllable was simplified in the context of an immediately preceding stressed syllable (Campbell 1959: §457): (23)

OE degemination gyl.den.ne > gyldene, 6.per.ra > 6pera, ώ/.ter.ra > sftera, d1.gel.lic > df.gelic, E.der.ric > Ederic, ^E.der.red > JEÖered, ώ.met.tig > ämetig, gy.den.ne > gydene, blic.cet.timg > bliccetung, reccend.dom > reccendom

Notice that, of the two geminates contained in each of the last two examples of (23), only the one occurring in the second unstressed syllable is actually allowed to undergo this simplification, a fact which follows naturally from our assumption that weakening processes such as degemination may in general be allowed to operate only in unstressed syllables.

8. Historical conspiracies Let us finally consider a series of complex changes which is comprised of both processes of strengthening in stressed syllables and those of weakening in unstressed syllables. Such complex changes can be regarded as a historical "conspiracy" (Kisseberth 1970), the individual processes of

On the fiinctional motivation of phonological

changes in English


which collectively "conspire" to attain the same aim of increasing salience in stressed syllables. The development of Primitive Germanic (PG) /{e9} in OE and ME provides us with a good case in point. This {e9} was first strengthened in late OE by being despirantized into [g] at the begining of a stressed syllable as in (24) (Nakao 1985: 387, Jordan - Crook 1974: 173). (24)



OE despirantization of PG /γ/ gold 'gold', gdst 'breath', grass 'grass', grdfan 'dig', glakd 'glad' Despirantization in secondarily stressed syllables w6r(i).gen 'wear out', OE syne.jian > süne.ge, sin(e).ge, myne.ßan 'to mention' > min.ge

The original consonant of PG, however, remained a spirant in the usual weakening environment, namely in the initial position of an unstressed syllable immediately after a stressed syllable, as is illustrated by the following examples (Prins 1972: 199): (25)

The OE reflex of PG /y/ in the weakening environment da.jas ' d a y s \ f l i o . j a n 'to fly', stf.jan 'to rise'

Then, in the ME period from 1200 onward, a velar spirant which stood in the initial position of a posttonic of a posttonic unstressed syllable came to be weakened into a labial glide [w] (Jordan - Crook 1974: 172-173): (26)

ME weakening of medial /y/ drd.jen > ME drd.wen 'to draw', ma.we 'maw', ha.we 'hawthorn', sä.we 'saw', ό.wen 'to own', do.we 'dough', bo.we 'bow', plo.wes 'ploughs', gdl.je 'gallows' > gdl.we, (OE hd l.jan >) ME hdl.jes 'saint' > hal.wes, swel.wen, swol.wen 'to swallow', fol.wen 'to follow', mor.we(n) 'morrow', bor.wen 'borrow'

Now, at the end of a stressed syllable, the velar spirant in question already in OE had been strengthened by devoicing (Jordan - Crook 1974: 173; for further examples, see (17)): (27)

OE final /y/-devoicing jenjen6h 'enough',


> burh



Norio Yamada

Seen from a functional point of view, the above-mentioned changes involving PG /γ/ are all obviously alike in that they are in toto aiming at maximizing perceptual salience in stressed syllables.

9. Conclusion To recapitulate, the above discussion on phonological change can be reduced to three main points. The first is that strengthening changes such as those in (3) are for the most part in complementary distribution with weakening changes like those in (4); namely the former changes take place only in stressed syllables while the latter changes occur only in unstressed syllables, though under certain predictable circumstances some vocalic changes, such as monophthongization and vowel shortening, may also occur in stressed syllables. The second point is that stressed syllables tend to be strengthened whereas unstressed syllables tend to be weakened, especially in the posttonic environment, and that these two processes are functionally equivalent in that both serve to enhance the salience of stressed syllables. At the outset we have posed three questions concerning the causality of phonological change: the repetition riddle (la), the causation riddle (lb), and the opposition riddle (lc). The first two questions can now be answered by saying that most phonological changes are perceptually motivated in the sense stated above. The third point is that there are two antithetic requirements on the use of language, which are constantly playing at a tug of war with each other. Thus one is the perceptual principle that stressed syllables be made as salient as possible, and the other is the productive requirements that segments be so modified as to become as easy to pronounce as possible. When one of the two requirements happens to redouble its force, leading to an ill-balanced situation, some counteractive change that may conform to the other requirement is expected to occur so as to restore equilibrium. Thus we see that the opposition riddle (lc) can also be solved by assuming that under certain conditions some counteractive changes may occur to keep the balance between these two requirements.

On the functional motivation of phonological changes in English


Note *

This article is a slightly revised and condensed version of Yamada (1994). Portions of this material were also presented at the Eleventh National Conference of the English Linguistic Society of Japan (November 1993), the Kyoto Conference on English Historical Linguistics and Philology (November 1995), and the Ninth International Conference on English Historical Linguistics (August 1996). I would like to thank the following people for their comments and questions: Toshio Nakao, Jacek Fisiak, Akio Oizumi, John Scahill, Theo Vennemann, Akiko Ueda, Shuji Chiba, Reiko Shimamura, and Yuriko Otsuka. I am also grateful to William Plain for suggesting stylistic improvements. Any errors or misconceptions in the paper are, of course, my own.

References Araki, Kazuo - Masatomo Ukaji 1984 Eigo-shi III A (History of English III A). Tokyo: Taishukan. Campbell, Alistair 1959 Old English grammar. Oxford: Oxford University Press. Chiba, Shuji et al. (eds.) 1991 Gendai eigogaku no shosou (Aspects of current English linguistics). Tokyo: Kaitakusha. Dobson, Eric J. 1957 English pronunciation 1500-1700. Vol. II. Oxford: Oxford University Press. [1968] [Second edition, Oxford: Oxford University Press]. Eaton, Roger - Olga Fischer - Willem Koopman - Frederike van der Leek (eds.) 1985 Papers from the 4th International Conference on English Historical Linguistics. (Current Issues in Linguistic Theory 41.) Amsterdam: John Benjamins. Gimson, Eric J. 1989 An introduction to the pronunciation of English. Fourth edition, London: Edward Arnold. Hayes, Bruce P. 1989 "Compensatory lengthening in moraic phonology", Linguistic Inquiry 20: 253-306. Hock, Hans H. 1986 "Compensatory lengthening: In defence of the concept 'mora'". Folia Linguistica xx: 431-460. Hooper, Joan B. 1976 An introduction to natural generative phonology. New York: Academic Press.


Norio Yamada

Johnston, Paul Α., Jr. 1992 "English vowel shifting: One great vowel shift or two small vowel shifts?", Diachronica IX/2: 189-226. Jones, Charles 1989 A history of English phonology. London: Longman. Jordan, Richard - Eugene J. Crook 1974 Handbook of Middle English grammar: Phonology. The Hague: Mouton. Kastovsky, Dieter - Gero Bauer (eds.) 1988 Luick revisited. Tübingen: Gunter Narr. Kim, Suksan 1991 "Lexical phonology and the fricative voicing rule", in: Shuji Chiba et al. (eds.), 541-558. Kisseberth, Charles W. 1970

"On the functional unity of phonological rules", Linguistics Inquiry 1: 291306. Kurylowicz, Jerzy 1966 "A remark on the Great Vowel Shift", Word 21: 183:187. Labov, William 1972 "The internal evolution of linguistic rules", in: Robert P. Stockwell - Ronald K. S. Macaulay (eds.), 101-171. 1994 Principles of linguistic change. Oxford - Cambridge: Blackwell. Lass, Roger 1985

"Minkova noch einmal: MEOSL and the resolved foot", Folia Linguistica Historica VI/2: 245-265.

Lutz, Angelika 1992 "Lexical and morphological consequences of phonotactic change in the history of English", in: Matti Rissanen - Ossi Ihalainen - Terttu Nevalainen Irma Taavitsainen (eds.), 156-166. Minkova, Donka 1982 "The environment for open syllable lengthening in Middle English", Folia Linguistica Historica III/l: 29-58. 1985 "Of rhyme and reason: Some foot-governed quantity changes in English", in: Roger Eaton - Olga Fischer - Willem Koopman - Frederike van der Leek (eds.), 163-178. 1991

The history of final vowels in English: The sound of muting. Berlin: Mouton de Gruyter.

Myers, Scott 1987 "Vowel shortening in English", Natural Language and Linguistic Theory 5: 485- 518. 1991 "Persistent rues", Linguistic Inquiry 22: 315-344.

On the functional motivation of phonological changes in English


Nakao, Toshio 1985

On-inshi (English historical phonology). Tokyo: Taishukan.

Prins, Anton A. 1972

A history of English phonemes from Indo-European

to Present-Day


Leiden: Leiden University Press. Rissanen, Matti - Ossi Ihalainen - Terttu Nevalainen - Irma Taavitsainen (eds.) 1992


of Englishes:

New methods





linguistics. Berlin: Mouton de Gruyter. Ritt, Nikolaus 1992

"Middle English vowel quantity reconsidered", in Matti Rissanen - Ossi Ihalainen - Terttu Nevalainen - Irma Taavitsainen (eds.), 207-222.


Quantity adjustment: English.


Vowel lengthening Studies

and shortening

in Linguistics:

in Early




Cambridge: Cambridge University Press. Samuels, Michael L. 1972

Linguistic evolution with special reference to English. Cambridge: Cambridge University Press.

Stockwell, Robert P. - Ronald K. S. Macaulay (eds.) 1972

Linguistic change and generative theory. Bloomington: Indiana University Press.

Stockwell, Robert P. - Donka Minkova 1988

"The English vowel shift: Problems of coherence and explanation", in Dieter Kastovsky - Gero Bauer (eds.), 355-394.

Sweet, Henry 1988

A history of English sounds from the earliest period. Oxford: Clarendon Press.

Wells, John C. 1982

Accents of English. 3 vols. Cambridge: Cambridge University Press.

Yamada, Norio 1994

Phonological changes in English: A functional explanation in historical phonology. Unpublished Ph.D. dissertation, Tsuda College.

On the productivity of the suffixes -ness and -ity: The case of Chaucer* Hiroshi Yonekura

1. Introduction Elaborate and scholarly investigations on the productivity of the synonymous suffixes -ness and -ity1 have been conducted by many scholars. Matthews (1974), Aronoff (1976), Randall (1980), Reichl (1982), Kiparsky (1983), Romaine (1983), van Marie (1985) and Shimamura (1991) have discussed the suffixes in Present-day English. Riddle (1984) and Romaine (1985) dealt with the productivity of the two competing suffixes from a historical point of view. Most of the research works mentioned above deal with the productivity of the suffixes -ness and -ity mainly on the basis of Aronoff (1976). The aim of the present paper is to discuss the productivity of the suffixes -ness and -ity in Chaucer's English from the point of view of morphology and semantics and to refer to the question whether or not what is observed in the Present-day English suffixes -ness and -ity may be true of the suffixes in Chaucer's English.

2. The distribution of the suffixes -ness and -ity The suffix -ness bears some relation to Gothic -assus / -inassues (e.g., piudinassus 'kingdom'). In Old English2 it is employed to form abstract feminine nouns, along with other suffixes like -had, -dom, -scipe (e.g., druncenhad, druncennesse, druncenscipe). On the other hand, the suffix ity, which is semantically cognate with the suffix -ness, makes its appearance in 14th and 15th century loanwords from Old French and, later, in loanwords from Latin. In short, the suffix -ness is derived from Germanic and the suffix -ity from Romance. Before going on to the productivity of the two suffixes -ness and -ity, it is necessary first to say a few words about their distribution in Chaucer 3 .


Hiroshi Yonekura

Table 1.

-ness -ity Total



132 (64%)

1065 (65%)


74 (36%)

570 (35%)




: Token : 8.07 : 7.70


Table 1 shows that the type/token nouns in -ness (64%/65%) have a higher frequency than nouns in -ity (36%/35%). From the ratio of token to type, however, it is obvious that the nouns in -ness (65% : 64%)4 occur roughly as often as the nouns in -ity (35% : 36%). According to Bybee (1985: 133) "productivity of morphological rules must be connected to high type frequency," and thus we may conclude that the suffix -ness is much more productive than the suffix -ity. Here our attention is directed to the tokenfrequency of the two suffixes both in Middle English and in Modern English. In comparing the versions of De Consolatione Philosophiae translated respectively by King Alfred, Chaucer and Queen Elizabeth I, Suzanne Romaine (1985: 459) gives the token-frequency of the suffixes ness and -ity. Table 2. De Consolatione Philosophiae ALFRED



-ness 87

-ness 59 (53%)

-ness 38 (50%)





In the translation by Chaucer the ration of the -ness form to the -ity form is 53% to 47%. In her translation Elizabeth uses -ness and -ity formation in equal proportions (50% of each). Although Chaucer uses the -ness words more often than the -ity words, the distribution of the -ity forms both in Chaucer and in Elizabeth shows that the nouns in -ity occur more frequently in the translation by Elizabeth.

3. An abstract morphological feature The notion that an abstract morphological feature [+latinate] contained in the base has a close relation with the productivity of the suffix has been suggested by Aronoff (1976: 51). According to him, the suffix -ness attaches freely to the bases containing either [+latinate] or [-latinate],

On the productivity of the suffixes -ness and -ity


while the suffix -ity is restricted only to the bases containing the feature [+latinate]. This morphological restriction exerts a great influence upon the difference in productivity between the two morphemes -ness and -ity. Table 3. -ness



98 (74%)


32 (24%)

1 d%) 73 (99%)

2 (2%)

0 (0%)

Obscure Origin Total



Another point that should be mentioned is that the bases containing [-latinate] predominate over those containing [+latinate] with the morpheme -ness.5 In Chaucer there is only one example in which the morpheme -ity attaches to the base containing the feature [-latinate] as in scantitee.6 but certes the superfluitee or disordinat scantitee of clothynge is reprevable. I.PS 431 Two words of which the ηbase is of obscure origin are used in Chaucer: tikelnesse and wrawnesse. Flee fro the prees and dwelle with sothfastnesse; Suffyce unto thy thing, though it be smal, For hord hath hate, and climbing tikelnesse, Truth 1-3 He dooth alle thyng with anoy, and with wrawnesse, slaknesse, and excusacioun, and with ydelnesse, and unlust; I.PS 680

4. A morphological restriction on the base type to which the suffix attaches This morphological restriction exerts a stronger influence upon the productivity of the two suffixes -ness and -ity than the abstract morpho-


Hiroshi Yonekura


logical feature [± latinate]. Aronoff (1976: 36) points out that the suffix -ness attaches more freely to various kinds of bases than the suffix -ity. The following list of words shows the occurrence of the morphemes -ness and -ity with different base types in Chaucer: -ness Xed Xfiil Xlees Xly Xous Xsom Xwis Xy

wikkednesse, wrechednesse blisfulnesse, welefulnesse, wilfiilnesse reccheleesneese liklynesse, semelynesse, unliklynesse graciousnesse, likerousnesse, preciousnesse fulsomnesse, hoolsomnesse rightwisnesse foolhardynesse, hardynesse, holynesse, gredynesse, lustynesse, unworthynesse, worthynesse

Xable Xal

immoevablete, notabilitee, perdurablete bestialite, comunalite, egalitee, sensualitee, universalite subtilitee contrarioustee, curiositee, fiimositee, mendicite


Xile Xous

From the list given above it is evident that the suffix -ness combines more freely with different kinds of bases than the suffix -ity. In particular, the morpheme -ity occurs with certain limited morphological classes of adjectives. This morphological restriction does not mean, however, that the suffix -ness occurs more productively with any kind of bases than the suffix -ity. In Present-day English, for example, bases ending -al / -able attach productively to the morpheme -ity.9 This is the case with the adjectives which themselves end in -al / -able. Xal / Xable10 bestial communal (MED) egal inmoevable notable perdurable sensual (MED) universal

-ity bestialite comunalite egalitee immoevablete notabilitee perdurablete sensualitee universalite


* * * *

* * * *

On the productivity of the suffixes -ness and -ity


In Chaucer we cannot find any example in which adjectives ending in -al / -able are connected with the morpheme -ness. This phenomenon is called 'the potentiation of affix* by affixy (in this case affixy must be placed before affix*) (Williams 1981: 249-250). In Chaucer -al or -able (affixy) potentiates -ity (affix*).11

5. Blocking Aronoff (1976: 43) points out that "there is a more direct connection between lexical listing and productivity." This connection is related to a phenomenon which is called 'blocking'.12 'Blocking' describes the phenomenon by which one form is precluded by the simple existence of another. The morpheme -ous can derive a new word from the combination with either -ness or -ity. However, if there exists in a given stem both an adjective with the suffix -ous and a semantically related abstract noun, then it is not possible to form the -ity derivative of the -ous adjective. Nominal -ity Xous * contrarious contrarioustee * curious curiosite * fumositee famous (MED) * grace gracious * lecherie likerous * preciouse preciosite (MED) prosperous (MED) prosperitee

-ness contrariousnesse (MED) 13 curiousnesse (MED) * graciousnesse likerousnesse preciousnesse Prosperousnesse (MED)

In Chaucer the already existing nouns grace and lecherie respectively block the new -ity derivative, while all the -ness derivatives of -ous adjectives are never blocked. We may, therefore, reasonably conclude that blocking plays a significant role in the choice between -ness and -ity in Chaucer.

6. Color words The claim that the morphemes -ness and -ity display a tendency to semantic differentiation may be made for color words, which only take the


Hiroshi Yonekura

morpheme -ness (Riddle 1985: 441). All the instances of color words used in Chaucer, though quite few, take only the -ness form. Of metals, whiche ye han herd me reherce, Consumed and wasted han my reednesse. G.CY 1099-100 Similarly G.CY 1097; Bo 1.p. 1.76 Or see your colour lyk the sonne bryght That of yelownesse hadde never pere. Purse 10-11 All the examples given above are of Old English origin.14

7. Count nouns The -ity words more frequently occur as count nouns than the -ness words. In short, the fact that the -ity nouns are regularly pluralized in order to express concrete meaning shows that the morpheme -ity is more productive than the morpheme -ness (Riddle 1985: 442). -ness (12 occurrences) besynesses (Tr 2.1174), bitternesses (Bo 2.p.4.119), blissefulnesses (Bo 4.p.l.54), dirknesses (Bo l.m.2.4.), goodnesses (I.PS 489), hardnesses (Bo 4.p.5.35), roundnesses (Bo 4.m.6.49), schrewednesses (Bo 4.p.2.218), seknesses (Bo 3.p.7.4.), unselynesses (Bo 4.p.4.32), wikkednesses (I.PS 275), woodnesses (Bo 2.m.4.18) -ity (19 occurrences) adversitees (B.ME 1566), auctoritees (D.FR 1276), benignitees (E.CL 827), comunalites (Bo l.p.4.28), dignytees (Bo 2.p.2.9), extremytees (Rom C 6526), familiarites (Bo 3.p.5.1), iniquitees (I.PS 442), mendicitees (Rom C 6525), nativites (Astr Π.4.1), necessites (Bo 3.p.9.110), propretes (Bo 3.m.l 1.17), prosperites (Bo 4.p.6.263), qualites (Bo 2.m.8.3), sensibilities (Bo 5.m.4.7), singularites (Bo 5.m.3.41), subtilitees (E.MC 2421), superfluytees (Bo 2.p.5.81), vanytees (B.NP 3091)

On the productivity of the suffixes -ness and -ity


The instances listed above show that the words ending in -ity takes precedence over the -ness words and that Boece most frequently uses the plural form both in -ness than in -ity nouns. Here pluralization may imply concretization. yif that he ne constreynede hem nat eftsones into roundnesses enclyned,... thei scholden departen from hir welle Bo 4.m.6.49 (roundnesses 'orbits' Cf. Latin: flexos ... in orbes) Us nedeth nat to speken but of game, And lete auctoritees, on Goddes name, D.FR 1275-6 (auctoritees 'authoritative texts or books')

8. Transparent and opaque Aronoff (1976: 39) states that "there is a direct link between semantic coherence and productivity." To put it differently, semantic transparency is a primary factor in the selection between the two suffixes -ness and -ity. The suffix -ness tends to denote an embodied attribute or trait, whereas the suffix -ity tends to denote an abstract or concrete entity. In the majority of cases, however, both -ness and -ity express the quality or state of the adjective to which the suffixes attach. For example, variousness as well as variety indicates the quality or state of being various, but variety sometimes refers to kind or sort as in the instance below: How many varieties of fish are there in the pond? (Aronoff 1976: 38) That is to say, the -ness word is semantically transparent, while the -ity word is sometimes opaque in meaning. In Aronoff s words, "the surer one is of what a word will mean, the more likely one is to use it" (1976: 39). In this sense, the morpheme -ness, which creates semantically transparent words, is more productive than the morpheme -ity, which forms semantically opaque words. The following examples show that both the -ness noun and the -ity noun are transparent:


Hiroshi Yonekura

Al were he fill of treson and falsnesse, F.SQ 506 (falsnesse 'deceitfulness') Somme seyde honour, somme seyde jolynesse, D.WB 926 (jolynesse 'pleasure') Thanne is ther constaunce, that is stablenesse of corage, I.PS 737 (stablenesse 'steadfastness') There are some instances where the -ity noun is opaque in meaning. And certes, sire, thogh noon auctoritee Were in no book, ye gentils of honour Seyn that men sholde an oold wight doon favour D.WB 1208-10 (auctoritee 'opinion') The ascendent sothly, as wel in alle nativites as in questions and eleccions of tymes, is a thing which that these astrologiens gretly observen. Astr 2.4.1 (nativites 'horoscope') Lo, whiche sleightes and subtilitees In wommen been! E.MC 2421-2 (subtilitees 'tricks') As is seen from these examples given above, all the -ness nouns are semantically coherent with the adjectives to which the suffix attaches, while some of the -ity nouns are semantically opaque. There are, however, some cases in which the semantics of the -ness derivatives is not nearly so coherent.15 And this thing was nat kept for holynesse, But al for verray vertu and clennesse, And for men schulde sette on hem no lak; LGW G 296-8 (holynesse 'religion') As he hadde seyn it chaunge bothe up and doun, Joye after wo, and wo after gladnesse, And shewed hem ensamples and liknesse. A.KN 2840-2 (liknesse 'parable') Levere in a forest that is rude and coold Goon ete wormes and swich wrecchednesse. H.MP 170-1 (wrecchednesse 'miserable food')

On the productivity of the suffixes -ness and -ity


In Chaucer, therefore, semantic incoherence is observed in the -ness word as well as in the -ity word. It may be interesting to note here that the transparency criterion does not always predict a preference for -ness rather than -ity}6

9. Final remarks The present paper has been primarily concerned with the productivity of the two competing suffixes. Finally, however, we would like to mention briefly some further problems of word formation in Chaucer. Fisiak (1965) has made an attempt at a detailed analysis of the morphemic structure of Chaucer's English. In his book he deals with several issues, including the productivity of all the affixes used in Chaucer. His work, however, is far from exhaustive. There is still no comprehensive monograph on the principles and practice of Chaucer's word formation.17 Oizumi's Chaucer Concordance (in ten volumes) contains a reverse-word list attached to each work of Chaucer, which should facilitate an advanced investigation of Chaucer's word formation. For an elaborate and scholarly examination of word formation in Middle English a computerized reverse Middle English Dictionary seems a must and should be made available to interested scholars as soon as possible.

Notes *

This is a revised version of Yonekura 1993. It was presented at the Kyoto Conference on English Historical Linguistics and Philology held on 3-4 November 1995. It is a great honor for me to express my thanks to Professor Jacek Fisiak at Adam Mickiewicz University, from whom I have received constant guidance. My sincere thanks are also due to Professor Toshio Nakao at Tsuda College for his useful suggestion some points of derivational morphology in Middle English and to Professor Ryuta Murakami at Seinan Gakuin University, who chaired my paper at the Kyoto Conference on English Historical Linguistics and Philology.

1. Koziol (1937: 203-4, 248-9), Jespersen (1942: 449), Marchand (1969: 227-8), Adams (1973: 119) and others remark that the suffixes -ness and -ity are synonymous. Pointing out the difference in meaning between -ness and -ity, however, Riddle (1985: 437) states as follows: "-ness tends to denote an embodied attribute or trait, while -ity tends to denote an abstract or concrete entity."


Hiroshi Yonekura

2. Kastovsky (198S: 244-6) gives a thorough morphological and semantic analysis of deverbal -ness formations in Old English. 3. Figures given in each Table are based on my examination of the Chaucer Concordance (Oizumi 1991). The question whether or not the three fragments of The Romaunt of the Rose may be attributed to Chaucer is not relevant here. 4. Masui (1964: 13) says that "the words ending in -nesse stand more frequently in rime than out of rime." 5. Romaine (1985: 462) states that "almost as soon as French words (and later Latin) were introduced into English, native prefixes and suffixes were added to them." Wyclif, for example, uses feerste and fersnesse for the same Lain word ferocitatem: Judith 3.11 Thei my3ten not swagen th & feerste [Latin ferocitatem] of his brest. Judith 3.11 Thei doynge these thingis myjten not swage the fersnesse [Latin ferocitatem] of his herte. 6. The adjective scant of this noun is derived from Old Norse (MED s.v. scant adj.). To the quotation in which scantitee occurs instead of scantnesse Donner (1978: 3) gives an explanation that "Chaucer tends to choose parallel morphologic forms." I sey nat that honestitee in clothynge of man or woman is uncovenable, but certes the superfluitee or disordinat scantitee of clothynge is reprevable. I.PS 431 The word scantitee is used in parallel with honestitee and superfluitee. Donner (1978: 3) also points out that "scantity is a strange word to find in Chaucer's vocabulary." 7. All the quotations are taken from The Riverside Chaucer. For the abbreviations used in this paper, see Oizumi (1991). 8. Aronoff (1980: 75) and Scalise (1984: 50). 9. This is pointed out in Marchand (1969: 238) and Williams (1981: 249-50). 10. The two words communal and sensual occur in the Middle English Dictionary but not in Chaucer. The asterisk (*) shows that the form in question is used neither in Chaucer nor in the Middle English Dictionary. 11. Romaine (1983: 182) gives formations in -ness and -ity in relation to base type which are observed in Present-day English.


-ness generousness

-ive -able

transitiveness reasonableness


-ity generosity


391 346

transitivity reasonability

96 200

On the productivity of the suffixes -ness and -ity

-al -ible

musicalness fallibleness

169 99

musicality fallibility

315 156

-ile -ic

fragileness domesticness

17 15

fragility domesticity

75 63


As is obvious from these figures in the Table, nouns ending -ness are most productively connected with bases ending in -ous, -ive, and -able, while the suffix -ity occurs most frequently with the morphemes -al, ible, ile, -ic. In Chaucer, however, the number of such examples is so small that we cannot draw any conclusions in the present paper. 12. This phenomenon is equivalent to 'pre-emption' mentioned in Clark and Clark (1979: 798). 13. (MED) indicates that the word does not occur in Chaucer. The asterisk (*) denotes the non-occurrence of the potential word in Chaucer. 14. According to Riddle (1985: 441), in Present-day English the vast majority of color words are of Romance origin. 15. As Professor Sadao Ando rightly commented on my earlier draft, all these examples are cited from literary sources so that rhetorical factors should also be taken into consideration in dealing with semantic coherence. 16. Attention should be paid to the question whether or not rhyming may have a strong influence upon the productivity of the two nominal suffixes in Chaucer. When we treat Chaucer's word formation it is natural that the problem arises of whether or not Chaucer's word formation may be related to rhyming positions. In this sense Oizumi and Yonekura's Rhyme Concordance to the Poetical Works of Geoffrey Chaucer will constitute a valuable contribution to the study of this question. 17. See Görlach (1978: 93), Sauer (1988), and Burnley (1992: 439-49).

References Adams, Valerie 1973

An introduction to modern English word-formation. (English Language Series 7.) London: Longman. Aronoff, Mark 1976 Word formation in generative grammar. (Linguistic Inquiry Monograph 1.) Cambridge, etc.: ΜΓΓ Press. 1980 "The relevance of productivity in a synchronic description of word formation", in: Jacek Fisiak (ed.), 71-82.


Hiroshi Yonekura

Benson, Larry Dean (ed.) 1987 The Riverside Chaucer, (3rd edition.) Based on The Works of Geoffrey Chaucer edited by Fred Noris Robinson. Boston: Houghton Mifflin Company. 1993

A glossarial concordance to the Riverside Chaucer. 2 vols. New York & London: Garland Publishing. Blake, Norman (ed.) 1992

The Cambridge history of the English language. Volume 2: 1066-1476. Cambridge: Cambridge University Press. Burnley, David 1992 "Lexis and semantics", in: Norman Blake (ed.), 409-499. Bybee, Joan L. 1985 Morphology: A study of the relation between meaning and form. Amsterdam: John Benjamins. Clark, Eve V. - Herbert H. Clark. 1979 "When nouns surface as verbs", Language 55: 767-811. Dalton-Puffer, Christiane 1992

"The status of word formation in Middle English: Approaching the question," in: Matti Rissanen (ed.), 465-482. Davis, Norman - Douglas Gray - Patricia Ingham - Anne Wallace-Hadrill (eds.) 1979 A Chaucer glossary. Oxford: Clarendon Press. Donner, Morton 1978 "Derived words in Chaucer's language," The Chaucer Review 13:1-15. Fisiak, Jacek 1965

Morphemic structure of Chaucer's English. (Alabama Linguistic and Philological Series 10.) Alabama: University of Alabama Press. Fisiak, Jacek (ed.) 1980 Historical morphology. (Trends in Linguistics. Studies and Monographs 17.) The Hague: Mouton. 1985a 1985b

Historical semantics: Historical word-formation. Berlin: Walter de Gruyter. Papers from the 6th International Conference on Historical Linguistics. Amsterdam: John Benjamins. Görlach, Manfred 1978

"Chaucer's English: what remains to be done," Arbeiten zu Anglistik, und Amerikanistik 4: 61-79. 1991 Introduction to early Modern English. Cambridge, etc.: Cambridge University Press. Jespersen, Otto 1942

A modern English grammar on historical principles. Part VI: Morphology. London: George Allen & Unwin.

On the productivity of the suffixes -ness and -ity


Kastovsky, Dieter 1982 Wortbildung und Semantik. (Studienreihe Englisch 14.) Bern und München: Francke Verlag. 1985

"Deverbal nouns in Old and Modern English: From stem-formation to wordformation," in: Jacek Fisiak ed. (1985a). 221-261.


"The interaction of semantic and formal structures in the lexicon," in: Jerzy Tomaszczyk and Barbara Lewandowska-Tomaszczyk (eds.), 75-91. Kiparsky, Paul 1983 "Word-formation and the lexicon", Proceedings of the 1982 Mid-America Linguistics Conference, (Lawrence, Kans.: Department of Linguistics, University of Kansas), 3-29. Koziol, Herbert 1937 Handbuch der englischen Wortbildungslehre. Zweite, neubearbeitete Auflage. (Germanische Bibliothek. Erst Reihe: Sprachwissenschaftliche Lehr- und Elementarbücher) Heidelberg: Carl Winter Universitätsverlag. Kurath, Hans - Sherman McAlister Kuhn - Robert E. Lewis (eds.) 1952Middle English dictionary. Ann Arbor: University of Michigan Press. [MED] Lindberg, Conrad (ed.) 1965 MS. Bodley 959: Genesis-Baruch 3.20 in the earlier version of the Wycliffite Bible. 5 vols. Stockholm: Almqvist & Wiksell. Marchand, Hans 1969

The categories and types of Present-day English word-formation: A synchronic-diachronic approach. (Second, completely revised and enlarged edition.) München: C.H. Beck'sehe Verlagsbuchhandlung. Markus, Manfred (ed.) 1988 Historical English. Innsbruck: Universität Innsbruck. Masui, Michio 1964 The structure of Chaucer's rime words: An exploration into the poetic language of Chaucer. Tokyo: Kenkyusha. Matthews, Peter H. 1974 Morphology: an introduction to the theory of word-structure (Cambridge Textbooks in Linguistics), Cambridge, etc.: Cambridge University Press. Oizumi, Akio (ed.) 1991


A complete concordance to the works of Geoffrey Chaucer. Programmed by Kunihiro Miki. 10 vols. Hildesheim, etc.: Olms-Weidmann. Oizumi, Akio - Hiroshi Yonekura (eds.) 1994

A rhyme concordance to the poetical works of Geoffrey Chaucer. Programmed by Kunihiro Miki. 2 vols. Hildesheim, etc.: Olms-Weidmann.


Hiroshi Yonekura

Randall, Janet H. 1980 "-ity: A study in word formation restrictions," Journal of Ρsycholinguistics Research 9: 523-534. Reichl, K. 1982

Categorial grammar and word-formation: The de-adjectival abstract noun in English. Tübingen: Max Niemeyer. Riddle, Elizabeth M. 1985

"A historical perspective on the productivity of the suffixes -ness and -ity" in: Jacek Fisiak ed. (1985a), 435-461.

Rissanen, Matti et al. (eds.) 1992 History of Englishes: New methods and interpretations linguistics. Berlin: Mouton de Gruyter.



Romaine, Suzanne 1983 "On the productivity of word formation rules and limits of variability in the lexicon," Australian Journal of Linguistics 3: 177-200. 1985 "Variability in word formation patterns and productivity in the history of English," in: Jacek Fisiak ed. (1985b), 451-465. Saito, Toshio - Mitsunori Imai (eds.) 1988 A concordance to Middle English romances. 2 vols., Frankfurt am Main, Bern, New York & Paris: Verlag Peter Lang. Sauer, Hans 1988 "Compounds and compounding in Early Middle English: problems, patterns, productivity," in: Manfred Markus (ed.), 186-209. Scalise, Sergio 1984 Generative morphology. Dordrecht: Foris. Shimamura, Reiko 1991 Eigo no gokeisei to sono seisansei. (English word formation and productivity). Toronto: Liber Press. Simpson, John A. - Edmund S.C. Weiner (eds.) 1989 The Oxford English dictionary. (2nd edition) Oxford: Clarendon Press. [OED] Tomaszczyk, Jerzy - Barbara Lewandowska-Tomaszczyk (eds.) 1990 Meaning and lexicography. Amsterdam: John Benjamins. van Marie, J. 1985 On the paradigmatic dimension of morphological creativity. Dordrecht: Foris. Williams, Edwin 1981 "On the notions 'lexically related' and 'head of a word'," Linguistic Inquiry 12: 245-274.

On the productivity of the suffixes -ness and -ity


Yonekura, Hiroshi 1993

"On the productivity of the suffixes -ness and -ity: from Chaucer to Shakespeare," Studies in Modern English 10: 1-25.

Index of names Abbott, Ε. Α. 330, 337 Ackerman, R. W. 5, 181,183 Adams, V. 32, 39,447,449 iElfric 21, 22,40,41, 56, 173,174,177, 178, 180-183,185, 187,188, 215224, 226, 228-304, 306, 307, 309, 341-346, 348, 349, 351-353, 355358,412 jEthelwold 205, 304, 306, 307, 357, 358 Alford, J. A. 411,414 Alfred, King 189,195,198, 199, 203207, 299, 304, 306, 327,407,440 Allen, C. L. 313, 314, 320 Anderson, G. K. 183,201,203 Anderson, J. M. 62,76, 87,148, 150, 209,213 Ando, Sadao 6, 8, 17,449 Ando, Shinsuke 16, 291 Andrew, S. O. 192,203 Apel.W. 183 Araki, K. 3 , 4 , 6 , 7 , 8, 9,17,423,435 Arndt, W. W. 414,416 Aronoff, M. 439,440,442,443,445, 448,449 Ascham, R. 168,169 Ash, J. 403 Asher, R. E. 261, 266 Attenborough, F. L. 189 Awaka, K. 13,17 Ayto, J. 405 Baba, A. 183,186, 397, 399 Bach, E. 76 Bäck, Η. 412,414 Bailey, C. J. N. 237, 261 Ball, C. 235 Bammesberger, A. 307, 310, 390, 398 Barnhart, R. K. 405,406,412 Barrack, C. M. 238, 258, 261 Barry, Μ. V. 264

Bately, J. M. 130, 192, 195, 203, 306, 307 Bauer, G. 307, 310,436,437 Baugh, A. C. 44, 51, 52, 53,195, 203 Bede 189,190,192,194-197,204,205,207, 301, 303, 304, 342, 344, 345, 347-349 Benskin, M. 88 Benson, L. D. 14, 52,53, 73,76, 108, 109, 293, 338,411,413,450 Bergien, A. 183 Berndt, R. 404,411,414 Bessinger, J. B., Jr. 36, 39 Bethurum, D. 173,178,182, 201, 204, 215, 225,227,412,413 Bischoff, B. 180,184 Bishop, Τ. A. M. 181,184 Blake, N. F. 7,12,15,16,49,53,91,109,194, 204,450 Bliss, A. J. 395, 398 Block, K. S. 411,413 Bloomfield, L. 252, 261 Board, C. 261,266 Boethius 9,189,199,204-206, 388, 392,407 B0gholm, N. 336, 338 Boitani, P. 227 Bonner, G. 195, 204 Bosworth, J. 114,125,126,130,133, 134, 136, 218, 226, 227, 387, 388, 395, 398, 399 Brenner, E. 305,307 Brewer, D. 227 Britton, D. 264 Brogyanyi, B. 261, 264 Brown, K. 235, 236 Brown, T.J. 180,184 Brown, W. H. 198,199, 204 Brunner, Κ. 25, 39, 337, 338 Buchloh, P. G. 307, 310 Bülbring, Κ. D. 412,414 Burnley, D. 91,93,108,448,449 Burnley, J. D. 12, 15,16 Buschinger, D. 16, 293, 294


Index of names

Bybee, J. L. 439,449 Byrnes, Ε. T. 337, 338 Caldwell, J. 180, 184 Callaway, M., Jr. 317, 321 Cameron, A. 184, 235, 398 Campbell, A. 23, 37, 38, 227, 305, 373, 395, 398,421, 427, 428, 430-432, 435 Capgrave, J. 186 Carlton, C. 25, 26, 29, 39, 201, 204 Carr, C. T. 398 Carroll, J. B. 254,261 Casaubon, M. 403 Cavalli-Sforza, L. L. 237, 261, 262, 264 Chambers, R. W. 412,413 Chambers, J. K. 236, 262 Chaucer, G. 3, 5, 11,12,14-17,43,46, 50, 52-55, 58, 71, 73, 76, 91,93, 101, 105-110, 169, 171, 176, 187, 188,206, 227, 234, 287, 288, 290-295, 323-339, 401,402, 409, 411,413, 439-444, 447-449,450,451,453 Chen, Μ. Y. 262 Chiba, S. 15, 17, 236, 384, 385,435,436 Chomsky, N. 259, 262, 320, 321 Clark, C. 192, 193,204 Clark, Ε. V. 449, 450 Clark, Η. H. 449,450 Clark, J. W. 79, 87 Cleasby, R. 235, 236 Clemoes, P. 180, 182, 184, 186, 204, 306, 307, 357, 358 Coates, J. 94, 95, 97,102,108 Cole, P. 76 Cook, A. S. 41, 305, 306, 307, 387, 397 Cooper, H. 13 Cosijn, P. J. 412,414 Costello, R. B. 412 Craddock, J. 257 Craigie, W. 235, 236

Cripin, A. 8,16, 293 Cromie, H. 287, 292, 293 Crook, E. J. 428,432,433, 436 Cröinin, D. 0 . 184 Curme, G. O. 53, 323, 335 Cynewulf 193,194, 397 Dahood, R. 181,183 Dalton-Puffer, C. 449 Danforth, D. 263 Darby, H. C. 249-252, 262 Date, T. 109 Daunt, M. 412, 413 Davenport, P. E. 395,410 Davis, N. 85-89,167, 169,184, 320, 337, 338, 450 Davis, R. H. C. 192, 204 Dawson, J. L. 54 Day, M. 182 Denison, D. 12, 168,169 Derolez, R. 204 D'Evelyn, C. 86, 87 Dickins, B. 307 Diensberg, B. 10,13, 412 Dobbie, E. van Kirk 40, 387, 388, 393, 395, 397 Dobson, E. J. 176,182,184,188, 258-260, 262,412, 415, 422-425,431,432, 435 Donaghey, B. S. 199, 204 Donaldson, D. 88 Donaldson, Ε. T. 413 Donaldson, W. D., Jr. 74, 76,411 Donner, Μ. 166,169,448,450 Donoghue, D. 11, 12, 313, 314, 321 Draat, P. F. van 195, 197, 204 Dressler, W. 384, 385 Dryden, J. 7, 9, 187, 339 Duckett, E. S. 192,195,196,198, 204 Duncan, Ε. H. 190, 196, 204 Dunn, C. W. 337, 338, 411

Index of names

Eaton, R. 435, 436 Eccles, M. 411,413 Eckert, P. 257 Ekwall, E. 251,262 Eliason, Ν. E. 182, 292, 293 Elizabeth, I. 440 Elliott, R. W. V. 12,109 Ellis, A. J. 258,262 Emonds, J. 167, 169, 324, 338 Enkvist, Ν. E. 109, 190, 204 Fehr, B. 178, 179,182, 215, 219, 225, 227,412 Feldman, Μ. W. 261 Fillmore, C. J. 57,58, 62,74, 75, 76 Fischer, 0 . 4 3 5 , 4 3 6 Fisiak, J. 19, 108,179, 184,185, 262, 308, 435,447,449, 450-452 Flexner, S. B. 412 Flom, G. T. 412, 415 Foley, J. 373, 384 Forshall, J. 235 Foster, R. 194, 205 Fowler, R. 192, 205, 225-227 Frank, R. 184 Franzen, C. 180,184 Friedrichsen, G. W. S. 404 Fries, C. C. 26,40, 321 Fujii.T. 185 Fujiwara, H. 6,10,12,17 Fujiwara, Y. 6, 17,21,27,40 Fukumura, T. 4, 17 Fukushima, O. 14,17 Fulk, R. D. 12, 82, 86, 87 Funaki, M. 57, 65, 74, 76, 77 Funke, Ο. 226, 227 Gaaf, W. van der 328, 337, 338 Gabriel, Κ. R. 262 Ganz, D. 184


Gerritsen, M. 185, 262 Giegerich, H. J. 384 Gimson, E. J. 422, 435 Gneuss, H. 185, 186,188, 192, 205, 301, 303, 306-308, 341,342,356-358 Godden, M. 114, 179 Goldsmith, J. 384 Gordon, Ε. V. 85-87, 89, 395, 397,411,413 Gordon, I. A. 205 Görlach, Μ. 449,450 Gorrell, J. H. 307 Götze, A. 411 Gower, J. 52, 54, 88 Gradon, P. Ο. E. 185,393, 397 Gräf, G. A. C. 43,45, 53 Graham, T. C. 179 Gray, D. 185,188, 449 Greenbaum, S. 170 Greenfield, S. B. 193, 205 Grinda, K. R. 305, 308,410,415 Gruber, L. C. 185, 186 Grünberg, Μ. 412,413 Hägerstrand, Τ. 237, 262 Hall, J. R. 137,150, 387, 388, 398 Halle, M. 259, 262 Halliday, W. J. 265 Hamel, M. 87 Hamilton, M. P. 411 Handley, R. 185 Harlow, C. G. 185 Harms, R. T. 76 Harris, D. P. 313, 321 Harsley, F. 305, 308 Harting, P. N. U. 196, 205 Hartmann, R. R. K. 52, 53 Härtung, A. E. 137,150,411,412,416 Hasegawa, K. 183, 397 Hashimoto, M. J. 238, 260, 262 Hayes, B. P. 384,420,435


Index of names

Healey, A. diP. 179, 305, 308, 311, 357, 358, 387, 388, 394, 395, 398,407,415 Hecht, Η. 189, 306 Hector, L. C. 174,185 Herbert, J. A. 182 Herrtage, S. J. 87 Herzog, Μ. I. 266 Heyne, Μ. 136 Hickes, G. 403 Hickey, R. 19, 262, 264 Hicks, C. 185, 188 Higashida, C. 17 Higuchi, M. 43,49, 53 Hill, J. 185 Hinckley, Η. B. 48, 53 Hirooka, H. 8, 17 Hirose, T. 10,17,213 Hoad, T. F. 404 Hock, Η. H. 420,435 Hoenigswald, Η. M. 263, 266 Hofstetter, W. 301-303, 305, 306, 308, 357,358 Hogg, R. M. 358, 398 Hollowell, I. M. 223, 224, 226, 227 Holt, R. 235 Holthausen, F. 138, 139, 150,404,407, 412,415 Hooper, J. B. 426,435 Hopper, H. 236 Horstmann, C. 412,413 Hotelling, H. 263 Howard-Hill, Τ. H. 292, 293 Hsieh, H-I. 263 Huddleston, R. 57, 62, 63, 75, 76,112, 113 Hughes, K. 204, 358 Hulst, H. van der 384 Hupp6, B. F. 228 Hyman, L. M. 384

Ichikawa, S. 2-4, 8, 17 Ihalainen, O. 436,437 Ikegami, Μ. T. 8,18 Ikegami, Y. 3,18,179 Imai.M. 11, 18,54,451 Imai, M. 11, 18,54,452 Inazumi, K. 9,18,168,169 Ingham, P. 449 Irie, K. 7 , 9 , 1 8 Irvine, S. 180,182 Ishibashi, K. 18 Ito, Ε. T. 5, 18,55 Ito, H. 7,15,18 Iwasaki, H. 5 , 7 , 1 2 , 1 5 , 1 8 Iyeiri, Y. 79, 86-88, 148 Jackendoff, R. 167,169 Jacobson, S. 25, 29,40, 320, 321 Jakobson, R. 288, 293 Jansen, F. 262 Jespersen, Ο. 53, 75, 76, 97, 109, 259, 263, 331,336-338,447,450 Jimura, A. 16, 18,91, 109 Jin, K. 13, 18, 111 Johnston, P. Α., Jr. 421,436 Jones, C. 137,138, 142, 144,145,147, 150, 165,166,169,420,436 Jordan, R. 259, 263, 428,432,433,436 Jost, Κ. 185, 222, 223, 225, 227, 306, 308, 357 Jucker, Α. Η. 180,185 Junius, F. 403 Kachru, Β. Β. 169,338 Kahane, Η. 338 Kahane, R. 338 Kane, G. 411,413 Karakida, S. 137,139,144,149, 151 Kastovsky, D. 185,186, 308, 309, 390, 398, 436,437,448,451 Katamba, F. 384

Index of names

Kato, Κ. 166, 169 Kato, Τ. 5, 7,18 Kawai, Μ. 13,18,102,107,109 Kemenade, A. van 320,321 Kendal, C. B. 395, 398 Kennedy, A. G. 11,14,54 Kenyon, J. S. 44-46,49, 53 Ker, N. R. 149,151,177, 180, 181. 183, 186, 306, 308 Kerkhof, J. 45,52, 53, 324 Kiefer, F. 76 Kihara, Κ. 17,18 Kim, S. 430,436 Kimmens, A. C. 302, 305, 308 Kinard, J. P. 225, 227 Kiparsky, P. 439,451 Kirk, J. M. 263, 264 Kisseberth, C. 432,436 Kittredge, G. L. 337,338 Kjellmer, G. 79, 88 Klaeber, F. 29,33,36,40,119,120,128, 134-136, 389, 396, 397 Klein, E. 404 Kluge, F. 302,404,411,415 Kobayashi, E. 3,5,18 Kohonen, V. 180,186 Koichi.J. 13, 18, 111 Kökeritz, Η. 258, 259, 263 Kolb, Ε. 257, 263 Koma, Ο. 153, 164,167, 169 Koopman, W. F. 235,435,436 Korhammer, M. 303, 308, 310 Koziot, H. 24, 40, 307, 447, 451 Krapp, G. P. 205, 388, 393, 397 Krishnamurti, B. 263 Kubouchi, T. 8, 11, 18, 171, 174, 176, 180, 181, 186, 308, 309, 164, 167-169, 173 Kuhn, H. 27,40 Kuhn, S. M. 151, 195, 205, 236, 305,


308, 338,451 Kurath, H. 151, 236, 338,415,450 Kurylowicz, J. 424,425,436 Labov, W. 237, 256, 257, 263, 264, 266, 420-422,436 Laing, M. 179 Lakoff, R. 92,96,97,109 Lakshmi, Β. B. 263, 264 Lane, G. S. 414 Langland, W. 54, 324, 332 Lapidge, M. 185,186, 188 Lass, R. 209,213,420,436 Layamon 56 Leech, G. Ν. 119, 170, 288, 293 Leek, F. van der 435,436 Lehman, W. P. 263, 266 Leimberg, I. 307, 310 Lemon, G. W. 403 Levin, S. R. 82, 83, 86, 88 Lewandowska-Tomaszczyk, B. 450,451 Lewis, R. E. 338,450 Liggins, Ε. M. 190, 205 Lightfoot, D. 167,169,314, 321 Lindberg, C. 450 Lindelöf, U. 305, 308 Lindenbaum, S. 411 Liuzza, R. M. 305, 309 Long, Τ. H. 404 Lucas, P. J. 186 Luick, K. 37,40,258, 259, 263,404,405,411, 415,436 Lumiansky, R. M. 411 Lutz, A. 384,426,436 Lutz, F. 404 Lyon, H. R. 182 Macaulay, G. C. 52, 54 Macaulay, R. K. S. 258, 261, 264, 266,436, 437


Index of names

Madden, F. 80, 88, 235 Maejima, G. 3,18 Maeno, S. 6,9, 18 Mäher, J. P. 262, 264 Malkiel, Y. 237, 263, 264, 266 Μ alone, K. 404,405 Manabe, K. 7,12,18,189,190,199, 203, 205 Manly, J. Μ. 11, 287, 292, 293, 294 Mann, J. 227 Marchand, Η. 24,40, 447,448,451 Marckwardt, A. H. 321 Markus, M. 450,451 Marie, J. van 439, 452 Marshall, I. 287, 294, 295 Masui, M. 2, 3, 11,13, 18, 107,109, 110, 287, 288, 292-294,448,451 Matsu, M. 15,18 Matsuda, T. 4, 10, 18 Matsunami, T. 4, 9,18, 196, 205 Matthes, H. C. 263 Matthews, J. 293 Matthews, P. H. 32,40, 439, 451 McColly, W. 79, 85, 88 Mcintosh, A. 86-88, 91,109,174, 186, 222, 226, 227 Mclntyre, E. A. 180, 186 Meid, W. 412, 415 Menozzi, P. 262, 264 Meritt, H. D. 302, 398, 412,415 Meyers, L. F. 264 Miki, K. 11, 14, 15, 18, 109, 294, 451 Mill, A. J. 86, 87 Miller, T. 2,47, 205 Milton, J. 180, 188 Minakawa, S. 9, 18 Minkova, D. 10,420-422, 436, 437 Minsheu, J. 259, 403 Mitchell, B. 21, 22, 25, 29,40,176,177, 179-181, 187, 225,227,235, 313, 317,

318, 321 Miyabe, K. 3, 5, 7,18,148,179,180, 310 Miyake, K. 5, 18 Mizobata, K. 12,18 Moessner, L. 12, 231, 236 Moffat, D. 388, 397 Mohanan, K. P. 380, 385 Mohrbutter, A. 225, 228 Moriya, Y. 15,18 Morris, R. 80, 82, 86, 88, 235,412,413 Morrison, S. 187 Morton, J. 182 Moses, L. 263 Mukai, T. 186,188 Müller, Ε. 404,411,415 Murakami, R. 447 Murphy, J. J. 187 Murray, J. A. H. 338,415 Murry, J. 136 Mustanoja, T. 4,44,45, 54, 56, 73, 76,153, 163, 167,169, 229, 236, 336-338 Myers, S. 436 Nagashima, D. 14, 18 Nakajima, F. 8, 18 Nakano, H. 7, 18 Nakao, T. 6, 7, 9, 10, 15, 18, 52, 54, 209, 213, 364, 376, 384, 385,420,421,423,426, 427, 430, 431,433,435,437,447 Nakao, Y. 5, 15, 16, 18 Nakashima, K. 7, 18 Napier, A. S. 121,182, 215, 228, 302, 303, 306, 309,412,414 Nevalainen, T. 436,437 Nichols, A. E. 220, 228 Nilsen, A. P. 57, 75, 76 Nilsen, D. L. F. 57, 75, 76 Niwa, Y. 5, 13, 17, 18 Nolloth, Η. E. 412, 414 Oda, Τ. 8, 18

Index of names

Oess, G. 305, 309 Ogawa, H. 10, 11,12, 18, 187,215,225, 228, 308, 309 Ogoshi, K. 109 Ogura, Michiko 7.10,12,18,180,187, 229,232,233, 235, 236 Ogura, Mieko 10, 12,18, 237, 240, 250, 256-258, 260, 261, 264 Ohkado, M. 157 Ohye, S. 224 Oizumi, A. v, vi, 1,4,10,12-17,19,73, 108, 109,179, 287, 288, 292, 294, 435, 447-449,451 Onions, C. T. 333, 338,404 Ono, H. 8-10,19 Ono, S. 5, 7,9,12,13,19, 297, 305-307, 309 Orton, H. 239, 258, 264, 265 Oshitari, K. 11, 19,151, 186, 187 Ostheeren, Κ. 306 Otsuka, Τ. 3, 8, 19,435 Page, R. I. 179 Palmer, F. R. 63, 76 Palmer, L. R. 252, 260, 265 Panton, G. A. 88 Paolo, M. di 235, 236 Papajewski, H. 307 Parkes, Μ. 171, 177, 179,180, 187 Partridge, Ε. 404 Paul, Η. 237 Pauli, I. 412,415 Payne, F. Α. 199,200,205 Pearsall, D. A. 52,54 Peters, R. A. 45, 54 Petti, A. G. 180, 187 Philippsen, M. 148, 150,151 Phillips, B. 254, 265 Piazza, A. 262, 264 Pickles, J. D. 54


Plain, W. 435 Plummer, C. 192, 206 Pokorny, J. 412,415 Polom6, E. C. 188 Pope, J. C. 40, 178, 182, 215, 228, 302, 306, 309, 395, 396, 398 Porter, L. 287, 294, 295 Potter, S. 3,190, 196-198, 206 Poutsma, H. 56, 57, 73, 75, 77, 337, 339 Prins, A. A. 265,429, 431-433,437 Puppel, S. 19, 262, 264 Quirk, R. 21, 24-26,40,157, 158,170,190,

206 Radar, J. 412 Randall, J. 439,452 Rankin, S. 188 Rauch, I. 180,188 Rauter, H. 307, 310 Raymo, R.P.412 Reddy, R. B. 263, 264 Reichl, K. 439,452 Richardson, C. 404 Rickert, Ε. 11, 287, 292, 293, 294 Riddle, E. 439,444,447,449,452 Ringbom, H. 109 Rissanen, M. 263, 265, 436,437, 449,451 Ritt, Ν. 420,437 Roberts, J. 412,414 Robinson, F. N. 16,44, 54,109,179,188, 259, 293, 294, 337, 339,406,407, 410,411,416, 450 Roeder, F. 305, 309 Romaine, S. 439,440,448,452 Roscow, G. H. 109 Rosier, J. L. 302, 305, 309 Sadock, J. M. 76 Saito, K. 15,19


Index of names

Saito, S. 2, 19 Saito.T. 11,19, 54,452 Salmon, V. 91, 109 Salu, Μ. B. 183 Samuels, M. L. 86, 88, 91, 110, 251, 252,265, 421,437 Sanderson, S. 264, 265 Sandved, A. O. 52,54 Sasaki, T. 19 Sato, H. 13, 19 Sato, M. 3,19 Sauer, Η. 398,449,452 Scahill, J. 179,435 Scalise, S. 448,452 Schabram, H. 304-306,308, 310, 357, 410,415 Schipper, J. 27,41 Schipper, W. 11,308, 309 Schlauch, Μ. 91, 110,199,206 Schleicher, A. 265 Schmidt, J. 265 Schofield, W. H. 79, 88 Scholar, C. 110 Schöningh, F. 235 Schuchardt, Η. 237, 265, 266 Schücking, L. L. 119,136 Scragg, D. G. 188, 303, 306, 310, 395, 397 Sebeok, T. A. 293, 294 Sedgefield, W. J. 189,199, 206 Seebold, E. 301, 305, 306, 310,412 S6guy, J. 257, 265 Severs, J. B. 411,412, 416 Seymour, T. 259 Shakespeare, W. 4, 16, 259, 263, 291, 330,402,403,411,453 Shannon, A. A. 193, 203, 206 Shepherd, G. 182 Shibatani, M. 77 Shimamura, R. 435,439,452

Shirley, C. G., Jr. 99,110 Shores, D. L. 29,41 Sievers, E. 27, 37,41, 235, 391, 395, 399 Simmons, T.F. 412,414 Simpson, Υ. Μ. Y. 261, 266, 399,451 Sisam, C. 305,310 Sisam, K. 305,310,385 Skeat, W. W. 41,43,46, 51, 52, 54,127,150, 151, 226, 235, 287, 288, 292-295, 302, 305, 310, 337, 339,404, 411, 412,414 Skinner, S. 403 Small, J. 89 Smith, J. J. 86, 88,91, 110, 259 Smith, Μ. E. 266 Smith, P. H., Jr. 36, 39 Smithers, G. V. 86, 89, 411,414 Sonoda, K. 166, 313 Soukhanov, A. 136 Spiewok, W. 293,294 Spreckel, C. 193, 206 Stanley, E. G. 4, 8,9,11,13,16,185,186, 188,205-207 Stanzel, F. K. 307, 310 Stein, J. 412 Stockwell, R. P. 258, 259, 261, 266,420-422, 436,437 Stork, F. C. 52, 53 Suzuki, E. 89 Suzuki, S. 5,12,19 Suzuki, T. 13,19, 186, 188 Svartvik, J. 170,190,206 Swanton, M. 120, 131, 134-136, 173, 183,193, 206, 221,228 Sweet, H. 52, 54,189, 206, 314, 320,407,421, 437 Szarmach, P. E. 228, 306, 310 Szwedek, A. 308, 309 Taavitsainen, 1. 436,437 Taguchi, M. 85, 89

Index of names

Tajima, Μ. 10,11, 15, 19, 79, 85, 89, 166, 170, 323, 324, 327, 329, 331, 334, 336, 339 Takahashi, Η. 108 Takahashi, S. 5,11,19 Takamiya, T. 13,16,19 Takeuchi, S. 341,355, 358 Talmy, L. 74, 75, 77 Tanaka, H. 76, 77 Taniguchi, J. 3,19 Tatlock, J. S. P. 14, 54 Taylor, D. 96,97, 99,105,110 Terajima, M. 10,19, 361, 385 Terasawa, J. 12,16,19, 387, 395, 396 Terasawa, Y. 7, 8,10, 11,19,148,151, 187, 309,310, 399,401 Thorpe, B. 113, 114,123,173,177,181, 183,306, 358 Tilling, P. M. 265 Timmer, Β. J. 387, 397 Tolkien, J. R. R. 13, 85,86, 87, 89,183, 184 Toller, Τ. N. 136, 218, 226, 227, 387, 388, 395, 398, 399 Tomaszczyk, J. 450, 451 Trager, G. L. 266 Traugott, Ε. C. 77, 234, 236 Travis, L. 319, 321 Treiman, R. 385 Treip, M. 180,188 Trevisa 234 Tripp, R. P. 185,186 Trnka, B. 337, 339 Trudgill, P. J. 236, 237, 262, 266 Tsuchiya, T. 6,19 Tyrwhitt, T. 411 Uchioke, S. 166 Ueda, A. 435 Ueda, M.4,19


Ukaji, M. 6, 8,19,423,435 Urdang, L. 136 Ure, J. M. 226,228 Vantuono, W. 411,414 Venezky, R. L. 305, 308, 311, 358, 387, 388, 394, 395, 398,407,415 Vennemann, Τ. 265, 266,435 Vigfusson, G. 235, 236 Visser, F. Τ. 43-45,47-49, 52, 53, 54, 56, 57, 66, 73, 74, 77,170, 325, 327, 330, 332, 333, 337, 339 Wadell, Μ. L. 133,136 Wakatabe, Η. 10,19 Wakelin, Μ. F. 265 Wallace-Hadrill, Α. 449 Wang, W. S-Y. 237, 257, 258, 260-262, 264, 266 Warner, A. 169, 170, 302 Waterhouse, E. R. 193,206 Webster, N. 403,404 Wedgwood, Η. Α. 404,405 Weekley, Ε. 404,412,416 Weier, D. 79, 85, 88 Weiner, A. S. C. 399,451 Weinreich, U. 266 Wells, J. E. 79, 89,420,422,437 Wells, J. C. 79, 89,420,422,437 Wenisch, F. 306,311 Wetzel, C. D. 305, 308,410,415 Whitbread, L. 181,188 White, R. M. 175,183, 235 Whitehall, Η. 404 Whitelock, D. 173,181, 183,192,195, 197-199, 206, 358 Whitney, W. D. 404 Widdowson, J. 264,265 Wiener, L. F. 263, 266 Wilbur, Τ. Η. 265, 266


Index of names

Wilcox, J. 188 Wildhagen, K. 305,311 Willard, R. 188 Williams, C. J. 87 Williams, E. 443,448,452 Williams, W. P. 13 Wilson, B. 225 Wilson, R. M. 5 Windeatt, B. A. 176, 188 Winter, W. vi Wollmann, Α. 307,310 Worcester, J. Ε. 404,405 Wormald, F. 358 Wrenn, C. L. 21, 24, 25, 26,40, 87, 88, 184, 193,195, 197, 207 Wright, Ε. M. 23,41 Wright, J. 23,41,180 Wulfstan 7, 56, 173, 174, 177-182, 185-188, 190, 192, 215-228, 304, 306,

308, 309, 345, 412-414 Wyclif, J. 235, 337,448 Wyld, H. C. 251, 258, 259, 266,404,405 Yamada, N. 17,19,417,435,437 Yamaguchi, H. 2, 3, 5, 6,9,19 Yamamoto, Τ. 1,2,19 Yamanouchi, K. 354, 359 Yamashita, H. 13,19 Yerkes, D. 306, 311 Yokoyama, S. 85, 89 Yonekura, H. 10,15,16,19, 292, 294,439, 447,449,451,453 Yoshida, H. 109, 110 Yoshikawa, M. 9,19 Yoshino, Y. 9,19 Zaic, F. 307, 310

Trends in Linguistics. Studies and Monographs Edited by Werner Winter Mouton de Gruyter · Berlin · New York 69 The Role of Theory in Language Description. Edited by William A. Foley. 1993. 70 Robert H. Robins, The Byzantine Grammarians. Their Place in History. 1993. 71 Trends in Romance Linguistics and Philology. Vol. 5: Bilingualism and Linguistic Conflict in Romance. Edited by Rebecca Posner and John N. Green. 1993. 72 Language Conflict and Language Planning. Edited by Ernst H. Jahr. 1993. 73 Language Change and Language Structure. Older Germanic Languages in a Comparative Perspective. Edited by Toril Swan and Olaf Jansen. 1994. 74 Allan R. Bomhard and John C. Kerns, The Nostratic Macrofamily. A Study in Distant Linguistic Relationship. 1994. 75 Tense and Aspect in Discourse. Edited by Co Vet and Carl Vetters. 1994. 76 Wolfgang U. Dressier and Lavinia M. Barbaresi, Morphopragmatics. Diminutives and Intensifies in Italian, German, and Other Languages. 1994. 77 Language Contact and Change in the Austronesian World. Edited by Tom Dutton and Darrell T. Tryon. 1994. 78 On Languages and Language. The Presidential Adresses of the 1991 Meeting of the Societas Linguistica Europaea. Edited by Werner Winter. 1995. 79 Medieval Dialectology. Edited by Jacek Fisiak. 1995. 80 Thomas V. Gamkrelidze and Vjacheslav V. Ivanov, Indo-European and IndoEuropeans. A Reconstruction and Historical Analysis of a Proto-Language and Proto-Culture. 1995. 81 Linguistic Change under Contact Conditions. Edited by Jacek Fisiak. 1995. 82 Language and the Cognitive Construal of the World. Edited by John R. Taylor and Robert E. MacLaury. 1995. 83 Insights in Germanic Linguistics I: Methodology in Transition. Edited by Irmengard Rauch and Gerald F. Carr. 1995. 84 Meaning as Explanation. Advances in Linguistic Sign Theory. Edited by Ellen Contini-Morava and Barbara S. Goldberg. 1995. 85 Autolexical Theory. Ideas and Methods. Edited by Eric Schiller, Elisa Steinberg and Barbara Need. 1996. 86 Creole Languages and Language Acquisition. Edited by Herman Wekker. 1996. 87 Peter Harder, Functional Semantics. A Theory of Meaning, Structure and Tense in English. 1996. 88 Language Contact in the Arctic. Northern Pidgins and Contact Languages. Edited by Ernst H. Jahr and Ingvild Broch. 1996.

89 A Bibliography on Writing and Written Language. Edited by Konrad Ehlich, Florian Coulmas and Gabriele Graefen. Compiled by Gabriele Graefen and Carl W. Wendland, in collaboration with Georg F. Meier and Reinhard Wenk. 1996. 90 Historical, Indo-European, and Lexicographical Studies. A Festschrift for Ladislav Zgusta on the Occasion of his 70th Birthday. Edited by Hans H. Hock. 1997. 91 Henning Andersen, Reconstructing Prehistorical Dialects. Initial Vowels in Slavic and Baltic. 1996. 92 Natural Phonology. The State of the Art. Edited by Bernhard Hurch and Richard A. Rhodes. 1996. 93 Hans H. Hock and Brian D. Joseph, Language History, Language Change, and Language Relationship. An Introduction to Historical and Comparative Linguistics. 1996. 94 Insights in Germanic Linguistics II: Classic and Contemporary. Edited by Irmengard Rauch and Gerald F. Carr. 1997. 95 Seiichi Suzuki, The Metrical Organization of Beowulf. Prototype and Isomorphism. 1996. 96 Linguistic Reconstruction and Typology. Edited by Jacek Fisiak. 1997. 97 Advances in Morphology. Edited by Wolfgang U. Dressier, Martin Prinzhorn and John R. Rennison. 1997. 98 Language Change and Functional Explanations. Edited by Jadranka Gvozdanovid. 1997. 99 Modality in Germanic Languages. Historical and Comparative Perspectives. Edited by Toril Swan and Olaf J. Westvik. 1997. 100 Language and its Ecology. Essays in Memory ofEinar Haugen. Edited by Stig Eliasson and Ernst H. Jahr. 1997. 101 Language History and Linguistic Modelling. A Festschrift for Jacek Fisiak on his 60th Birthday. Edited by Raymond Hickey and Stanislav Puppel. 1997. 102 Robert S. Bauer and Paul K. Benedict, Modern Cantonese Phonology. 1997. 103 Studies in Middle English Linguistics. Edited by Jacek Fisiak. 1997. 104 Culture and Styles of Academic Discourse. Edited by Anna Duszak. 1997. 105 New Approaches to Chinese Word Formation. Morphology, Phonology and the Lexicon in Modern and Ancient Chinese. Edited by Jerome L. Packard. 1997. 106 Codeswitching Worldwide. Edited by Rodolfo Jacobson. 1997. 107 Salish Languages and Linguistics. Theoretical and Descriptive Perspectives. Edited by Ewa Czaykowska-Higgins and M. Dale Kinkade. 1997. 108 The Life of Language. Papers in Linguistics in Honor of William Bright. Edited by Jane H. Hill, P. J. Mistry and Lyle Campbell. 1997. 109 English Historical Linguistics and Philology in Japan. Edited by Jacek Fisiak and Akio Oizumi. 1998.